Volume 21 Issue 1 by IFAMA

International Food and Agribusiness Management Review

Official Journal of the International Food and Agribusiness Management Association

Volume 21 Issue 1 2018

EVERY GENERATION NEEDS ITS LEADERS. As a farmer-owned cooperative, we help lead the agricultural industry by promoting the most innovative, responsible farming practices available today. And we want to make sure that the next generation of leaders is well trained for tomorrow. That’s why we proudly support youth leadership programs like FFA. It’s also why we’ve invested in scholarships and countless land-grant university initiatives across the country. At CHS, we know every generation needs its leaders. And every potential leader needs a little help becoming one. Visit chsinc.com/stewardship to learn more about our ongoing commitment to leadership.

International Food and Agribusiness Management Review

Editorial Staff Executive Editor

Gerhard Schiefer, University of Bonn, Germany

Regional Managing Editors Asia, Australia, and New Zealand

Derek Baker, UNE, Australia Kim Bryceson, University of Queensland, Australia Kevin Chen, IFPRI-Bejing, China Jeff Jia, University of Exeter, United Kingdom Nicola M. Shadbolt, Massey University, New Zealand

Europe

Pegah Amani, Technical Institute of Sweden, Sweden Vera Bitsch, Technical University of Munich, Germany Laura Carraresi, University of Bonn, Germany Alessio Cavicchi, University of Macerata, Italy Hans De Steur, Ghent University, Belgium Loic Sauvee, UniLaSalle, Beauvais, France Cristina Santini, University San Raffaele, Italy Jacques Trienekens, Wageningen University, The Netherlands

North America

Ram Acharya, New Mexico State University, USA Yuliya Bolotova, Clemson University, USA Michael Gunderson, Purdue University, USA William Nganje, North Dakota State, USA R. Brent Ross, Michigan State University, USA Aleksan Shanoyan, Kansas State University, USA David Van Fleet, Arizona State University, USA Nicole Olynk Widmar, Purdue University, USA Cheryl Wachenheim, North Dakota State University, USA

South America

Aziz da Silva Júnior, Federal University of Vicosa, Brazil Jose Vincente Caixeta Filho, University of Sao Paulo, Brazil Emilio Morales, University of New England, Australia

Africa

Nick Vink, Stellenbosch University, South Africa

Editorial Board Filippo Arfini, Universita’ di Parma, Italy Stefano Boccaletti, Universita’ Cattolica, Italy Michael Boehlje, Purdue University, USA Dennis Conley, University of Nebraska - Lincoln, USA Francis Declerck, ESSEC Business School, France Jay Lillywhite, New Mexico State University, USA Woody Maijers, INHOLLAND University, The Netherlands

Marcos Fava Neves, FEA / USP / PENSA, Brazil Onno Omta, Wageningen University, The Netherlands Hernán Palau, Buenos Aires University, Argentina Christopher Peterson, Michigan State University, USA Thomas Reardon, Michigan State University, USA Mary Shelman, (Retired) Harvard Business School, USA Johan van Rooyen, University of Stellenbosch, South Africa

The IFAMR (ISSN #: 1559-2448) is published quarterly and the archived library is available at http://www.ifama.org. For copyright and publishing information, please contact: Marijn van der Gaag, Administrative Editor Wageningen Academic Publishers • P.O. Box 220 6700 AE Wageningen • The Netherlands • Tel: +31 317 476511 Fax: +31 317 453417 • Email: ifamr@wageningenacademic.com • Web: http://www.wageningenacademic.com/loi/ifamr.

International Food and Agribusiness Management Review Volume 21 Issue 1, 2018

Research Articles 1.

TABLE OF CONTENTS

Vertical integration in the Brazilian orange juice sector: power and transaction costs Nobuiuki Costa Ito and Decio Zylbersztajn

Information flow in the Sino-Brazilian beef trade

Seeding eastern Africa’s maize revolution in the post-structural adjustment era: a review and comparative analysis of the formal maize seed sector

Uncertainty in milk production by smallholders in Tanzania and its implications for investment

Towards high value markets: a case study of smallholder vegetable farmers in Indonesia

Competitiveness of local food: an empirical analysis of the tomato market dynamics

Susanne Knoll, Antonio Domingos Padula, Mariane Crespolini dos Santos, Guilherme Pumi, Shudong Zhou, Funing Zhong, and Júlio Otávio Jardim Barcellos

Olaf Erenstein and Girma Tesfahun Kassie

Edgar E. Twine, Amos Omore, and Julius Githinji

Marcus Maspaitella, Elena Garnevska, Muhammad I. Siddique, and Nicola Shadbolt

Xin Fang, Hui Huang, and PingSun Leung

Efficient farming options for German apple growers under risk – a stochastic dominance approach

101

Lighting the flame of entrepreneurship among agribusiness students

121

Maren B.K. Röhrig, Bernd Hardeweg, and Wolfgang Lentz

Lindsey M. Higgins, Christiane Schroeter, and Carlyn Wright

International Food and Agribusiness Management Review

Table of contents

Volume 21, Issue 1, 2018

Case Studies 9.

Supply chain re-engineering: a case study of the Tonghui Agricultural Cooperative in Inner Mongolia Qianyu Zhu, Cheryl J. Wachenheim, Zhiyao Ma, and Cong Zhu

10. Growth strategies for a commercial farm: the AgroPastoril Campanelli case study

Roberto Fava Scare, Allan Wayne Gray, Rodrigo Lourenรงo Farinha, Erin Chelsea Fullerton, and Marcos Fava Neves

International Food and Agribusiness Management Review

133

161

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2016.0071 Received: 14 March 2016 / Accepted: 3 February 2017

Vertical integration in the Brazilian orange juice sector: power and transaction costs RESEARCH ARTICLE Nobuiuki Costa Ito

and Decio Zylbersztajnb

aResearcher

and bProfessor of Economics of Organizations, School of Economics, Business and Accounting, University of São Paulo, Center of Studies in Agro-Industrial Relations (PENSA), Avenida Professor Luciano Gualberto, 908 sala C-18, Cidade Universitária, 05508-900 São Paulo, SP, Brazil

Abstract This paper aims to investigate the role of market power on vertical integration choices in the Brazilian orange juice sector. The main hypothesis states that market power, along with economizing drivers, has had an important role in the economic organization of the orange juice sector. In order to accomplish this task, we examine the make-or-buy decision of juice-processing firms in procurement of inputs (fruits), through a 15-year panel of five firms (more than 90% of the market). Empirical results offer new theoretical insights on role of market power in vertical integration decision. From these theoretical advancements, as implications for practitioners, market power raises attention for critical strategizing issues and antitrust remedies. Keywords: transaction costs, power, vertical integration, orange juice JEL code: D23, L22, L49 Corresponding: nobuiuki@usp.br

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

1. Introduction The presence of juice processors in the citrus growing activity has been a characteristic in the Brazilian orange juice sector since its beginning (Hasse, 1987; Neves and Lopes, 2005). However, during the 1990s, this presence became increasingly large, which led to citrus growers’ accusation that vertical integration was harming competition. The Administrative Council of Economic Defense, the Brazilian antitrust office, received citrus growers’ complaints, but, in the 1990s, antitrust intervention in the sector did not address any constraints to vertical integration. Nonetheless, accusations of market power through vertical integration are persistent in the sector throughout the 2000s and the 2010s. In fact, partial backward vertical integration of juice processors into orange production increased, inasmuch as the production of oranges in the farms owned by juice processors scaled up from more than 20 million boxes in 2002 to 130 million boxes in 2012 (CADE, 2014). The research question is: how power affects juice processor’s vertical integration choice in orange juice sector after 1990? The distinction of economic power from efficiency arguments in vertical integration decision is an ambiguous matter in economic theory (Lafontaine and Slade, 2008). Transaction Costs Economics (TCE) literature points to economizing on transaction costs as the main case and applied to all situations, while power explanations are applied to small number situations (Williamson, 1991a,b, 1995). Recognition of power as determinant of vertical integration decision is limited and no efforts have been made to investigate which are the special case situations in which power is relevant. According to Joskow (2002: 105): essentially no effort has been made to harmonize the large body of theoretical and empirical work in the TCE tradition that is relevant to understanding why specific governance arrangements emerge, and for performing any trade-offs that may arise between increases in market power and reduction in the costs of transacting à la Williamson. The paper aims to analyze vertical integration choice of juice processors in the orange juice chain, in order to highlight the determinants of the increasing in vertical integration after 1993. One of the hypotheses posits that vertical integration is more likely to occur as the relationship specific investments deepen through time. A second hypothesis states that inefficiencies caused by antitrust intervention contributed to increasing vertical integration. Finally, a third hypothesis states that power has an important role in the economic organization of the orange juice chain due to structural changes in the sector. Thus, the hypotheses are grounded in the economizing on transaction costs and strategizing through market power, in order to explain the vertical integration path of the orange juice sector. Nevertheless, a simultaneous effect of economizing and strategizing is expected (Williamson, 1991b). The empirical evidence contributes to understand the special case situations in which power is relevant, with implications to business practice and antitrust authorities. The paper is organized in six sections including this introduction. The second section presents the transaction costs arguments to vertical integration decision and role of power under this theoretical lens. The third section analyzes the history of orange juice sector in São Paulo state, located in Brazil, emphasizing the previous events that have led to increasing vertical integration after 1993. The fourth section presents data and methods and, in the fifth section, results are discussed. Finally, in the sixth section, concluding remarks follow.

2. Vertical integration choice It is possible to split the production process in many technologically separable activities and the firm is a combination of these activities to transform inputs in outputs. Vertical integration occurs when the firm internalizes technologically separable activity that was originally carried on outside the firm, i.e. through the market. In the standard view of industrial organization theorists, the firm combines economic activities that are technologically similar or activities that are clearly physically related, in order to minimize production costs (Joskow, 2005). Considering this perspective, there is no point in joining two industrial plants in the absence of technological relationship. International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

As consequence of this traditional approach of vertical relationships among firms, those non-standard vertical integrations were usually seen as some kind of monopoly power. According to Coase (1972: 67): one important result of this preoccupation with the monopoly problem is that if an economist finds something – a business practice of one sort or other – that he does not understand, he looks for a monopoly explanation. And as in this field we are very ignorant, the number of ununderstandable practices tends to be rather large, and the reliance on a monopoly explanation, frequent. Nevertheless, the firm’s vertical boundaries are not determined solely by technical criteria such as economies of scale and production technology (Williamson, 1971). TCE (Williamson, 1975, 1985) exploits organizational choices of vertical boundaries that are determined by relational features among economic agents. The perspective that market organization and firm’s internal organization can be substitute rather than complementary was conceptualized by the seminal work of Ronald H. Coase (1937). Neoclassical economic theory disseminates the idea that markets are the coordination mechanism that maximizes resources allocation. Thus, if it is true, why do firms exist? Or conversely, why does not just one big firm exist? Coase’s (1937) answer to these questions states that there is a cost to use the price mechanism, called transaction costs. Firms can economize those transaction costs and, therefore, market and non-market organizations can be compared in terms of costs of transacting. In TCE framework, the vertical integration decision depends upon the transaction costs involved on the comparative analysis of different forms to organize a transaction; which are under market, contracts, and bureaucracy within firms. Thus, transaction is the ultimate unit of analysis in TCE (Williamson, 1975, 1985), since it can capture the essence of the decision of produce in-house versus buying in the market. The choice for these types of governance structures occurs taking into account three transaction attributes: (1) uncertainty, which are disturbances in quantity, quality, or prices; (2) asset specificity, which are the relationship-specific investments that lose value in alternative uses; and (3) frequency, in terms of transaction recurrence (Williamson, 1985). Among transaction attributes, asset specificity received more attention in both theoretical models (Williamson, 1985, 1991a) and empirical analyses (Masten and Saussier, 2000). The main hypothesis states that as relationship-specific investments deepen, risks of renegotiation and quasi-rent appropriation become higher (Klein et al., 1978) and safeguards are required, i.e. in presence of asset specificity, long length contracts and vertical integration are more likely to be adopted – the alignment hypothesis (Williamson, 1991a). TCE hypothesis finds incontrovertible empirical support (Masten, 1993, 1996; Masten and Saussier, 2000; Shelanski and Klein, 1995). TCE’s empirical tests use measurements of asset specificity (independent variables) and correlate it to observed governance structure – usually using limited dependent variable (Maddala, 1983). Parameter shifts in TCE’s empirical inquiries were disregarded. Parameter shifts are factors that affect governance costs, such as property rights, contract law, uncertainty, and reputation (Williamson, 1991a). Property rights are related to value expropriation by the firms’ lack of capacity to protect property rights against Government, rivals, suppliers or buyers. Changes in contract law can affect governance costs, because it alters contract enforcement. Uncertainty is disturbances in factors that affect the transaction. Finally, reputation represents a behavioral standard that can or cannot guarantee contracts. Furthermore, empirical research in TCE tradition has limited efforts in testing power explanations for economic organization along with transaction costs variables. Nevertheless, power is not disregarded in transaction costs lens. As mentioned before, TCE framework emerged in a context in which power explanations for non-standard organizational forms prevail (Coase, 1972). TCE had an important role in emphasizing the efficiency explanations that were disregarded. In doing so, theorists of TCE assigned higher weight to efficiency purpose organizations, in order to counterbalance the previous perspective of monopoly power. Williamson (1985: 17) made it clear asserting that: International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

the economic institutions of capitalism have the main purpose and effect of economizing on transaction costs. The main purpose is not, however, to be confused with solely purpose. Complex institutions commonly serve a variety of objectives. This is no less true here. The inordinate weight that I assign to transaction costs economizing is a device by which to redress a condition of previous neglect and undervaluation. Power is one of the factors composing these complex institutions, but Williamson (1995) contrasts and describes transactions using efficiency and power purposes. When contracting is voluntary, knowledgeable, and farsighted, there is no room for power. Actually, power is present in the opposite situation, when contracting is involuntary, uninformed, and myopic. Williamson (1995) claims that contracting processes are frequently voluntary, knowledgeable, and farsighted, especially in intermediate product markets, and economizing lens is a useful place to start the analysis. Williamson (1995) points to another two problems when power is considered in economic analysis. First, its concept is vaguely defined and fails to identify the critical dimensions in which power makes difference. Definition of power ends in ex post rationalization, i.e. an artifice that explains the event after the occurrence. Second, still related to the first one, the concept of power is tautological. It is a concept that explains everything and, consequently, explains nothing. The concept of transaction costs also faced this tautological problem, but defining the transaction as unit of analysis, distinguishing transactions by its attributes, and measuring those attributes could solve the problem. Despite this, power can be an attribute of the transactions, since it characterizes a relationship; it is not possible to distinguish transactions by power features because robust measurements are not available. Williamson (1991b) also considers the concept of power in the Strategic Management literature. Among several streams of thoughts, Strategic Management field can be grouped into two main rationales: strategizing and economizing. The first one, strategizing, is rooted in market power and it is the most disseminated in the field. The second is the economizing, which is more theoretically refined – in price theory and the firm as production function – but has received less attention. In order to understand the differences in the strategizing and economizing perspectives, it is necessary to understand the economies of first and second orders. The source of these economies is the potential unnecessary bureaucracies and waste due to maladaptation problems, which is neglected by a firm as production function model. Maladaptation problems are related to bad choices in organizational forms. Transactions differ in terms of attributes and governance structures distinguish by costs and competences. In effect, misalignments in transaction features and the chosen organization – maladaptation problems – create waste. First and second order economies are the result of the solution of these maladaptation problems. Using a simple model of partial equilibrium (Williamson, 1991b), represented in Figure 1, it is possible to visualize the effects of first and second order economies. Assuming a firm selling a given product in the quantity q1 for the price of p1, which is higher than average costs c0. In this situation, due to maladaptation problems, there is an excess in costs equal to b that results in production and transaction costs equal to c0+b. If the activity reorganization removes maladaptation problems, therefore, the excess b is eliminated. Holding prices constant, social gains are promoted by the elimination of waste W=bq1. This is the first order economizing in which changes in governance structure eliminates the waste, W. In the case of change in price from p1 to p2, reducing the price to the level of average costs c0, we add benefits from allocative efficiency equal to L=½bΔq, where Δq= q2–q1. This is the second order economizing, given by L. Although the second order economizing is also significant, it is less than first order economizing. Williamson’s (1991b) critics consist that economists usually focus on second order economizing, while first order is neglected, even whether the waste is greater than the deadweight losses (triangles). In effect, the dominant perspective emphasizing in market power and strategizing could be substituted by the refinements of economizing perspective, suggesting that ‘between economizing and strategizing, economizing is much International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

p1 b

p2 = co

D q2

Figure 1. First and second order economizing (adapted from Williamson, 1991b). more fundamentalâ&#x20AC;&#x2122; (Williamson, 1991b: 75). Finally, the author argues that strategizing and economizing are not mutually exclusive drivers for economic organization, but strategizing is relevant to a small number of situations, while economizing is relevant to all situations. Azevedo (1996), in turn, uses bargaining power as the main factor to explain vertical integration choice. Using game theory, Azevedo (1996) analyzed the partial vertical integration in the Brazilian orange juice sector during the 1990s. Taking into account the presence of specific investments and barriers to forward vertical integration of citrus growers, vertical integration can serve as a bargaining instrument over surpluses. The model showed that an optimal degree of vertical integration is chosen when marginal benefit is equal to marginal cost of organizing this activity under contracts. Nevertheless, the model also showed how juice processors could extrapolate the optimal degree of vertical integration in order to appropriate part of the surplus generated from cooperation. In fact, when higher than optimal vertical integration degree is found, welfare losses will take place according to the author. Thus, vertical integration can be used as strategic means to redistribute wealth through bargaining power with implications to welfare. Transaction costs arguments highlight non-technological determinants to vertical integration, represented by transaction attributes. This is a significant contribution, since in standard neoclassical theory, combination of economic activities under non-technological features were solely viewed as market power exertion. Nevertheless, investigations in TCE traditionally did not debate a possible trade-off between economizing on transaction costs and power explanations. Indeed, debates between power and efficiency arguments to vertical integration are ambiguous. Vertical integration is efficient when avoiding the double marginalization in a sequence of two monopolists (Tirole, 1988) or when mitigating risks from opportunistic renegotiation in contracts (Klein et al., 1978). However, under power perspective, vertical integration creates barriers to entry, raises rivals costs, affects prices (Joskow, 2005; Riordan, 1998), or serves as a bargaining instrument (Azevedo, 1996). Despite TCE provides additional information about efficiency reasons to vertical integration, it fails to recognize and give sufficient attention to special case situations in which power is relevant (Dorward, 2001). There is a risk in conclusions for efficiencies purpose in every non-standard â&#x20AC;&#x201C; different from minimize production cost â&#x20AC;&#x201C; vertical integration, because power is also potentially present.

3. Orange juice sector: transaction costs and power The large-scale production and exportation of high quality orange juice in Sao Paulo initiated in 1960s, after an intense freeze that devastated orange crops in Florida. The origins of Brazilian juice processing firms are related to large orange producers, which were able to employ their fruits excess in an alternative product. International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

Initially complementary, juice production in Brazil presented rapid growth during the 1970s and 1980s. Exportation of orange juice scaled up from 531 tons in 1963 to more than 33 thousand tons in 1970 and, then, to more than 401 thousand tons in 1980. The sector also became more specialized in juice production, as 2% of the orange production in São Paulo was used to produce juice in 1970, while this figure changed up to 81% in 1980. As a result, large citrus growers became large juice-processing firms. The industry structure is concentrated since its beginning, given that the top four processors controlled 97% of the market in 1967, 80% in 1990 (Hasse, 1987; Maia, 1996), and 96% in 20101. These large juice processing-firm depended upon thousands citrus growers, which had heterogeneous features. In effect, given that juice-processing industry structure is persistently more concentrated than citrus industry structure, market power has been potentially present throughout the history of the sector. Additionally, partial vertical integration has been present in the organization of the orange juice sector since 1963, since large juice-processing firm kept a small amount of their previous main occupation, i.e. orange production. However, starting in the beginning of 1990s, there was a remarkable acceleration in the growth rate of backward vertical integration of juice-processing firms. In other words, large juice firms were producing more oranges, as Figure 2 shows. The traditional TCE analysis would look for efficiency explanations in terms of transaction costs. Relationshipspecific investments are the most important dimension to be empirically analyzed under transaction costs arguments and there are several relationship specific investments in the juice production chain. Regarding citrus production, first, it is a perennial crop and new trees remain unproductive for long periods; second, it is an immobile investment in terms of geographic location; and third, types of oranges destined to juice production are not appreciated by fresh fruit consumers. In juice processing, the production plants used to process orange juice are requiring very specific investments, which often times minimizes the value of the asset beyond the citrus grower and juice processor’s relationship. Analyzing the relationship between citrus growers and juice processors, the distance between farms and processor plants are called site specificity. According to Williamson (1985), ‘cheek-by-jowl’ relationships are more specific, due to redeployment and set up costs. Distances between farms and production plants reduce transportation and coordination costs and reduce transportation time. There is also a temporal specificity, because oranges are perishable and it must be processed quickly after harvest. According to Masten et al. Calculated by IEA, the Agricultural Economics Institute. The institute is the research branch of São Paulo state government and it focuses on agriculture. 1

40,000,000

Firm B

Firm A

30,000,000

20,000,000

10,000,000

International Food and Agribusiness Management Review

Figure 2. Two largest juice processing firm’s investment in orange production.

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

(1991), temporal specificity occurs when threats of delays are conditions to extract price concessions, which is the case in the orange juice production chain. Thus, as distances between production plants and farms decrease, more specific are the investments, in terms of site and temporal specificity, and more vertical integration is expected. The relationship between citrus growers and juice processors also presents physical specificity. According to Williamson (1985), physical specificity is related to investments in equipment, machines, and other physical assets with characteristics that are designed to a specific transaction. In the citrus growersâ&#x20AC;&#x2122; activity, trees are physically specific because orange type destined to juice production loses value in alternative use, i.e. for fresh fruit consumers. The investment in trees requires time â&#x20AC;&#x201C; around 4 years â&#x20AC;&#x201C; to be fully productive and redeployment of land to other uses is quite restricted. Thus, as physical asset specificity increases, greater will be transaction costs associated and more vertical integration will be expected. Thus, as the relative quantity of oranges delivered in juice-processing plants grows (compared to oranges destined to fresh fruit consumer), more specific are the investments in orange production. Joskow (1987) used a similar relative quantity of supply as asset specificity measurement in coal mining sector. Finally, in juice production side, juice-processing plants have no alternative use other than orange juice production. It is possible to note that investments in juice distribution such as ports terminals, trucks and ships would follow investments in processing plants. Thus, as the number of plants increase, more specific are those investments. Hypothesis 1 is then a hypothesis of economizing on transaction costs. H1: (a) site specificity; (b) physical specificity in orange crop; and (c) physical specificity in processingplants are positively associated to vertical integration. Relationship-specific investments create a situation of small number bargaining, i.e. bilateral monopoly between citrus growers and juice processors (Williamson, 1985). This asset specificity analysis, however, is a cross-sectional view of transaction that does not include the process perspective (Zajac and Olsen, 1993). Structural changes in the sector could affect the relationship between citrus growers and juice processors after specific investments. The immediate effect of this situation is the change in governance costs, even keeping the asset specificity constant, In other words, under TCE lens, parameter shifts could affect governance costs and, therefore, vertical integration decision. In the case under analysis, the change occurred in contract law. In the orange juice sector, there were an important antitrust contentious in the 1990s. In this contentious, citrus growers were complaining about some of the terms in the standard contract in force at that time. The standard contract was a result of learning and collective negotiations between citrus growers and juice-processing firms, which was pursued for more than 15 years in the sector. The contract design proposed attached orange prices to juice prices in the international markets, and it was voluntarily and massively adopted by the parties. Despite the spontaneous adoption, the accusation targeted the methodology to calculate orange prices, claiming that it is an instrument to raise profits, i.e. oranges would be underpriced. Antitrust Council accepted the accusations and initiated the legal process, but citrus growers and juice processors decided to sign an agreement, called commitment term to conduct cessation. Antitrust authorities suspended the use of a standard contract and collective negotiations were forbidden. In fact, citrus growers were not asking for the termination of standard contracts, but advocating for punctuated changes in its provisions. Antitrust Council, in turn, decided to not interfere in contractual practices, and extinguished the standard contract as an attempt to re-establish more competitive conditions. The rationale that underlies antitrust decision counted on the competition and individual negotiations as a mechanism to achieve efficiency through multiple contracting practices. Furthermore, antitrust office disregarded the vertical integration, which was part of the complaints made by citrus growers, i.e. juice-processing firms backward vertical integration would harm competition. In practice, the major organizational arrangement at the time was forbidden and did not leave a viable means to carry out the transactions. This lack of means moved up the hybrid governance costs, which turned the contracting costlier and led to higher degrees of vertical integration. International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

(H2): antitrust intervention and standard contract prohibition is positively associated to vertical integration. Still regarding dynamic aspects of the organizing process (Zajac and Olsen, 1993), there were technological changes in the sector that affected not only the industrial organization but also the viable safeguards available. During the 1980s, the distribution of frozen orange juice using steel drums was substituted by a bulk system. This innovation required large investments in special-purposes trucks and ships, as well as dedicated port terminals, to transport the production from Brazil to Europa and the USA. Yet immense investments were necessary, the cost reduction was huge, and all large juice processing-firms adopted the new technology. As consequence, the barriers to entry increased in the sector of juice-processing, as new players must invest not only in production plants, but also create an expensive and complex logistic structure, dealing with the saturated port system in Brazil. Beyond initial investments, incumbent juice-processing firms developed close ties to buyers in international markets, which create another barrier to entry by difficult access to buyers. The barrier to entry introduced by the new distribution system transformed the industrial organization. The largest juice-processing firm consolidated their positions, as concentration increased from 74% in 1997 to 96% in 2010 – considering the sum of market share of the four largest orange juice producers. This concentration growth per se creates condition to market power. However, there is another effect from the technological transformation, impacting the safeguards to transactions. Before the 1990s, citrus growers could protect their specific investments through vertical integration and they did it, as well as these juice processors maintained low levels of investments in orange production. Nevertheless, changes in the relationship and market context led to a situation in which only juice processors could protect their specific investments through vertical integration. Citrus growers remained vulnerable to opportunistic price renegotiations after the transformations caused by the technological innovations. The year of 1993 marked this disadvantageous context to citrus growers, when Frutesp, one of top four juice-processing firms and the last controlled by citrus grower’s co-operative, was acquired by an international trading company. From this point, there were no forward vertical integration options to farmers. The contracting is no longer voluntary, knowledgeable, and farsighted (Williamson, 1995), because nonpredictable changes occurred after specific investments were made. The barriers to entry in the juice processing industry and the absence of citrus growers’ efforts in forward vertical integration are transformations in the evolution process of the relationship. In the presence of barriers to farmer’s forward vertical integration, due to barriers to entry, juice-processing firms had increasing market power. In this context, juice processingfirms could take advantage of market power and choose higher degrees of vertical integration to achieve bargaining power over citrus growers (Azevedo, 1996). As juice processors control greater market share, the market power will higher and, then, more vertical integration will be expected. The hypothesis 3 is then a hypothesis of power explanations for vertical integration. (H3): market share of juice-processing firms are positively associated to vertical integration. Economizing and strategizing are frequently treated as mutually exclusive in the literature, i.e. there is a trade-off between transaction costs and market power (Joskow, 2002; Williamson, 1991b). Using TCE’s lens, backward vertical integration could be the solution to coordinate orange transactions to economize on transaction costs. Another explanation is consistent with power purposes to vertical integration, as conditions of voluntary, knowledgeable and farsighted contracting are not satisfied. Instead of taking one side or another, economizing or strategizing, in our hypotheses setting, we claim that both drivers are simultaneously found in the economic organization of orange juice sector.

International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

4. Data and methods The following functional form represents the basic model: VI = f (K, SHIFTPAR, POWER, LAND, SUGC) where, VIit = Investments in orange production made by juice processing firms, given by the quantity of orange boxes produced by ith firm in the tth period (million boxes; data provided by Associtrus). K= Asset Specificity: K_CROPt = physical specificity2 of orange crop, sum of the quantity of orange boxes processed by all firms in the tth period divided by total boxes produced in São Paulo state in the tth period. This measure is not firm specific (data provided by CitrusBR3 and Agricultural Economics Institute (IEA)); K_PLANTit = physical specificity of processing plants, dummy variable where 0 denotes the period before installation of new processing plant by ith firm and 1 denotes the period after installation of new processing plant by ith firm (data provided by Associtrus); and K_SITEit = site specificity4, given by the concentration ratio of orange production around juice-processing plants. It is the average of the quotient of quantity of oranges produced in jth city over distance between the jth city and the city where juice-processing plant is located (adapted from IEA data). SHIFTPAR = Parameter shifts: ANTIT t = changes in contract law due to antitrust intervention through standard contract prohibition. It is a dummy variable where 0 denotes period before antitrust intervention and 1 denotes the period after intervention; UNCERTt = uncertainty, standard deviation of orange boxes annual prices in the last five years for tth period (data provided by Informa Economics (FNP)5 and CEPEA6); and PROP_Rt = overall score of economic freedom index for tth period, which measures elements such as property rights, business freedom, labor freedom, among others. (adapted from Heritage Foundation; http://tinyurl.com/zgdx9wd). POWERit = power, given by market share of ith firm in the tth period (adapted from IEA, FNP, and Sabes, 2010). LANDit = Land value, average price of land in the region of the ith firm in the tth period. Prices deflated using Índice Geral de Preços Disponibilidade Interna (IGP-DI) index (data provided by IEA). SUGC = Influence of sugarcane sector in orange juice sector: SUGC_Pit = average value of tenancy for sugarcane production in the region of ith firm in the tth period. Prices deflated using IGP-DI index (data provided by IEA); and SUGC_Ait = production area of sugarcane in the region of ith firm in the tth period (data provided by IEA). Initially, some commentaries about dependent variable are needed, which is a measure for vertical integration. Backward vertical integration is the production of oranges by juice-processing firms. One can say that some This variable also captures temporal specificity effects. CitrusBR is a representative association of the biggest Brazilian producers and exporters of citrus juices and derivatives; http://www.citrusbr.com. 4 Concentration ratio indicates the extent in which orange production is geographically concentrated around cities where processing plants are installed in each firm and each year. 5 Informa Economics (FNP) is a consulting group focused on agribusiness information; http://www.informaecon-fnp.com. 6 CEPEA is a research center located in the agronomics school of University of São Paulo; https://www.cepea.esalq. usp.br/br. 2 3

International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

kind of degree of vertical integration is better than the dependent variable used in this study. Nevertheless, there is no available data about the number of orange boxes processed by each juice-processing firm. Despite the lack of available data, the chosen dependent variable can capture the increase of vertical integration. Indeed, it is possible to state that the chosen variable is more accurate in terms of underlying rationality of decision-makers, because it is more likely that people decide over quantities rather than percentages. It is not the relative quantity of inside production versus outside procurement, but the simple fact that juiceprocessing firms are increasing the quantity of inside orange production of oranges, regardless the quantity of juice produced. Conversely, if managers in juice-processing firms decided to reduce juice production, it is not necessarily true that they also decided to increase vertical integration. Interest variables are asset specificity (H1), antitrust intervention (H2), and power (H3). In addition, consistent with transaction costs propositions, parameter shifts can influence the choice for governance structures. Whereas TCE’s theoretical propositions do not specify expected effects for parameter shifts, it is expected that uncertainty is positively associated to vertical integration because more hierarchical coordination deals with uncertainty. Regarding property rights, it is expected that increases in the quality of property rights index are negatively associated to vertical integration, because it is less costly to avoid value expropriation through markets or contracts. Parameter shifts variables are not firm specific. Control variables include: land value and sugarcane influence over orange sector. It is expected that land value is negatively associated to vertical integration, since increases in land prices can inhibit expansion of vertical integration. The sugarcane crop is located in the same region as the orange crop and these two products compete for land. Thus, price of land tenancy to sugarcane production can attract citrus growers to change their crop from orange to sugarcane. Furthermore, increases in the production area of sugarcane can also threaten orange supply. We expect an ambiguous effect, because, on the one hand, sugarcane crop is positively associated to vertical integration, because juice processors will seek to guarantee their supply of fruits. On the other hand, sugarcane can increase the overall land opportunity cost, with negative influence to vertical integration. This is an exploratory study. This study advances in empirical inquiry under TCE lens using longitudinal data of just one transaction, which is not the traditional approach. Inclusion of parameter shifter effects is not frequent in empirics of TCE as well, which is made in this study. The model tests power explanations along with TCE’s traditional measures of asset specificity and institutional changes like antitrust intervention, which is also not frequent in this kind of investigation. The sample considered a 15-years period, from 1993 to 2007, for the five largest juice-processing firms. This is an unbalanced panel data, since there was a juice processor that closed its operations during this period and there is no available information about market share in 1993 for one of juice processors. Hence, a total of 70 observations are available for major part of variables. Property rights index, land prices and tenancy prices for sugarcane are available between 1995 and 2007. Table 1 presents descriptive statistics and correlation matrix. VIit = β0 + β1 K_CROPt + β2 K_SITEt + β3 K_PLANTit + β4 UNCERTt + β5 PROP_Rt (+) (+) (+) (+) (-) + β6 POWER + β7 LANDit +β7 SUGC_Pit + β8 SUGC_Ait + β 9 ANTITt (+) (-) (+/-) (+/-) (+)

The econometric model is specified as follow: The dataset is organized in a panel data. This is a long panel, since it has relatively many time periods and few firms. Regarding estimation techniques, the main distinction in those datasets is random or fixed effects (Cameron and Trivedi, 2009). As the five largest juice-processing firms constitute the sample, it is reasonable to assume that fixed effects estimates are suitable to data characteristics. Nevertheless, estimations using fixed effects and random effects were assessed by the Hausman test, indicating that fixed effects provide International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

Table 1. Descriptive statistics and correlation matrix.1 Variable

N Average Std. dev.

1. VI 2. K_ CROP 3. K_SITE 4. K_PLANT 5. UNCERT 6. PROP_R 7. POWER 8. LAND 9. SUGC_P 10. SUGC_A 11. ANTIT

70 70 70 70 70 61 70 61 61 70 70

11.26 0.77 15,500 ND** 1.04 57.89 0.17 10,427 603 131,219 ND**

= significant at P≤0.05;

9.92 1 0.08 0.34* 2,945 -0.31* ND** 0.20 0.26 0.04 5.11 0.43* 0.09 0.77* 3,745 0.14 100 0.06 24,061 0.11 ND** 0.47*

1 -0.06 -0.27* -0.10 0.16 0.28* -0.17 -0.05 0.17 0.38*

1 0.25* 1 0.05 -0.07 1 0.01 0.23 0.66* 1 -0.11 -0.03 0.08 0.49* 1 -0.01 -0.09 -0.27* 0.19 0.15 1 -0.29* 0.15 -0.33* -0.15 -0.13 -0.03 1 0.17 0.08 0.084 0.27* 0.34* -0.23 -0.06 -0.01 0.28* -0.10 0.37* 0.35* -0.41* -0.10

10.

11.

1 0.39*

= dummy variable.

more consistent estimators. Finally, natural log transformations of both dependent and independent variables were performed, exception made for dummy variables.

5. Results Table 2 presents the results. In general, all three hypotheses are supported: economizing on transaction costs as asset specificity deepens (H1); antitrust intervention as a shifter parameter (H2); and power exertion (H3). Regarding the traditional TCE’s alignment hypothesis, physical specificity plays an important role influencing the vertical integration trend in the analyzed period because it presents a positive and significant effect (H1b,c). Site specificity, in turn, has no statistically significant effect over vertical integration (H1a). This last result does not mean that economizing on transaction costs has limited influence in vertical integration Table 2. Results of panel data analysis. Variables

VI 1993-2007

VI 1995-2007

VI1 1993-2007

VI1 1995-2007

K_CROP K_SITE K_PLANT UNCERT PROP_R POWER LAND SUGC_P SUGC_A ANTIT Constant Observations Adjusted R-squared

0.664* (0.367) -0.183 (0.287) 0.576*** (0.101) 0.387*** (0.131)

0.694** (0.314) -0.485 (0.340) 0.556*** (0.0928) 0.501*** (0.184) 1.715*** (0.559) 0.320*** (0.102) -0.165 (0.143) 0.154 (0.211) -0.507** (0.229)

0.664 (0.333) -0.183 (0.476) 0.576*** (0.0470) 0.387** (0.130)

0.694* (0.302) -0.485 (0.382) 0.556*** (0.116) 0.501** (0.174) 1.715** (0.529) 0.320 (0.217) -0.165 (0.134) 0.154 (0.149) -0.507* (0.194)

0.339*** (0.0644)

-0.530** (0.260) 0.813*** (0.104) 9,932** (4.001) 70 0.827

6,899 (5.723) 61 0.826

0.339*** (0.0253)

-0.530** (0.161) 0.813*** (0.116) 9,932 (5,344) 70 0.838

Using cluster robust. = significant at P≤0.01; ** = significant at P≤0.05; *= significant at P≤0.10; t-statistic in parentheses.

***

International Food and Agribusiness Management Review

6,899 (4,943) 61 0.840

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

in this case because our focus is vertical integration decision made by juice-processors. In this sense, since physical specificity in juice-processing plants is highly significant in all estimated models, results point to strong evidence that relationship specific investment affected the vertical integration decision. Antitrust intervention also contributed to increasing vertical integration of firms during the analyzed period. Antitrust decision influence was positive and statically significant (H2). Given the prohibition of a contract that was a voluntary and privately negotiated contract type, the hybrid governance curve became more expensive due to negotiations of juice-processing firms with thousands of different suppliers. In effect, contracting became more expensive after antitrust intervention, which moved up the hybrid governance curve. In fact, antitrust intervention created inefficiencies in terms of transaction costs minimization, since it substituted contracts by greater internal production. Supporting the third hypothesis, vertical integration path has also a power component. In other words, changes in market share have a positive and statistically significant association to vertical integration (H3). This result indicates that gains in bargaining power can play an important role in the decision of vertically integrate the production. The empirical evidence is consistent to Azevedoâ&#x20AC;&#x2122;s (1996) arguments, suggesting that, in the presence of specific investments and barrier to forward vertical integration of the supplier, the buyers tend to pursuit backwards vertical integration. The high level of partial vertical integration would allow the buyers appropriate part of the surplus generated in the transaction. The evidence is also consistent with farmersâ&#x20AC;&#x2122; formal accusations in the antitrust contentious during the 1990s. Also related to TCEâ&#x20AC;&#x2122;s prediction, the other parameter shifts presented positive and statistically significant association to vertical integration. Uncertainty about orange prices contributed to a vertical integration trend, as expected, because internal production is better to deal with prices changes. However, property rights also influenced the vertical integration trend but in the opposite direction. It was expected that better quality institutional environment would induce more contracts in the sector. This is a longitudinal analysis and these variables controlled potential changes in competitive environment, which serve as parameter shifts for governance curves. Finally, the control variable of land price has no statistically significant association to the dependent variable, while competition with sugarcane production presents negative and statistically significant association in terms of area used to produce sugarcane. This evidence, on one hand, indicates that sugarcane production is raising the opportunity cost for the land, which is reducing the expansion of orange production by juiceprocessing firms. On the other hand, value tenancy for sugarcane production is not a threat to juice-processing firms, i.e. citrus growers are not attracted to change their production to sugarcane.

6. Conclusions and implications The paper aimed to analyze the partial backward vertical integration trend of juice-processing firms into orange production after 1993, in order to highlight the main determinants of this path. In this endeavor, beyond traditional determinants of vertical integration in TCE literature, such as asset specificity, we also included institutional features like antitrust intervention and market power as influencing factors in makeor-buy decisions. In fact, the discussion can approach both theoretical and practical implications, as follow. 6.1 Implications for theory building It is not simple to investigate power as determinant of vertical integration, since it is an ambiguous matter in economic theory and it has a limited role in TCE models. Power, in TCE literature, is present only in special case situations, whereas economizing on transaction costs is much more recurrent and it is the starting point for the analysis. This study advances in the exploration of those special case situations in which power is relevant, more specifically, showing that the market dynamics can change conditions to build safeguards to relationship-specific investments. The barrier to entry in juice-processing activity, accompanied to barriers International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

to grower’s forward vertical integration, are the mechanisms that determined these unpredictable conditions to market power. Investigations into which are the special case situations where power is relevant are rare and this paper is an attempt in this direction. Theory suggests that: (1) when contracting process take place under voluntary, knowledgeable, and farsighted situations, there is no room for market power; and (2) economizing is more frequent than strategizing in real world transactions. Nonetheless, theory also assumes that bounded rationality and the consequent incomplete contracting for the transactions. Thus, in presence of incomplete contracts, there is some limited foreseen situations when relationship specific investments are made, which make room for market power. Firms can actively introduce changes in the environment, affecting the bilateral dependence from those specific assets. Thus, this study also contributes to providing a dynamic view of the process involved in the vertical integration decision, especially in how changes in relevant factors across time can help to understand organizational evolution. Another advancement in this study is the explicit consideration of parameter shifts in the econometric model. This procedure is not frequent in empirical investigation of vertical integration under TCE lens, since investigations use cross-sectional data – e.g. Masten et al. (1991) and Monteverde and Teece (1982). The inclusion of parameter shifts proved to be fruitful, when antitrust intervention showed important role in the formation of the new arrangement for transactions. In this sense, cross-sectional investigations using asset specificity can produce inaccurate estimations, as vertical integration trend are also affected by environmental dynamic dimensions; i.e. keeping asset specificity constant, the increasing or decreasing vertical integration can be related to changes in contract law, for instance. We emphasize that this balance between economizing and strategizing is the core element in antitrust analysis. Failures in addressing this issue can lead to mistakes in the identification of the anticompetitive conduct and/or its remedy, which practical implications is explored in the following subsection. After analyzing those factors that determined vertical integration choice between 1993 and 2007 in the Brazilian juice sector, econometric evidence showed that economizing on transaction costs, antitrust intervention and power are relevant factors to explain vertical integration in this context. The results showed that economizing on transaction costs is not necessarily the starting point to organizational analysis, therefore, it is critical to take into account the balance between economizing and strategizing. 6.2 Implications for practice The fact of distinctions from efficiency and strategic purposes are blurry in this situation creates challenges for antitrust analysis. As it was difficult to anticipate indirect effects from the antitrust decision, there is a risk to produce more inefficiency after interventions than the previous situation. This problem occurred in the case of orange juice sector, as the antitrust intervention contributed to increasing vertical integration of juice-processing firms. The consequence was the adoption of more expensive arrangement for transactions, in term of governance costs, and generated no beneficial results to citrus growers. In fact, the intervention can harm growers’ conditions when the consequences to firm strategy are taking into account. Therefore, it is possible to indicate a relationship between hypotheses 2 and 3, since antitrust intervention not only increased transaction costs, but also was not capable of avoiding the juice-processor’s bargaining power. The results also indicate practical implications for firm competitive strategy. The role of vertical integration to bargaining power in the buyer-supplier relationship is explored by competitive strategy model, more specifically the five competitive forces (Porter, 1980). Thus, in the case of orange juice sector, after no antitrust office disregarded any actions against vertical integration path in the sector, juice-processing firms increased the partial backward vertical integration. Despite the econometric analysis cannot evaluate the financial results from the vertical integration strategy, the five competitive forces model suggests that this kind of strategy create conditions to price or concessions renegotiations, which has positive impacts on juice-processing firms performance. International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

6.3 Limitations First, from the empirical evidence it is possible to indicate the concomitant role of market power and transaction costs in the explanation of vertical integration decision. However, it is not possible to estimate the size of the effect these different mechanisms, and, therefore, it is not possible to conclude about the welfare effects of the increasing vertical integration in the sector. Second, another limitation of this study is related to the dependent variable, as data on new farms owned by juice-processing firms was obtained from public information in press or firm’s reports. Thus, it is possible that data underestimate the amount of internal fruit productions of juice firms. Finally, third, constraints for forward vertical integration of citrus growers in 1993 is temporally close to antitrust intervention in 1994. This proximity can influence the interpretation of the results of the dummy variable labeled as ANTIT, because constraints for forward vertical integration of citrus growers can increase the bargaining power of juice-processing firms through backward vertical integration.

References Azevedo, P.F. 1996. Integração vertical e barganha. University of São Paulo, São Paulo, Brazil. Cameron, A.C. and P.K. Trivedi. 2009. Mircoeconometrics using stata. Stata Press, College Station, TX, USA. Coase, R.H. 1937. The nature of the firm. Economica 4(16): 386-405. Coase, R.H. 1972. Industrial organization: a proposal for research. In: Economic research: retrospect and prospect, volume 3, edited by V.R. Fuchs. National Bureau of Economic Research, New York, NY, USA. Conselho Administrativo de Defesa Econômiccade (CADE). 2014. Voto do conselheiro Ricardo Machado Ruiz, ato de concentração no 08012.0036065/2012-21. CADE, Brasília, Brazil. Dorward, A. 2001. The effects of transaction costs, power and risk on contractual arrangements: a conceptual framework for quantitative analysis. Journal of Agricultural Economics 52(2): 59-73. Hasse, G. 1987. A laranja no Brasil: a história da agroindústria citrícola Brasileira, dos quintais coloniais às fábricas exportadora de suco do século XX. Duprat e Lope Propaganda, São Paulo, Brazil. Joskow, P.L. 1987. Contract duration and relationship-specific investment: empirical evidence from coal markets. American Economic Review 77(1): 168-185. Joskow, P.L. 2002. Transaction cost economics, antitrust rules, and remedies. Journal of Law, Economics and Organization 18(1): 95-116. Joskow, P. 2005. Vertical integration. In: Handbook of new institutional economics, edited by C. Ménard and M.M. Shirley. Springer, Dordrecht, the Netherlands, pp. 319-348. Klein, B., R.G. Crawford and A.A. Alchian. 1978. Vertical integration, appropriable rents, and the competitive contracting process. Journal of Law and Economics 21(2): 297-326. Lafontaine, F. and M. Slade. 2008. Exclusive contracts and vertical restraints: empirical evidence and public policy. In: Handbook of antitrust economics, edited by P. Buccirossi. MIT Press, Cambridge, UK. Maddala, G.S. 1983. Limited-dependent and qualitative variables in econometrics. Cambridge University Press, Cambridge, UK. Maia, M.L. 1996. Citricultura paulista: evolução, estrutura e acordo de preços. IEA, São Paulo, Brazil. Masten, S.E. 1993. Transaction costs, mistakes, and performance: assessing the importance of governance. Managerial and Decision Economics 14(2): 119-129. Masten, S.E. 1996. Empirical research in transaction costs economics: challenges, progress, directions. In: Transaction cost economics and beyond, edited by J. Groenewegen. Kluwer Academic Publishers, Norwell, MA, USA. Masten, S.E., J.W. Meehan and E.A. Snyder. 1991. The costs of organization. Journal of Law, Economics, and Organization 7(1): 1-25. Masten, S.E. and S. Saussier. 2000. Econometrics of contracts: an assessment of developments in the empirical literature on contracting. Revue d’Economie Industrielle 92: 215-236. Monteverde, K. and D.J. Teece. 1982. Supplier switching costs and vertical integration in the automobile industry. The Bell Journal of Economics 13(1): 206-213. Neves, M.F., and Lopes, F.F. 2005. Estratégias para a Laranja no Brasil. São Paulo: Atlas. International Food and Agribusiness Management Review

Ito and Zylbersztajn

Volume 21, Issue 1, 2018

Porter, M.E. 1980. Competitive Strategy. Free Press, New York, NY, USA. Riordan, M.H. 1998. Anticompetitive vertical integration by a dominant firm. American Economic Review 88(5): 1232-1248. Sabes, J.J.S. 2010. medidas de concentração no processamento de laranja no estado de São Paulo, no período de 2000/01 a 2007/08. Available at: http://tinyurl.com/huklru6. Shelanski, H.A. and P.G. Klein. 1995. Empirical research in transaction cost economics : a review and assessment. Journal of Law, Economics, and Organization 11(2): 335-361. Tirole, J. 1988. The theory of industrial organization. MIT Press, Cambridge, UK. Williamson, O.E. 1971. Vertical integration of production: market failure considerations. American Economic Review 61(2): 112-123. Williamson, O.E. 1975. Market and hierarchies: analysis and antitrust implications. Free Press, New York, NY, USA. Williamson, O.E. 1985. The economic institutions of capitalism. Free Press, New York, NY, USA. Williamson, O.E. 1991a. Comparative economic organization: the analysis of discrete structural alternatives. Administrative Science Quarterly 36(2): 269-296. Williamson, O.E. 1991b. Strategizing, economizing, and economic organization. Strategic Management Journal 12(S2): 75-94. Williamson, O.E. 1995. Hierarchies, markets and power in the economy: an economic perspective. Industrial and Corporate Change 4(1): 21-49. Zajac, J. and P. Olsen. 1993. From transaction cost to transactional value analysis: implications for the study of interorganizational strategies. Journal of Management Studies 30(1): 131-145.

International Food and Agribusiness Management Review

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2017.0018 Received: 13 February 2017 / Accepted: 21 August 2017

Information flow in the Sino-Brazilian beef trade RESEARCH ARTICLE Susanne

Knolla, Antonio

Domingos Padula b, Mariane Crespolini dos Santosc, Guilherme Pumid, Shudong Zhoue, Funing Zhonge, and Júlio Otávio Jardim Barcellosf

aResearcher,

Center of Agribusiness Studies – CEPAN, Universidade Federal do Rio Grande do Sul, Porto Alegre (UFRGS), Avenida Bento Gonçalves 7712, CEP 91540-000, Porto Alegre, RS, Brazil bProfessor,

School of Administration and Center of Agribusinnes Studies – CEPAN, Universidade Federal do Rio Grande do Sul, Porto Alegre (UFRGS), Ruat Washington Luiz 855, Centro Histórico 90010-460, Porto Alegre, RS, Brazil cPh.D.

Candidate, Institute of Economics, Universidade de Campinas (UNICAMP), Campinas, Avenida 31 de Março, 1001 Canadá-44, CEP 13424-305, Piracicaba, SP, Brazil

dProfessor,

Department of Pure and Applied Mathematics, Universidade Federal do Rio Grande do Sul, Porto Alegre (UFRGS), Bento Gonçalves Avenue 9500, Building 43-111, Agronomia 91509-900, Porto Alegre, RS, Brazil eProfessor,

College of Economics and Management, Nanjing Agricultural University, Weigang 1, Xuanwu District, Nanjing 210095, China P.R.

fProfessor,

Department of Animal Science (Zootecnia) and Center of Agribusinnes Studies – CEPAN, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Av. Bento Gonçalves 7712, CEP 91540-000, Porto Alegre, RS, Brazil

Abstract Considering the opportunities offered by the Chinese beef market and the fragilities in the Sino-Brazilian beef trade supply chain, this study aims to identify the main sources and channels used by Brazilian beef packers to obtain information on the Chinese market. The results reveal that the Brazilian beef packers and institutions within the export sector have little knowledge regarding the Chinese market. Neither the size nor the export experience of the beef packing firms and the foreign direct investment seem to significantly influence the quality of the knowledge they hold on the Chinese beef market. The sector has neither an integrated database containing the essential information on the Chinese market nor a unified traceability system in place that could facilitate the information flow among the agents within the beef supply chain. Consequently, firms need to dedicate substantial management resources (time, financial and human) to collect information from various sources. Keywords: Sino-Brazilian beef trade, Chinese beef market, Brazilian beef packers, market information, market knowledge JEL code: M16 Corresponding author: antonio.padula@ufrgs.br © 2017 Knoll et al.

Knoll et al.

Volume 21, Issue 1, 2018

1. Introduction Since the 1990s, food safety has become an important factor for an increasingly urbanized and socially diversified China. Food safety scandals have become the unwanted consequences of a highly fragmented food production and supply chain (Lam et al., 2013). Despite government efforts to minimize such events (Waldron et al., 2010), the Chinese food safety regulatory framework and state operated certification and controlling bodies have proven to be of limited effect (Linhai et al., 2013). In many cases, Chinese consumers trust foreign food safety controlling bodies more than the Chinese inspection systems, which are perceived as less reliable (Bloomberg Businessweek, 2016, 2017a,b; Sun et al., 2014). Consequently, imported beef is an attractive alternative for Chinese consumers, not only due to the diversification the new products offer and the prestige related to their consumption, but also because they are perceived as a healthier alternative to pork (Longworth et al., 2001). The need to meet an estimated demand of 4.5 kg/per capita/ year in China (Youyang8, 2015) has led to a strong dependence on imported beef (legally or illegally), due to the impossibility of China attaining production self-sufficiency (Waldron et al., 2015). In 2017, Chinese beef imports are expected to reach 950 thousand metric tons, up 17% from 2016. The major beef suppliers to China in 2016 were Brazil (29%); Uruguay (27%); Australia (19%); New Zealand (12%) and Argentina (9%) (MICA, 2017). It is worth presenting a picture of the meat smuggling activity in the mainland Chines market. China is the world’s third largest meat consumer but demand has outstripped domestic production, creating an opportunity for smugglers (Reuters, 2015). Although the import of beef into Hong Kong is legal, its transfer to mainland China is considered illegal by the Chinese authorities. In addition, a considerable amount of beef imported by Vietnam find its way across the border into China (Global Times, 2015). Up to two million tons of beef were smuggled into China in 2012 and 2013 (South China Morning Post, 2015). In June 2015, Chinese authorities seized 100,000 tons of frozen US meat from smugglers (Business Insider, 2015). Smuggling has a detrimental effect on the legitimate export traders from countries such as Brazil, Uruguay, Australia, New Zealand and Argentina. Due to the import duties in mainland China, smuggled beef can be sold 30 to 60% cheaper that legitimately imported beef (Reuters, 2015). Despite the problem of smuggling, the Chinese beef market presents considerable opportunities for big beef exporting countries, such as Brazil, to expand their markets and revenues. In 2016 Brazil exported 165,619 tons of beef to China (Beefcentral, 2017). The trade between Brazil and China is expected to grow in line with their respective gross domestic products (GDPs) (Squartini and Garlaschelli, 2013). Nevertheless, as shown by Tinbergen (1962) in his ‘gravity model’, although the trade flow between two countries is proportional their GDPs, as in the case for China and Brazil, it is inversely proportional to the distance between them (almost 17,000 km) and the cultural and linguistic differences (The Economist, 2016). The gravity model is currently considered a good predictor of trade relationships (Dhingra, 2013; Ghemawat, 2001; Goh et al., 2013). This means that Brazil will face strong competition from Australia and New Zealand, which are geographically closer to China and are strongly associated with beef safety, quality and taste (MLA, 2016). Thus, Brazilian producers face the dual challenge of meeting the high demand presented by the Chinese beef market at a competitive cost and, simultaneously, ensuring quality and product safety levels that conform to the specific preferences of Chinese consumers. To highlight the huge market potential for the of Brazilian beef, it is worth noting that the 165,619 tons Brazil exported to Mainland China represent approximately 30% of China’s total beef imports in 2016 (Beefcentral, 2017; MDIC, 2016). Notably, although this trade only started in June 2015, this volume is already among the largest international beef flows. After conducting systemic mapping of the Sino-Brazilian beef supply chain, and its shortcomings, Knoll et al. (2017) found that the Brazilian traceability system is based on fiscal/ commercial documents, rather than product follow-up, the information flow between the stakeholders is dysfunctional, and that opportunistic business behavior dominate the whole supply chain. Research also reveals that aligning the information flow and the production and trade practices among the different members is a basic necessity for a supply chain to be effective and competitive (Cooper et al., 1997; Jie et al., 2013). A well International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

organized and well managed supply chain encourages cooperation and trust-based relationships among its members and improves the effectiveness of the whole supply chain in identifying and responding to market opportunities (Ding et al., 2014; Jie et al., 2013; Lambert et al., 1998; Simatupang et al., 2002). Considering these aspects, the present research focuses on the specific case of what defines information availability and actual market knowledge of the Brazilian beef packers regarding the Chinese market. Furthermore, considering the opportunities offered by the Chinese beef market and the fragility identified in the SinoBrazilian beef trade supply chain (Knoll et al., 2017), this research explores the experience of Brazilian beef packers that export or are applying for accreditation to export to Mainland China. This research aims to identify their information sources and channels employed in organizing, governing and managing their supply chain. The in-depth analysis of the information flow and its peculiarities provided by this research represents a first step towards improving chain coordination and governance, which, in turn, contributes to ensure a safer, healthier and more attractive business environment for all the stakeholders within the chain, including the Chinese consumer. By evaluating the information sources, content and flow, it is possible to assess the extent to which the absence of a unified traceability system (Knoll et al., 2017) impacts the current availability of information in the chain. Based on this assessment, more specific suggestions regarding the construction of an information sharing and processing system can be made, which may be adapted to the Sino-Brazilian situation. The article is organized as follows: Section 2 presents a description of the theoretical framework used to support the methodology; Section 3 describes the methodological approach adopted; Section 4 includes a detailed discussion of the results obtained from the applied questionnaire and interviews; Section 5 contains the concluding remarks and highlights the theoretical and managerial implications, while section 6 points out the study limitations.

2. Theoretical framework In building the theoretical foundation for this research, firstly, the role of information and knowledge sharing on supply chain organization and management is discussed and analyzed. Then, the relationship between firm size and exporting experience is explored. The aim is to identify and evaluate the different information sources and the knowledge currently held by the Brazilian beef exporters regarding the Chinese beef market. 2.1 Supply chain management and information flow An efficient supply chain requires the integration of organizational units involved in the flow of products, information and finance in order to attend consumer demands and add value to the chain shareholders (Lambert et al., 1998; Stadtler, 2005). This is particularly challenging when dealing with food supply chains, considering the associated product and process specificities (perishability, contamination, food safety and security, shelf life, etc. (Van Donk et al., 2008)). Food supply chain structuring and management analysis has grown in importance due to its ability to track vulnerabilities and risks that endanger food safety (Ding et al., 2014; Van der Vorst and Beulens, 2002) and interdependence in business transactions, which can affect every stakeholder in the chain (Gereffi et al., 1994). Certain practices, namely, strategic partnerships with suppliers, continuous process flow, outsourcing, cycle time compression, quality certification, customer relationships and the use of inter-organizational systems such as electronic data interchange and the internet can help prevent pitfalls in the supply chain, leading to more effective chain management (Alvarado and Kotzab, 2001; Lambert et al., 1998; Tan, 2001). A considerable amount of literature highlights the importance of information accessibility, quality and sharing between the different stakeholders in a supply chain (Ding et al., 2014; Jie, et al., 2013; Min and Mentzer, 2004; Prajogo and Olhager, 2012; Tan et al., 2002). Information sharing is essential for company success, especially when it comes to transnational operations (Bartlett and Ghoshal, 1989) and positively impacts a firmâ&#x20AC;&#x2122;s operational performance (Frohlich and Westbrook, 2001; Jie et al., 2013; Prajogo and Olhager, 2012; Zhou, 2007). On the other hand, from a supply chain point of view, a lack of coordination between stakeholders can yield negative consequences such as higher inventory and transportation costs, longer delivery times, higher levels of product loss, customer service inefficiency, International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

and imbalance between supply and demand forecasting, etc. (Akerlof, 1970; Lambert et al., 1998; Lee et al., 1997; Simatupang et al., 2002). Information is a strategic and often costly asset. Thus it is crucial for firms to decide whether or not to acquire expensive information and, if so, what kind of information is needed (Fu and Zhu, 2010). Nowadays, internet and other web-based technologies have a positive impact on the maximization of such demands (Gimenez and Sierra, 2013). When no integrated electronic information exchange mechanism is available, one of the most essential knowledge exchange mechanisms between firms is partnering among employees. However, this is less likely to happen among geographically distant firms since, generally, distance has a negative impact on information flow (Morosini et al., 1998). Also, information exchange among geographically distant places presents additional hurdles such as different time zones and long transmission channels, which, altogether, have a negative impact on the amount and quality of information exchanged. However, distance can be a motivator for the development of new solutions to solve information transfer issues. Thus, distance does not always inhibit the smooth flow of knowledge. Nevertheless, it certainly impacts the effectiveness of certain information transfer mechanisms (Ambos and Ambos, 2009; Tihanyi et al., 2005). 2.2 Firm size, industry structure and internationalization Several studies have indicated the existence of a positive relationship between firm size and internationalization (Baird et al., 1994; Calof, 1993). According to Bonaccorsi (1992), larger firms have a competitive advantage when it comes to dealing with foreign markets, mainly due to their greater managerial and financial resources and access to information. Small companies tend to perform less well when it comes to internationalization, mainly due to the insufficiency of financial and human resources to proceed with the internationalization process or to acquire knowledge and understand the targeted foreign market (Etemad, 2004; Knight and Kim, 2009; Pangarkar, 2008). Additionally, Julien and Ramangalahy (2003) found that small firms seem to have difficulty with some core competences, such as distribution, pricing, and monitoring foreign markets, etc. On the other hand, the structure of small and medium-sized companies allows for greater flexibility and speedier decision making (Cretoiu, 2010), due to the absence of internal bureaucratic hurdles, processes and protocols (Knight and Kim, 2009). This can be seen as a competitive advantage in the context of rapidly changing market conditions, like those found in China. Another element that can influence the internationalization process is the industry structure (Fernhaber et al., 2007; Gao et al., 2010), which is characterized by the number and concentration of firms producing close substitute products in a certain market, the level of product differentiation and the intensity of competition among those firms (Porter, 1980). Industry concentration is measured by calculating the sales volumes or employment accounted for by the largest four or eight firms in the industry. Firm concentration ratio is an indicator of the relative power of firms in an industry (Fernhaber et al., 2007). In a concentrated industry, dominant firms are able to compete based on advantages achieved through high economies of scale (Besanko et al., 1996). Oviatt and McDougall (1994) found that the industrial structure influences (1) the organization of international transactions; (2) the reliance on alternative or hybrid governance structures (mode of internationalization (Bucley and Casson, 1998)); (3) the creation of foreign location advantages (industry concentration and industry evolution not only allow for the formation of new international ventures, but also add to their foreign location advantage (Fernhaber et al., 2007)); and (4) the control over unique resources (knowledge appropriability, for example (Hoenen and Hansen, 2009)). The mode of internationalization can vary from direct exporting, licensing, joint venture to equity investment (Foreign Direct Investment â&#x20AC;&#x201C; FDI (Buckerly and Casson, 1998)). Direct exporting is the quickest for firms to enter foreign markets and involves the least risk. Licensing and FDI require more organizational resources (equity and marketing the product abroad) and involve higher risks than other modes of entering foreign markets (Gao et al., 2010; Yuan et al., 2016). Emerging economies are important destinations of outward FDI (Bandeira-de-Mello et al., 2016; Goh et al., 2013; Yuan et al., 2016). Lu et al. (2010) studied the internationalization of firms in an emerging economy (China) and identified that a firmâ&#x20AC;&#x2122;s ability to coordinate, International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

recombine, and allocate organizational resources to meet the different requirements of foreign markets influences its international performance. Firms that made direct investment in China increased their export to that country (Liu et al., 2001). These findings corroborate those reported by Pfaffermayr (1994), who studied the internationalization process of firms in rich countries. Large enterprises have the organizational and financial resources to explore opportunities in foreign markets. The ability of a firm to collect, absorb, and integrate information to understand customer needs, market opportunities and regulatory requirements in a foreign economy is fundamental to achieving superior international performance and building a sustainable competitive advantage. The development of managerial ties through FDI can also help firms in emerging economies overcome the liabilities associated with foreignness and newness in host countries (Lu et al., 2010; Powel and Rhee, 2016). 2.3 Internationalization and information flow The market itself is a network of relationships wherein firms are linked to each other in a variety of ways (Johanson and Vahlne, 2009). Thus, being part of a relevant network opens possibilities, through the development of knowledge and trust among the different network partners, which can provide the basis for a solid partnership in operations involving a sustainable flow of business. Internationalization generally has a positive effect on a firm’s performance and its access to information (Barkema et al., 1996; Li, 1995). Cunningham and Homse (1986) argue that during a firm’s internationalization process, managers in both home and foreign markets develop valuable contacts based on social relations and routine communication, which promotes fruitful information flows. This not only allows for the construction of knowledge regarding their respective markets and processes, but also promotes trust, a valuable asset that may lead to greater commitment among the partners (Hunt and Morgan, 1994). Common ties can increase a firm’s performance and productivity while facilitating closer cooperation within a partnership network. The closer the relations a firm maintains with its foreign business partners (suppliers, customers, distributors), the more likely it is that the firm will have stable relations with them (Lu et al., 2010). The stakeholders may eventually develop a certain type of mutual knowledge that would present opportunities unavailable to those who do not cooperate to the same extent (Zajac and Olsen, 1993). Direct networking may not be the only option though. Intermediaries can also facilitate the flow of goods and information (Root, 1987). Often such intermediaries are specialized service providers, which, for the exporter company, serve as an outsourced export department and information source regarding a foreign market (Peng et al., 2008). The literature also discusses how smaller firms, in particular, tend to employ third party market intelligence to facilitate access to a foreign market (Hessels and Terjesen, 2010; Terjesen et al., 2008). The situations in which a firm contracts an export intermediary mostly depend on the firm’s ability to handle the foreign market’s size, the expected financial risk and the cultural difference assessed by the local firm’s management (Felbermayr and Jung, 2011). It is also relevant to note that, although intermediaries can facilitate trade, employing them may lead the exporter to lose control over its exportation processes (Blomstermo et al., 2006). Finally, enterprises can count on government agencies to support them to get information and knowledge about foreign market conditions, customer needs and regulatory requirements. It is a challenge for firms in an emerging economy to obtain information related to foreign markets. Government can have an important role in the provision of information concerning foreign markets. ‘By engaging in government export programs, firms can construct and build appropriate skills, routines, knowledge, and procedures in scanning and identifying useful information, thereby enhancing their information acquisition capability’ (Lu et al., 2010). To support the Brazilian enterprises in their internationalization process, the Federal Government created the National Export Promotion Agency (APEX in Portuguese).

International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

3. Research goals and method In this research it is assumed that knowledge of the targeted foreign market improves the export supply chain performance and leads to a more effective attendance of the foreign customer’s requirements (Katsikeas, 1994; Lu et al., 2010; Pfaffermayr, 1994). Thus, as stated in the paper´s introduction (Section 1), the research aims are to identify the information sources most commonly used by the Brazilian beef exporters to China (direct, network, intermediary-based or FDI), to explore the relationship between firm size and the kind of information sources, and to assess the extent to which it affects the firm’s market knowledge. Therefore, the following questions are pertinent: ■■ How do Brazilian beef packers obtain information regarding the Chinese beef market? ■■ To what extent are Brazilian beef packers aware of the recent tendencies in the Chinese market related to choice of cuts, market segmentation and business behavior? ■■ Is the level of knowledge held by Brazilian beef packers related to company size or the level of internationalization? 3.1 Methodological design and its justification Firstly, the choice of a specific sample to answer the research questions is explained. After, the methodological approach applied to obtain the answers from the interviewees is described. Interviewees directly involved in the Sino-Brazilian beef trade with knowledge of specific consumer needs and the legal requirements of the Chinese market were sought (Knoll et al., 2017). Thus, in Brazil, the beef supply chain stakeholders interviewed can be assumed to be those most directly affected by tendencies within the Chinese beef market. Consequently, within the Brazilian section of the supply chain, the Brazilian beef packers are expected to have the broadest and most in-depth information and knowledge regarding the Chinese market. Although well-known for the decentralized nature of its cattle raising and beef processing sector (Jank et al., 2001), the scale, productive capacity and the degree of centralization in the Brazilian agro-export sector has grown in the recent years. The ongoing centralization is reflected in the fact that only four firms detain more than 60% of the whole beef slaughtering, processing and exporting business in Brazil (Vieira and Traill, 2008). Hence, although Brazil is a large beef exporter, only a few firms, owning several meat-packing plants, are relevant in the export sector. Market and export related decisions are not made at the level of the individual meat-packing plant, but at the central headquarters and within the export departments of the holding firms. Thus, we focus on getting information from the central headquarters of the holding firms instead of contacting individual meat-packing plants. The aim of the interviews was to get evidence from a specific target group of beef processing firms, namely, those with experience of exporting directly to Mainland China or to Hong Kong (which serves as an entrance to the Mainland Chinese market (Knoll et al., 2017; MLA, 2016; MICA, 2017; U.S. Meat Export Federation, 2014; Waldron et al., 2015)), those that are accredited, or applying for accreditation, to export to Mainland China. To get the largest possible sample of such firms, we sought the support of the Brazilian Beef Processors and Exporters Association (Associação Brasileira das Indústrias Exportadoras de Carnes (ABIEC; http://www.abiec.com.br)). The approach adopted in the data collection is illustrated in Figure 1. ABIEC is currently the most widely recognized export association within the Brazilian beef sector, consisting of 26 member companies and representing around 96 (or 39%) of the total number of Brazilian beef slaughterhouses certified by the Brazilian Federal Inspection (Serviço de Inspeção Federal (SIF) – Brazilian Ministry of Agriculture Livestock and Supply). SIF certification is mandatory for every slaughterhouse seeking to distribute its products throughout the Brazilian territory, or to export to destinations with basic requirements such as China, Hong Kong, Uruguay, Argentina, Vietnam, Peru, Venezuela, Israel, Egypt, etc. However, besides the SIF certification, accreditation for export to Mainland China requires an additional set of requirements, part of an agreement between Brazilian and Chinese State authorities (Knoll et al., 2017). International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Contacting associations specifically formed for export purposes

ABIEC Number of associated companies: 26 Number of slaughterhouses associated through the companies: 96 Average number of slaughterhouses per company: 3.69 From all Brazilian slaughterhouses with SIF, 39% are members of ABIEC From all 16 accredited beef exporting slaughterhouses , 15 are members of ABIEC From all 8 companies which have plants accredited to China, 7 are members of ABIEC 13 company representatives replied

Response rate 50%. Too small number of responses!

6 interviews with company representatives at SIAL Shanghai

Figure 1. Methodological design. ABIEC = Brazilian Beef Processors and Exporters Association (Associação Brasileira das Indústrias Exportadoras de Carnes); SIF = Brazilian Federal Inspection (Serviço de Inspeção Federal); SIAL = Salon International de l’alimentation. Of the 96 slaughterhouses associated to the ABIEC, only 15 are accredited to export to Mainland China, while only one such slaughterhouse is not a member of the ABIEC. Currently, all the slaughterhouses associated to the ABIEC are accredited to export to Hong Kong, since accreditation only requires SIF certification, and at least 23 are in the process of applying for accreditation to supply Mainland China. To provide the most complete and accurate picture of the research object, and considering the nature of the sector under analysis, this study is based on exploratory and descriptive data analysis (Miles et al., 2013). Firstly, a questionnaire was prepared and delivered to the target group of Brazilian beef processing firms. As only a small number of responses were received from that group, interviews were held to achieve a workable response rate and a more holistic understanding of the subject. 3.2 Development of the questionnaire The questionnaire was designed to provide analyzable data regarding the research goal (Miles et al., 2013). The questions were reviewed and discussed by our team of researchers and five international industry professionals (with at least five years of experience in beef trading with China). The questions were chosen so that the responses could be used as proxies to achieve the research goals. The questionnaire was designed to take no longer than ten minutes to answer and was distributed and filled by an online survey tool (Surveymonkey; https://www.surveymonkey.com). The link to the online platform, with the research explanation, was forwarded by ABIEC to all its members, with a special request for cooperation. Responses were expected to come from either the export or administrative department, or from the chief executive officer’s office of the 26 firms associated to ABIEC. It was hoped to achieve a statistically analyzable census (response rate over 80%) due to the direct involvement of ABIEC in forwarding the questionnaires. However, only 13 companies responded to the questionnaire. Of the respondents, 10 had at least one slaughterhouse in the process of applying for accreditation to export to China and only one company had a single slaughterhouses accredited to export to Mainland China. Two companies had not applied for accreditation to export to mainland China, but exported to Hong Kong. Responses were received between April 3rd and May 11th, 2016. International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

The questionnaire consisted of three main parts, labeled Part A, B and C. Part A seeks to collect information regarding the firm’s size and export profile; Part B seeks to assess the information held by the Brazilian firms regarding the Chinese beef industry; Part C aims to measure the firm’s current knowledge regarding the Chinese beef market (Questionnaire is attached as Supplementary methods S1 (in Portuguese) and Supplementary methods S2 (Part C – in English). 3.3 Development of interviews Due to the small number of responses to the questionnaire (13), it was decided to boost the data by conducting short interviews with slaughterhouse representatives that had not answered the questionnaire. To ensure the right population of respondents, members of the research team attended the Salon International de l’alimentation, or SIAL, in Shanghai 2016 (May 5th-7th, 2016), the biggest networking event and food trade show in Asia. Fourteen companies that are members of ABIEC attended the SIAL Shanghai 2016. When contacting each one of them, a general introduction to the research was given, and the representatives were asked whether the company had answered the questionnaire sent through ABIEC. If the answer was negative, the representative was invited to participate in an interview. In the end, six interviews were conducted. Due to the business nature of the event, and the busy schedules of the representatives, the time available was very limited and, on average, the interviews (conducted in Portuguese) lasted five minutes per company. Initially, applying the Delphi method (Dalkey and Helmer, 1963) was considered, however, due to the limited time available and the business setting, it was thought unfeasible. Instead, a less than ideal, but nonetheless meaningful approach was adopted, whereby the interviews were prepared based on the research goal (Denzin and Lincoln, 2011) and the dynamic circumstances in which they were conducted. The semistructured interviews were open-ended, which allowed the interviewers to ask follow-up questions. Out of respect for specific requests made by some interviewees and the trade fair’s sensitive business environment, the interviews were not recorded, but rather notes were taken during the conversations. The interviewees represent the largest Brazilian slaughterhouses and beef exporters, as shown in Table 2. When interpreting the interviews, the approach proposed by Livesey (2006) was adopted. Two questions were asked regarding part B of the questionnaire (in free translation): ■■ A. What do you know about the Chinese market that makes it a desirable choice for your firm? ■■ B. What are your main source of information regarding the Chinese market? Based on the information contained on their business cards and the company homepage, company profiles were elaborated (corresponding to part A of the questionnaire). However, there was no opportunity to assess the level of knowledge held regarding the Chinese market (corresponding to part C of the questionnaire), due to the time limitation and the sensitive nature of the situation.

4. Results and discussion Altogether, of the 26 firms associated to ABIEC, 13 completed the questionnaires and 6 accepted to be interviewed. The profile of the questionnaire respondents and their firms is presented in Table 1. The profile of the interviewees and their firms is presented in Table 2. 4.1 General profile of the companies In Brazil, particularly in sectors where natural resources are dominant, larger firms are more likely to be involved in exportation than small and medium-sized companies (Fleury et al., 2007), as suggested by Besanko et al. (1996), Fernhaber et al. (2007), Gao et al. (2010) and Oviatt and McDougall (1994). No standard definition of firm size was found regarding the slaughterhouse sector in Brazil. For the purposes of the present research, firms with a capacity to slaughter between 500 and 800 animals per day were considered International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Table 1. Profile of respondents to the questionnaire (position/company size/accreditation). Responsibility in company

Number of packers

Size of biggest packer Joint investments/ Number Accredited (animal per day) headquarters of export and/or in outside of Brazil? destinations process

Between 500 and 800

In process

Between 500 and 800

Accredited

Between 500 and 800

4 or more Between 800 and 2,000 Yes

In process

1 3 4 or more 1 2 3 4 or more

No Yes No No Yes No Yes

3 2 2 5 3 4 5

Quest 12 Export manager 4 or more Between 800 and 2,000 Yes Quest 13 Financial and office 4 or more More than 2,000 Yes administrator

1 5

In process In process In process In process In process In process Accredited and in process In process No

Quest 1

Federal agricultural inspector Quest 2 Financial and office administrator Quest 3 Quality guarantee manager Quest 4 Quality guarantee manager Quest 5 Export manager Quest 6 Export manager Quest 7 Export manager Quest 8 Export manager Quest 9 Production manager Quest 10 General manager Quest 11 Export manager

Between 500 and 800 Between 800 and 2,000 More than 2,000 Between 500 and 800 Between 500 and 800 Between 500 and 800 Between 800 and 2,000

Table 2. Profile of the interviewees (position/company size/accreditation). Responsibility in company

Number Size of biggest packer Joint investments/ Number Accredited of packers (animal per day) headquarters of export and/or in outside of Brazil? destinations process

Int 1 Export manager

4 or more

Over 2,000

Yes

Int 2 Exportation manager Int 3 Chief executive officer Int 4 Export manager

Over 2,000

Between 500 and 800

Over 2,000

Int 5 Director of international businesses Int 6 Export manager

Between 800 and 2,000 Yes

Accredited and in process Accredited and in process Accredited and in process Accredited and in process In process

Over 2,000

In process

large and those able to slaughter over 800, very large. The four biggest slaughterhouses presented in Table 1 and 2 were among the 32 biggest Brazilian exporters in 2015 (Exame, 2016). All the companies in the research can be regarded as either large or very large (Table 1 and 2), and most companies own more than one packer. The sector is quite centralized, with only four firms detaining more than 60% of the total beef International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

slaughtering, processing and exporting business in Brazil (Vieira and Traill, 2008). Hence, although Brazil is a major beef exporter, only a few firms, each owning several packers, are relevant in the export sector. About half of the companies have international joint ventures. This seems to be a high percentage of FDI, given that in Brazil FDI represents only 1% of the gross fixed capital formation, whereas the world average is 8.3% (United Nations Conference on Trade and Development, 2016). FDI is also one of the most commonly adopted ways for US food-processing companies to enter foreign markets, the strategy being to acquire a maximum share of 10% in the external markets of business partners (Bolling and Somwaru, 2001). This is a particularly advantageous and opportune way of overcoming trade barriers and to obtain first-hand information from the external market (Lu et al., 2010; Pfafermayr, 1994). This research reveals that among the Brazilian beef packing companies that have international joint ventures, the average number of countries to which the companies export is 3.7, while among the companies that do not have joint ventures, the number is 2.6. Thus, internationalization via FDI seems to be an efficient gate opener for the general internationalization of Brazilian beef packers, as suggested by Barkema et al. (1996) and Li (1995). However, it is important to note that different slaughterhouses might have different roles in relation to exportation, depending on the overall strategy of the holding company. 4.2 Information sources The research revealed that Brazilian slaughterhouses have various means for gathering information on the Chinese market. The diversity of information sources may be due to the absence of a clear and unified traceability system (Knoll et al., 2017). Thus, stakeholders in the meat packing sector need to be resourceful to obtain valuable and current information on the Chinese markets, making it a crucial competitive factor. The results from the questionnaires (Supplementary methods S2– questionnaire responses) show that for the slaughterhouses, traders (also known as intermediaries) are the primary source information. Although such sources may provide the most up-to-date information, there is also a greater likelihood of opportunism, especially when the Brazilian slaughterhouse does not possess sufficient and trustworthy background information on the foreign market. This can be a problem, as reflected in the comments from Int. 3, who states: ‘we receive a lot of information from traders, much more than we can process and verify its trustworthiness.’ Int. 4 adds, ‘we receive a lot of requests via email from Chinese. We never know exactly who we are dealing with, but we take the risk’. Thus, it seems information from the channels most used by Brazilian beef packing firms might not be completely reliable. However, the lack of reliable information from other sources (or the lack of investment to obtain it) induces firms to take the information from potentially unreliable sources into consideration, even when it presents risk. Information is a strategic and often costly asset. Thus, it is crucial for firms to decide whether to acquire information, what kind of information is needed and what costs are related to its query (Fu and Zhu, 2010). The present research reveals how the Brazilian beef packers tend to invest in the acquisition of information. According to the survey results, private consultants are the second largest source of market information. Interestingly, the biggest beef packing firms decided to leave this option blank. As suggested by Besanko et al. (1996) and Fernhaber et al. (2007), it was assumed that some companies, especially the larger ones, consider it a strategically relevant issue and prefer not to share it with competitors. Similarly, the interviewees opted not to answer questions on this issue. This behavior highlights the competitive importance given to information on the Chinese market by the Brazilian beef packing firms and reveals the phenomenon known as ‘follow-the-leader’ (Hoenen and Hansen, 2009), which is commonly seen among companies operating in a concentrated industry. This concurs with the findings reported by Thomé and Vieira (2012), who found that, in the case of Brazilian beef packers’ knowledge on the Russian market, any flow of information between the firms is generally perceived as being against the interest of the company. The third most important source of information on the Chinese market is the trade associations and government and state information communication, as suggested by Lu et al. (2010). Int. 3 agrees with Int. 2 who states that ‘we constantly receive information from the ABIEC and sometimes from the Agricultural attaché’. International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Since 2001, the ABIEC together with the APEX have undertaken an aggressive marketing plan to establish a common brand: Brazilian Beef on a global level (Steiger, 2006). Their goal is to improve horizontal partnership in the processing and slaughtering sector to decrease opportunistic behavior and increase the competitiveness of industrial exports. However, one of the biggest challenges ABIEC faces is to prepare its members to develop supply chain sustainability through improved information sharing (Vieira and Traill, 2008). The Agricultural attaché’s responsibility, on the other hand, is more of a technical nature, focusing on safety matters and the plant accreditation process. Thus, by being a member of ABIEC, the Brazilian packing firms receive reports focused on market movements and tendencies, whereas they receive more technicalrelated information from the Official Federal Government database. Although the questionnaire respondents and interviewees consider both information sources trustworthy, the information quality regarding volume and timeliness is perceived as quite low. Interviewees also mentioned the importance of the FDI regarding their acquisition of market knowledge, as observed by Gao et al. (2010), Lu et al. (2010) and Pfafermayr (1994). Int. 6 states, ‘I am open to capital investments from Chinese companies to get better access to information on Chinese demand, and I am also looking forward to investing in their brands or companies’. The importance of investing in networking as a market strategy regarding information sourcing is also mentioned by Int. 5: ‘I help Chinese from different areas to do business in Brazil. Then they help me in China...I also propose cooperative product development to the Chinese.’ Thus, such collaboration between firms, either through direct investment or the exchange of favors, can create an essential knowledge exchange mechanism between firms (Cunningham and Homse, 1986). This kind of information sourcing is a managerial decision involving long-term planning and considerable resource investment that will probably pay off through the acquisition of unique first-hand information, which can ultimately provide unique opportunities. As suggested by Lu et al. (2010), the development of managerial ties through FDI can help firms overcome the liabilities of foreignness and newness in host countries. Experience also proves to be a valuable tool for foreign market information sourcing (Reuber and Fischer, 1997), as mentioned by Int. 1.: ‘we have precise information regarding the specificity of the Chinese market since we have one packer which is accredited to China and dedicates almost 100% of their production to the Chinese market.’ Barigozzi and Garlaschelli (2010) and Lu et al. (2010) state that the information flow on a specific market (in our case China) becomes more specific and broader if the firm has a wider international network and experience in the market. Thus, knowledge sharing processes can also be used from inside the organization (Becerra-Fernandez and Sabherwal, 2001; Hakanson and Nobel, 2001). However, according to the research findings, solid and reliable inside information seems to be the least frequent source of knowledge. This might be because the Chinese market still represents a new challenge for the Brazilian beef processors, thus only few firms (such as Int. 1), have first-hand, inside experience of the Chinese market. Last, but not least, according to the interviewees and the questionnaire respondents, news and media sources seem to have the least impact when it comes to reliable information on the Chinese market. This might be related to the generally poor perception of the trustworthiness of the media, but it might also indicate the need for the media to adopt a more unified media channel regarding exports, free from the influence from group interests and sensationalist approaches. In March 2017, the ‘Operation Weak Flesh’, led by the Brazilian Federal Police, revealed inspectors in the Brazilian beef-sanitation control system were being bribed. This event was intensively communicated by the Brazilian and international media and sparked trade bans ranging from China to Europe. Some of the biggest Brazilian meat exporters were among dozens of firms targeted by ‘Operation Weak Flesh’ (Reuters, 2017a). Meat industry representatives and the Brazilian government proved to the international beef importers that there were only a few isolated cases of wrongdoing and China and other importers lifted the restrictions and bans (Reuters, 2017b). This event provoked a review of the Brazilian beef traceability and sanitary control systems. On the other hand, the Chinese media and consumers are very critical of the sanitary control system in China (Bloomberg Businessweek, 2016, 2017a,b). Chinese consumers consider the control systems used by foreign exporters more reliable than those adopted by the domestic companies. This situation could be seen as an opportunity for Brazilian beef exporters to communicate to the Chinese consumers the reliability of the Brazilian sanitary control and product traceability International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

systems, which can ensure the quality and safety of the products in accordance with the specific preferences of the Chinese consumer. 4.3 Testing the knowledge about the Chinese market The knowledge of the Brazilian packing firms regarding the Chinese beef market was assessed using a ‘knowledge test’. At the same time, the existence of a possible relation between the level of knowledge, the firm’s size and the information sourcing was investigated. Eleven questions were formulated to cover the most relevant information expected to be known by companies aiming to export to the Chinese market. Five international industry professionals aided in the process of formulating and choosing the questions. Eriksson et al. (1997) highlight the importance of institutional knowledge as taxation, consumer channels and their particular features, rules, laws and business practices. Closs et al. (1997) and Moberg et al. (2002) emphasize the relevance of the accuracy and timeliness of knowledge. Berg’s (2001) suggestion regarding the value of listening to the opinion of experts was followed. Questions related to Chinese general beef import volumes, the most desired product types at the research moment and the ports most frequently used to enter Mainland China beef market. Multiple choice questions were applied (Questionnaire Part C – Supplementary methods S1), usually containing five alternatives from which the respondent was expected to mark only one. The alternatives usually consisted of two of absurd answers (out of context), two very likely answers and one ‘I do not know’. Responses are classified according to the following criteria: ■■ 80-100% of correctly answered questions: very well informed firm; ■■ 60-79% of correctly answered questions: well informed firm; ■■ 40-59% of correctly answered questions: reasonably informed firm; ■■ 20-39% of correctly answered questions: somewhat informed firm; ■■ 0-19% of correctly answered questions: poorly informed firm. These results represent the majority of Brazilian beef companies currently exporting or seeking to export to Mainland/mainland China. Overall, the highest rate of correct answers was six out of eleven (54.5% of the correct response rate), which correlates to a company that would be ‘reasonably informed’ on the Chinese market, obtained by Quest 5 (Where do the Chinese consumers buy beef?). Interestingly, this score was obtained by a company that only owns one medium-size beef-packer, with only three different export destinations, and no foreign investment. Although the firm exports to Hong Kong, which could serve as a proxy for greater in-depth knowledge of the Chinese market, its export share to Hong Kong is not outstanding. However, in relation to the other companies, their export departments use inside information on the Chinese market from all the sources listed in the questionnaire, namely, government and state information communication channels, trade associations, traders, private consultants, colleagues/friends and the media. Thus, it can be said that in the specific case of the Chinese market, the knowledge held by Brazilian firms and the whole Brazilian slaughterhouse sector is so limited that none has or wants to share valuable market information. This suggests that almost any kind of knowledge on the Chinese beef market is given strategic value by Brazilian companies, which regard such information as an important asset to increase their competitiveness in the Chinese beef market. As suggested by Fernaber et al. 2007, Lu et al. 2010, Luo et al. (2011), Powel and Rhee (2016), this point of view can be expected given the size of the Chinese market, its recent opening to Brazilian firms and the fast-changing market environment encountered in China. No clear evidence was found on a possible relation between foreign headquarters (FDI), number of export destinations and difference between firm sizes. However, it is interesting to note the companies that scored the highest in the knowledge test used the maximum or nearly the maximum number of information sources available to them. But, it should be noted that the firms that scored five points did not, in general, mention International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

whether they obtained information through private consultancy, and two out of three did not state the amount of information they receive from government organizations. Due to the small number of responses, it is not possible to confirm the extent to which private consultancy and government information sources are of strategic importance to firms. It is also interesting to note that, upon dividing the questionnaires according to processing capacity, the average number of correctly answered questions among the larger and very large companies (slaughter capacity over 800 animals per day) is 38.2%, while among smaller companies, the average is only 21.6%. Although they seem somewhat distant from each other, we cannot consider this proportion statistically different at any reasonable significance level, due to the small sample size and consequent low power of a two sample test. The number of joint ventures outside of Brazil (FDI) does not seem to influence the company scores in Part C of the questionnaire. The average percentage of correctly answered questions among companies that have international joint ventures is 28.8% against 27.3% for those that do not. The internationalization competences of firms tend to increase with the existence of a foreign headquarters, but may later stagnate if they fail to accompany developments in the foreign markets (Sull and Escobari, 2004). This might be the situation in the case of Brazilian slaughterhouses. Thus, it might be inferred that, while the foreign headquarters of company Nr. 2 are reasonably well-informed regarding the latest events in the Chinese market, the processing and communication of that information to the Brazilian headquarters, or perhaps the learning and knowledge transforming process in the Brazilian headquarters, has weak points. Similarly, the number of countries to which the company exports does not seem to affect the score: the average percentage of correctly answered questions among the companies that export to four or more different countries is 30.3%, while for those who export to less than four countries it is 26%. In contrast to the results found by Bandeira-de-Mello et al. (2016), Lu et al. (2010) and Powell and Rhee (2016), experience with export destinations does not radically improve the knowledge levels of these firms. The question most often answered correctly was the one regarding typical meat cuts sold in China (Quest 1), correctly answered by 69% of the respondents. According to specialists, shoulder steak, flank, bottom round are the beef cuts used for hot pot, barbeque and the increasingly popular Korean dishes (Brown et al., 2013). Surprisingly, the research team found no reliable public information on this question, thus it can be inferred this information is exchanged among business networks in China and Brazil (Johanson and Vahlne, 2009). The second question most frequently correctly answered was related to the Chinese regions with the current highest and greatest potential for future consumption expansion (Quest 9), correctly answered by 53.8% of the respondents. The correct answer is that all major cities in China have high rates of beef consumption and potential for growth. As suggested by Lu et al. (2010), this kind of information can be widely available through government agencies (APEX-Brazil) and the major export associations (ABIEC) and (Association of Brazilian Slaughterhouses (ABRAFRIGO) – Associação Brasileira dos Frigoríficos). These agencies have access to this information and can forward it to their members (Frischtak et al., 2015). Any urban hotspot is an attractive market for beef sales, especially those of imported beef, because, in general, the Chinese are more likely to trust international food safety regulations than their own production and certification systems (Bloomberg Businessweek, 2016, 2017a,b; Whitehead, 2014). Additionally, it is important to note that consumption in Northern China, especially those regions with a large Muslim population, also plays a relevant role, as commented by the interviewees. Questions 7 and 8, related to the import quantity and import peaks, were both correctly answered by five companies. This information has been reported in several sector specific media (for example Canal Rural, 2016; Dinheiro Rural, 2016; Sistema Faep, 2015). Questions related to bureaucracy, logistics and business culture (Quest 2, 3, 10 and 11) also present low correct response rates (three, three and four, respectively), showing that few companies would be able to act International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

efficiently without well-informed traders. The questions related to import duties were correctly answered by only one company representative, although this information is publicly available at http://tariffdata.wto. org and is fundamental when calculating revenues from a business transaction. Weakness on the part of the company export departments would seem to be the most plausible explanation for this result, which might reflect the high turnover of employees in Brazilian companies. Information on marketing channels and the size of the premium market is only available to Chinese market insiders and, accordingly, only one respondent correctly answered the related question. Such information is only available to those with firsthand experience of the market, or with an extremely well-established business networks in a highly-trusted environment, as suggested by Hunt and Morgan (1994), Knoll et al. (2017) and Lu et al. (2010). Surprisingly, none of the respondents seemed to know that there is no long-standing tradition of cooking beef in China (except among the Muslim minorities), thus the Chinese rarely cook it at home and consequently the percentage of sales through supermarkets is small. This information seems to be a well-published in several Brazilian magazines and organizations that have highlighted the importance of the Chinese catering sector (BeefPoint, 2013; Bloomberg Brasil, 2016; GloboRural, 2015). This would suggest that firms do not trust the information available in the Brazilian media.

5. Concluding remarks and managerial and policy implications The research results reveal that the interviewees, who are in leading positions in the headquarters of Brazilian beef packing companies or institutions operating in the export sector, have a low level of knowledge about the Chinese market. The findings show that neither firm size nor experience of beef exporting seem to influence the level of knowledge regarding the Chinese beef market. Similarly, no evidence was found to support the assumption that big multinationals have greater knowledge on the subject than smaller companies These findings appear to contradict the literature, which for the most part, finds that big companies with more experience in the international market have more access to information and knowledge about their importing markets (Fernhaber et al., 2007; Hoenen and Hansen, 2009; Lu et al., 2010). Regarding the possible sources of information and knowledge, the Brazilian beef export sector has neither an integrated data base containing essential information on the Chinese market, nor a unified traceability system in place that could facilitate the flow of information among the agents across the beef supply chain (Knoll et al., 2017). Thus, firms need to dedicate substantial management resources (time, financial, human) to collecting information and knowledge from several sources. Considering that information and knowledge about the Chinese market are a competitive factor, a fact firms seem to be well-aware of, leading firms that have the resources to invest in information gathering activities can obtain a competitive edge in the Chinese market. Due to the â&#x20AC;&#x2DC;follow-the-leaderâ&#x20AC;&#x2122; behavior (Hoenen and Hansen, 2009), whereby competitors adopt the strategy of the leader or first company to move into the new foreign market, a typical behavior of firms operating in a concentrated industry (four firms detain more than 60% of the total beef slaughtering, processing and exporting industry in Brazil), the leader firms try to protect this strategic resource. Unfortunately, although the situation presented above might benefit individual companies with the resources to tackle the challenges related to information gathering, organizing and learning, it inhibits the successful promotion of Brazilian Beef as a brand on the Chinese market. Based on previous research conducted by this research team (Knoll et al., 2017), companies that have been successful in the emerging Chinese beef market are known to promote and exploit the country of origin (Uruguay, New Zealand, Australia, etc.) as one of their strong points in marketing communication strategies. For Chinese consumers, it is much easier to relate the quality of beef to the country of origin than to an unknown company or brand name. In China, Australian beef is strongly associated with safety, quality and taste (MLA, 2016). In countries such as Uruguay, New Zealand and Australia, farmers and packers recognized the importance of the geographic indication, and are now, together with government support and the private sector, positioning themselves International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

accordingly. Additionally, these countries transfer a lot of information on their products to the Chinese consumers (Bloomberg Businessweek, 2016; Ortega et al., 2016), which makes their products even more desirable due to the consumer trust in them. Brazilian exporters could do well to make greater use of the APEX structure and information systems to communicate the quality of â&#x20AC;&#x2DC;Brazilian beefâ&#x20AC;&#x2122; to the Chinese consumers. It was found that firms only partially penetrate the Chinese market, even with the supposed benefits of FDI, experience with exports or traders and third parties. Instead, crucial knowledge seems to be acquired from multiple sources which, according to Sull and Escobari (2004), is another means of obtaining valuable information on a foreign market (customer needs, market opportunities and strategic partners, for example). There might be several reasons for this. One of which is time (Johanson and Vahlne, 2009), since the Chinese beef market has only recently re-opened, factors such as experience and foreign investment have not yet paid off, and might need at least five years to do so (Hohenthal, 2006). Cultural differences in dealing with business networks might prolong that period, as suggested by Dhingra (2013). On the other hand, the mere fact of having the company headquarters located abroad does not ensure that Brazilian beef packers acquire quality and timely information, since, in order to do so, the team must do its job adequately. Because Brazilian beef is essentially a commodity product, with no specification for the Chinese market, Chinese buyers only pass on the information that is absolutely necessary, such as quantity needed, price, and delivery time, without signaling any further information regarding the market tendencies (Knoll et al., 2017). When it comes to inside information about the Chinese market, the beef packers rarely develop their sources and scarcely consider such information. This might be because they do not have well-equipped specialists with reliable and valuable knowledge. Therefore, it can be assumed that even in the largest and leading export companies, the managers do not have the whole picture on the specific foreign market. Another explanation could be the high staff turnover in the Brazilian work force, which inhibits employees from focusing on consolidating their expertise on the Chinese beef market. This situation obliges companies to use specialist traders and special exporting channels, thus reducing the role of the export manager to managing the in-firm or out-firm networks and partnerships. The research results highlight the need for the Brazilian beef export sector to develop and implement information systems capable of ensuring the traceability of the beef from the farm to the Chinese consumer and that provide every segment within the supply chain with the framework to access at least the most basic information needed to ensure sustainable supply chain management (Buhr, 2003). Efficacy and efficiency in generating and delivering information across the supply chain is a big issue for the food system in emerging countries (Cunningham and Homse, 1986; Dhingra, 2013; Hunt and Morgan, 1994; Lu et al., 2010). Accordingly, more coordinated efforts on the part of the Brazilian Federal System of Quality Control (SIF), the APEX, the ABIEC and the ABRAFRIGO to homogenize and integrate information to a certain extent regarding the Chinese market and import requirements might be called for. Depending on the cuts exported, Brazilian beef products to China are up to 50% cheaper than those supplied from Australia and Uruguay. Brazil is considered one of the cheapest geographical origins, while it has less direct access to Mainland China. To add value and reach the middle-class Chinese consumer, a traceability system is recommended with the technological, institutional and informational infrastructure capable of producing and delivering weekly reports by the Brazilian and Chinese governments, export and import agents and sectorial associations to ensure and communicate food safety and supply and demand needs at a relatively low cost. Thus, a system is suggested in which the supply chain agents (both governmental and those private institutions already existing and active in the chain) and stakeholders should report, at least weekly, to an online database, to which every registered chain member would have access. Such a database would contain the origin, production and processing and storage capacity of the enterprises. It could signal demand and supply quantity and quality on weekly, monthly and half-yearly bases. By accessing such information and knowledge the supply chain member would know exactly who to contact and which stakeholder to network with. Such a platform could become a reliable traceability system based on agreements or contracts among the agents involved. International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Finally, it is relevant to point out that the conclusions and suggestion regarding how to improve the export and traceability systems reported above are based on the results obtained from the analysis of the questionnaires and interviews, as well as the theoretical foundations of a traceability system debated in the theoretical background (Section 2) (Buhr, 2003; Cooper et al., 1997; Cunningham and Homse, 1986; Dhingra, 2013; Hunt and Morgan, 1994; Jie et al., 2013; Knoll et al. 2017; Lu et al., 2010).

6. Limitations One of the research limitations is the low number of respondents to the questionnaire (13), although they represent a relatively high percentage of the total volume of beef exported from Brazil to China in 2016. To enrich the research, interviews were conducted with Brazilian and Chinese traders and export/import agents. It should be noted, however, that there were limitations on time, contacts and resources when conducting the interviews. Besides the brief nature of the interviews, it is important to acknowledge that only one person from each firm was interviewed, so the responses carry this limitation. The same is true for the case of the questionnaires. The position of the particular respondent in the firm should also be taken into account, as one would expect an export manager to know more about the Chinese market than a general manager, for instance. The questions in part C of the questionnaire are related to basic knowledge of the Chinese market. However, it would be an exaggeration to claim that someone has or does not have a complete overview of the Chinese market because they correctly or incorrectly answered 11 questions. Hence, while the data collected offer valuable insights into the knowledge held by the Brazilian beef industry regarding the Chinese market, those insights should be taken as evidence rather than statistical conclusions, since, due to the small sample size, it was not pertinent to apply any statistical test to check any hypothesis. In the future, a census in the sector could produce richer and statistically representative knowledge about the information flow in the Sino-Brazilian beef supply chain.

Acknowledgments The authors would like to thank the Brazilian National Council for Science and Technological Development (CNPq), the Brazilian Coordination for Improvement of Personnel in Higher Education (CAPES), and the Chinese Scholarship Council (CSC) for their financial support. Additionally, the authors would like to thank the interviewees and the survey participants for their cooperation. The authors would also like to express gratitude to the ABIEC for their constant availability to answer research related questions and to the three anonymous reviewers for their encouraging comments and suggestions.

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2017.0018. Methods S1. Questionnaire (in Portuguese). Methods S2. Questionnaire responses (in English).

References Akerlof, G.A. 1970. The market for â&#x20AC;&#x2DC;lemons â&#x20AC;&#x2122;: quality uncertainty and the market mechanism. The Quarterly Journal of Economics 84(3): 488-500. Alvarado, U.Y. and H. Kotzab. 2001. Supply chain management: the integration of logistics in marketing. Industrial Marketing Management 30(2): 183-98. Ambos, T.C. and B. Ambos. 2009. The Impact of distance on knowledge transfer effectiveness in multinational corporations. Journal of International Management 15(1): 1-14. Baird, I., A.M. Lyres and J.B. Orris. 1994. The choice of international strategies by small business. Journal of Small Business Management 31(4): 48-59.

International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Bandeira-de-Mello, R., M.T.L. Fleury, C.E.S. Aveline and M.A.B. Gama. 2016. Unpacking the ambidexterity implementation process in the internationalization of emerging market multinational. Journal of Business Research 69: 2005-2017. Barigozzi, M. and D. Garlaschelli. 2010. Multinetwork of international trade: a commodity-specific analysis. Physical Review E 81(4). Barkema, H.G., J.H.J. Bell and J.M. Pennings. 1996. Foreign entry, cultural barriers, and learning. Strategic Management Journal 17(2): 151-166. Bartlett, C. and S. Ghoshal. 1989. Managing across borders: the transnational solution. Harvard Business School Press, Boston, MA, USA. Becerra-Fernandez, I. and R. Sabherwal. 2001. Organizational knowledge management: a contingency perspective. Journal of Management Information Systems 18(1): 23-55. Beefcentral. 2017. JBS slaches Brazilian processing capacity in wake of market crisis. Available at: http:// tinyurl.com/y7k9eg76. BeefPoint. 2013. Pecuária de corte e mercado da carne bovina na China: mercado crescente, importações explodindo, oportunidades para o Brasil – Relatório completo do Rabobank. Available at: http:// tinyurl.com/yct5y8l3. Berg, B.L. 2001. Qualitative research methods for the social sciences. Person Education Limited, Essex, UK. Besanko, D., D. Dranove and M. Shanley. 1996. Economics of strategy. John Wiley and Sons Inc, New York, NY, USA. Blomstermo, A., D. Deo Sharma and J. Sallis. 2006. Choice of foreign market entry mode in service firms. International Marketing Review 23(2): 211-229. Bloomberg Brasil. 2016. Apetite Chinês por carne bovina impulsiona embarques do Brasil. Available at: http://tinyurl.com/yc4rgz6s. Bloomberg Businessweek. 2016. On Chinese acquaculture farms, as are the pigs, whose waste feeds the fish. Issue December 19th, 2016: 38-43. Bloomberg Businessweek. 2017a. A case of chicken vs machine. Issue January 16th, 2017: 18-20. Bloomberg Businessweek. 2017b. China´s Foodmakers try new growth recipes. Issue May 27th – June 4th, 2017: 20-24. Bolling, C.H. and A. Somwaru. 2001. U.S. food companies access foreign markets though direct investment. FoodReview 24(3): 23-28. Bonaccorsi, A. 1992. On the relationship between firm size and export intensity. Journal of International Business Studies 23(4): 605-636. Brown, C.G., S.A. Waldron and J.W. Longworth. 2013. A diachronic analysis of the beef industry. In: The political economy of agro-food markets in China: the social construction of the markets in an era of globalization. Palgrave Macmillan, Basingstoke, UK, pp. 127-152. Buckley, P.J. and M. Casson. 1998. Analyzing foreign market entry strategy: extending the internationalization approach. Journal of international Business Studies 29(3): 539-561. Buhr, B.L. 2003. Traceability and information technology in the meat supply chain: implications for firm organization and market structure. Journal of Food Distribution Research 34(3): 13-26. Business Insider. 2015. People are smuggling 40-year-old meat into China and selling it on the street. Available at: http://tinyurl.com/ybdgt253. Calof, J.L. 1993. The impact of size on internationalization. Journal of Small Business Management 31(4): 60-70. Canal Rural. 2016. Exportações de carne bovina brasileira atingem US$2,8 no primeiro semestre. Available at: http://tinyurl.com/y8o76yyx. Closs, D.J., T.J. Goldsby and S.R. Clinton. 1997. Information technology influences on world class logistics capability. International Journal of Physical Distribution and Logistics Management 27(1): 4-17. Cooper, M.C., D.M. Lambert and J.D. Pagh. 1997. Supply chain management: more than a new name for logistics. The International Journal of Logistics Management 8(1): 1-14. Cretoiu, S.L. 2010. Internacionalização de pequenas e médias empresas: 2000-2008. Revista Ibero-Americana de Estratégia – RIAE 9(3): 112-38.

International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Cunningham, M.T. and E. Homse. 1986. Controlling the marketing-purchasing interface: resource development and organisational implications. Industrial Marketing and Purchasing 1(2): 3-27. Dalkey, N. and O. Helmer. 1963. An experimental application of the Delphi method to the use of experts. Management Science 9(3): 458-467. Denzin, N.K. and Y.S. Lincoln. 2011. The SAGE Handbook of Qualitative Research. SAGE Publications, Los Angeles, CA, USA. Dhingra, S. 2013. Trading away wide brands for cheap brands. The American Economic Review 103(6): 2554-2584. Ding, M.J., F. Jie, K.A. Parton and M.J. Matanda. 2014. Relationships between quality of information sharing and supply chain food quality in Australian beef processing industry. The International Journal of Logistics Management 25(1): 85-108. Dinheiro Rural. 2016. Carne para o mundo. Available at: http://tinyurl.com/y9xkzdd6. Eriksson, K.A., J. Johanson, A. Majkgard and D. D. Sharma. 1997. Experiential knowledge and cost in the internationalization process. Journal of International Business Studies 28(2): 337-360. Etemad, H. 2004. Internationalization of small and medium- sized enterprises: a grounded theoretical framework and an overview. Canadian Journal of Administrative Sciences 1(21): 1-21. Exame. 2016. As 40 maiores exportadoras do Brasil em 2015. Available at: http://tinyurl.com/yd2resrw. Felbermayr, G. and B. Jung. 2011. Trade intermediation and the organization of exporters. Review of International Economics 19(4): 634-648. Fernhaber, S.A., P.P. McDougal and B.M. Oviatt. 2007. Exploring the role of industry structure in new venture internationalization. Entrepreneurship: Theory and Practice 31(4): 517-526. Fleury, M.T.L., F.M. Borini, A. Fleury and M.M. de Oliveira Junior. 2007. Internationalization and performance: a comparison of Brazilian exporters versus Brazilian multinationals. Economia E Gestao 7(14): 1-178. Frischtak, C., A. Soares, T. Cariello, C.F. Orth, C. Santos and P. Steffen. 2015. Oportunidades de Comércio e Investimento na China Para Setores Selecionados. Available at: http://tinyurl.com/yakpba9q. Frohlich, M.T. and R. Westbrook. 2001. Arcs of integration: an international study of supply chain strategies. Journal of Operations Management 19(2): 185-200. Fu, Q. and K. Zhu. 2010. Endogenous information acquisition in supply chain management. European Journal of Operational Research 201(2): 454-462. Gao, G.Y., J.Y. Murray, M. Kotabe and J. Lu. 2010. A ‘strategy tripod’ perspective on export behaviors: evidence from domestic and foreign firms based in an emerging economy. Journal of International Business Studies 41: 377-396. Gereffi, G., M. Korzeniewicz and R.P. Korzeniewicz. 1994. Introduction: global commodity chains. In: Commodity chains and global capitalism, edited by G. Gereffi and M. Korzeniewicz. Greenwood Press, Westport, Ireland, pp. 1-14. Ghemawat, P. 2001. Distance still matters. Harvard Business Review 79(8): 137-147. Gimenez, C. and V. Sierra. 2013. Sustainable supply chains : governance mechanisms to greening suppliers. Journal of Business Ethics 116: 189-203. Global Times. 2015. Smuggled meat came via Vietnam: official. Global Times. Available at: http://www. globaltimes.cn/content/928998.shtml. GloboRural. 2015. Carne é setor de maior potencial de investimento na China. Available at: http://tinyurl. com/ychsnzcy. Goh, S.K., K.N. Wong and S.Y. Tahm. 2013. Trade linkages of inward and outward FDI: evidence from Malaysia. Economic Modeling 35: 224-230. Hakanson, L. and R. Nobel. 2001. Organizational characteristics and reverse knowledge transfer. Management International Review 41(4): 395-420. Hessels, J. and S. Terjesen. 2010. Resource dependency and institutional theory perspectives on direct and indirect export choices. Small Business Economics 34: 203-220. Hoenen, A.K. and M.W. Hansen. 2009 Oligopolistic competition and foreign direct investment, (Re) Integrating the strategic management perspective in the theory of multinational corporations. Copenhagen Business School, Centre for Business and Development Studies, CBDS Working Series, Working Paper Nr. 10, 2009. Available at: http://tinyurl.com/ybt488a9. International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Hohenthal, J. 2006. Managing interdependent business relationships in SME internationalization. In: Business networks and international marketing, edited by A. Hadjikhani, J.-W. Lee and J. Johanson. Doo Yang Publishing, Seoul, South Korea, pp. 209-222. Hunt, S.D. and R.M. Morgan. 1994. Relationship marketing in the era of network competition. Marketing Management 3(1): 18-28. Jank, M.S., M.F.P. Leme, A. Nassar and P.E. Filho. 2001. Concentration and internationalization of Brazilian agribusiness exporters. International Food and Agribusiness Management Review 2(3/4): 359-374. Jie, F., K.A. Parton and R.J. Cox. 2013. Linking supply chain practices to competitive advantage: an example from Australian agribusiness. British Food Journal 115(7): 1003-1024. Johanson, J. and J.E. Vahlne. 2009. The Uppsala internationalization process model revisited: from liability of foreignness to liability of outsidership. Journal of International Business Studies 40(9): 1411-1431. Julien, P.A., and C. Ramangalahy. 2003. Competitive strategy and performance in exporting SMEs: an empirical investigation of their export information search and competencies. Entrepreneurship: Theory and Practice 27(3): 227-245. Katsikeas, C.S. 1994. Export competitive advantages: the relevance of firm characteristics. International Marketing Review 11(3): 33-53. Knight, G. and D. Kim. 2009. International business competence and the contemporary firm. Journal of International Business Studies 40(2): 255-273. Knoll, S., C.S.S. Marques, J. Liu, F. Zhong, A.D. Padula and J.O.J. Barcellos. 2017. The Sino-Brazilian beef supply chain: mapping and risk detection. British Food Journal 119(1): 164-80. Lam, H.M., J. Remais, M.C. Fung, L. Xu and S.S.M. Sun. 2013. Food supply and food safety issues in China. The Lancet 381(9882): 2044-2053. Lambert, D.M., M.C. Cooper and J.D. Pagh. 1998. Supply chain management: implementation issues and research opportunities. The International Journal of Logistics Management 9(2): 1-20. Lee, H.L., V. Padmanabhan and S. Whang. 1997. Information distortion in a supply chain: the bullwhip effect. Management Science 43(4): 546-558. Li, J. 1995. Foreign entry and survival: effect of strategical choices on performance in international market. Strategic Management Journal 19(3): 333-352. Linhai, W., W. Shuxian and X. Lingling. 2013. The study of consumer demand in traceable food market: the case of traceable pork. Journal of Public Management 10(3): 119-128. Liu, X, C. Wang and Y. Wei. 2001. Casual links between foreign direct investment and trade in China. China Economic Review 12: 190-202. Livesey, C. 2006. The relationship between positivism, interpretivism, and sociological research methods. In: As Sociology for AQA: 1-5. Available at: http://www.sociology.org.uk/notes/revgrm5.pdf Longworth, J.W., C.G. Brown and S.A. Waldron. 2001. Beef in China; agribussines opportinities and challenges. The University of Queensland Press, Queensland, Australia. Lu, Y., L. Zhou, G. Bruton and W. Li. 2010. Capabilities as a mediator linking resources and the international performance of entrepreneurial firms in an emerging economy. Journal of International Business Studies 41: 419-436. Luo, Y., H. Zhao, Y. Wand and Y. Xi. 2011. Venturing abroad by emerging market enterprises. Management International Review 51: 433-459. MDIC. 2016. Brazilian ministry of development industry and foreign trade (MDIC in Portuguese). Exportações Brasileiras de carne bovina in natura. Available at: http://tinyurl.com/yaa5d36w. MICA. 2017. Meat Import Council of America. China and global beef market. Available at: http://tinyurl. com/y7flrkxl. Miles, M.B., A.M. Huberman and J. Saldana. 2013. Qualitative data analysis: a methods sourcebook. SAGE Publications, Thousand Oaks, CA, USA. Min, S. and J.T. Mentzer. 2004. Developing and measuring supply chain management concepts. Journal of Business Logistics 25(1): 63-99. MLA. 2016. Meat and Livestock Australia. Insights China. Available at: http://tinyurl.com/ybw7lzwm.

International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Moberg, C.R., B.D. Cutler, A. Gross and T.W. Speh. 2002. Identifying antecedents of information exchange within supply chains. International Journal of Physical Distribution and Logistics Management 32(9): 755-770. Morosini, P., S. Shane and H. Singh. 1998. National cultural distance and cross border acquisition performance. Journal of International Business Studies 29(1): 137-158. Ortega, D.L, S. Jeong, H.H. Wang and L. Wu. 2016. Emerging markets for imported beef in China : results from a consumer choice experiment in Beijing. Meat Science 121: 317-323. Oviatt, B.M. and P.O. McDougall. 1994. Toward a theory of international new ventures. Journal of International Business Studies 25: 45-65. Pangarkar, N. 2008. Internationalization and performance of small- and medium sized enterprise. Journal of World Business 43(4): 475-485. Peng, M.W, D.Y.L. Wang and Y. Jiang. 2008. An institution-based view of international business strategy : a focus on emerging economies. Journal of International Business Studies 39: 920-936. Pfaffermayr, M. 1994. Foreign investment and exports: a time series approach. Applied Economics 26(4): 337-351. Porter, M.E. 1980. Competitive strategy: techniques for analyzing industries and competitors. The Free Press, New York, NY, USA. Powell, K.S. and M. Rhee. 2016. Experience in different institutional environments and foreign subsidiary ownership structure. Journal of Management 42(6): 1434-1461. Prajogo, D. and J. Olhager. 2012. Supply chain integration and performance: the effects of long-term relationships, information technology and sharing, and logistics integration. International Journal of Production Economics 135(1): 514-522. Reuber, A.R., and E. Fischer. 1997. The influence of the management team’s international experience on the internationalization behaviors of SMEs. Journal of International Business Studies 28(4): 807-825. Reuters. 2015. China meat smuggling crackdown stokes risky underground trade. Available at: http://tinyurl. com/ybzxqsxs. Reuters. 2017a. Operation weak flesh takes bite out of Brazil´s meat exporters. Available at: http://tinyurl. com/l4u63sj. Reuters. 2017b. China, others lift ban on meat imports in boost for Brazil. Available at: http://tinyurl.com/ ych6nrz8. Root, F.J. 1987. Foreign market entry strategies. Lexington Books, Lexington, MA, USA. Simatupang, T.M., A.C. Wright and R. Sridharan. 2002. The knowledge of coordination for supply chain integration. Business Process Management Journal 8(3): 289-308. Sistema Faep. 2015. Exportação de carne bovina indica retomada. Available at: http://tinyurl.com/yb87c3jj. South China Morning Post. 2015. Illegal smuggling routes’ exposed after rotting meat from the 1970s seized by Chinese customs. Available at: http://tinyurl.com/ycom4e3x. Squartini, T. and D. Garlaschelli. 2013. Economic networks in and out of equilibrium. Available at: http:// tinyurl.com/y8ye532e. Stadtler, H. 2005. Supply chain management and advanced planning – basics, overview and challenges. European Journal of Operational Research 63: 575-588. Steiger, C. 2006. Modern beef production in Brazil and Argentina. Choices 21(2): 105-110. Sull, D.N. and M. Escobari. 2004. Sucesso made in Brazil 3rd ed. Elsevier, Rio de Janeiro, RJ, Brazil. Sun, C.H., W.Y. Li, C. Zhou, M. Li, Z.T. Ji and X.T. Yang. 2014. Anti-counterfeit code for aquatic product identification for traceability and supervision in China. Food Control 37: 126-134. Tan, C. 2001. A framework of supply chain management literature. European Journal of Purchasing and Supply Management 7: 39-48. Tan, K.C., S.B. Lyman and J.D. Wisner. 2002. Supply chain management: a strategic perspective. International Journal of Operations and Production Management 22(6): 614-631. Terjesen, S., C. O’Gorman and Z.J. Acs. 2008. Intermediated mode of internationalization: new software ventures in Ireland and India. Entrepreneurship and Regional Development 20(1): 89-109. The Economist. 2016. Free exchange. Brexiteers need to respect gravity models of international trade. Issue October 1st 2016: 73. Available at: http://tinyurl.com/ydarq6ef. International Food and Agribusiness Management Review

Knoll et al.

Volume 21, Issue 1, 2018

Thomé, K.M. and L.M. Vieira. 2012. Internationalization among emerging countries: insights from BrazilianRussian beef network. Journal on Chain and Network Science 12(3): 231-241. Tihanyi, L., D.A. Griffith and C.J. Russell. 2005. The effect of cultural distance on entry mode choice, international diversification, and MNE performance: a meta- analysis. Journal of International Business Studies 36(3): 270-283. Tinbergen, J. 1962. Shaping the world economy; suggestions for an international economic policy. Twentieth Century Fund, New York, NY, USA. U.S. Meat Export Federation. 2014. Beef shortage, surging consumption bring more imports to Vietnam. Available at: http://tinyurl.com/ybwau5k3. United Nations Conference on Trade and Development. 2016. World investment report 2016 – country fact sheet Brazil. Available at: http://tinyurl.com/y8svec8a. Van der Vorst, J.G.A.J. and A.J.M. Beulens. 2002. Identifying sources of uncertainty to generate supply chain redesign strategies. International Journal of Physical Distribution and Logistics Management 32(6): 409-430. Van Donk, D.P., R. Akkerman and T. van der Vaart. 2008. Opportunities and realities of supply chain integration: the case of food manufacturers. British Food Journal 110(2): 218-235. Vieira, L.M., and W.B. Traill. 2008. Trust and governance of global value chains: the case of a Brazilian beef processor. British Food Journal 110(4/5): 460-473. Waldron, S., C. Brown and J. Longworth. 2010. A critique of high-value supply chains as a means of modernising agriculture in China: the case of the beef industry. Food Policy 35(5): 479-487. Waldron, S., J. Wang, H. Zhang, X. Dong and M. Wang. 2015. The Chinese beef industry. In: Regional workshop on beef markets and trade in southeast Asian and China. Ben Tre, Vietnam. Available at: http://tinyurl.com/ycmaswfo. Whitehead, M. 2014. China’s great beef challenge – a golden opportunity for the Australian beef sector. Available at: http://tinyurl.com/y7xydzau. Youyang8. 2015. China ‘s beef cattle industry development report 2014-2015 (Mandarin Chinese: 2014-2015 年度中国肉牛业发展报告). Available at: http://www.youyang8.com/?id=490. Yuan, L., X. Qian and N. Pangarkar. 2016. Market timing and internationalization decisions: a contingency perspective. Journal of Management Studies 53(4): 497-519. Zajac, E.J. and C.P. Olsen. 1993. From transaction cost to transaction value analysis: implication for the study of organizational strategies. Journal of Management Studies 39(1): 131-145. Zhou, H. 2007. Supply chain practice and information sharing. Journal of Operations Management 25(6): 1348-1365.

International Food and Agribusiness Management Review

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2016.0086 Received: 18 April 2016 / Accepted: 4 October 2017

Seeding eastern Africa’s maize revolution in the post-structural adjustment era: a review and comparative analysis of the formal maize seed sector RESEARCH ARTICLE Olaf Erenstein

and Girma Tesfahun Kassieb

aAgricultural

Economist, International Maize and Wheat Improvement Center (CIMMYT), Carr. Mex-Ver Km. 45, El Batan, Texcoco, Mexico

bAgricultural

Market Economist, International Center for Agricultural Research in the Dry Areas (ICARDA), P.O. Box 5689, Addis Ababa, Ethiopia

Abstract Improved maize seed is instrumental to deliver an Asian-style ‘green revolution’ for Africa. The paper reviews and makes a comparative analysis of the maize (corn) seed sector and its evolution in Kenya, Tanzania, Uganda and Ethiopia drawing from seed sector surveys and secondary data. Enhancing farmers’ access to and use of new maize varieties still presents a number of challenges in eastern Africa – not least due to a number of policy and institutional impediments to the development of the seed sector. The regional seed sectors also show some remarkable contrasts: they have evolved at different speeds and in different directions, driven by diverging agricultural growth opportunities and varying degrees of regulation, liberalization and restructuring. The paper reiterates calls for an enabling environment for private seed companies to evolve in order to serve the diverse farmer communities so that they benefit from existing and future improved maize seed opportunities. Keywords: seed supply, seed business, corn, improved varieties, structural adjustment JEL code: Q13, O13 Corresponding author: o.erenstein@cgiar.org

Erenstein and Kassie

Volume 21, Issue 1, 2018

1. Introduction The need to increase agricultural productivity to enhance food security and reduce poverty in Africa is widely acknowledged (IBRD, 2007; McArthur, 2015). Improved crop varieties play a critical role in agricultural intensification (Evenson and Gollin, 2003; Walker and Alwang, 2015) – particularly when combined with external inputs and a conducive policy environment as exemplified by the Asian Green Revolution. There has been considerable interest in an African green revolution (Byerlee and Eicher, 1997; Otsuka and Larson, 2016; Scoones and Thompson, 2011; Smale et al., 2011; Toenniessen et al., 2008). Improved maize (or corn, Zea mays L.) varieties play a potential pivotal role to revolutionize Sub-Saharan Africa’s (Africa hereafter) agriculture, enhance food security and reduce poverty. Maize is a strategically important crop for food security and economic growth in Africa. Maize is a key food crop across large swathes of Africa, and Africa’s largest and most widely cultivated cereal with over 30 million ha cultivated and 50 million tons produced annually. Currently maize productivity is a major determinant of food security from the household to the regional level. Improved maize seed is a key to boost productivity so as to generate the surpluses to lift rural incomes and feed burgeoning rural and urban populations. Unlike in many other parts of the world, maize is a key staple crop across many eastern and southern African countries and fast growing in importance in West Africa. Hybrid maize seed is the key to viable seed industries, further enabled by the structural adjustment induced market liberalization and privatization. Maize seed has thereby long been viewed as instrumental to deliver an Asian-style ‘green revolution’ for Africa. Yet, despite improved maize seed’s promise and substantial progress made in recent years in Africa – progress has been slow, intermittent and uneven (Kassie et al., 2013; Langyintuo et al., 2008, 2010; Morris, 1998; Rusike and Eicher, 1997; Timothy et al., 1988). Enhancing farmers’ access to and use of new maize varieties still presents a number of challenges in the region – not least due to a number of policy and institutional impediments to the development of the seed sector. This is particularly important given that most of the maize in Africa is grown by small-scale farmers facing asset, institutional and policy constraints in accessing new technology and markets. A vibrant, efficient and functioning seed sector is critical in ensuring widespread distribution of new maize varieties to farmers for increasing productivity and reducing food prices for consumers. While improved seed supply is critical in many regions for improving access and use, demand is also crucial for igniting productivity change and stimulating growth in the seed industry. The institutional and policy impediments are important on both supply and demand sides of this process. The seed sector in Africa also shows some remarkable contrasts: it has evolved at different speeds and in different directions, driven by diverging agricultural growth opportunities and varying degrees of regulation, liberalization and restructuring. The objective of the paper is to assess and characterize the maize seed sector and its evolution in Kenya, Tanzania, Uganda and Ethiopia in the post-structural adjustment era. Specifically, the paper first reviews the state of the formal maize seed sector in eastern Africa based on secondary data and the literature. The paper then makes a comparative analysis of the maize seed sector stakeholders drawing from seed sector surveys in each of the four study countries. The premise is that we need to understand seed systems in order to enable further seed sector development and a maize revolution; whereas their inherent context specificity and complexity calls for comparative analysis using a similar approach across national borders within eastern Africa. This study aims to start filling the knowledge gap given the importance of maize in the region and that of the seed sector in enhancing the development and use of improved crop varieties. The paper is structured as follows. The following section provides an analytical framework. The third section introduces the material and methods including the underlying seed sector surveys. The fourth section reviews the eastern Africa’s maize seed sector. The fifth presents a comparative analysis of the seed sector stakeholders. The paper then discusses the associated challenges and opportunities for the sub-region’s maize seed sector development, before concluding.

International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

2. Analytical framework Seed systems can be broadly categorized into formal and informal – with the formal referring to the organized seed sector including institutionalized seed producers and companies, be it private or public. The informal seed sector is non-institutionalized, encompassing seed saving, seed exchange and seed production by farmers and is often highly localized. The informal sector is the major source of seed of all crops in Africa (e.g. Louwaars and De Boef, 2012), with an estimated seed share above three-quarters across eastern Africa (e.g. 80% in Kenya (Wulff et al., 2006), 90% in Tanzania (Ngwediagi et al., 2009), 92.5% in Uganda (Kabeere and Wulff, 2008), and 96.5% in Ethiopia (Atilaw and Korbu, 2011)). The relative shares also vary by crop with the formal share being substantially higher for maize, although estimates vary considerably across eastern Africa. For instance, in the case of Ethiopia, seed from the formal sector is estimated to cover 19% of the maize area (Alemu, 2011), whereas it approximates three-quarters in Kenya (Langyintuo et al., 2010). The formal and informal seed systems can play diverging but also complementary roles (Almekinders and Louwaars, 2002), also in view of the significant market failures for seed research and development in developing countries (Kremer and Zwane, 2005). This paper purposively focuses on the formal seed sector in view of its critical role in agricultural intensification through the introduction of improved seed. The paper particularly focuses on maize seed as the formal seed sector in eastern Africa primarily revolves around maize. This reflects the role maize plays as a major staple crop in the sub-region and the inherent characteristics of maize seed. Maize is a cross-pollinating crop and thereby opened the prospects of hybridization and commercialization through the associated potential demand for recurrent seed renewal further aided by its high multiplication factor. Seed system analysis needs to go beyond identifying the formal and informal sectors. Seed systems are complex and dynamic – not least due to inherent variations in products, supply and demand side factors and their interrelated associations with the policy environment. Various analytical frameworks have been put forward to analyze seed systems (i.e. Almekinders and Louwaars, 2002; Cromwell et al., 1992; Louwaars and De Boef, 2012; Van Mele et al., 2011a). Some have emphasized the dynamic nature of seed systems and perceived development paths from ‘basic’ to ‘mature’ seed systems (Morris, 1998; Pray and Ramaswami, 1991; Rusike and Eicher, 1997) with a decreasing role for the informal sector and corresponding increasing role for the formal sector and corresponding stage dependent policy implications. The associated path dependency of seed sector development from preindustrial to emergence to expansion to maturity is appealing for its simplicity and apparent logic. But others have argued that reality has proved more complex and that there is no blueprint seed policy implying that there is a need to understand seed systems in order to develop the corresponding seed policy options (Louwaars, 2002). This study builds on this premise – and seeks to understand seed systems within their local context. The context specificity of seed systems and their inherent complexity have led to numerous case studies of (sub)national seed systems with a diversity of approaches and emphasis – with seed systems in eastern Africa being a case in point (Kenya: Misiko et al., 2011; Muhammad et al., 2003; Muthoni and Nyamongo, 2008;Wulff et al., 2006; Tanzania: Ngwediagi et al., 2009; Uganda: Kabeere and Wulff, 2008; Larson and Mbowa, 2004; Muhhuku, 2002; Van Mele et al., 2011b; Ethiopia: Alemu, 2011; Alemu et al., 2010; Atilaw and Korbu, 2011; Bishaw et al., 2008). These diverse stand-alone case studies make it difficult to make any comparative analysis or synthesis and this challenge is further compounded by the diverging points of departure and timing and extent of liberalization. The current paper addresses this gap by providing a comparative analysis using a similar approach across national borders within eastern Africa.

3. Material and methods The study draws on a structured questionnaire to generate data and information on the maize seed industries in eastern Africa for a comparative analysis. The questionnaire based interview included respondents representing seed companies, national research institutes, and seed traders’ associations in each of the International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

countries (Ethiopia, Kenya, Tanzania and Uganda). The survey was not intended as a census but aimed and succeeded to at least include all the main established stakeholders willing and able to cooperate (Table 1). The sample includes some relatively independent subsidiary seed companies as separate entries; i.e. the instrument focusing on their immediate operational domain and not on their respective parent companies. The survey was complemented with a literature review and secondary data where available. However, unless specified otherwise, data presented here were collected during the 2010 seed sector survey.

4. A review of eastern Africa’s maize seed sector The formal seed sector in Africa typically originated from a similar public sector dominance and till the turn of the century the commercial seed sector had generally been slow to develop (Tripp and Rohrbach, 2001). Indeed, seed sector development in Africa has long been hampered by regulatory frameworks that favored parastatal enterprises and inhibited commercial innovation (Tripp and Rohrbach, 2001). The regulatory frameworks variously emphasized control, cooperation and/or competition (Louwaars, 2002). Seed commercialization by the formal sector normally implies the need for varietal registration and seed certification in each country (GRAIN, 2005; Setimela et al., 2009; Waithaka et al., 2011). Despite similar origins, in the 1990s there was already considerable variation in seed sector development across countries: e.g. in eastern Africa, only Kenya was perceived as having a rather effective and diversified seed industry, with still rather limited progress in Tanzania, Uganda and Ethiopia (Lanteri and Quagliotti, 1997). Some of the current seed sector characteristics reiterate the differential start, with Kenya being the first to have a national seed trader association, having substantially more active seed companies, more registered maize varieties, more certified maize seed sales and more adoption of improved maize varieties (Table 2). The last decade has seen on-going efforts to harmonize and rationalize the seed policy environment across eastern Africa. Maize dominates the national formal seed sales with shares in excess of 70% in all eastern African countries, except in Ethiopia where maize comprises 38% (Table 2). Maize also dominates the national variety lists – with for instance more than a third of entries in both Kenya’s national variety list and Tanzania’s plant varieties officially released for commercialization between 1995-2008 (Ngwediagi et al., 2009). These selected indicators however also mask further regional variations. For instance, the reported average national volumes of certified maize seed have been relatively stagnant over the last few years in Kenya and Tanzania, but have seen a steady increase in Uganda (Waithaka et al., 2011). There has also been a varying liberalization of the seed sector. Uganda stands out with the seed industry having moved from a public monopoly of seed production to a completely private seed sector in the wake of policies to liberalize and privatize the economy that started in the early 1990s (Kabeere and Wulff, 2008). Kenya liberalized the seed sector in 1996 (Wulff et al., 2006) with numerous seed companies rapidly emerging in the following years to a total of 74 in 2009. However, despite the rapidly increasing number of private seed companies, the Kenya Seed Company (KSC) remains the largest company that still dominates the national seed market (e.g. nearly 75% of certified seed sales in 2005 (Wulff et al., 2006)) while being publicly owned. In Tanzania, the former seed parastatal was privatized – with the private sector primarily responsible for production and marketing of certified seeds although up to recently the public Agricultural Seed Agency remained responsible for the production of basic seed of public varieties and even marketed some certified seeds (Ngwediagi et al., 2009). Table 1. Number of formal maize seed sector stakeholders interviewed, 2010. Number of seed companies Number of other stakeholders Total

Kenya

Tanzania

Uganda

12 2 14

10 1 11

8 2 10

International Food and Agribusiness Management Review

Ethiopia 5 1 6

Total 35 6 41

Erenstein and Kassie

Volume 21, Issue 1, 2018

Table 2. Selected characteristics of seed sector from secondary data.1 Kenya

Tanzania

Number of active seed companies a 74 31 % Share of maize seed in national 87 (2009) 71 (2009) seed sales a,b Certified maize seed sales 25-30 (2005-08) 7-7.5 (2005-08) a,b (×1000 ton per year) Number of registered maize varieties 164 (2009) 75 (2008) (year) c Estimated adoption rates of improved maize varieties (% maize area) d,e 2007 72 18 1997 71 4 Establishment national seed trader 1982 2002 association (year) f

Uganda

Ethiopia

20 75 (2009)

30 38 (2007-08)

8.5 (2009)

8.6 (2007-08)

36 (2010)

42 (2009)

35 9 1999

19 8 2006

1 Adapted from: (a) Waithaka et al., 2011; (b) MoARD, personal communication; (c) data obtained from: Variety Release Committees

of Ministry of Agriculture and Rural Development, Addis Ababa, Ethiopia; Kenya Plant Health Inspectorate Services (KEPHIS), Nairobi, Kenya; Ministry of Agriculture and Food Security (MAFS), Morogoro, Tanzania; Ministry of Agriculture, Animal Industry and Fisheries, Entebbe, Uganda; (d) Langyintuo et al., 2010; (e) Hassan et al., 2001; (f) National seed trader associations (http:// afsta.org/memberships/afsta-members).

In Ethiopia, the Ethiopian Seed Enterprise is a public enterprise that still dominates the national seed market. In the 1990s, Pioneer Ethiopia was the only other operational seed company, and only in the 2000s did the number of seed companies start to increase. But Ethiopia stands out for its ambivalent attitude towards economic liberalization and the private sector (Alemu, 2011). Indeed, Ethiopia is the only case study country where the number of public seed enterprises has been increasing, with the recent establishment of regional seed enterprises with the aim of addressing the seed demand at the regional level. Private seed companies contributed only 12% to the volume of formal seed produced in Ethiopia (across all crops, 2006-08), with pubic seed companies contributing 79% and others (including state farms, unions, research centers) contributing 10% (Alemu et al., 2010). The role of private seed companies is somewhat more substantial in the case of hybrid maize seed production (33%, ibid.).

5. A comparative analysis of eastern Africa’s maize seed sector stakeholders Liberalization opened the seed sector to many new entrants. The startup year of the surveyed seed companies averages late 1990s, reflecting their relatively recent nature and their increasing number, with most seed companies having worked with maize seed from their inception (Table 3). Seed typically was the core business of the surveyed seed companies contributing 89% of their gross revenue, with two-thirds of companies having an exclusive seed focus (Table 3). Most of the seed companies had a diversified seed portfolio in terms of crops, although maize seed is the most important accounting for about 64% of the gross revenue (Table 3) – reiterating the relative profitability of maize seed vis-à-vis other crops. Ethiopia stands out for having the least diversified seed companies – reflecting that most private seed companies focus on hybrid maize seed. Surveyed seed companies in Tanzania and Ethiopia catered purely for their domestic market, whereas at least a third of the surveyed seed companies in Kenya and Uganda also reported regional cross-border operations. About half the surveyed seed companies produced publicly available maize varieties, with 29% of surveyed seed companies reporting having their own breeding unit (to develop their own varieties) and 14% relying on a breeding unit with their parent company (Table 3). In-house breeding units were markedly more common in Kenya and Uganda, implying their relatively more mature private seed sector. An increasing share of the released maize varieties now is proprietary material owned by private seed companies (Setimela et al., 2009). Surveyed seed companies typically produce their International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

Table 3. Surveyed seed company characteristics. Startup year (average) Seed company Maize seed Seed focus Exclusive (% seed companies (SC)) Share gross revenue (%) Maize seed focus Exclusive (% SC) Share gross revenue (%) Geographic scope (% SC) Sub-national National Regional Number of countries covered In-house (% SC) Breeding unit (+ with parent Co.) Seed quality control lab In-country seed multiplication (% SC) % irrigated % area contract farmers

Kenya

Tanzania

Uganda

Ethiopia

Overall

1990 1995

2002 2002

2001 2004

1998 1998

1997 1999

58 84

70 87

63 95

80 95

66 89

8 65

0 71

0 47

40 79

9 64

33 33 33 1.9

20 80 0 1.0

0 50 50 1.9

40 60 0 1.0

23 54 23 1.5

42 (+17) 30 83 37 71

10 (+10) 44 90 17 62

38 (+13) 71 88 6 85

20 (+20) 20 100 29 30

29 (+14) 42 89 23 65

seed in-country â&#x20AC;&#x201C; only a few companies bring their seed from elsewhere (Table 3). The seed multiplication is mainly rainfed and hence subject to the vagaries of weather. About two-thirds of the companiesâ&#x20AC;&#x2122; maize seed multiplication area is contracted from seed out-growers while one third of the area is owned by the companies themselves â&#x20AC;&#x201C; the latter particularly more important in Ethiopia apparently linked to the land tenure system (Table 3). The surveyed seed companies reported an average portfolio of five current maize varieties (Table 4). The maize portfolio is heavily biased towards hybrids in Kenya and Ethiopia whereas open pollinated varieties (OPV) still dominated in Tanzania and Uganda at the time of the survey (Table 4). The volume of maize seed sold in the domestic market averages 1,500 tons per surveyed company, with Uganda and Tanzanian companies selling half that and Kenyan companies selling nearly two times the average (2,700 tons; Table 4). These country averages are, however, influenced by the small sample size and the substantial volumes handled by the large public seed companies. Indeed, across the region the surveyed seed companies are primarily small private seed companies with seed volumes below 1,500 tons per year. On average, a quarter of the seed volume handled by each company was classified as short duration and 22% as drought tolerant (Table 4). These two categories overlap to some extent but typically reflect maize seed targeted for the drier drought prone environments. The retail price for maize seed averaged 1.8 and 1.1 US$/kg for hybrid and OPV, respectively (Table 4). As expected, OPV seed was substantially cheaper than hybrid maize, although it was relatively expensive in Kenya in part associated with the limited OPV volumes handled and correspondingly small sample size. Both types of maize seed were relatively cheap in US$ terms in Ethiopia compared to the region as whole; whereas hybrid maize seed was relatively expensive in Tanzania. The maize seed prices expressed relative to the rural grain prices oscillated around 8 for hybrid and 5 for OPV maize seed (Table 4). Maize seed is primarily targeted at small-scale farmers and is thus predominantly marketed in small bag sizes with the 2 kg bag being the most common (Table 5). A notable exception is Ethiopia where all maize seed was sold in mandatory 12.5 kg bags at the time of the survey. Maize seed is mainly sold through independent private retail outlets (or agro-dealers, especially in Kenya and Tanzania), with smaller shares reported as International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

Table 4. Seed company seed portfolio. Maize seed portfolio Number of current varieties % hybrid entries Average maize seed volume sold in domestic market (× 1000 ton per year) % hybrids % short duration % drought tolerant Retail prices (US$/kg) Hybrid maize seed OPV1 maize seed Maize grain price (US$/kg) Relative seed price (seed:grain) Hybrid seed OPV seed 1

Kenya

Tanzania

Uganda

Ethiopia

Overall

5.2 91 2.7

5.0 17 0.7

4.3 42 0.8

5.4 92 1.6

4.9 58 1.5

89 28 22

17 24 36

26 36 16

94 3 4

55 25 22

1.91 1.84 0.25

2.22 1.08 0.23

1.69 0.85 0.20

1.27 0.44 0.16

1.77 1.05 0.22

7.8 7.5

9.8 4.8

8.6 4.4

8.0 2.7

8.1 4.8

Tanzania

Uganda

Ethiopia

Overall

10 90 40 20 0

25 88 75 63 13

0 0 0 100 0

18 79 35 47 15

76 8 16 0

46 25 29 0

1 0 3 96

57 8 17 18

OPV = open pollinated varieties.

Table 5. Seed company maize seed distribution. Kenya Main bag sizes maize seed (in kg, % SC reporting)1 0.5-1 27 2 100 5 18 10-12.5 36 25-50 36 Maize seed distribution outlets (% share sales per year) Private retail 76 Own seed company retail 0 Direct sales 14 Other sales network 10 1

Columns may not sum 100% as multiple responses are possible; SC = seed companies.

direct sales (usually larger volumes sold directly by the seed company to individuals and/or institutions) or sales through the seed companies own retail outlets (Table 5). Ethiopia is an exception again that seed sales are centrally organized and maize seed demand is first estimated and maize seed subsequently distributed through farmer cooperatives and unions. The surveyed formal maize seed sector stakeholders were queried on their perceptions of the seed policy environment – with about a third perceiving an overall increase in seed regulatory processes in their respective countries (Table 6). The time to release a new variety was generally perceived to have decreased to an average of 2.7 years. The decrease was generally associated with the harmonization efforts and the average was relatively uniform across the case study countries. After release it would take on average an additional two years to have the seed available in ‘significant quantities’, although this showed more regional variation depending on the seed company and country specificities. Kenya’s seed certification system is relatively well developed and used as a model for the other case study countries. As a result, most stakeholders in Kenya

International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

Table 6. Perception on seed policy environment by surveyed formal maize seed sector stakeholders (2010).1 Perceived increase in seed regulatory processes (% SC, n=34) Time to release new variety (yrs, n=33) Perceived decrease in duration varietal registration (% SC, n=34) Time to have seed available (yrs from release, n=33) Perceived change in strictness of seed certification (% SC, n=40) Decrease Same Increase Public support to private seed companies Maize seed sales Price controls Subsidies Sales restrictions 1

Kenya

Tanzania

Uganda

Ethiopia

Overall

2.9 57

2.8 63

2.6 67

2.3 67

2.7 62

2.8

1.3

1.5

2.5

2.1

0 79 21 (+)

10 20 70

50 30 20 +

33 50 17 ++

20 48 33

(R)

(+) +

++ ++

SC = seed companies; ++ = substantial; + = some; () = only selected; R = relief.

did not report any perceived change in strictness of seed certification, whereas in Tanzania most perceived it to have increased. Earlier studies have called for a greater emphasis on regional strategies for public plant breeding and private seed marketing (Tripp and Rohrbach, 2001). The import and export of maize germplasm and seed is generally possible in the sub-region provided the regulatory procedures are followed â&#x20AC;&#x201C; including phyto-sanitary and other bureaucratic procedures (e.g. in Uganda germplasm imports require permission from National Agricultural Research Organization (NARO)). Imported seed also needs to have been registered in the importing country and excludes genetically modified material. Occasionally, additional ad-hoc barriers are imposed. For instance, in Kenya it was reported that in times of domestic seed shortage seed export certificates may not be issued in time, or in times of grain shortage grain exports may be banned and exporters need to prove that the product really is seed. In Uganda, there were reports of earlier restrictions on seed imports/exports, but these have subsequently been eased. In general, imports/exports seem to have benefitted from the regional harmonization, although some non-Kenyan stakeholders reported the difficulty of meeting Kenyaâ&#x20AC;&#x2122;s stringent requirements, resulting in largely one-way seed flows out of Kenya to Uganda and Tanzania. Surveyed stakeholders generally acknowledged the generic public support to private seed companies through the public role in germplasm development/access and certification services. There was however considerable divergence in terms of the perceived specific public support, in part associated with the relative roles played by the public and private sector. Only in Ethiopia was such public support perceived as substantial (e.g. in terms of land access) and to a lesser extent in Uganda (e.g. in terms of training and information). In the case of Kenya public support was perceived to be primarily directed towards KSC. The maize seed sales were perceived to be most heavily regulated in Ethiopia, both in terms of price controls and sales restrictions. In Kenya, only the seed prices of KSC are controlled, but being the largest player this has a marked influence on overall seed prices. Such price controls for the public sector tend to become particularly prominent in times of seed shortage like in the aftermath of a drought. At the time of the survey, subsidy schemes with targeted seed vouchers were operational in both Kenya and Tanzania. In Uganda, no subsidy scheme was operational as such, but there were reports of institutional agents like NGOs distributing free relief seed in selected target areas. Market based voucher instruments are now generally preferred over the distribution of International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

large amounts of free or subsidized seed that undermine seed enterprise development (Tripp and Rohrbach, 2001). However, voucher-based subsidies can still ‘crowd out’ the private sector and be fiscally unsustainable (Smale et al., 2011). In Ethiopia, no subsidies were reported – but the seed sector regulation and price controls did imply the lowest retail prices for maize seed in US$ terms in the study region. There is a general lack of reliable maize seed market data across the region – be it in terms of potential market demand (e.g. maize area and its characteristics like maize ecologies and drought incidence), actual market demand (e.g. maize varietal use and seed renewal) and supply (e.g. maize seed volumes produced/traded). Most of the surveyed stakeholders did not have such information when queried and thus have corresponding difficulties in assessing their relative market shares or identifying market opportunities. Although an individual seed company’s seed volumes may be guarded as a trade secret, the overall maize seed sector would benefit from reliable aggregate market indicators. For instance, whereas formal maize seed tends to be certified – the actual aggregate national volumes of certified maize seed per year are not systematically reported in the public domain nor otherwise easily available. It would thereby be a relatively simple but beneficial step to annually publish the certified seed volumes of maize (and other crops) on the internet. A number of additional market failures hamper maize seed markets in the sub-region. Although some seed companies have developed their own or market proprietary maize varieties, most emerging private seed companies still rely on public maize varieties. With these public varieties being public goods, the companies have limited incentives to invest in marketing/promoting them. Public support and investment in the promotion of public varieties may thus be needed as well as help to stimulate commercial seed production (Tripp and Rohrbach, 2001). Such generic market failures are further compounded by country specific aspects. In the case of Tanzania the public Agricultural Seed Agency was the only authorized producer of foundation seed for public varieties up to the time of the seed sector survey. As a result, the surveyed companies variously complained about the insufficient availability of foundation seed to meet their needs and having their seed portfolios dictated by foundation seed availability and not actual market demand. In the subsequent years some seed companies started to circumvent these restrictions by releasing their own proprietary varieties. The case of Ethiopia is extreme with all but one company producing and emphasizing the same narrow portfolio of public material at the time of the survey (including a few popular public hybrids). At the time any incentive for seed companies to invest in the marketing of their products was further curtailed by Ethiopia’s attempt to centrally compile seed demand through cooperatives/farmer unions and centrally distribute seed accordingly. Central planning of seed markets also can only go so far, with Ethiopia moving from a severe structural shortage of maize seed to a maize seed glut in the aftermath of a program to stimulate seed production in 2009/10 (Atilaw and Korbu, 2011). Such market failures pose serious constraints to the effective development of the private maize seed sector. At the same time they impose significant barriers to the introduction of innovations such as promising new drought tolerant maize germplasm. The surveyed seed companies were queried as to their business model expectations for the next five years (Table 7). The majority of seed companies see the importance of maize in their respective countries increasing – often in relation to population growth and maize being a competitive staple food. Worryingly though, an even larger majority expects the incidence of drought to increase in their respective countries, except amongst those surveyed in Ethiopia who already reported a widespread incidence of drought. The expected drought increase is often associated with climate change and increasingly erratic rain patterns. Reflecting the drought incidence expectations, the majority of seed companies across countries expected drought tolerant maize to play an increasing role in their maize seed portfolio. Indeed, all surveyed seed companies saw drought tolerant maize as an opportunity, be it in relation to agro-ecological changes/opportunities, market opportunities/ farmer demand, maize areas already being drought prone or technological characteristics of drought tolerant maize varieties. However, at the same time nearly all seed companies listed potential constraints linked to drought tolerant maize. The constraints for drought tolerant maize comprised a diverse set, but most widely reported were seed market/regulation issues.

International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

Table 7. Business model expectations for the next five years by surveyed seed companies (% seed companies reporting, 2010). % seed companies Increase in country scenario Importance of maize Incidence of drought Increased role in seed portfolio Drought tolerant maize Hybrid maize Maize (vs other) seed specialization Decrease Same Increase Seed (vs non-seed) specialization Decrease Same Increase

Kenya

Tanzania

Uganda

Ethiopia

Overall

50 75

80 90

63 88

80 0

66 76

83 25

80 90

88 88

80 40

83 60

58 17 25

10 80 10

25 25 50

0 80 20

29 46 26

0 67 33

30 40 30

63 38 0

40 60 0

29 51 20

Seed companies, particularly in Tanzania and Uganda where OPVs still play a prominent role in current portfolios, also expect to increase (or maintain) the role of hybrids in their maize seed portfolio (Table 7). Surveyed seed companies were somewhat divided in terms of their future planned maize seed specialization vis-à-vis other crops (Table 7). Most commonly they expected to maintain the current dominant role of maize seed – particularly in Tanzania and Ethiopia. In Kenya, there was an inclination towards diversifying seed away from maize – reflecting the relatively mature maize seed market. In contrast, in Uganda inclinations were towards increasing maize seed specialization reflecting the rapid recent growth in maize seed markets relative to other seed. The seed companies were also somewhat divided in terms of their future seed specialization vis-à-vis non-seed business, although most commonly they expected to maintain the current seed specialization, except in Uganda where there was a preference towards diversifying with non-seed business (e.g. other agro-inputs such as fertilizer (Table 7)).

6. Discussion: challenges and opportunities for maize seed sector development Seed sector development in Africa has long been a challenge. Indeed up to the turn of the century, progress has often been very limited in spite of substantial investments and assistance (Lanteri and Quagliotti, 1997). The foregoing sub-regional review and comparative analysis thus helps to identify both country specific and wider lessons for seed sector development and some of the associated policy implications. It is now widely and increasingly acknowledged that the private sector can effectively carry out many seed production and distribution activities in the sub-region – but this requires a favorable policy environment, including (1) a clear regulatory framework; (2) fair competition; (3) access to germplasm from national or international research centers; and (4) limits on the distribution of free emergency seed (Minot et al., 2007). Fair competition includes the assurance that private seed companies will not have to compete with subsidized public enterprises – an issue that still hampers the Kenyan and Ethiopian seed sectors and to a lesser extent Tanzanian. Ethiopia also stands out as most reliant on public seed enterprises and having an overly regulated environment that reduces flexibility and business viability as well as raising barriers to entry. It is also increasingly clear though that there is no blue-print to seed sector development in the sub-region and each country would do best in being open minded in assimilating the best-practices appropriate to their respective situation. This calls for an inclusive seed policy debate amongst the numerous stakeholders (Louwaars, 2002). Indeed, debates on technological and/or market dimensions of seed sector development can easily miss-out on the political economy dimensions – whereas an examination of the latter can reshape International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

the terms of the debate and open up alternative pathways to more sustainable and socially just seed systems (Scoones and Thompson, 2011). Political economy dimensions can indeed have far reaching consequences for Africa’s agricultural development – be it in terms of undermining Kenya’s 2004 strategy for revitalizing agriculture (Poulton and Kanyinga, 2014), determining the roles of Ethiopia’s extensive agricultural extension program (Berhanu and Poulton, 2014), and shaping Malawi’s farm input subsidy program (Chinsinga and Poulton, 2014) and southern Africa’s seed industry (Kassie et al., 2013). The overarching vision should be to create an enabling environment for private seed companies to evolve and thrive and thereby service the diverse farmer communities for them to be able to adopt and benefit from existing and future improved maize seed opportunities. Indeed, it is common knowledge that private seed companies pursue market opportunities – with a corresponding bias towards the higher potential agricultural areas (Odame and Muange, 2011). Targeted policies could help incentivize a more comprehensive coverage of lower potential areas, including drought prone maize environments. For instance, policies could strengthen farmers’ capacities to assess new varieties and to be effective consumers of commercial inputs (Tripp and Rohrbach, 2001); policies could support the development of new varieties appropriate for less potential areas; and/or policies could enhance the demand for such varieties through the targeted use of seed vouchers for cash strapped small holders. The seed sector in eastern Africa features unprecedented growth in private enterprise and the popularization of hybrid maize. Still, the contrasting seed sector landscape presents different implications for seed business management (MacRobert, 2009). The domestic market still presents a major determinant for further seed business investments and growth. Kenya thereby presents the most developed seed sector but also is the most crowded. Ethiopia presents the strongest public sector interference. Uganda presents the most liberal sector but it is landlocked with a relatively small maize seed market. Tanzania has the biggest untapped growth potential, but also presents some of the biggest operational challenges. The regional market offers potential beyond the domestic market base, but has been hampered somewhat by the diversity and slow harmonization progress – with Kenya still as the most likely base for seed businesses with regional aspirations. Despite the challenges for seed sector development it is promising to note that new products like drought tolerant maize have successfully entered the market and present growth opportunities across the region (Fisher et al., 2015). Still the contrasting outcomes across the region pose particular challenges. Indeed, Kenya despite having the most developed seed sector and wide-spread improved maize adoption has seen relatively stagnant national maize yields. This shows that maize seed sector development may be a necessary but not sufficient condition for Africa’s maize revolution. Indeed, there is a need to complement the provision and use of improved maize seed with complementary crop management practices and functional market linkages (both input and output markets – e.g. Dixon et al., 2007). In the post-structural adjustment era it is perhaps somewhat surprising that Ethiopia – despite its public sector dominance – has made significant strides in realizing its maize revolution over the last decades (Abate et al., 2015; Zeng et al., 2015). Ethiopia stands out as a maize revolution success story country with increasing national maize yields and with maize having become the largest staple in terms of production – reiterating the need for improved maize seed to be complemented with improved crop management and functional markets.

7. Conclusions Despite the prevalence of maize seed as the core business for seed companies in eastern Africa, the subregional seed sector also shows some remarkable contrasts between the case study countries. Although each country comes from a similar public sector dominance of the seed sector, they have evolved at different speeds and in somewhat different directions and present different seed business management challenges and growth opportunities. Driving the seed industry dynamics are diverging agricultural growth opportunities and varying degrees of regulation, liberalization and restructuring of the seed sector.

International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

Diverging policy implementation between countries has resulted in varying degrees of market concentration and public-private sector roles. In Ethiopia the seed sector remains dominated by the government. Kenya has a well-developed and virtually saturated market, but with a single public enterprise still dominating the national seed market. In Tanzania and Uganda the policy environment has allowed the private seed sector to become the main producer and marketer of maize seed, albeit still largely dependent on OPVs. Despite the regional diversity, there are also some similarities including a proliferation of private seed companies, an increasing emphasis on hybrid maize seed and the emergence of national seed traders’ associations to help organize the increasingly complex and evolving maize seed sector. There is, however, no blue-print to maize seed sector development in the sub-region and each country would do well in being open minded in assimilating the best-practices appropriate to their respective situation. The overarching vision should be to create an enabling environment for private seed companies to evolve in order to service the diverse farmer communities for them to adopt and benefit from existing and future improved maize seed opportunities.

Acknowledgement The present paper draws from work implemented by the authors as CIMMYT staff based in Africa under the Drought Tolerant Maize for Africa (DTMA) initiative which was supported by the Bill & Melinda Gates Foundation (Grant number OPPGDI39O); and also received support from CGIAR Research Program on Maize (MAIZE). The authors gratefully acknowledge all those that contributed to the study, in particular the various seed companies and other stakeholders surveyed and reviewers. We also would like to acknowledge the late Wilfred Mwangi in particular – who provided constructive inputs and guidance but passed away in 2014. The views expressed here are those of authors and do not necessarily reflect the views of the funders or associated institutions, or any of the stakeholders surveyed. The usual disclaimer applies and the authors are responsible for any remaining errors and inferences.

References Abate, T., B. Shiferaw, A. Menkir, D. Wegary, Y. Kebede, K. Tesfaye, M. Kassie, G. Bogale, B. Tadesse and T. Keno. 2015. Factors that transformed maize productivity in Ethiopia. Food Security 7(5): 965-981. Alemu, D. 2011. The political economy of Ethiopian cereal seed systems: state control, market liberalisation and decentralisation. IDS Bulletin 42(4): 69-77. Alemu, D., S. Rashid and R. Tripp. 2010. Seed system potential in Ethiopia: constraints and opportunities for enhancing the system. IFPRI, Addis Ababa, Ethiopia. Almekinders, C.J.M. and N.P. Louwaars. 2002. The importance of the farmers’ seed systems in a functional national seed sector. Journal of new seeds 4(1-2): 15-33. Atilaw, A. and L. Korbu. 2011. Recent development of seed systems of Ethiopia. In: Improving farmers’ access to seed, edited by D. Alemu, S. Kiyoshi and A. Kirub. EIAR – JICA, Addis Ababa, Ethiopia, 13-30. Berhanu, K. and C. Poulton. 2014. The political economy of agricultural extension policy in Ethiopia: economic growth and political control. Development Policy Review 32 (s2): s197-s213. Bishaw, Z., Y. Sahlu and B. Simane. 2008. The status of the Ethiopian seed industry. In: Farmers, seeds and varieties: supporting informal seed supply in Ethiopia, edited by M.H. Thijssen, Z. Bishaw, A. Beshir and W.S. de Boef. Available at: http://edepot.wur.nl/18448. Byerlee, D., and C.K. Eicher. 1997. Africa’s emerging maize revolution. Lynne Rienner Publishers, Boulder, CO, USA. Chinsinga, B., and C. Poulton. 2014. Beyond technocratic debates: the significance and transience of political incentives in the Malawi farm input subsidy programme (FISP). Development Policy Review 32(s2): s123-s150. Cromwell, E., E. Friis-Hansen and M. Turner. 1992. The seed sector in developing countries: a framework for performance analysis. Working paper 65. Overseas Development Institute, London, UK. Dixon, J., J. Hellin, O. Erenstein and P. Kosina. 2007. U-impact pathway for diagnosis and impact assessment of crop improvement. Journal of agricultural science 145(3): 195-206.

International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

Evenson, R.E. and D. Gollin. 2003. Crop variety improvement and its effect on productivity: the impact of international agricultural research. CABI, Wallingford, UK. Fisher, M., T. Abate, R.W. Lunduka, W. Asnake, Y. Alemayehu and R.B Madulu. 2015. Drought tolerant maize for farmer adaptation to drought in sub-Saharan Africa: determinants of adoption in eastern and southern Africa. Climatic Change 133(2): 283-299. GRAIN. 2005. Africa’s seed laws: red carpet for the corporations. Seedling (7): 28-35. Hassan, R.M., M. Mekuria and W. Mwangi. 2001. Maize breeding research in eastern and southern Africa: current status and impacts of past investments made by the public and private sectors 1966-97. Available at: http://tinyurl.com/yd9oflrq. IBRD. 2007. World development report 2008: agriculture for development. World Bank, Washington, DC, USA. Kabeere, F. and E. Wulff. 2008. Seed sector country profile: Uganda. Overview of seed supply systems and seed health issues. Danish Seed Health Centre for Developing Countries, Copenhagen, Denmark Kassie, G.T., O. Erenstein, W. Mwangi, J. MacRobert, P. Setimela and B. Shiferaw. 2013. Political and economic features of the maize seed industry in southern Africa. Agrekon 52(2): 104-127. Kremer, M. and A.P. Zwane. 2005. Encouraging private sector research for tropical agriculture. World Development 33(1): 87-105. Langyintuo, A.S., W. Mwangi, A.O. Diallo, J. MacRobert, J. Dixon and M. Banziger. 2008. An analysis of the bottlenecks affecting the production and deployment of maize seed in eastern and southern Africa. CIMMYT, Harare, Zimbabwe. Langyintuo, A.S., W. Mwangi, A.O. Diallo, J. MacRobert, J. Dixon and M. Bänziger. 2010. Challenges of the maize seed industry in eastern and southern Africa: a compelling case for private-public intervention to promote growth. Food Policy 35(4): 323-331. Lanteri, S. and L. Quagliotti. 1997. Problems related to seed production in the African region. Euphytica 96(1): 173-183. Larson, D.W. and S. Mbowa. 2004. Strategic marketing problems in the Uganda maize seed industry. International Food and Agribusiness Management Review 7(4): 86-93. Louwaars, N.P. 2002. Seed policy, legislation and law: widening a narrow focus. Journal of new seeds 4(12): 1-14. Louwaars, N.P. and W.S. de Boef. 2012. Integrated seed sector development in Africa: a conceptual framework for creating coherence between practices, programs, and policies. Journal of Crop Improvement 26(1): 39-59. MacRobert, J.F. 2009. Seed business management in Africa. CIMMYT, Harare, Zimbabwe. McArthur, J. 2015. Agriculture’s role in ending extreme poverty. In: The last mile in ending extreme poverty, edited by L. Chandy, H. Kato and H.J. Kharas. Brookings Institution Press, Washington, DC, USA, 175-218. Minot, N., M. Smale, C. Eicher, T.S. Jayne, J. Kling, D. Horna and R. Myers. 2007. Seed development programs in sub-Saharan Africa: a review of experiences. Available at: http://tinyurl.com/ydhthy2o. Misiko, M., C. Almekinders, I. Barker, D. Borus, J. Oggema and J. Mukalama. 2011. Kenya: a company, a cooperative and a family. In: African seed enterprises: sowing the seeds of food security, edited by P. Van Mele, J.W. Bentley and R.G. Guei. CABI, Wallingford, UK, pp. 142-155. Morris, M.L. 1998. Maize seed industries in developing countries. Lynne Rienner Publishers, Boulder, CO, USA. Muhammad, L., K. Njoroge, C. Bett, W. Mwangi, H. Verkuijl and H. de Groote. 2003. The seed industry for dryland crops in Eastern Kenya. Available at: http://tinyurl.com/yc2xm6r2. Muhhuku, F. 2002. Seed industry development and seed legislation in Uganda. Journal of new seeds 4 (12): 165-176. Muthoni, J. and D.O. Nyamongo. 2008. Seed systems in Kenya and their relationship to on-farm conservation of food crops. Journal of new seeds 9(4): 330-342. Ngwediagi, P., E. Maeda, H. Kimomwe, R. Kamara, S. Massawe, H.B. Akonaay and L.N.D. Mapunda. 2009. Tanzania report on the state of plant genetic resources for food and agriculture. Available at: http://tinyurl.com/y7q39kx6. International Food and Agribusiness Management Review

Erenstein and Kassie

Volume 21, Issue 1, 2018

Odame, H. and E. Muange. 2011. Can agro-dealers deliver the green revolution in Kenya? IDS Bulletin 42(4): 78-89. Otsuka, K. and D.F. Larsons. 2016. In pursuit of an African green revolution: views from rice and maize farmers’ fields, Natural Resource Management and Policy volume 48. Springer Japan, Tokyo, Japan. Poulton, C. and K. Kanyinga. 2014. The politics of revitalising agriculture in Kenya. Development Policy Review 32(s2): s151-s172. Pray, C.E. and B. Ramaswami. 1991. A framework for seed policy analysis in developing countries. In: Occassional Paper 18. International Food Policy Research Institute, Washington, DC, USA. Rusike, J. and C.K. Eicher. 1997. Institutional innovations in the maize seed industry. In: Africa’s emerging maize revolution, edited by D. Byerlee and C. Eicher. Lynne Rienner Publishers, Boulder, CO, USA, pp. 173-192. Scoones, I. and J. Thompson. 2011. The politics of seed in Africa’s green revolution: alternative narratives and competing pathways. IDS Bulletin 42(4): 1-23. Setimela, P.S., B. Badu-Apraku and W. Mwangi. 2009. Variety testing and release approaches in DTMA project countries in sub-Saharan Africa. International Maize and Wheat Improvement Cente, Harare, Zimbabwe. Smale, M., D. Byerlee and T. Jayne. 2011. Maize revolutions in Sub-Saharan Africa. Policy Research working paper 5659. World Bank, Washington, DC, USA. Timothy, D.H., P.H. Harvey and C.R. Doswell. 1988. Development and spread of improved maize varieties and hybrids in developing countries. Bureau for Science and Technology, Agency for International Development, Washington, DC, USA. Toenniessen, G., A. Adesina and J. DeVries. 2008. Building an alliance for a green revolution in Africa. Annals of the New York Academy of Sciences 1136(1): 233-242. Tripp, R. and D.D. Rohrbach. 2001. Policies for African seed enterprise development. Food Policy 26: 147-161. Van Mele, P., J.W. Bentley and R.G. Guei. 2011. In: African seed enterprises: sowing the seeds of food security. CABI, Wallingford, UK. Van Mele, P., M.A. Ugen, D. Wanyama, R. Anyang, J.C. Rubyogo and L. Sperling. 2011. Uganda: dreams of starting a company. In: African seed enterprises: sowing the seeds of food security, edited by P. van Mele, J.W. Bentley and R.G. Guei. CABI, Wallingford, UK, pp. 156-180. Waithaka, M, J. Nzuma, M. Kyotalimye and O. Nyachae. 2011. Impacts of an improved seed policy environment in Eastern and Central Africa. Available at: http://tinyurl.com/y96jpdla. Walker, T. S. and J. Alwang. 2015. Crop improvement, adoption, and impact of improved varieties in food crops in sub-Saharan Africa. CABI, Wallingford, UK. Wulff, E., L. Bodker and J. Torp. 2006. Seed sector country profile: Kenya. In: Overview of seed supply systems and seed health issues. Overview of seed supply systems and seed health issues. Danish Seed Health Centre for Developing Countries, Copenhagen, Denmark. Zeng, Di, J. Alwang, G.W. Norton, B. Shiferaw, M. Jaleta and C. Yirga. 2015. Ex post impacts of improved maize varieties on poverty in rural Ethiopia. Agricultural Economics 46(4): 515-526.

International Food and Agribusiness Management Review

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2017.0028 Received: 15 March 2017 / Accepted: 26 October 2017

Uncertainty in milk production by smallholders in Tanzania and its implications for investment RESEARCH ARTICLE Edgar E. Twine a, Amos Omoreb, and Julius Githinjic aScientist,

and bSenior Scientist, International Livestock Research Institute, C/O IITA East Africa Hub, P.O. Box 34441, Dar es Salaam, Tanzania

cResearch

Assistant, International Livestock Research Institute, P.O. Box 30709, Nairobi, Kenya

Abstract The study evaluates the impact of risk on enterprises of male, female and young farmers operating in the formal and informal smallholder dairy value chains in Tanzania. It also examines the effect of uncertainty on the decision to invest in milk production in the two value chains. Results indicate that youths in the informal dairy value chain face the greatest level of risk followed by men in the formal value chain, and then men in the informal value chain. Women in both value chains and youths in the formal value chain face relatively low risk. Overall, milk production in the informal value chain is found to be substantially riskier than production in the formal chain. Optimal investment triggers are found to be much larger than the conventional triggers and are sensitive to volatility of returns. The resultsâ&#x20AC;&#x2122; managerial and policy implications for inclusive dairy industry development in Tanzania are highlighted. Keywords: dairy production, risk, investment, real options, Tanzania JEL code: Q12, Q14 Corresponding author: e.twine@cgiar.org

ÂŠ 2017 Twine et al.

Twine et al.

Volume 21, Issue 1, 2018

1. Introduction Generally, farmers face uncertainty in the biophysical and economic environments in which they operate. Uncertainty refers to random events whose probabilities of occurrence are not completely known. A concept that is closely related to uncertainty and one that Antle (1983: 1099) aptly labels ‘the farmer’s perennial problem’, is risk. Risk refers to random events whose probabilities of occurrence can be quantified. Therefore in simple terms, both uncertainty and risk refer to randomness, with uncertainty being a necessary but not sufficient condition for risk (Gough, 1988). In other words, risk is uncertainty with real consequences. When randomness enters a farmer’s optimization problem through, for instance, input prices, output prices, and technology, it renders the farmer incapable of behaving optimally (Antle, 1983). This is because optimality conditions that hold in a deterministic world might not necessarily hold with random variables in the optimization problem, and this could lead to sub-optimal production and investment decisions. Investment decisions in smallholder milk production in Tanzania would likely concern investments in dairy breeds, feeds, proper animal husbandry practices and milk handling. Risk is especially challenging for the resource-constrained or risk-averse farmer that is either excluded from the financial market or operates in an environment devoid of one. This means that insurance against risk is not so much of an option for such a farmer. Hella et al. (2001) and Baker et al. (2015) document the existence of uncertainty and risk in livestock production in Tanzania. The country has a total population of 25.8 million heads of cattle, of which over 95% are reared for both milk and beef. Cattle are considered the most economically and socially important type of livestock. Uncertainty and risk are major concerns particularly for the dairy industry, which is seen as having relatively great potential to reduce poverty, improve nutrition and foster inclusive development. This is because milk production at the household level is for the most part a female preserve (Ministry of Livestock and Fisheries Development, 2016). In addition 30% of livestock’s contribution to agricultural gross domestic product is from dairying, which takes place in five milk sheds, namely, Eastern (Pwani, Morogoro, Tanga), Northern (Arusha, Kilimanjaro, Manyara), Lake Victoria (Mwanza, Mara, Kagera), Central (Tabora) and Southern Highlands (Mbeya, Iringa, Njombe). To ensure that uncertainty and risk do not impede the industry from realizing its potential, there is need to identify and quantify their various sources and recommend farm management and investment strategies that account for uncertainty and risk in the economic environment. Also important is the need to identify risk mitigation strategies that can benefit from public-private investments. An example of where public investment has complimented private investment in mitigating risk is the index-based livestock insurance scheme that insures Kenyan pastoralists against losses due to adverse drought conditions (Chelang’a et al., 2015). There has been virtually no empirical work on the effects of uncertainty and risk on milk production in Tanzania and elsewhere in Africa. Existing literature on the effects of the two phenomena on enterprise profitability and feasibility of investment decisions has largely been in the context of developed-country agriculture. Therefore the study’s contribution to the literature is to fill this information gap and demonstrate the application of the methods to a developing country context. This study has two objectives: the first objective is to identify the risk factors and their economic impacts in Tanzanian milk production, and the second one is to determine the effect of uncertainty on the decision to invest in milk production. The two objectives are related in the sense that the first objective provides parameters relevant to achieving the second objective. Specifically, the study seeks to identify the various sources of risk faced by milk producers, quantify their impact, and generate a single measure of risk in milk production. The study then uses the consolidated risk measure to estimate a risk-adjusted discount rate and hence the optimal investment trigger (Dixit, 1992; Dixit and Pindyck, 1994) if producers are to account for uncertainty and irreversibility of investment in their investment decisions. There are three important considerations in this study. First, to the extent that the government of Tanzania views the dairy industry as being crucial to poverty alleviation and improving food and nutrition security, the analytical approach is intended to provide evidence relevant to inclusive value chain development. Inclusive value chain development is an approach to value chain development that not only focusses on the inclusion of smallholder farmers in value chains, but also recognizes the vulnerability of different categories International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

of smallholder farmers. In Tanzania, the vulnerable are mainly women and the youths (United Republic of Tanzania, 2003). Therefore the study undertakes a disaggregated analysis of the risks that men, women, and the youths face in milk production. Second, the study recognizes the two types of value chains that exist in the Tanzanian dairy industry; the formal value chain where milk is processed and often packaged before selling it to the final consumer, and the informal value chain where milk is sold to the final consumer in its raw form. Producers in the formal value chain sell their milk either directly to milk collection centers or to traders who in turn supply the milk to the collection centers. The centers are operated by individual agents, producer groups, cooperatives, or processing companies. Price discovery mechanisms and relationships between agents are different in the two value chains, and so are the prices and their fluctuations. For instance, although milk prices in the formal value chain are relatively low, they tend to be more stable than prices in the informal chain. This implies different levels of output price risk exposure for milk producers in the two value chains. Therefore for each of the three producer categories, the analysis is undertaken for the two value chains. Third, the study recognizes seasonality in milk production as a permanent feature of the industry in Tanzania. But seasonality per se is not a source of risk. Rather, it is its effects on regularity of feed supply and hence unpredictable fluctuation in some production and price variables within each season that causes risk. In simulating the impact of risk, the study therefore accounts for fluctuations in some of the risk variables during the dry and wet seasons. The rest of the paper is organized as follows: the second section presents a review of previous studies. The primary purpose of our literature review is to gain insights into the theory and empirical methods and how they can be substantively applied to a developing country context. The review is focused on issues pertinent to mathematical modeling and simulation of the effect of risk on a farm enterprise, and modeling the effect of uncertainty on the investment decision. The third section discusses the different sources of risk in milk production in Tanzania. This is followed by a conceptual discussion of the likely differential impact of risk on enterprises of male, female and young milk producers. Section five presents the analytical methods including a discussion of the representative farm approach, and data for examining risk and incorporating uncertainty and irreversibility of investment in the investment decision. Results are presented in section six and the last section concludes the paper.

2. Related literature Modeling the effect of uncertainty and risk on a dairy farm enterprise necessitates accounting for biological and economic factors that affect milk production. In smallholder production systems, environmental factors too are an important consideration since milk production is highly weather-dependent. Seasonality affects prices and quantities of outputs and inputs including non-marketed inputs such as water, forage and some types of fodder. Le Gal et al. (2013) demonstrate how a whole dairy farm simulation model that captures seasonal variation in availability of feed can be used to understand farmersâ&#x20AC;&#x2122; decisions. The model was applied to six smallholder dairy farmers in Brazil. Livestock enterprise simulation studies can be based either on total herds or individual animals or both. Herd-based models consider the entire herd as a single unit and use aggregate herd data. Therefore they are relatively simple to construct and simulate. Individual animal-based models treat and follow each animal in the herd separately. This makes them more realistic but also more computationally demanding than herdbased models. However, in smallholder dairy production systems where farmers typically rear one to five animals (with only one or two animals lactating at any given point in time), individual animal-based models can easily be constructed for the farmerâ&#x20AC;&#x2122;s entire herd. Humphry et al. (2005) used the case of an outbreak of bovine viral diarrhea in a beef herd to compare a stochastic herd-based model with stochastic individual animal-based models. They concluded that although individual animal-based models are structurally relatively

International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

complex, they are theoretically more credible and hence more appropriate for research and decision support than a herd-based model. Applying the individual animal-based model to a smallholder dairy farm situation with limited information on animal performance would likely require drawing upon already empirically established relationships. For instance, in incorporating production risk in smallholder dairying in Tanzania, instead of attempting to separately model changes in body condition scores and daily milk yield, it would, in the absence of data, suffice to consider only the latter since it has been found to be dependent on the former (Msangi et al., 2005). Bewley et al. (2010) constructed a stochastic model for a typical U.S dairy farm to determine the feasibility of investing in precision dairy farming technologies, and the effect of input and output price risk on the costs associated with culling, days open, and disease. Available data enabled them to construct a herd-based model (1000 milking cows) that also contained an individual animal-based module for an average cow. The individual animal module was used to calculate, on a daily basis, several parameters including daily milk yield, dry matter intake, costs and revenues, which were fed into modules for body condition scoring, herd demographics, and retention pay-off. The model was then simulated over a ten-year horizon. Stochastic input and output prices were found to impact costs of culling, days open, and disease. The study is appealing insofar as it is able to integrate an individual animal-based module into a herd-based model. But such an exercise requires availability of good and sufficient data, and attempting to apply it to the context of smallholder dairy farmers in Tanzania is bound to be onerous and yield a level of complexity that would not necessarily accurately capture real farm-level management decisions. Hyde and Engel (2002) analyzed the feasibility of investing in a robotic milking system by a dairy farm in the U.S. They used Monte Carlo simulations to incorporate randomness in milk output, milk price, maintenance costs, the parlorâ&#x20AC;&#x2122;s useful life, inflation rate, and salvage value, considering three herd sizes. The metric used was the breakeven value of the robotic milking system, defined as the maximum amount of money that could be paid for the system given costs of alternative milking equipment and other factors. In two other studies, Engel and Hyde (2003) and Hyde et al. (2003) used the real options approach to analyze the feasibility of replacing a traditional milking system with an automatic one on a dairy farm with 60 cows. They opted for the real options approach rather than traditional capital budgeting techniques in order to account for the uncertainty and irreversibility inherent in such an investment decision. A study that is perhaps the most closely related in spirit to our study is by Tauer (2006). He applied the real options approach to analyze entry and exit decisions of New York dairy farmers. He found the optimal investment trigger (price of milk) to be $17.52 per hundredweight. Our study adopts this approach to analyze the option of a prospective Tanzanian smallholder dairy farmer to enter dairy farming. However, unlike previous studies that have considered output price as the only source of uncertainty, our study considers additional sources of uncertainty. They include input prices, input quantities and output quantity. As in Twine et al. (2016), we use Monte Carlo simulations to obtain a combined measure of uncertainty.

3. Sources of risk in milk production Generally, farm enterprises face two broad types of risks, namely, business risk and financial risk, also known as leverage risk (Unterschultz, 2000). Business risk is risk that arises directly from production and marketing activities of an enterprise and can therefore be sub-divided into production risk and market (price) risk. Financial risk stems from an enterpriseâ&#x20AC;&#x2122;s association with the financial market and it refers to the level of indebtedness of the enterprise. Unterschultz (2000) notes that the two broad types of risks are related in that an increase in business risk could lead to greater indebtedness of the enterprise. Covarrubias et al. (2012) and Twine et al. (2015) have found the incidence of credit to be considerably low among cattle keepers in Tanzania and therefore this study disregards financial risk. However, the results have implications for lending.

International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

Milk producers face both production risk and price risk. Production risk is fluctuation in output and is usually caused by variation in weather conditions, hence variation in availability of water and feed, and variation in animal health status due to diseases. Hella et al. (2001) attribute the highly risky nature of livestock production in the semi-arid region of Dodoma to the large variation in the amount of rainfall. Changes in herd health due to disease can be severe and result in death loss. Swai et al. (2010) estimate dairy cattle mortality rates to be 8.5 and 14.2 per 100 cattle years at risk1 for Tanga and Iringa regions, respectively, and are mainly due to East Coast fever, a tick-borne disease. Ultimately, production risk manifests itself in fluctuations in daily milk yield or milk yield per lactation period, quality of milk produced and herd size. Quality of milk produced and sold also depends on milking and milk handling practices, which could be considered an internal source of risk. Milking and milk handling practices could be dictated by attitudes and cultural norms, but are also likely to vary depending on the cost of inputs used to avoid contamination before, during and after milking. Price risk is fluctuation in output and input prices, with fluctuation in output prices being mostly seasonal. Even though producers are aware of seasonality, the real source of risk is the unpredictable fluctuation in seasonal patterns such as the variation in the onset and duration of different seasons, which may in turn change the variance of prices. Input price risk is associated with the cost of the animal, the cost of labor, animal health services and feed. Compound feeds are in the form of maize bran, cotton seed cake and sunflower seed cake, and their prices closely follow prices of the respective raw materials. Heavy dependence on maize for concentrate feed is a serious concern for the industry because of the large fluctuation in maize prices (Geerts, 2014). Maize production in Tanzania is heavily dependent on rainfall whose variability during the growing season has been found to reduce maize yields (Rowhani et al., 2011).

4. Differential impact of risk The notion that risk has different impacts on different gender groups of milk producers derives from observed differences in the degree of vulnerability of the different groups to external shocks. For instance, although the different sources of production risk are largely external, their impact may be exacerbated by internal risk factors such as the farmerâ&#x20AC;&#x2122;s animal husbandry practices, level of education, access to social networks and resources including credit, and extent of participation in making decisions. These factors influence the degree of vulnerability and tend to vary with gender. Individuals lacking formal education may not know enough about risk to incorporate it into their decision making, and even if they do, they may not be able to institute adequate risk mitigation measures if they are severely constrained by resources. This is the case for Tanzaniaâ&#x20AC;&#x2122;s women and youths, whose vulnerability is attributed to lack of or limited access to land (Tsikata, 2003). Without access to land, the ability of these groups to undertake long-term economic activities that require land is constrained, and so is their access to credit. In addition to lack of or limited access to land, the vulnerability of women entrepreneurs in particular can be attributed to some of the challenges they face in trying to establish their businesses (Mori, 2014): competition for their time between their reproductive roles and business activities, and lack of adequate support from their spouses to formalize their businesses. The agricultural and livestock policy of 1997 explicitly recognized the limited access by women to productive resources, business services (such as credit) and agricultural income, and hence the need to promote their access and that of the youths to land, credit, education and information (Ministry of Agriculture and Cooperative Development, 1997). Given the perceived vulnerability of women and the youths, this study assesses the impact of risk on their dairy enterprises relative to the impact on male-owned enterprises.

This is an epidemiological measure of risk of mortality and is different from the measure used in this empirical analysis.

International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

5. Methods 5.1 The representative farm approach The study relies on representative smallholder dairy farms as the study units. The representative farm approach has been used in previous empirical studies on, inter alia, supply response and adoption of beneficial management practices (e.g. Cortus, 2005; Koeckhoven, 2008; Sharples, 1969; Taylor et al., 1992; Trautman, 2012; Yang, 2009). Ideally, representative farms should embody characteristics of actual farms in the study area, and should be selected based on the factors that have the most influence on the problem of interest (Nuthall, 2011). Some studies have constructed hypothetical representative farms based on data on only the physical characteristics of a farm population. But Nuthall (2011) cautions that such hypothetical units might not be representative of real farms, and that instead, actual farms should be used. According to Plaxico and Tweeten (1963), non-physical factors such as human resources and institutions greatly impact farm organization, efficiency and revenues, but these factors are difficult to quantify and it is difficult to determine their distribution within a farm population. They observe that such factors are usually conveniently ignored in selecting representative farms. In this study, we did not have information on the most important factors affecting uncertainty and risks faced by dairy farmers. As such, we were unable to survey and stratify all farms into homogeneous groups and construct hypothetical farms. Instead, we use real smallholder farms, and we caution that our analysis is normative rather than predictive. The primary criteria for defining a representative milk producer for each gender category were that the producer undertakes commercial milk production and owns the dairy enterprise. In this regard, internal risk due to inability to make decisions regarding the enterprise does not arise. In addition, producers were selected based on their willingness and ability to provide accurate and substantial enterprise data. 5.2 Study area and data The study was undertaken in August 2016 in Lushoto district, located in the northern part of Tanga region in the Eastern milk shed. 75% of the district is covered by the Western Usambara Mountains. The topography allows for only intensive dairy cattle feeding, and therefore the results and conclusions of the study might not apply to milk production in the countryâ&#x20AC;&#x2122;s extensive cattle feeding system. Lushoto district was selected on the basis of having a large number of typical smallholder dairy farmers (Mgeni et al., 2013). They have historically benefited from most of Tanzaniaâ&#x20AC;&#x2122;s smallholder dairy development projects and as a result, they keep improved dairy breeds and milk production tends to be market-oriented. In addition, the district is one of the four districts of the More Milk in Tanzania (MoreMilkiT) project that is led by the International Livestock Research Institute. This research for development project aims to generate evidence on mechanisms for improving the participation of smallholder dairy farmers in input and output markets. Lushoto district is the only MoreMilkiT project district where dairy farmers practice intensive cattle feeding. Farm sizes, herd structures and geography in Lushoto are typical of locations elsewhere in Tanzania where smallholder intensive dairy farms are found such as in the northern zone near Kilimanjaro and Arusha and in southern highlands. It is estimated that some 220,000 households keep about 700,000 heads of improved dairy cattle on such farms in Tanzania currently; typically, about 2 cows and their followers per farm. Data on variables related to milk production cash flows were obtained from representative milk producers in three gender categories: men, women and the youths. The Tanzanian government defines youths as persons from the age of 15 years up to 35 years (Ministry of Labour, Employment and Youth Development, 2007). Following this definition, the study analyzes the dairy enterprise of a male youth as there are hardly any female youths in the study area that own dairy enterprises. Non-youth male and female farmers were above 35 years of age. For each gender category, two representative producers were selected: one producer sells milk into the formal value chain and the other sells into the informal value chain. Therefore data were collected from a total of six dairy enterprises. We estimate the percentage of dairy farms represented by each group using data from the 2014 monitoring and evaluation survey of the MoreMilkiT project sites (International International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

Livestock Research Institute, 2015); the proportions of dairy enterprises individually owned and controlled by men, women, and male youths in Lushoto district are 20, 31 and 6%, respectively. The rest are jointly owned by spouses. Each selected producer provided data (Table 1, with the exception of death loss) on their best performing cow that was lactating or had finished lactating in the last one year. Clearly, the data are typical of a low-input low-output production system. Since data for each category of producers were obtained from a single farm, we examine how reliable they are by comparing them with values found in the literature. But most of the existing information on the variables of interest is not necessarily disaggregated by gender, type of value chain and season. Nonetheless, comparison with aggregate averages enables us to somehow conceptualize some of the variations among the groups analyzed in this study and establish a better context for evaluating our findings. However, keep in mind that for each representative farm, it is the fluctuation in values of its parameters (within and between seasons) rather than their levels that is important in this type of risk analysis. Generally, farms selected for this study are representative of the general population of market-oriented smallholder dairy farms in Lushoto district. For instance, Njehu and Omore, (2014) report average daily yield (ADY) across all producer categories, value chains and seasons in Lushoto district to be 4.2 liters per day, of which an average of 60% was sold. The average yields across seasons for farms in this study appear to be close to this average except for male- and youth-owned farms operating in the formal value chain; theirs are substantially above 4.2 liters and quite close to each other. Dairy farmers in Lushoto district rear crossbred cattle2 and the composition of their feed that includes fodder and forage does not vary significantly. In a study undertaken in North-East Tanzania, Msangi et al. (2005) found milk yield to be mainly determined by the body condition score of cows at calving, which is in turn determined by the ability of the farmer to cut and carry fodder. It is reasonable to expect dairy farmers in the formal value chain to be able to ensure a steady supply of cut fodder given the assured nature of the market for their milk. But this is perhaps not the case for female farmers as they are likely to be relatively more constrained by insufficient land and labor for producing fodder. Also notice that there is a relatively large variation in daily yield across seasons 2

The degree of genetic variation among the animals on these farms is not fully known. The phenotypic expressions that farmers usually rely on for breed selection may not necessarily reflect the true genetic composition of the animals. But in the intensive feeding system, cross-bred cattle are generally products of the exotic (Bos taurus) breeds and indigenous (Bos indicus) breeds (Ojango et al., 2016).

Table 1. Data on some of the parameters used in the cash flow models.1 Parameter

ADY – wet season (liters/day) ADY – dry season (liters/day) % of ADY sold – wet season % of ADY sold – dry season Av. price of milk– wet season ($/liter) Av. price of milk – dry season ($/liter) Av. quantity of feed – wet season (kg/day) Av. quantity of feed – dry season (kg/day) Av. price of feed – wet season ($/kg) Av. price of feed – dry season ($/kg) Av. cost of medicines ($/day) Annual death loss (%)

4.00 3.00 75.00 66.70 0.35 0.35 1.50 1.50 0.09 0.09 0.00 9.20

6.00 2.50 83.30 80.00 0.23 0.23 0.00 0.00 NA NA 0.00 9.20

4.00 2.75 50.00 54.50 0.55 0.55 1.43 1.43 0.12 0.12 0.00 9.20

12.50 4.50 84.00 67.00 0.23 0.23 2.00 1.00 0.20 0.16 0.008 9.20

6.00 6.00 67.00 67.00 0.23 0.35 1.30 0.40 0.15 0.31 0.00 9.20

10.00 9.00 80.00 78.00 0.23 0.23 0.00 0.00 NA NA 0.00 9.20

1 ADY denotes

average daily yield, while FI, FF, MI, MF, YI, and YF denote producer categories and the value chains they operate in as follows: female informal, female formal, male informal, male formal, youth informal and youth formal, respectively. NA denotes not applicable as these farmers did not purchase compound feed over the reference period. However, they did indicate use of compound feed in previous lactation periods. International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

for male and female farmers in the formal value chain; their milk production more than doubles in the wet season. We do not know why this is the case, but we suspect it is caused by a confluence of internal risk and production risk factors. The proportion of ADY sold by five of the six producer categories is higher than the average reported by Njehu and Omore (2014) probably because one of our criteria for selecting representative farms emphasized market-oriented production. Non-marketed milk is consumed by the household, and on some farms, some of it is bucket-fed to calves. Although this milk has economic value, it does not constitute actual cash receipts that directly determine the farms’ liquidity. Therefore for purposes of this analysis, it would not be instructive to include non-marketed milk in calculating farm revenues. However, as shown below, it still is encapsulated in ADY, whose variability we simulate to capture the impact of production risk. Milk prices in each value chain depend on the type of market outlet and are comparable to those in the literature. Njehu and Omore (2014) report average prices ranging from USD 0.34 per liter (received from milk traders) to USD 0.54 per liter (restaurants) in the informal value chain. In the formal value chain, processors pay an average price of USD 0.23 per liter (International Livestock Research Institute, 2015). The type of compound feed purchased by farmers in this study is maize bran. Although it is the preferred feed supplement in Lushoto district (Paul et al., 2016), it is used by only 30% of dairy farmers (Mangesho et al., 2013; Shikuku et al., 2017). These are likely to be the highly market-oriented farmers. Previous studies have not been able to establish a reliable estimate of the quantity of maize bran given to animals in Lushoto district and elsewhere, but it could be anywhere between 0.5 kg (Shikuku et al., 2017) and 1.6 kg per cow per day (Faida MaLi, 2014). Faida MaLi (2014) reports an average price of USD 0.14 per kg. These figures are comparable to those used in this study. The cost of veterinary medicines seems to be relatively small. Baseline information for the MoreMilkiT project indicates that farms in Lushoto district incurred the lowest cost of veterinary drugs estimated at USD 10.10 annually (CGIAR Research Program on Livestock and Fish, 2014). With an average of 1.2 cows per farm, this translates to about USD 6.92 per cow per lactation or USD 0.023 per cow per day. 5.3 Examining risk Following Twine et al. (2016), the impact of risk is examined using a Monte Carlo cash flow model of milk production by a single cow for one lactation period (300 days). A cash flow model illustrates the flow of cash in and out of the dairy enterprise and can therefore be used to predict the enterprise’s financial performance and any imminent financial constraints. Because a cash flow statement reveals the enterprise’s liquidity (ability to pay its bills), it is an important tool for assessing the short-term viability of milk production. An individual animal-based cash flow model is constructed for each representative producer. Conceptually, the potential cash flow for each producer in any given month is calculated as: CF = (P× Q) – CH – (W × X) – ∑OC – OHC – DL (1) where CF is cash flow in $ (USD), P is average price per liter of milk ($), Q is quantity of milk sold in liters, CH is cost of an in-calf heifer ($), W is average price of feed per kilogram ($), X is average quantity of feed, OC are other operating costs ($), OHC are overhead costs ($), and DL is death loss ($). Q in any given month for a given season is calculated as the percentage of ADY sold in that season multiplied by ADY for that season multiplied by 30 days. That is, Q=% of ADY sold×ADY×30. For CH, it is assumed that the animal is purchased with a loan and loan repayment is half of monthly revenues. This is the practice by Covenant Bank, which offers dairy cattle loans to smallholder farmers. All selected farmers reported that OC mainly comprised of the cost of veterinary medicines although only one representative farmer had incurred the cost in the last twelve months. For all farmers, there were zero OHC. Regarding DL, it is not necessarily a cash outflow but because it represents loss in cash inflows in the event of death of the animal, International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

it enables accounting for production risk due to death. Mortality rates are used to calculate the amount of milk lost that would have been sold. Production risk is incorporated in the producer’s cash flow model using DL, fluctuations in ADY, and fluctuations in X. Price risk is captured through fluctuations in P and W. The cash flow model in equation (1) is simulated using Monte Carlo simulation in which triangular distributions are specified for P, ADY, W, X and DL. Values of parameters of the triangular distributions were obtained from the producers. In essence, the variables P, Q, W, X and DL are made stochastic, implying a stochastic rather than deterministic cash flow model. Cash flows are obtained after 10,000 iterations. Cash flow at risk (CFaR) is used to quantify the effect of risk on cash flows in the dry (Jan and Feb; Jun to Sep) and wet (Mar to May; Oct to Dec) seasons. CFaR of the enterprise is defined at a given confidence level, c, as the probability that the future cash flow value, cf, is less than or equal to a given cash flow value CF* and is at most (1-c). As specified in Jorion (2001), P(cf ≤ CF* ) = 1 – c = m (2) It is either the probability, m, for a given CF*or the CF* at a given probability, m. In order to obtain a combined measure of risk from the different sources of risk, we use cash flows to calculate the monthly volatility of returns from milk production, σm. This is the standard deviation of the average monthly return on investment. Following Copeland and Antikarov (2001) and Hull (2005), the annual volatility, σa, is then calculated as: σa = σm × √12 (3) 5.4 Examining investment in milk production The decision to invest in milk production can be analyzed using traditional capital budgeting methods such as net present value, adjusted present value, internal rate of return, modified internal rate of return, accounting rate of return, payback period, and cost-benefit analysis. However, these methods do not account for uncertainty in the economic environment and irreversibility of investment decisions. There is considerable uncertainty in smallholder milk production in Tanzania, which is exacerbated by the fact that investments in milk production are generally sunk costs and hence irreversible. Irreversibility means that once an investment has been made, it cannot be easily reversed; milk production technology is industry-specific, and even if it were not, it would fetch less than its original value on a secondary market. Given uncertainty and irreversibility, waiting to invest until more information becomes available to the decision maker might be of value. Therefore in the face of uncertainty and irreversibility, the decision is not only about whether or not to invest but when to invest. This study employs the real options approach to capital budgeting. The approach emanates from the theory of financial options analysis. Pioneered by Myers (1977), the approach has had a wide range of empirical application as outlined in Lander and Pinches (1998), including natural resources, real estate, manufacturing and research and development. Unlike traditional capital budgeting approaches, the real options approach is able to account for uncertainty and a producer’s flexibility in making investment decisions. For instance, a farmer might alter the timing of an investment in dairy production or the size of a dairy project already in progress as he or she obtains better market information. An option is a decision maker’s right but not obligation to undertake a business transaction. If the transaction involves a tangible asset, the option is real. If the transaction involves a traded asset such as a stock or bond, the option is financial. Essentially, a real option is an opportunity to invest in a physical asset. For instance, a farmer might have an opportunity to invest in milk production if they own unutilized land or are knowledgeable about dairy cattle husbandry and dairy business management. Common types of real options include the option to defer an investment, time to build option, option to adjust the scale of operation, option to abandon an investment project, option to switch, growth options, and multiple interacting options (Trigeorgis, 1993). Usually, a decision maker International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

will tend to exercise the option to delay an investment (also known as the option to wait or deferral option) because of uncertainty. Therefore a deferred investment opportunity is not necessarily a missed opportunity. The three common approaches to analyzing real options are the Black-Scholes option pricing model (Black and Scholes, 1973), the Binomial option pricing model (Cox et al., 1979) and the Dixit-Pindyck model (Dixit and Pindyck, 1994). The Dixit-Pindyck model has been widely applied in agricultural and resource economics studies, and it is the preferred framework for this study. It modifies the Black-Scholes model to analyze investment decisions under irreversibility and uncertainty. The model provides an analytic (closedform) solution to an option pricing problem and as such, it is easy to perform a sensitivity analysis. Studies that have applied the Dixit-Pindyck model include Carey and Zilberman, 2002; Elmer et al., 2001; Engel and Hyde, 2003; Insley, 2002; Isik et al., 2001; Lempiö, 1997; Price and Wetzstein, 1999; Purvis et al., 1995; Richards and Patterson, 1998; Salin, 2000; Steigert and Hertel, 1997; Tauer, 2006. Following Dixit (1992) and Dixit and Pindyck (1994), consider an individual that intends to invest in milk production. Let I denote the sunk cost that they would incur, and V the flow of net operating revenues per unit time that lasts in perpetuity. Uncertainty means that future milk revenues are not exactly known, but in each time period, it is assumed that V follows a geometric Brownian motion3. The individual aims to maximize the expected (average) present value of profits, and therefore future revenues are to be discounted at a positive discount rate, ρ, equal to the opportunity cost of riskless capital. The Marshallian criterion for the decision to either invest now and get V/ρ-I or not investing at all and thus get 0 is that investment should occur (or that the option should be exercised) if V/ρ>I. The prospective investor will be indifferent between investing now and not investing at all if M = ρI

(4)

where M is the Marshallian investment trigger – the borderline level of the current revenue flow. Traditional investment analysis would recommend investing when current flow of revenue exceeds M. At M, waiting is better than either investing immediately or not investing at all, and will remain better for initial values of V slightly greater that M. When current revenue exceeds a certain level, H, investment then becomes optimal. We refer to H as the critical or trigger level of current revenue flows. It is larger than M and it shows that the farmer benefits from waiting for some time before investing. The optimal investment decision can be illustrated graphically (Figure 1) when H is exogenously given, and when it is endogenously determined by the farmer. Both the value of investing immediately (V/ρ-I) and the value of the option to wait are denoted by P, and are plotted against revenues, V. If the project is undertaken yet V=0, then the farmer loses I. As revenues increase, so does the value of investing immediately as shown by the straight line i1i2. The point at which the line i1i2 crosses the horizontal axis is the Marshallian trigger, M. The optimal investment decision when H is exogenously given occurs where the value of the option to wait as given by the convex curve w1w2 intersects i1i2. The value of the option to wait is the segment w1h. Beyond this point, the option to wait has no value. If the investment trigger H is to be optimally determined by the farmer, it has to be increased above the value it had when it was exogenously given. This requires shifting the graph of the value of waiting until it is tangent to the line i1i2 as illustrated by the dotted curve. This is called the smooth pasting condition. It is a condition where the slope of the value of waiting is equal to the slope of the value of investing.

This is a continuous-time stochastic process (also known as a Wiener process or standard Brownian motion) that is exponentiated to ensure that it is always positive. That is, V can trend upward and downward in equal proportions and the distribution of its logarithm is approximately normal (i.e. lognormal).

International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

i2 w2

h w1

-I

H’

Figure 1. Optimal investment decision (adapted from Dixit, 1992). After some calculus and algebra, the optimal investment trigger chosen by the farmer is given as: H = (β / (β – 1)) ρI (5) where β = 0.5 (1 + √(1 + (8ρ / σa)) (6) The optimal investment trigger can be expressed in a manner similar to the Marshallian trigger in equation (4) as: H = ρrI where ρr = (β / (β – 1)) ρ

(7)

is the discount rate adjusted for the value of waiting. It is also known as the hurdle rate. A discount rate of 0.135, which was the Government of Tanzania risk-free interest rate on treasury bonds issued on December 7, 2016 (Bank of Tanzania, 2016) is applied to the model. Other data used to implement the model are obtained from the cash flow model. We examine the sensitivity of the hurdle rate and optimal investment trigger to changes in volatility and discount rate.

6. Results and discussion 6.1 Impact of risk on cash flows Average cash flows and their standard deviations are calculated for each month and are noncumulative across months (Table 2). Positive cash flows are obtained for all producers in each month except for youths in the informal value chain who obtain negative cash flows in the wet seasons. The representative youth reported that in order to induce animals to drink water in the wet season, they feed the animal with more concentrate feed. This does not seem to be economically feasible considering that unlike the other categories of producers, the price the young farmer receives for his milk declines in the wet season. Cash flows in the formal value chain are higher than those in the informal chain except for male milk producers in the dry seasons. Overall, youths in the formal value chain have the largest cash flows in both seasons, and International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

Table 2. Cash flows of dairy enterprises by gender and type of value chain.1 CF ($) Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

5.05 (0.02) 5.67 (0.01) 5.78 (0.22) 4.89 (0.49) 15.03 (2.20) 19.85 (0.05)

8.24 (0.75) 14.18 (0.03) 7.52 (1.22) 20.26 (0.40) -0.63 (4.25) 22.69 (0.05)

5.05 (0.02) 5.67 (0.01) 5.78 (0.22) 4.89 (0.49) 15.03 (2.20) 19.85 (0.05)

8.24 (0.75) 14.18 (0.03) 7.52 (1.22) 20.26 (0.40) -0.63 (4.25) 22.69 (0.05)

FF MI MF YI YF

5.05 (0.02) 5.67 (0.01) 5.78 (0.22) 4.89 (0.49) 15.03 (2.20) 19.85 (0.05)

Figures in parentheses are standard deviations; while FI, FF, MI, MF, YI, and YF denote producer categories and the value chains they operate in as follows: female informal, female formal, male informal, male formal, youth informal and youth formal, respectively; CF = cash flow.

whereas female formal value chain producers have slightly higher cash flows than their male counterparts in the dry seasons, the latter have considerably larger cash flows than the former in the wet seasons. In the informal value chain, youths have significantly higher cash flows than male and female producers in the dry season, but have negative cash flows in the wet season. Cash flows for male and female informal chain producers are comparable. Therefore regarding liquidity, the key finding that could be of concern is that youths in the informal value chain do not feasibly produce milk during the wet seasons. However, their cash flows in the dry seasons seem to be large enough to offset the negative cash flows in the wet seasons. Next is a quantification of the impact of risk on the cash flows of milk producers. This is done by calculating the 20% CFaR values and the probability of obtaining net cash flows that are less than their seasonal averages (Table 3). CFaR values at 20% are a realistic measure that indicates likely losses to the enterprise for one in five chances. At the 20% level, losses are observed only for youths in the informal value chain during the wet seasons; there is one chance in five that a loss of $4.37 or more will occur. The probabilities of cash flows falling below their seasonal averages do not vary much across the different producer categories and seasons. For instance, in the informal value chain, the probability of youthsâ&#x20AC;&#x2122; cash flows being less than their seasonal average is about 45% for both seasons and is nearly the same for male producers in both seasons and for female producers in the wet seasons. In the dry season, the probability increases to about 51% for women. In the formal value chain, the probabilities are slightly higher but quite invariant across seasons; about 51% Table 3. Cash flow at risk values by gender, value chain and season.1 CF at 20% FI FF MI MF YI YF

Prob CF<seasonal average

Dry season ($)

Wet season ($)

Dry season (%)

Wet season (%)

5.03 5.66 5.57 4.61 13.20 19.81

7.55 14.15 6.40 20.06 -4.37 22.64

50.5 50.5 44.2 55.7 44.9 50.5

44.5 50.5 44.6 55.3 44.6 50.5

CF = cash flow; FI, FF, MI, MF, YI, and YF denote producer categories and the value chains they operate in as follows: female informal, female formal, male informal, male formal, youth informal and youth formal, respectively. International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

for youths and female producers, and 56% for male producers. The probability of cash flows falling below their seasonal average suggests insignificant seasonal variation in risk for each producer category and among producer categories in each value chain. In fact an examination of the risk variables with the largest effect on cash flows reveals that for four of the six producer categories, the same risk variable has the largest impact on cash flows in both seasons (Table 3). Holding other factors constant, fluctuation in quantity of concentrate feed given to the animal accounts for the largest variation in cash flows of youths in the informal value chain (Table 4). Availability of concentrate feed varies seasonally because most of it is locally produced from maize. Supplies are low during the wet season when the maize crop is still in farmersâ&#x20AC;&#x2122; fields and they are high in the dry season after harvest. However, young milk producers opt to feed animals with more concentrates in the wet season, a practice that can be avoided. As such, they expose themselves to greater risk. This is a typical case of external risk being compounded by a producerâ&#x20AC;&#x2122;s internal risk factors, which in this case is the producerâ&#x20AC;&#x2122;s husbandry practices. For youths in the formal value chain, death loss is the greatest risk factor. Likewise, death loss is the greatest risk factor for women in the formal value chain, and for those in the informal value chain, it features prominently in the dry seasons. The finding that death loss is a major risk for women and the youths can be explained by the finding of Swai et al. (2010); cattle mortality is lower among farmers that receive training in animal husbandry than among those that do not. Data collected by the authors from a recent survey in the study area shows that a smaller proportion of women and the youths have received training on dairy husbandry than men. Fluctuation in ADY is important for men and women in the informal chain, while fluctuation in feed prices and quantity are important for men in the formal chain. Interestingly, fluctuation in milk prices is not a major source of risk for any of the producer categories. Overall, these results point to the need to tailor risk mitigation measures to individual categories of producers to reflect the specific sources of important risks they face. In order to obtain a better comprehension of the magnitude of risk faced by the different gender categories, a consolidated measure of risk that accounts for all the risks faced by each category of producers is calculated (Table 5). The measure is based on returns to milk production and is calculated on an annual basis. Youths in the informal value chain are found to face the highest annual volatility of returns to milk production of 35.14% compared to only 1.60% obtained for their counterparts in the formal value chain. Men in the formal value chain experience the second highest level of volatility of 10.02% followed by men in the informal value chain (7.90%). Contrary to what was expected a priori, female milk producers in either value chain face relatively low levels of risk. This could be attributed to women generally having more experience in milk production than men and youths as mentioned earlier. In fact Kaliba et al. (1997) found that the probability of women in Tanzania participating in intensive dairy farming was 27.5% higher than that of men. Table 4. Risk variables with the largest effect on cash flows.1 FI FF MI MF YI YF

Wet season

Dry season

ADY (6.80-9.23) Death loss (14.12-14.24) ADY (5.19-9.10) Feed price (19.75-21.01) Feed quantity (-8.17-4.40) Death loss (22.60-22.79)

Death loss (5.02-5.09) Death loss (5.65-5.70) ADY (5.35-6.06) Feed quantity (4.26-5.83) Feed quantity (12.08-16.89) Death loss (19.77-19.94)

Figures in parenthesis are ranges of cash flows in USD; ADY = average daily yield; FI, FF, MI, MF, YI, and YF denote producer categories and the value chains they operate in as follows: female informal, female formal, male informal, male formal, youth informal and youth formal, respectively.

International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

Table 5. Annual volatility of returns to milk production. Youths Men Women Combined

Formal value chain (%)

Informal value chain (%)

1.60 10.02 1.60 4.41

35.15 7.90 4.03 15.69

We now depart from gender disaggregation in order to focus on the value chains as a whole and compute values of parameters necessary for evaluating the effect of uncertainty on investment. Combining all producer categories in each value chain, we find greater risk in the informal value chain than in the formal one, with annual volatilities of 15.69 and 4.41%, respectively. That milk production in the formal value chain is significantly less risky than production in the informal chain is to be expected. Since the mid-1970s when the Government of Tanzania started supporting commercialization of smallholder dairying, emphasis has been on the formal value chain4. In the study area in particular, farmers operating in the formal value chain are relatively well-linked to input and output markets and extension services, and have benefited from donorsupported dairy development programs courtesy of their membership in primary dairy cooperatives. Several of these cooperatives constitute the Tanga Dairies Cooperative Union, a secondary cooperative that owns Tanga Fresh Ltd., the largest dairy processor in the country. Through the companyâ&#x20AC;&#x2122;s projects such as the Modern Dairy Services Network, producers have gained access to risk mitigating services and technologies including information, better dairy breeds, milk collection centers, and credit. The preceding analysis has provided values of parameters (Table 6), except the risk-free discount rate, that are relevant to analyzing the effect of uncertainty and irreversibility on the decision to invest in milk production. The cost of investing in the formal value chain is about a half of the cost of investing in the informal value chain. This is because of the relative ease with which a prospective formal value chain producer is able to access the necessary support from the organizational infrastructure that already exists in the value chain. Moreover, the countryâ&#x20AC;&#x2122;s milk processing capacity utilization is only 26% of total installed capacity mainly because of supply-side constraints. As such, milk processors are generally supportive of smallholder farmers willing to enter the formal value chain. However, the analysis undertaken thus far raises a fundamental question: if milk production in the formal value chain is relatively less risky and investing in the value chain is less costly than investing in the informal value chain, why does the majority of smallholder farmers operate in the informal value chain, supplying 97% of the milk consumed in the country? The answer to this question can best be provided by an analysis of producersâ&#x20AC;&#x2122; risk preferences. Such an analysis, however, is beyond the scope of this study. But disregarding risk preferences, a probable answer lies in the importance that farmers attach to high milk prices given the low-input low-output nature of smallholder milk production. Milk prices received by producers in the informal value chain are higher than (sometimes twice as high as) prices in the formal value chain. And indeed, Rao 4

However, Quaedackers et al. (2009) contend that government support for the development of the formal value chain has been less than sufficient.

Table 6. Data on parameters used in the real options model. Volatility of returns (%) Risk-free discount rate Beta Investment cost ($/liter)

Formal value chain

Informal value chain

4.41 0.135 1.06 0.13

15.69 0.135 1.02 0.27

International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

et al. (2016) have found that most smallholder milk producers prefer marketing arrangements that offer the highest milk price possible to those that do not, even though the latter might have other economically beneficial attributes. 6.2 Effect of uncertainty on the investment decision The real options model yields hurdle rates that are substantially larger than the conventional discount rate (Table 7). The resulting optimal investment triggers of $0.33 and $2.15 per liter of milk for the formal and informal value chains, respectively, are much larger than the Marshallian investment triggers. Therefore owing to the uncertainty that currently exists in the dairy industry, the option to wait to invest in milk production is of value. For the formal value chain, the current price of milk of $0.23 per liter (Table 1) has to increase by $0.10 before waiting to invest ceases to be optimal. This, however, is much less than the increase in price that is needed to make investment in the informal value chain optimal. The current farm gate price of milk in the informal value chain, averaged across the three producer categories, is $0.38 per liter. It would have to increase nearly six-fold to make investing in the informal value chain optimal. Notice that if a prospective milk producer is to disregard uncertainty and go by the Marshallian criterion, they should invest immediately since current farm gate prices in both value chains are way greater than the Marshallian triggers. But anecdotal evidence indicates farmers are reluctant to adhere to the Marshallian criterion. This study was undertaken in Tanga region where the authors were involved in implementing a research-for-development project that supported greater investment in milk production. In the course of project implementation, farmers consistently argued that the milk prices they receive are low and discourage further investment in milk production. These results suggest that the farmers are right and are perhaps aware of the risks and uncertainty they face. 6.3 Sensitivity analysis Sensitivity of the hurdle rate and optimal investment trigger are examined by increasing and decreasing the discount rate and volatility of returns each by 10%. Generally, the hurdle rate and optimal investment trigger are not sensitive to changes in the risk-free discount rate (Table 8). For instance, a 10% increase in the discount rate, holding other factors constant, does not increase the optimal trigger for the formal value chain, and only does so by a mere 0.5% for the informal value chain. However, the two parameters do respond to changes in volatility by nearly the same degree; for instance, a 10% increase in volatility, holding other factors constant, causes a 9.1 and 9.8% increase in the optimal investment trigger for the formal and informal value chains, respectively. Similarly, reduction in volatility by 10% lead to almost proportional reduction in optimal investment triggers. Table 7. Hurdle rates, optimal and Marshallian investment triggers. Hurdle rate Optimal investment trigger ($/liter) Marshallian investment trigger ($/liter)

Formal value chain

Informal value chain

2.47 0.33 0.02

8.11 2.15 0.04

International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

Table 8. Hurdle rates and optimal triggers for different discount rates and volatility levels.

Discount rate 0.149 0.122 Volatility (formal chain) 4.85% 3.97% Volatility (informal chain) 17.26% 14.12%

Formal value chain

Informal value chain

Optimal trigger Hurdle rate

$0.33 $0.32

2.50 2.44

$2.16 $2.15

8.14 8.09

$0.36 $0.30

2.69 2.25 $2.36 $1.95

8.90 7.33

7. Concluding remarks It is well-known that the persistently low level of dairy farm productivity in Tanzania has been perpetuated mainly by insufficient and unstable supply of animal health services, breeding services, and feed (International Livestock Research Institute, 2013). The sources of production risk addressed in this study are directly associated with these three factors, and whereas these factors could be mitigated through policy interventions in output and factor markets, the markets themselves are inherently risky. The end result, as established by this study, is a level of uncertainty that discourages private investment in milk production in both the formal and informal value chains, but more so in the latter. The studyâ&#x20AC;&#x2122;s findings have four major implications for the management of dairy enterprises. First, existing and prospective farm managers should equip themselves with proper knowledge and skills in animal husbandry to be able to adequately deal with the various sources of production risk. For instance, they need to understand the use of body condition scoring as a dairy herd management tool to minimize fluctuations in daily milk yield. Also, given their limited access to animal health services, managers ought to improve their capacity to prevent animal diseases and hence minimize death loss. Second, managers need to ensure feed self-sufficiency. This could be achieved through allocating enough land to produce fodder including maize specifically for maize bran. Third, in the absence of market-based insurance as is the case in rural Tanzania, self-insurance through income diversification would help dairy enterprises cope with negative shocks. Last but not least, since it is impossible to completely eliminate uncertainty in the economic environment, managers should be flexible in decision making. This would require them to have access to market information and knowledge of the implications of alternative production decisions. Milk producers in rural Tanzania have limited access to credit (Twine et al., 2015), another hindrance to achieving greater farm productivity (International Livestock Research Institute, 2013). Although the study disregarded credit risk, the results have implications for lending. The overall finding that the two value chains face significantly different levels of risk exposure means that lenders to smallholder farmers should be cognizant of this fact. The different levels of risk exposure, in conjunction with the individual producersâ&#x20AC;&#x2122; risk profiles, should enable lenders determine the appropriate risk (insurance) premiums and hence interest rates to charge on cattle and other loans. Currently, the practice by lenders that provide dairy cattle loans is to charge uniform risk premiums that account neither for the individual producerâ&#x20AC;&#x2122;s risk profile nor for the riskiness of the value chain they operate in. It is possible for cash flow-based lenders to use cash flow models such as the one developed in this study to determine appropriate risk premiums. From a policy perspective, attaining inclusive dairy industry development will necessitate that the Tanzanian government recognizes that the impacts of uncertainty vary by gender of producer and type of value chain. Therefore assuming that smallholder farmers are risk averse and government is to support implementation of International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

risk mitigation measures, it should ensure that the measures are tailored to the needs of the different gender categories and value chains. An important subject for future research is to examine and compare the distribution of risk preferences of milk producers in the formal and informal value chains. The goal would be to determine which producers are risk-averse, risk-neutral or risk-loving, and what factors influence their risk preferences. The results would be instructive in targeting producers with respect to interventions for risk mitigation.

Acknowledgement The authors gratefully acknowledge support from NWO-WOTRO through the LIQUID project and Irish Aid in Tanzania through the MoreMilkiT project. We also thank two anonymous referees of this journal for their valuable comments.

References Antle, J.M. 1983. Incorporating risk in production analysis. American Journal of Agricultural Economics 65(5): 1099-1106. Baker, D., J. Cadilhon and W. Ochola. 2015. Identification and analysis of smallholder producers’ constraints: applications to Tanzania and Uganda. Development in Practice 25(2): 204-220. Bank of Tanzania. 2016. Financial markets. Available at: http://tinyurl.com/ya6a7pa3. Bewley, J.M., M.D. Boehlje, A.W. Gray, H. Hogeveen, S.J. Kenyon, S.D. Eicher and M.M. Schutz. 2010. Stochastic simulation using @Risk for dairy business investment decisions. Agricultural Finance Review 70(1): 97-125. Black, F and M. Scholes. 1973. Pricing of options and corporate liabilities. Journal of Political Economy 81(3): 637-654. Carey, J.M and D. Zilberman. 2002. A model of investment under uncertainty: modern irrigation technology and emerging markets in water. American Journal of Agricultural Economics 84(1): 171-183. CGIAR Research Program on Livestock and Fish. 2014. The MoreMilkiT project: baseline report. ILRI, Nairobi, Kenya. Available at: http://tinyurl.com/ybpc6vzw. Chelang’a, P.K., R. Banerjee and A. Mude. 2015. Index-based livestock insurance (IBLI) – Lessons in extension and outreach: a case of Wajir County. ILRI Research Brief 39. Available at: http://tinyurl. com/yae7gu8z. Copeland, T and V. Antikarov. 2001. Real options: a practitioner’s guide. TEXERE Publishing Ltd., New York, NY, USA. Cortus, B.G. 2005. The economics of wetland drainage: a case study in Canada’s Prairie Pothole Region. MSc. Thesis. University of Alberta, Edmonton, Canada. Covarrubias, K., Nsiima, L. and Zezza, A. 2012. Livestock and livelihoods in rural Tanzania: a descriptive analysis of the 2009 national panel survey. Available at: http://tinyurl.com/yamxjfhk. Cox, J.C., S.A. Ross and M. Rubenstein. 1979. Option pricing: a simplified approach. Journal of Financial Economics 7: 229-263. Dixit, A. 1992. Investment and hysteresis. Journal of Economic Perspectives 6(1): 107-132. Dixit, A.K and R.S. Pindyck. 1994. Investment under uncertainty. Princeton University Press, Princeton, NJ, USA. Elmer, N.A., A.P. Thurow, J.L. Johnson and C.P. Rosson, III. 2001. An ex ante assessment of investments in Texas grapefruit under uncertainty. Journal of Agricultural and Applied Economics 33(3): 391-401. Engel, P.D and J. Hyde. 2003. A real options analysis of automatic milking systems. Agricultural and Resource Economics Review 32(2): 282-294. Faida MaLi. 2014. Gross margin calculation – Vitii, Lushoto. Faida Market Linkage, Arusha, Tanzania. Geerts, A. 2014. An evaluation of the compound feeds manufactured in Tanzania. MSc. Thesis, University of Reading, Reading, UK.

International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

Gough, J. 1988. Risk and uncertainty. Information paper No. 10. Centre for Resource Management, University of Canterbury and Lincoln College. Available at: http://tinyurl.com/y9afnzj9. Hella, J.P., N.S. Mdoe, G. Van Huylenbroeck, L. D’Haese and P. Chilonda. 2001. Characterization of smallholders’ livestock production and marketing strategies in semi-arid areas of Tanzania. Outlook on Agriculture 30(4): 267-274. Hull, J.C. 2005. Fundamentals of Futures and Options Markets. Fifth Edition. Pearson Prentice Hall, New Jersey, NJ, USA. Humphry, R.W., A.W. Stott and G.J. Gunn. 2005. Modelling BVD at herd level compared with individual animal level. Preventive Veterinary Medicine 72(1): 169-175. Hyde, J and P. Engel. 2002. Investing in a robotic milking system: a Monte Carlo simulation analysis. Journal of Dairy Science 85(9): 2207-2214. Hyde, J., J.R. Stokes and P.D. Engel. 2003. Optimal investment in an automatic milking system: an application of real options. Agricultural Finance Review 63(1): 75-92. Insley, M. 2002. A real options approach to the valuation of a forestry investment. Journal of Environmental Economics and Management 44: 471-492. International Livestock Research Institute. 2013. More milk by and for the poor: adapting dairy market hubs for pro-poor smallholder value chains in Tanzania. A proposal to fund collaborative research and development between the International Livestock Research Institute and R&D partners in Tanzania. ILRI, Nairobi, Kenya. International Livestock Research Institute. 2015. Tanzania dairy value chain: MoreMilkiT project monitoring and evaluation report. ILRI, Nairobi, Kenya. Isik, M., M. Khanna and A. Winter-Nelson. 2001. Sequential investment in site-specific crop management under output price uncertainty. Journal of Agricultural and Resource Economics 26(1): 212-229. Jorion, P. 2001. Value at risk: the new benchmark for managing financial risk. Second Edition. McGrawHill, New York, NY, USA. Kaliba, A.R.M., A.M. Featherstone and D.W. Norman. 1997. A stall-feeding management for improved cattle in semiarid central Tanzania: factors influencing adoption. Agricultural Economics 17(2-3): 133-146. Koeckhoven, S.W.J. 2008. Economics of agricultural best management practices in the lower Little Bow watershed. MSc. Thesis. University of Alberta, Edmonton, Canada. Lander, D.M and G.E. Pinches. 1998. Challenges to the practical implementation of modeling and valuing real options. Quarterly Review of Economics and Finance 38: 537-567. Le Gal, P-Y., J. Bernard and C-H. Moulin. 2013. Supporting strategic thinking of smallholder dairy farmers using a whole farm simulation tool. Tropical Animal Health and Production 45(5): 1119-1129. Lempiö, P. 1997. Farm investments under uncertainty. Research Reports 219, Agricultural Economics Research Institute, Finland. Available at: http://tinyurl.com/y6v97tvm. Mangesho, W., R. Loina, J. Bwire, B.L. Maass and B. Lukuyu. 2013. Report of a livestock feed assessment in Lushoto district, Tanga Region, the United Republic of Tanzania. CIAT, Nairobi, Kenya. Mgeni, C.P., S. Nandonde, A. Omore, A. and I. Baltenweck. 2013. Targeting dairy value chains in Tanzania: process towards benchmark survey. ILRI, Nairobi, Kenya. Available at: http://hdl.handle. net/10568/34340. Ministry of Agriculture and Cooperative Development. 1997. Agricultural and livestock policy, 1997. Dar es Salaam, Tanzania. Available at: http://tinyurl.com/y88ljtyh. Ministry of Labour, Employment and Youth Development. 2007. National youth development policy. Dar es Salaam, Tanzania. Available at: http://tinyurl.com/y9monx9g. Ministry of Livestock and Fisheries Development. 2016. Tanzania livestock modernization initiative. Dar es Salaam, Tanzania. Available at: https://cgspace.cgiar.org/handle/10568/67749. Mori, N. 2014. Women’s entrepreneurship development in Tanzania: insights and recommendations. International Labour Office, Geneva, Switzerland. Available at: http://tinyurl.com/yamxdwwq. Msangi, B.S.J., M.J. Bryant and P.J. Thorne. 2005. Some factors affecting variation in milk yield in crossbred dairy cows on smallholder dairy farms in North-East Tanzania. Tropical Animal Health and Production 37(5): 403-412. Myers, S.C. 1977. Determinants of corporate borrowing. Journal of Financial Economics 5: 147-175. International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

Njehu, A and A. Omore. 2014. Milk production, utilization and marketing channels in Tanga and Morogoro regions of Tanzania. Brief 8, CGIAR Research Program on Livestock and Fish. ILRI, Nairobi, Kenya. Available at: http://hdl.handle.net/10568/64462. Nuthall, P.L. 2011. Farm business management: analysis of farming systems. CAB International, Oxfordshire, UK. Ojango, J.M.K., C.B. Wasike, D.K. Enahoro and A.M. Okeyo. 2016. Dairy production systems and the adoption of genetic and breeding technologies in Tanzania, Kenya, India and Nicaragua. Animal Genetic Resources 59: 81-95. Paul, B., S. Heemskerk, J. Bwire, B. Nzongela, P. Tittonell and J.C.J. Groot. 2016. Potential impacts of increased Napier cultivation in Lushoto, Tanzania. Available at: http://tinyurl.com/y8vxaedr. Plaxico, J.S and L.G. Tweeten. 1963. Representative farms for policy and projection research. Journal of Farm Economics 45(5): 1458-1465. Price, T.J and M.E. Wetzstein. 1999. Irreversible investment decisions in perennial crops with yield and price uncertainty. Journal of Agricultural and Resource Economics 24(1): 173-185. Purvis, A., W.G. Boggess, C.B. Moss and J. Holt. 1995. Technology adoption decisions under irreversibility and uncertainty: an ‘ex ante’ approach. American Journal of Agricultural Economics 77(3): 541-551. Quaedackers, P., V. van der Linden, and D. de Boer. 2009. Organizing milk collection in the Tanzania dairy sector: an analysis of milk collection centres in Tanzania. Dar es Salaam, TAMPA and SNV. Available at: http://tinyurl.com/y8juycgh. Rao, E.J.O., N. Mtimet, E. Twine, I. Baltenweck and A. Omore. 2016. Farmers’ preference for bundled input-output markets and implications for adapted dairy hubs in Tanzania – A choice experiment. 5th International Conference of the African Association of Agricultural Economists, September 23-26, 2016, Addis Ababa, Ethiopia. Available at: http://tinyurl.com/y9a3mxbg. Richards, T.J and P.M. Patterson. 1998. Hysteresis and the shortage of agricultural labor. American Journal of Agricultural Economics 80(4): 683-695. Rowhani, P., D.B. Lobell, M. Linderman and N. Ramankutty. 2011. Climate variability and crop production in Tanzania. Agricultural and Forest Meteorology 151(4): 449-460. Salin, V. 2000. A real option approach to valuing food safety risks. In: Economics of HAACP: costs and benefits, edited by L.J. Unnevehr. Eagan Press, St. Paul, MN, USA. Sharples, J.A. 1969. The representative farm approach to estimation of supply response. American Journal of Agricultural Economics 51(2): 353-361. Shikuku, K.M., R.O. Valdivia, B.K. Paul, C. Mwongera, L. Winowiecki, P. Läderach, M. Herrero and S. Silvestri. 2017. Prioritizing climate-smart livestock technologies in rural Tanzania: a minimum data approach. Agricultural Systems 151(2017): 204-216. Steigert, K.W and T.W. Hertel. 1997. Optimal capacity in the anhydrous ammonia industry. American Journal of Agricultural Economics 79(4): 1096-1107. Swai, E.S., E.D. Karimuribo and D.M. Kambarage. 2010. Risk factors for smallholder dairy cattle mortality in Tanzania. Journal of the South African Veterinary Association 81(4): 241-246. Tauer, L.W. 2006. When to get in and out of dairy farming: a real option analysis. Agricultural and Resource Economics Review 35(2): 339-347. Taylor, M.L, R.M. Adams, and S.F. Miller. 1992. Farm level response to agricultural effluent control strategies: the case of the Willamette Valley. Journal of Agricultural and Resource Economics 17: 173-185. Trautman, D.E. 2012. The economics of beneficial management practices adoption on representative Alberta crop farms. MSc. Thesis. University of Alberta, Edmonton, Canada. Trigeorgis, L. 1993. Real options and interactions with financial flexibility. Financial Management 22(3): 202-224. Tsikata, D. 2003. Securing women’s interest within land tenure reforms: recent debates in Tanzania. Journal of Agrarian Change 3(1-2): 149-183. Twine, E.E., E.J.O. Rao, I. Baltenweck, and A.O. Omore. 2015. Credit, technology adoption and collective action in Tanzania’s smallholder dairy sector. Available at: http://ageconsearch.umn.edu/handle/204198. Twine, E.E., J. Unterschultz and J. Rude. 2016. Evaluating Alberta cattle feeders’ loan guarantee program. Agricultural Finance Review 76(2): 190-210. International Food and Agribusiness Management Review

Twine et al.

Volume 21, Issue 1, 2018

United Republic of Tanzania. 2003. Sub-programme for womenâ&#x20AC;&#x2122;s and gender advancement. Ministry of Community Development, Women Affairs and Children, Dar es Salaam, Tanzania. Available at: http://ihi.eprints.org/494/1/ihi_(105).pdf. Unterschultz, J.R. 2000. Managing market risk in Western Canadian agriculture. Canadian Journal of Agricultural Economics 48(4): 527-537. Yang, D. 2009. Farm level economics of winter wheat production in Canadian Prairies. MSc. Thesis. University of Alberta, Edmonton, Canada.

International Food and Agribusiness Management Review

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2017.0011 Received: 1 February 2017 / Accepted: 20 June 2017

Towards high value markets: a case study of smallholder vegetable farmers in Indonesia RESEARCH ARTICLE Marcus Maspaitellaa, Elena Garnevska b, Muhammad I. Siddiquec, and Nicola Shadboltd aLecturer,

Department of Development Economics, Papua University, Manokwari, Indonesia

bSenior

Lecturer, cPostdoctoral Fellow, and dProfessor Farm and Agribusiness management, Institute of Agriculture and Environment, College of Science, Private Bag 11222, Massey University, Palmerston North 4442, New Zealand

Abstract The expansion of modern markets has significant implications for agriculture in many developing countries that provides both opportunities and challenges for smallholder farmers. The purpose of this paper is to analyse key determinants affecting farmersâ&#x20AC;&#x2122; participation in high value markets, compared to traditional market. Face to face interviews based on a questionnaire were conducted with a sample of 126 smallholder vegetable farmers in the Manokwari region. Binary logistic regression and bivariate correlation analysis were used in this study. The results suggested that age, education level, vegetables cultivated area and membership in farmer groups/cooperatives were the key determinants that had significant effects on the smallholder farmersâ&#x20AC;&#x2122; decision about marketing channel participation. In addition, the income generated from vegetable farming was positively correlated to high value market participation. Some implications that need to be prioritized in agricultural development strategies include improving technical innovations and empowering collective actions through cooperatives or farmer groups. Keywords: Indonesia, high value market, market channel decisions, traditional market, vegetable farmers JEL code: Q13 Corresponding author: e.v.garnevska@massey.ac.nz

ÂŠ 2017 Maspaitella et al.

Maspaitella et al.

Volume 21, Issue 1, 2018

1. Introduction Agrifood systems in developing countries, including Indonesia, are rapidly changing towards high value markets (Imami et al., 2013; Reardon et al., 2012). The development of global food retailers has taken place in this country for more than two decades. Moreover, modern food markets in Indonesia are currently not only being developed in major cities, but have also been mushrooming into provincial cities, and reaching rural and distant communities (Suryadarma et al., 2010). For small-scale farmers, this phenomenon could present better economic opportunities (e.g. increased incomes, productivity and welfare) however it can also bring some challenges (e.g. higher product standards and quality requirements). A number of previous studies have attempted to investigate smallholder participation in modern market channels. While some studies concluded that smallholder farmers would get obvious economic opportunities from being linked to high value markets (Hernandez et al., 2007; Miyata et al., 2009; Rao and Qaim, 2011), other studies found that there were challenges limiting smallholder farmers’ participation (Boselie et al., 2003; Reardon et al., 2009). There have been no clear conclusions about whether smallholder farmers can effectively participate in high value market chains. According to Reardon et al. (2009), in the dual-scale case, modern food markets are likely to source from commercial and large farmers, and exclude smallholder farmers. Linking smallholder farmers to high value markets is crucial for the Indonesian economic development agenda. This is because the majority of Indonesian people depend on agriculture for their living. According to the 2013 Indonesian Agriculture Census, total agricultural households had reached 25.75 million, and 55.33% were engaged in small scale farming activities (Statistics Indonesia, 2014). In addition, approximately 85.14% of smallholder farmers in Indonesia live in rural areas (Ministry of Agriculture, 2014) with associated problems such as limited access to farm assets, infrastructure, markets, and institutional support. These fundamental issues often reduce smallholder farmers’ abilities to escape from poverty. Thus, considering the growth of agrifood market transformation throughout the country, participation of smallholder farmers in high value markets can be a significant alternative for rural development and poverty alleviation strategies. Literature regarding smallholder farmer participation in high value markets in the Indonesian context is still limited. Previous studies have investigated the importance of smallholder farmer participation, and its implication for farmer welfare (Simmons et al., 2005), and changes and consequences of the emergence of modern food retailers for the agricultural supply chain (Chowdhury et al., 2005). However, these studies took place in West Indonesian regions that are more developed in terms of accessibility of production inputs and basic infrastructure, such as transportation, information, and communication technologies. There is very limited information available regarding the linkages between modern food markets and smallholder farmers in underdeveloped regions, especially in the eastern part of Indonesia. Therefore, this study focuses on smallholder vegetable farmers in the Manokwari region, Papua Barat province of eastern Indonesia. In this region, even though traditional market channels are still dominant, modern retail store formats are also emerging rapidly. Modern market formats in the Manokwari region have emerged since the early 2000s, despite being dominated only by home-grown supermarkets and food stores (Ministry of Agriculture, 2014). Since 2010, the modern food retail sector, taking the format of supermarkets, has started growing in the city of Manokwari. This situation brings new opportunities for smallholder farmers to be involved in the growing modern market channels. Therefore, the purpose of this research is to analyse the key factors affecting smallholder farmers’ participation in high value markets, compared to the traditional market in Indonesia and Manokwari region in particular. This study also describes the current situation of Indonesian vegetable growers and markets, and analyses the impact of market participation on farmers’ income.

International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

2. Literature review Various studies have investigated a variety of determinants that affect smallholder farmers’ decisions to be involved in modern market channels (Neven et al., 2009; Reardon et al., 2009; Schipmann and Qaim, 2010). These studies have conceptualised the decisions of smallholder farmers to participate in modern market chains as ‘technology adoption of product marketing’. Schipmann and Qaim (2010) identified three possible aspects/factors that influence farmers’ decision making to participate in high value markets, including the personal and household aspect, the farm and household aspect and the contextual aspect. Personal and household aspects relate to the demographic variables of farmers such as education, age, farming experience and household size (Miyata et al., 2009; Schipmann and Qaim, 2010). The influence of demographic variables, incentives and capacity on smallholder farmers’ participation in high value markets has not been uniforms across different industries and countries (Blandon et al., 2009; Miyata et al., 2009; Neven et al., 2009; Rao et al., 2012). While some studies show that farmers who supply high value markets have a higher education level than traditional market suppliers (Neven et al., 2009; Rao and Qaim, 2011; Schipmann and Qaim, 2010), others found that there was no correlation between level of education and market participation (Hernandez et al., 2007; Blandon et al., 2010). Regarding farmer age, some studies provide information that younger farmers tend to be modern market suppliers (Blandon et al., 2010; Hernandez et al., 2007; Schipmann and Qaim, 2010). However, Neven et al. (2009) claimed that there is no association between age and market channel choice of smallholder farmers. Household was another aspect in determining farmer marketing decision. Hernandez et al. (2007) and Rao and Qaim (2011) found that household size has a negative correlation with the farmers’ adoption of modern market chains. Miyata et al. (2009) however found that household size was not different between modern and traditional market suppliers. Farm aspects include farm size, land ownership and irrigation (Schipmann and Qaim, 2010). Neven et al. (2009) found that smallholder farmers who owned relatively large farms are likely to sell produce to supermarkets. This finding is similar to cases such as sweet peppers in Thailand (Schipmann and Qaim, 2010), vegetable growers in China (Wang, Zhang, and Wu, 2011), and vegetable farmers in Kenya (Ismail et al., 2013). Conversely, in some cases, farm size has no significant effect on the decision of smallholder farmers to participate in high value markets, such as the tomato growers in Guatemala (Hernandez et al., 2007), apple growers in China (Miyata et al., 2009) and fresh fruit and vegetable farmers in Honduras (Blandon et al., 2009). Having a larger farm area allows farmers to cultivate larger crop areas for selling to modern market chains. The influence of irrigation on the marketing decision of smallholder farmers is found to be various in different studies. Hernandez et al. (2007) state that irrigation technology applied by smallholder tomato growers in Guatemala correlated to the decision to participate in modern market supply chains. This is similar to the study of vegetable farmers in Kenya (Neven et al., 2009), indicating that the irrigation infrastructure has a significant effect on market channel adoption. Conversely, studies of Miyata et al. (2009) in China, Blandon et al. (2009) in Honduras, and Rao et al. (2012) in Kenya found that irrigation technology had no influence on farmers’ decisions to participate in high value market chains. Contextual aspects relate to access to services and road conditions (Schipmann and Qaim, 2010). In the developing countries, the distance to marketplaces is also an important factor for farmers in terms of product delivery. Some studies have looked at how the distance or location of a farm can encourage smallholder farmers to participate in high value market chains. Miyata et al. (2009) found that distance is a strong explanatory variable determining smallholder farmers’ decisions to participate in such chains. Smallholder producers who live near the major village significantly tend to sell to high value markets. Similarly, Rao and Qaim (2011), incorporating farmers’ access to the main road as a predictor in their analysis, found that it gave an advantage for vegetable producers to supply supermarkets that demand a stricter schedule of delivery.

International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

Reardon et al. (2009) and Pascucci (2011) provide alternative frameworks categorising the determinants of market decisions into (1) the incentives in the modern market channels, and (2) the capacity of smallholder farmers to adopt the technology. Regarding the incentive factors, there are two aspects that should be considered by smallholder farmers. The first aspect relates to the net premium prices paid by high value markets, which are relatively higher than the price paid by wholesalers in traditional market channels (Reardon et al., 2009). For example, Neven et al. (2009) found that supermarkets in Kenya paid horticultural suppliers about 10-20% more than what they got in the traditional market. Likewise, nearly 60% of smallholder vegetable producers supplying supermarket chains in Honduras received higher prices than from traditional market channels (Blandon et al., 2009). The second aspect of incentive factors is the relative risk and cost. Reardon et al. (2009) emphasised that farmers should also consider the possibilities of risk and the cost of farm production and post-harvest handling technologies to deal with the quality and transactional requirements needed by modern market channels. Blandon et al. (2009) included the farmers’ perception of risk as an independent variable in the farmers’ participation model, and revealed that the perceived risk of low quality causing product rejections significantly influenced smallholder farmers’ market decisions. However, the perceived risk of product losses due to bad weather or pests was not found to be an important factor. Moreover, smallholder farmers often experience additional costs derived from barriers of entry to high value markets. Reardon et al. (2009) highlighted that these costs reduce smallholder farmers’ choice of participation in supermarket channels. In addition, transaction costs derived from poor transportation and communication conditions can also affect smallholder farmers’ adoption of modern market channels (Rao and Qaim, 2011). Smallholder farmers living further from urban areas and cities, with poor access to transportation and communication, face high additional costs and are less likely to be offered contracts by modern food markets (Barrett et al., 2011). The second set of determinants of farmers’ marketing decision is farmers’ capacity. The capacity variables refer to investments of various forms of capital by farmers to access high value markets, including physical farm assets, collective capital, and institutional capital (Reardon et al., 2009). Physical capital can include land and non-land assets, such as equipment and irrigation that is needed to meet quality and consistency requirements of the high value markets. The emergence of new procurement practices of high value markets forces actors along the supply channels, including farmers, to make investments in social or collective capital. Farmers’ organizations or cooperatives can also play crucial roles in facilitating smallholder farmers to gain access to modern markets by investing in collective capital such as warehouses and vehicles (Pascucci, 2011; Reardon et al., 2009). These collective investments can help smallholder farmers to reduce transaction costs (Hellin et al., 2009). The involvement in farmer organizations provided a higher chance for smallholder vegetable growers to access supermarket channels in Kenya (Ismail et al., 2013). However, the case of tomato growers in Guatemala (Hernandez et al., 2007) provided an opposite finding. The effect of farmer organization was significant, but negative. In this case these organizations were not marketing organizations, but just provided technical assistance and training. Furthermore, smallholder farmers also need to invest in institutional capital. This capital is associated with the embodied relationships between farmers and institutions such as companies, non-government organizations (NGOs) and the government (Reardon et al., 2005). Previous studies have also identified several key challenges for farmers’ participation in high value markets such as limited access to market information, poor basic infrastructure (transport and communication) in rural areas, low bargaining position due to the low volumes of outputs supplied and lack of physical, financial and human capital (Barrett et al., 2011; Berdegue et al., 2005; Blandon et al., 2009; Irianto and Herwanto, 2009; Neven et al., 2009; Reardon et al., 2009). These barriers, contributing to the exclusion of smallholder farmers, can vary from case to case. International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

In line with the literature, the framework used in this study captured farmers’ demographics (age, education, experience, family size), farm characteristics (farm size, irrigation, livestock ownership), marketing characteristics (average price, income, market information, transportation, distance to supermarkets) and institutional characteristics (access to credit, extension services, farmer’s organization membership) as independent variables in analysing the determinants affecting market channel choice of smallholder vegetable farmers in eastern Indonesia. The theoretical framework is shown in Figure 1.

3. Methodology 3.1 Data and methods This study utilised a quantitative approach to achieve the research aim. A structured survey was used as the primary data collection that was carried out over the period September to October 2014. The survey was conducted in three sub-districts; Prafi, Manokwari Selatan and Manokwari Barat (Figure 2). These subdistricts were selected because of their high production of vegetables complemented with agro-climatic conditions that were favourable for growing an array of non-perennial vegetables.

Farmer demographics

Farm characteristics

Market channel choice: • Traditional market channel • Supermarket channel

Marketing characteristics

Institutional characteristics

Figure 1. Theoretical framework.

Figure 2. Map of Manokwari region showing the study sites. International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

Therefore, a two stage non-probability sampling method was used for data collection i.e. purposive and convenience sampling techniques. Purposive sampling helped identified the sub-districts where high production of vegetables was concentrated. Within the selected sub-districts, convenience sampling technique was used for the selection of respondents (Poole et al., 2003; Siddique and Garnevska, in press; Teddlie and Yu, 2007). This technique builds a sample on the basis of finding convenient or available respondents in the studies districts (Ruane, 2005). The main consideration for that was the unavailability of the population of vegetable farmers. However, with convenience sampling there is an issue of non-response bias that can be addressed through several ways like pilot testing, pre-scheduled meetings, long field times, and visiting in the fields/ work place (Fogelman and Comber, 2007). A pilot study, scheduled meetings and visiting respondents at their work place helped reduced non-response bias for this study. The face to face interviews, based on a structured questionnaire, was administered during visits to vegetable farms or farmer homes in the three sub-districts of east Indonesia. The data was collected in a short interval of time, from September to October 2014, and a total of 135 respondents were interviewed. The data was incomplete for 9 respondents and were removed and a total of 126 smallholder vegetable farmers’ data were included in the final analysis. It included both types of farmers i.e. supplying high value markets/supermarkets and farmers supplying traditional markets. The sampling technique, small sample size, limited time and resources are the limitations of this study; however, it still provides meaningful insight about the farmers’ participation in modern supply chains in Indonesia and Manokwari region in particular. These limitations of the study also render its scope, lack of generalization and essentially make it area specific. Data was analysed using the Statistical Package for Social Sciences (SPSS version 22, IBM, Armonk, NY, USA). Initially a cleaning process was performed to ensure its completeness and validity. This process included checking for logical inconsistencies, outliers and missing values. In order to avoid these data problems, the values of means and standard deviations of variables were produced. Based on these values, there was no missing values found, but some outliers were identified. The outliers were treated by replacing them with the mean values of each variable. This study utilized binary logistic regression analysis to examine potential factors affecting farmers’ decisions about market channel participation. A binary logistic regression is a type of regression models in which the dependent variable is a categorical dichotomy that takes only two values; zero and one (Wooldridge, 2013). 3.2 Model specification and variable description The binary logistic regression model was used for the analysis since the probability of farmer responses were assumed as a binary choice due to the availability of only two marketing channels i.e. traditional and high value marketing channels. The dependent variable measured the choice of market channels (either modern supermarkets or traditional markets), while a set of independent variables were derived from farmer demographics, farm characteristics, marketing and institutional factors. These variables included in the analysis were in line with the literature and were also pre-tested before final data collection. Initially, the number of variables was more than what specified in the model. A pre-testing of questionnaire with all these variables was conducted with 8 respondents whose response was not included in the final analysis. A number of variables like gender, off-farm employment, total land, means of transportation, etc. were removed after pre-testing. All the respondents in pre-testing identified and responded to the following variables that were included in the final model showing that these variables play a decisive role in the choice selection between traditional and high value markets. The empirical model for analysis in this research can be expressed as:

Prob( MC  1 | x)   0  1 AGE   2 EDU   3 EXPRNCE   4 FMLY _ SZ   5VEG _ AREA   6 IRR   7 LVSTOCK   8 DIST _ SM   9TRNS _ COST  10 AV _ PRICE  11EXTN  12CREDIT  13MKT _ INFO  14 FARMER _ GR  

International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

The farmers’ demographic variables included age of the farmers, education level, farming experiences, and number of family members. Farm characteristic variables included vegetable farm size, irrigation system and livestock ownership. Marketing aspects included in the model were distance from vegetable farms to supermarkets, estimated transportation cost, and average price received by farmers. Institutional factors included farmers’ attendance of farming extension service, access to credit, access to market information, and farmer group membership (Table 1). ‘Age’ of the farmer represented the age of the vegetable farmer in years. It was claimed that younger farmers were expected to be more adventurous and more risk takers than older farmers. Thus it was expected to be negatively correlated with high value market participation. Education of the farmer that was measured in years of schooling was expected to have a positive effect on supermarket participation (Gong et al., 2007). ‘Farming experience’ was predicted to have a positive influence on modern market participation (Ouma et al., 2010; Shiimi et al., 2012). ‘Family size’ was predicted to be negatively associated with high value market participation (Balint and Wobst, 2006; Hernandez et al., 2007; Rao and Qaim, 2011). These studies argued that larger households tend to have more dependents and their production activities might be more subsistence oriented. ‘Vegetable area’ was hypothesised to have a positive influence on the marketing decision to sell at modern markets. Having larger cultivated areas could allow the household to have a surplus in production and be in position to sell (Balint and Wobst, 2006). ‘Irrigation’ was essential for commercial agriculture. Having irrigated area could increase farmers’ possibility to participate in high value markets. This variable was set as a dummy variable taking the value of one if the farmer had irrigation system and zero otherwise. It was expected that irrigation has a positive effect on the dependent variable (Hernandez et al., 2007; Neven et al., 2009). ‘Livestock ownership’ was set as a dummy variable which took the value one if the household owned livestock, or zero otherwise. It was predicted to have a negative correlation with modern market participation (Hernandez et al., 2007). Table 1. Variable definition, unit of measurement and expected signs. Variable code

Variable name

Measurement

Dependent variable MC Market channel participation Independent variables AGE Age of farmer EDU Education of farmer EXPRNCE Farming experience FMLY_SZ Family size VEG_AREA Vegetable cultivated area IRR Irrigation LVSTOCK Livestock ownership DIST_SM Distance to supermarket TRNS_COST Transportation cost AV_PRICE Average price received by farmers EXTN Attendance of extension meetings CREDIT Access to credit MKT_INFO Access to market information FARMER_GR Membership of farmer groups

Expected signs

1 supermarket, 0 traditional market Number of years Years of schooling Number of years Numbers Hectares 1 if yes, 0 otherwise 1 if yes, 0 otherwise Kilometers IDR1 IDR 1 if yes, 0 otherwise 1 if yes, 0 otherwise 1 if yes, 0 otherwise 1 if yes, 0 otherwise

– + + – + + – – – + + + + +

1 IDR = Indonesian Rupiah. 1 US Dollar equaled to approximately 12,200 IDR = 1 US, calculated on the basis of the exchange rate

on October 2014; http://tinyurl.com/pbxmqku.

International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

‘Distance to supermarkets’ was expected to exert a negative effect on high value market participation since it related to transaction costs that farmers would pay (Hernandez et al., 2007; Miyata et al., 2009). Another variable relating to distance from farm to marketplace was ‘transportation cost’. This variable was expected to have a negative effect on supermarket channel decision (Shiimi et al., 2012). ‘Average price’ of vegetables received by farmers was hypothesized to influence modern market channel positively (Alene et al., 2008; Balint and Wobst, 2006). ‘Attendance of extension meetings’ was set as a dummy that took the value one and zero otherwise. The expected sign of this variable was positive (Jari and Fraser, 2009). ‘Access to market information’ was measured through the ability of the farmer to access market information and to comprehend it. This variable was allocated dummy values taking the value one if a farmer had access to market information and zero otherwise, and was expected to be positively associated with the dependent variable (Jari and Fraser, 2009; Ouma et al., 2010; Panda and Sreekumar, 2012). ‘Access to credit’ provides financial capital that might encourage farmers to participate in vegetable market channels. This variable was set as a dummy variable, and hypothesized to have a positive effect on high value market channel (Shiimi et al., 2012). ‘Membership of farmer groups’ can enable farmers to create economies of scale in production and to compete effectively in markets. This variable was set as a dummy variable, and hypothesised to influence supermarket participation positively (Alene et al., 2008; Shiimi et al., 2012). Finally, bivariate correlation analysis was conducted to test the correlation between market channel participation and vegetable income to explain the effect of supermarket channels on farmer’s income.

4. Results and discussion 4.1 Descriptive analysis The majority of the respondents (about 70%) from the sample were over 35 years old and had completed at least elementary school. About 60% of the respondents had less than 10 years’ experience in farming with an average of 13 years. The majority (over 85%) of the sample had four or less than four people in their family and relied on incomes from farming. There were no differences between the demographic characteristics of the farmers supplying different market channels except the result that one third of the farmers supplying traditional markets were less than 35 years old (Supplementary Table S1). Over 80% of the respondents had farm size of less than 1 ha (over 60% had less than 0.5 ha) and did not have access to any kind of irrigation systems. Over 60% of the respondents grew water spinach as their main vegetable, followed by long yard beans and vegetable amaranth, at 45.2 and 44.7%, respectively. Respondents (over 50%) supplying high value markets had a farms size of over 1 ha (Supplementary Table S2). About 40% of the sample had incomes over 6.500 IDR per kilogram of vegetable sold, transported their produce over 40 km on their own motorcycle. The majority of sampled farmers (77%) spent less than 500,000 IDR for transportation per year. Farmers who were supplying high value markets (75%) paid over 500,000 IDR for their transportation cost per year and 75% of them received over 6.500 IDR per kilogram of vegetable sold (Supplementary Table S3). The majority of the interviewees did not have access to credit (over 90%), extension services (74%), or market information (90%). About 45% of them were members of cooperatives and other farmers’ organisations. A significant share (about 80%) of the growers supplying the high value markets were members of cooperatives and as a result had better market information (Supplementary Table S4).

International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

4.2 Factors affecting market channel participation: binary logistic regression analysis Table 2 presents the results of the binary logistic regression estimating the factors that influence the marketing channel participation of smallholder farmers. The table showed the estimated coefficients (B), standard error (S.E.), significance value (sig.) and odds ratios of the explanatory variables in the model. According to Gujarati (2004), the coefficients (B values) estimate the change probability of the dependent variable for a unit change in the corresponding predictor, other predictors being equal. The sign of the coefficient values indicates the direction of influence of the explanatory variable. A positive sign, therefore, implies an increase in the likelihood of changing from selling through the traditional marketing channel to selling through the supermarket channel. The results showed that the model was highly significant in estimating the factors influencing farmers’ choice of vegetable marketing channels (P-value of 0.000). Among the explanatory variables, age, education level, vegetable farm area and farmer group participation were statistically significant in determining supermarket channel participation. However, farming experience, family size, irrigation methods, livestock ownership, distance to supermarket, transportation cost, average price, attendance of extension meetings, access to credit, and access to market information were not statistically significant. Furthermore, the signs of the estimated coefficients of some independent variables were consistent with the a priori expectations whereas others were contrary to expectations. The variable ‘age of farmer’ was found to significantly influence modern market participation. The beta coefficient of this variable was 0.059, with an associated P-value of 0.064. Unexpectedly, the effect of age on supermarket channel participation was positive. This positive relationship was contradictory to the Table 2. Binary logistic regression results on market channel participation.1,2 Variables

S.E.

Sig.

Exp(B)

AGE EDU EXPRNCE FMLY_SZ VEG_AREA IRR LVSTOCK DIST_SM TRNS_COST AV_PRICE EXTN_SERV CREDIT MKT_INFO FARMER_GR Constant Number of observations Log likelihood Chi square (14) Sig. chi square Nagelkerke R2 % correct predictions

0.059 0.239 0.014 -0.288 2.429 -0.152 0.344 0.036 0.000 0.000 -0.279 1.556 0.097 2.289 -10.417 126 82.209 46.079 0.000 0.480 84.1

0.032 0.116 0.035 0.230 1.227 0.937 0.701 0.036 0.000 0.000 0.729 1.134 1.131 0.748 2.771

0.064*** 0.039** 0.684 0.211 0.048** 0.871 0.624 0.308 0.273 0.183 0.702 0.170 0.931 0.002* 0.000

1.061 1.270 1.014 0.750 11.351 0.859 1.410 1.037 1.000 1.000 0.756 4.741 1.102 9.865 0.000

B = coefficients; S.E. = standard error; sig. = significance value; Exp(B) = odds ratio. Significant at the 1% level; ** Significant at the 5% level; *** Significant at the 10% level.

International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

studies done by Hernandez et al. (2007) and Alene et al. (2008). These studies reported that due to the reluctance of older farmers in adopting new technologies, younger farmers were more likely to participate in supermarket channels. The odds ratio of 1.061 supported that when farmers were more mature, the possibility of participation in supermarket channels was higher than in traditional market channels. This may correlate with how long farmers are involved in marketing relationships. Older farmers may have repeat contracts gained through long-term marketing relationships, which can enhance trust between farmers and their main buyers including supermarkets. The level of education of sampled farmers had a positive effect on high value market participation. The beta coefficient of this variable was 0.239, with a significance P-value of 0.039. This explained that better education for smallholder farmers may result in households shifting from selling through traditional market channels to selling through supermarket channels. This result was in line with the previous studies conducted by Neven et al. (2009), Rao and Qaim (2011), and Ismael et al. (2013) which concluded that farmers who supply high value markets have a higher education level than traditional market suppliers. The result, however, did not coincide with the findings of Hernandez et al. (2007) who reported no significant effect on supermarket channel participation was made by the education background of the farmers in Guatemala. Although most vegetable farmers in the Manokwari region were categorised with a low level education as presented in the descriptive analysis, the odds ratio value (Exp(B)=1.270), however, showed that vegetable farmers were likely to choose supermarket channels with an increased level of education. More educated farmers were expected to have a better understanding not only of the production process, but also of marketing and business aspects, such as supply requirements and price negotiations. Another variable that significantly affected supermarket channel participation was vegetable farm area (hectares). The outcome showed that the coefficient of this variable (2.429) had a significance P-value of 0.048. The positive sign on its coefficient indicated that an increase of farm size may result in a higher probability for smallholder farmers to participate in modern market channels. This positive relationship was consistent with the a priori expectations, which also confirmed the results of various studies done in other countries by researchers such as Balint and Wobst (2006), Neven et al. (2009), Schipmann and Qaim (2011) and Wang et al. (2011). Furthermore, despite the fact that the average vegetable farm size in the Manokwari region was relatively small, the larger odds ratio of 11.351 indicated that when there was a unit (hectare) increase of land area under vegetables, the probability of participation in supermarket channels would increase about 11.2 times greater than in traditional market channels. The results also showed that membership of cooperatives or farmer groups had a significant influence on vegetable marketing channel choice. The coefficient value of this variable was 2.289, with the significance level (P-value) of 0.002. The positive relationship between farmer group membership and market channel participation was consistent with the a priori expectations, and supported previous studies (Blandon et al., 2009; Jari and Fraser, 2009; Panda and Sreekumar, 2012). The larger value of the odds ratio (9.865) provided evidence that when farmers participated in cooperatives/farmer groups as members, there was a higher possibility of participating in high value market channels. The plausible explanation to this was that through the farmer groups, individual farmers had access to technical assistance, market information and training that enabled them to meet production thresholds for market participation and increase marketed supply. Farming experience had a non-significant effect on market channel participation. So farmers who had longer farming experience did not necessarily have a higher possibility of selling through supermarket channels. This result did not support the previous studies such as Gong et al. (2007), Ouma et al. (2010) and Shiimi et al. (2012) who concluded that experience of farming was a strong explanatory variable in determining farmersâ&#x20AC;&#x2122; participation in high value markets. Moreover, this study also found that the effect of family size on market channel participation was statistically not significant. This confirmed the findings of Gong et al. (2007), Hernandez et al. (2007) and Neven et al. (2009) that the number of family members had no relationship with what market channel farmers participated in.

International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

The variable ‘Irrigation methods applied’ demonstrated a negative and statistically non-significant effect on marketing channel participation. This finding did not support the previous results reported by Hernandez et al. (2007) and Neven et al. (2009). As most sampled farmers run their vegetable farms on a small scale, it would therefore be costly for them to apply advanced irrigation systems. The effect of livestock ownership on modern market participation was found positive, but not significant. The explanation to this may be associated with the fact that most farmers included in the sample run their livestock farms in a relatively small scale, for the purpose of household consumption. Therefore, whether smallholder farmers had livestock or not, it would not affect their participation in modern market channels. All marketing aspects included in the logistic regression model demonstrated non-significant effects. Distance to supermarkets had a positive, but not significant influence on market channel participation. This explained that the location of vegetable farms was not an important factor determining supermarket participation, despite some respondents expressing the location (in terms of distance to big cities) of farms as a problem limiting supermarket channel participation. The result was contradictory to the study of Miyata et al. (2009) emphasising that the distance variable was a strong explanatory variable, affecting farmers’ decisions to participate in high value markets. Similarly, transportation costs had a positive but not significant effect on marketing channel participation. This finding did not confirm the result of the study done by Shiimi et al. (2012) who concluded that transportation costs had a negative influence on market channel decision to sell to supermarkets. This was probably because the studied districts have relatively good roads and adequate public transport that facilitated farmers to have contact with marketplaces in the city of Manokwari including supermarkets. Average prices of vegetables received by farmers had a positive, but not significant, influence on modern market channel participation. This relationship did not coincide with the previous studies conducted by Balint and Wobst (2006) and Alene et al. (2008). These studies concluded that the relative prices that farmers received for the agricultural produce they sold could motivate them to increase their participation in supermarket channels. Institutional factors included in the model demonstrated non-significant effects on market channel participation, except the variable ‘membership of farmer groups’. The results showed that the attendance of extension meetings had a negative and non-significant effect on high value market channel participation. This variable was found to be inconsistent with the study by Alene et al. (2008), who found that extension services played an important role in encouraging smallholder farmers to participate in the supermarket channels. Access to credit had a similar effect on market channel participation, which was positive and non-significant. The plausible reason to this relationship was that most smallholder farmers did not have access to financial institutions such as banks. Some respondents borrowed some amount of money from other farmers or family, but not for investing in farming. This result was also similar to the studies of Rao and Qaim (2011) and Ismail et al. (2013). These studies found that access to credits had a non-significant effect on supermarket channel participation. The variable access to market information had a positive and non-significant effect on modern market channel participation. This finding was contradictory to the studies of Jari and Fraser (2009), and Panda and Sreekumar (2012), who found that access to market information, increased the possibility of smallholder farmers participating in high value markets. The plausible reason for this was probably because there was no viable market information service in the research area. In some cases, farmers had to find information regarding prices and new marketplaces by asking the local traders or going physically to local assembly markets. In the Manokwari region, food price information was often published by the local government through the radio and newspapers, and only for major items such as rice, sugar, and some vegetables such as cabbages and chilies. Overall, the binary logistic regression has provided information regarding key variables that significantly influenced smallholder farmers’ decisions about vegetable market channel participation. Of the farmers’ demographics, only education was the most important factor affecting their decision to sell to supermarket International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

channels. Farm size, representing farm characteristics, also became an important predictor explaining modern market channel participation. All marketing-related variables included in the logistic regression model showed non-significant effects on market channel participation. Most importantly, participation in farmersâ&#x20AC;&#x2122; groups was a strong predictor determining their participation in the supermarket channels. This also indicated that acting collectively can enable smallholder farmers to reach the high value markets. 4.3 Impact of market participation on vegetable income The bivariate correlation analysis was conducted under the null hypothesis, stating that there was no correlation between supermarket channel participation and vegetable income, and the alternative hypothesis stating that there was a correlation between these two variables. The result of the correlation analysis regarding the relationship between market channel participation and income was presented in Table 3. The Pearson correlation coefficient (r) was 0.258, with a related significance level of 0.004. Based on these outcomes it can be concluded that the association between supermarket channel participation and income generated from vegetable farms was positive and statistically significant. In addition, this relationship was categorised between small and moderate correlation (Cohen, 1988, as cited in Corder and Foreman, 2014). This finding implied that differences in vegetable marketing practices contributed to essential differences in profitability between traditional market suppliers and supermarket channel suppliers. This study identified difference in income received from different market channel that ranged from 20-40% depending on type of vegetable cultivated. Furthermore, this finding was also in line with results from previous studies revealing that participation in the modern market channels was associated with relatively higher household income (Miyata et al., 2009; Neven and Reardon, 2006; Rao and Qaim, 2011). For the context of the Manokwari region, where the supermarket channels have been growing, the result suggested that this marketing channel mode can be beneficial for smallholder farmers in relation to providing alternative marketplaces and economic potential.

5. Conclusions and recommendations The development of modern food markets, including supermarkets, has been taking place in Indonesia for more than two decades. The presence of supermarkets, with new procurement practices, has affected all actors in the supply chains including smallholder farmers. Participation in supermarket channels can bring opportunities for smallholder farmers in gaining economic advantages. However, smallholder farmers are also facing constraints regarding higher standard requirements that might potentially limit participation possibilities. The results of this study indicated that age and education level were the only significant farmer demographic factor for high value market participation. Better educated smallholder farmers between 35-55 years old have better ability to analyse market situations and participate in high value market channel. Farm size was another important factor to enable smallholder farmers to participate in supermarket channel. Farmers with farm size of over 1 ha have better opportunities in dealing with supermarket requirements. However, the majority of vegetable farmers in Manokwari region owned relatively small areas under vegetables, which also became a main challenge in dealing with supermarkets requirements and standards. This study also Table 3. Correlation between market channel and vegetable income.1 Vegetable income Market channel

1 **

Pearson correlation Sig. (2-tailed) n

0.258** 0.004 126

Correlation is significant at the 0.01 level (2-tailed). International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

reported that in some cases, modern food markets source from smallholder farmers, even when the access to larger farms is available. Farmersâ&#x20AC;&#x2122; participation in organisations/groups was another critical factor that helped smallholder farmers to access high value market channels as well as to increase production. Furthermore, the results revealed that the smallholder vegetable farmers that were selling through supermarket channels received higher incomes, compared to those who marketed through the traditional market channels. Smallholder farmersâ&#x20AC;&#x2122; participation in modern market channels in Eastern Indonesia is challenging, due to the fact that the majority of agricultural products, including fresh fruits and vegetables, supplied to supermarkets are mostly delivered from outside the Manokwari region. Another challenge that smallholder farmers are facing is dealing with the quality and continuity of products supplied to supermarkets or food stores. Because smallholder farmers are characterized by small farm size and low productivity, it is usually difficult to meet the basic requirements regarding quality and consistent supply. Hence, instead of participating in modern food retail and wholesale markets, smallholder farmers prefer to sell products through the traditional market channels that are considerably free of binding contracts. In addition, interventions of the supporting institutions, such as the government, NGOs, cooperatives and associations, regarding agricultural development are more likely to engage with technical aspects, whereas aspects relating to value addition and marketing receive very little attention. Based on the results of this study several recommendations are drawn for the farmers, farmers groups, policy makers and other institutional players. Policies that encourage smallholder farmers to build their capacity through sharing knowledge and information are essential to be reinforced. Having higher level of technical skills can help smallholder farmers to produce better quality vegetables and increase productivity that can enable them to sell to high value markets. Another recommendation relates to encouraging and strengthening collective action through farmer groups/cooperatives. The presence of cooperatives and farmer organisations not only helps smallholder farmers in sharing knowledge and information but also strengthens their market position with supermarkets. The final recommendation for the local and regional government is to initiate an integrated systems approach, in order to address the institutional issues such as lack of market information, standards and grades, credit access, and extension services. Moreover, infrastructure developments, such as roads, transportation, market outlets, and vegetable based industries, should also be improved in order to provide a positive environment for the small scale vegetable farms in the Manokwari region of Indonesia. Further investigation in other regions of Indonesia, larger sample size, larger number of farmers supplying high value markets would provide more comprehensive and comparative information about the market channel participation.

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2017.0011. Table S1. Demographic characteristics of the respondents. Table S2. Farm characteristics. Table S3. Marketing characteristics. Table S4. Institutional characteristics.

International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

References Alene, A.D., V.M Manyong, G. Omanya, H.D. Mignouna, M. Bokanga and C. Odhiambo. 2008. Smallholder market participation under transaction costs: maize supply and fertilizer demand in Kenya. Food Policy 33: 318-328. Balint, B. and P. Wobst. 2006. Institutional factors and market participation by individual farmers: the case of Romania. Post-Communist Economies 18(1): 101-121. Barrett, C.B., M.E. Bachke, M.F. Bellemare, H.C. Michelson, S. Narayanan and T.F. Walker. 2011. Smallholder participation in contract farming: comparative evidence from five countries. World Development 40(4): 715-730. Berdegue, J.A., F. Balsevich, L. Flores and T. Reardon. 2005. Central American supermarkets’ private standards of quality and safety in the procurement of fresh fruits and vegetables. Food Policy 30: 254-269. Blandon, J., S. Henson and J. Cranfield. 2009. Small-scale farmer participation in new agri-food supply chains: case of the supermarket supply chain for fresh fruit and vegetables in Honduras. Journal of International Development 21: 971-984. Boselie, D., S. Henson and D. Weatherspoon. 2003. Supermarket procurement practices in developing countries: redefining the roles of the public and private sectors. American Journal of Agricultural Economics 85(5): 1155-1161. Chowdhury, S., A. Gulati and E. Gumbira-Sa’id. 2005. The rise of supermarkets and vertical relationships in the Indonesian food value chain: causes and consequences. Asian Journal of Agriculture and Development 2: 39-48. Cohen, J. 1988. Statistical power analysis for the behavioural sciences. Academic, New York. NY, USA. Corder, G.W. and D.I. Foreman. 2014. Nonparametric statistics: a step-by-step approach. (2nd ed.). John Wiley and Sons, Inc, New Jersey, NJ, USA. Fogelman, K. and C. Comber. 2007. Surveys and sampling. In: Research methods in educational leadership and management, edited by A.R. Briggs and M. Coleman. SAGE Publications Ltd, Londen, UK. Gong, W., K. Parton, R.J. Cox and Z. Zhou. 2007. Transaction costs and cattle farmers’ choice of marketing channels in China: a tobit analysis. Management Research News 30(1): 47-56. Gujarati, D. 2004. Basic econometrics (4th ed.). McGraw Hill, New York, NY, USA. Hellin, J., M. Lundy and M. Meijer. 2009. Farmer organization, collective action and market access in MesoAmerica. Food Policy 34: 16-22. Hernandez, R., T. Reardon and J. Berdegue. 2007. Supermarkets, wholesalers, and tomato growers in Guatemala. Agricultural Economics 36: 281-290. Imami, D., E. Zhllima, D. Viaggi and W. Bokelmann. 2013. Between weak markets and weak regulations: determinants of contracting in orchard farming in Albania. Journal on Chain and Network Science 13(1): 37-46. Irianto, B., and Herwanto. 2009. Linking small banana producers in Lumajang district to better markets. In: Proceedings of the Sixteenth International Symposium on Horticultural Economics and Management, edited by P.P. Oppenheim. Acta Hort, Leuven, Belgium, pp. 115-122. Ismail, M., M.M. Kavoi and B.K. Eric. 2013. Factors influencing the choice of supermarket channel by smallholder vegetable farmer suppliers in Nairobi and Kiambu Counties, Kenya. Journal of Agricultural Economics and Development 2(9): 333-344. Jari, B. and G.C.G. Fraser. 2009. An analysis of institutional and technical factors influencing agricultural marketing amongst smallholder farmers in the Kat River Valley, Eastern Cape Province, South Africa. African Journal of Agricultural Research 4(11): 1129-1137. Ministry of Agriculture. 2014. Statistik ketenagakerjaan sektor pertanian tahun 2013. Was available at: http:// pusdatin.setjen.pertanian.go.id/tinymcpuk/gambar/file/Statistik_Tenaga_Kerja_Pertanian_2013.pdf [link obsolete]. Miyata, S., N. Minot and D. Hu. 2009. Impact of contract farming on income: linking small farmers, packers, and supermarkets in China. World Development 37(11): 1781-1790. Neven, D., M.M. Odera, T. Reardon and H. Wang. 2009. Kenyan supermarkets, emerging middle-class horticultural farmers, and employment impact on rural poor. World Development 37(11): 1802-1811. International Food and Agribusiness Management Review

Maspaitella et al.

Volume 21, Issue 1, 2018

Neven, D. and T. Reardon. 2006. Farmer response to the rise of supermarkets in Kenya’s fresh fruit and vegetables supply system. Journal of Food Distribution Research 37(1): 120-124. Ouma, E., J. Jagwe, G.A. Obare and S. Abele. 2010. Determinants of smallholder farmers’ participation in banana market in Central Africa: the role of transaction costs. Agricultural Economics 41: 111-122. Panda, R.K. and Sreekumar. 2012. Marketing channel choice and marketing efficiency assessment in agribusiness. Journal of International Food and Agribusiness Marketing 243: 213-230. Pascucci, S. 2011. Factors affecting farmers’ networking decisions. Journal on Chain and Network Science 11(1): 7-17. Poole, N., A.W. Seini and V. Heh. 2003. Improving agri-food marketing in developing economies: contractual vegetable markets in Ghana. Development in Practice 13(5): 551-557. Rao, E.J.O., B. Brummer and M. Qaim. 2012. Farmer participation in supermarket channels, production technology, and efficiency: the case of vegetables in Kenya. American Journal of Agricultural Economics 94(4): 891-912. Rao, E.J.O. and M. Qaim. 2011. Supermarkets, farm household income, and poverty: insights from Kenya. World Development 39(5): 784-796. Reardon, T., C.B. Barrett, J.A. Berdegue and J.F.M. Swinnen. 2009. Agrifood industry transformation and small farmers in developing countries. World Development 37(11): 1717-1727. Reardon, T., J.A. Berdegue, C.P. Timmer, T. Cabot, D. Mainville, L.F. Flores, R. Hernandez, D. Neven and F. Balsevich. 2005. Links among supermarkets, wholesalers, and small farmers in developing countries: conceptualization and emerging evidence. The future of small farms 45. Reardon, T., C.P. Timmer and B. Minten. 2012. Supermarket revolution in Asia and emerging development strategies to include small farmers. PNAS 109(31): 12332-12337. Ruane, J.M. 2005. Essentials of research methods. Blackwell Publishing, Victoria, Australia. Schipmann, C. and M. Qaim. 2010. Spillovers from modern supply chains to traditional markets: product innovation and adoption by smallholders. Agricultural Economics 41: 361-371. Schipmann, C. and M. Qaim. 2011. Supply chain differentiation, contract agriculture, and farmers’ marketing preferences: the case of sweet pepper in Thailand. Food Policy 36: 667-677. Shiimi, T., T.P.R. Taljaard and H. Jordaan. 2012. Transaction cost and cattle farmers’ choice of marketing channel in North Central Namibia. Agrekon 51 (1): 42-58. Siddique, M.I. and E. Garnevska. In press. Factors affecting marketing channel choice decisions of smallholder citrus growers. Journal of Agribusiness in Developing and Emerging Economies. (forthcoming) Simmons, P., P. Winters and I. Patrick. 2005. An analysis of contract farming in East Java, Bali, and Lombok, Indonesia. Agricultural Economics 33: 513-525. Statistics Indonesia. 2014. Jumlah rumah tangga usaha pertanian pengguna lahan dan rumah tangga petani gurem menurut wilayah tahun 2003 dan 2013. Available at: http://tinyurl.com/yc3qw6e2. Suryadarma, D., A. Poesoro, Akhmadi, S. Budiyati, M. Rosfadhila and A. Suryahadi. 2010. Traditional food traders in developing countries and competition from supermarkets: evidence from Indonesia. Food Policy 35: 79-86. Teddlie, C. and F. Yu. 2007. Mixed methods sampling: a typology with examples. Journal of mixed methods research 1(1): 77-100. Wang, H.H., Y. Zhang and L. Wu. 2011. Is contract farming a risk management instrument for Chinese farmers?: evidence from a survey of vegetable farmers in Shandong. China Agriculture Economic Review 3(4): 489-505. Wooldridge, J.M. 2013. Introductory econometrics: a modern approach (5th ed.). South-Western Ohio, OH, USA.

International Food and Agribusiness Management Review

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2016.0139 Received: 15 August 2016 / Accepted: 18 October 2017

Competitiveness of local food: an empirical analysis of the tomato market dynamics RESEARCH ARTICLE Xin Fang a, Hui Huangb, and PingSun Leungc aAssociate

Professor and bAssistant Professor, Department of Financial Economics and Information Systems, Hawaii Pacific University, 900 Fort Street Mall, Suite 600, Honolulu, HI 96813, USA cProfessor,

Department of Natural Resources and Environmental Management, University of Hawaii at Manoa, 3050 Maile Way, Gilmore 111, Honolulu, HI 96822, USA

Abstract Local food has become the foci of food self-sufficiency and sustainable development. This work explores the dynamic interactions between local and imported food in Hawaii, which is a typical small, open economy and an ideal market for a local food system study. Retail scanner data of three major grocery chain stores are used to construct a time series of prices and quantities for local and imported grape and cherry tomatoes in one year (52 weeks). Vector Autoregressive model and impulse response functions are used to investigate the correlations and Granger causalities. Empirical results provide implications for the competitive status of local suppliers and the dynamics of the local food market. Keywords: VAR model, market dynamics, shocks, local food, imported food JEL code: Q10, Q13 Corresponding author: xfang@hpu.edu

ÂŠ 2017 Fang et al.

Fang et al.

Volume 21, Issue 1, 2018

1. Introduction Local food systems have been attracting increased attention due to their role in promoting sustainable development. Because of the small scale of operations, local producers are unable to achieve economies of scale (Low and Vogel, 2011) and the high cost of local production could contribute to high food prices in the local community (Pretty et al., 2005). The inflated costs divert diet choices towards lower priced, energy-dense food items (Salois, 2012). This diet bias is more severe for underrepresented groups who are more sensitive to food prices (Powell and Chaloupka, 2009), thus creating an inequity within the local community. As a major concern, the long-distance shipping of imports imposes great pressure on the ecosystem by increased fossil fuel consumption and greenhouse gas emissions (Coley et al., 2009). With increasing volumes of imported products, leakage in the system becomes greater and restricts regional economic and labor multipliers (Martinez, 2010; Swenson, 2009). Moreover, changes in energy prices, trade policies, world food supply and demand, and climate change all threaten food security of the local community (Feenstra, 1997; Martinez, 2010). Local agriculture has the potential to improve the structure of the food system, leading to community food security, fewer food deserts, and improvements in consumer health that are linked to eating more fresh and unprocessed foods (Adams and Salois, 2010). In addition, local farmers’ green efforts from promoting locally produced products can provide economic or social benefits in terms of reduced energy consumption and pollution (Amato, 2007). Shoppers prefer locally-grown food for the freshness, taste, quality, and convenience, as well as the sense of being environmentally friendly and supportive of local farmers (Eastwood et al., 1999; Govindasamy et al., 1998; Keeling-Bond et al., 2009; Kezis et al., 1984; Wolf, 1997; Wolf et al., 2005; Zepeda and LevitenReid, 2004). However, the demographics of consumers, such as income, education, and ethnicity which are closely related to consumer preferences, play an important role in consumption choices and market demand (Brooker and Eastwood, 1989; Brown, 2003; Eastwood, 1996; Eastwood et al., 1999; Govindasamy et al., 1998; Keeling-Bond et al., 2009; Zepeda and Li, 2006). Several empirical studies show consumers’ willingness to pay a price premium for local food products, including seafood, lettuce, milk, eggs and tomatoes (Davidson et al., 2012; Geslani et al., 2015; Keahiolalo, 2013; Loke et al., 2015, 2016; Ulupono Initiative, 2011; Xu et al., 2015a,b). The questions that remain unanswered are: does the ability to command a price premium create a competitive advantage for local food over imported food? What implications can we derive from the interactions between local and imported food for the local farmers’ competitive strategies and sustainability of the local food system? The local tomato market in Hawaii is a suitable context to explore the research questions stated above given its geographic location. Local tomatoes ‘are picked ripe and have a shinier and brighter red outside, thinner skin, redder and juicier inside, and taste sweeter than Mainland tomatoes’, and consumers are willing to pay about $2.50 per pound for the ‘local’ feature (Ulupono Initiative, 2011). Local tomatoes seem to have a dominant status in Hawaii, accounting for about 77% of the market supply (Xu et al., 2015a). A cost study could be helpful to identify the profit levels and a demand system estimation would shed light on price-cost margins, that is, the market power of local suppliers. In the absence of sufficient data, however, we propose a study of the market dynamics to provide insights into competition patterns. By evaluating the status of local suppliers, we could draw implications regarding the sustainability of the local food industry and provide policy or strategic suggestions for local farmers. There are no organic tomatoes from local suppliers in the dataset. Given that the feature of ‘organic’ could be as appealing as ‘local’ or, alternatively, imported nonorganic tomatoes might be comparable with local tomatoes that are also nonorganic, we divide the imported categories into imported organic and imported nonorganic to better determine the intercorrelation between local and imported tomatoes. We expect imported organic tomatoes to be a better substitute and a closer competitor for local tomatoes, therefore stronger interactions between these two categories are predicted. Compared to national brands, local food producers are characterized by small-scale operations, but are less affected by transportation cost fluctuations (Martinez, 2010). Competition among imported food and local International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

food producers might exhibit a unique pattern. Limited empirical work has been done to investigate the dynamic interactions between local and imported goods and for a particular type of food. Prices, as well as volumes sold, of local and imported food are linked through the market, and the local market might also be subject to outside shocks, such as seasonality and transportation costs. The interactions between local and non-local suppliers can shed light on consumers’ attitudes towards local food, with reference to the substitutability between local and imported food as measured by the cross-price elasticity. Consumers take a critical part in creating sustainable solutions through responsible consumption (Knight, 2004). Green market retail campaigns and brands could potentially increase sustainable consumption, although their effectiveness is not proven (Sullivan et al., 2013). Xu et al. (2015a) find local tomatoes are quite substitutable to import tomatoes, which suggests local tomatoes might not be so differentiated in the view of consumers. Our findings would further reveal consumers’ recognition of the ‘local’ feature by examining the response of local tomatoes’ prices to strategic changes made by the imported (organic or nonorganic) tomatoes. The Vector Autoregressive (VAR) Model chosen in our proposed study would provide a theory-free approach where the endogeneity problem between price and quantities could be solved (cf. De Crombrugghe et al., 1997; Wang and Bessler, 2006). Although the interpretation in terms of causal relationships in VAR Model is controversial, most researchers regard VAR as useful method of summarizing time series ‘facts’ and assess dynamic influences of structural shocks on observables (Hamilton, 1994; Holtz-Eakin et al., 1988). This study intends to draw a picture of the dynamic interrelationship between local and imported food and provide insights into pricing and competitive strategies for local farmers in response to the strategies of non-local competitors. The results would also uncover the degree of differentiation and market power of local farmers, as well as their reactions to the outside shocks, such as weather, seasonality, and transportation costs. The paper will be organized as follows: we first discuss the relevant literature, followed by the data source and variables used in the estimation. We then describe the econometric modeling and provide tests and rationale for the methodology chosen. After analyzing the parameter estimates, we draw conclusions and discuss policy implications and extensions.

2. Literature review The current trend of ‘buy local’ is consistent with the findings about consumers’ desire to support the local economy and consumer’s choice of locally grown products because of high quality and its value relative to the costs (Andreatta and Wickliffe, 2002; Darby et al., 2006; Eastwood et al., 1999; Wolf, 1997). Bonini and Oppenheim (2008) found that 87% of consumers in eight major global economies think about the environmental and social impact of their purchases, with around one third of these consumers purchased or plan to purchase a green product. A typical grocery store consumer tends to pay $0.64 more for a product labeled as grown in their home state (Darby et al., 2006). However, consumers’ preferences are based on their knowledge and experiences that influence how much they value the products (Day, 1994; Woodruff and Gardial, 1996; Zeithaml, 1988). Branding and logos are observed to be critical in food marketing, and shoppers tend to buy state-labeled products (Eastman et al., 1999; Guthrie et al., 2006). Some studies show consumers could be skeptical of the quality of green products and environmental claims over these goods. The study of Sullivan et al. (2013) look at the demand for Hawaiian avocados, which makes up about 27% of the total demand for tropical fruits in 2005. The findings indicate that about 49% of local avocados do not make it to market due to insufficient efforts of marketing and unawareness about consumer preferences (Krishnakumar et al., 2007). Hawaii imported about two million avocados in 2005 (National Agricultural Statistics Service, 2005). The significant reliance on imported avocados is found to meet consumer demand, even though local producers are able to satisfy the needs. Supply chain and marketing strategies are claimed to be insufficient for local food sellers.

International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

Demand for a product during sales depends on the time since the last sale and past prices (Pesendorfer, 2002). Intuitively, current demand tends to be higher if previous price is higher (Pesendorfer, 2002; Warner and Barsky, 1995) and a stock of goodwill, representing consumers’ brand loyalty, habit information or product awareness, would accumulate (or erode) when a seller charges low (or high) prices (Slade, 1998). The inventory conditions would be related to the price changes as well. It is found that high inventory leads to low prices, more frequent price reductions, as well as the probability of sales (Aguirregabiria, 1999; Levin et al., 2009; Pesendorfer, 2002). Therefore, the literature suggests potential intertemporal correlations between prices and quantities in several dimensions: prices and quantities of one seller would be correlated over time through state of demand or expected demand and inventory conditions (Sobel, 1984). Cross-seller correlations, however, such as the interaction between prices and quantities among different sellers, would depend on market competition between local and imported tomato suppliers and their strategies. We examine the competitiveness of local suppliers in the market of grape and cherry tomatoes in Hawaii from a different perspective by adopting a dynamic analysis about the interactions of prices and quantities for local and imported tomatoes. Research on competition among local and imported food is a study of differentiated market, where product sales depend on demand conditions and prices respond to intertemporal demand (Nevo and Wolfram, 2002). Demand of differentiated markets have been extensively studied with different demand function specifications (Berry, 2004; Berry et al., 1994; Hausman et al., 1994; Pinkse et al., 2002; Slade, 2004). Among those studies, own-price elasticities are used to derive price-cost margins as an indicator for market power, while cross-price elasticities are used to assess the interactions among competitors by estimating the movement of consumers’ demand when price changes, therefore showing the impact of one product’s demand in terms of another. Localized competition with differentiated competitors has been studied in spatial models. Some researchers assume symmetry of cross-price elasticities (cf. Deaton and Muellbauer, 1980), but others use unrestricted approaches (cf. Hausman et al., 1994). Models of monopolistic competition as described in Chamberlin (1933) suggest global but symmetric competition, while one-dimensional-spatial models (Gabszewicz and Thisse, 1979; Hotelling, 1929; Salop, 1979) predict fluctuations in prices of more distant competitors have no effect on own sales, conditional on neighbor prices. In the work of Pinkse et al. (2002), restrictions of symmetry have also been removed, and competitors could engage in multi-dimensional local competition or asymmetric global competition. The interactions between local and imported tomatoes in our proposed model do not have to be symmetric, especially when competition from imported organic and imported nonorganic are divided. The competition between different pairs of the three categories could be different. Competition among local producers would be absorbed, thus not analyzed. Based on the models and empirical evidence in the literature, we predict sales and prices of local food might not only depend on the rival prices, but on the ‘locality’ or identity of the sellers. VAR is selected for the intertemporal analysis for the presence of simultaneities and bi-directional influences. Granger causality tests will be run to check potential causal relationships between interested variables, as done in similar dynamic panel studies (Awokuse, 2005; Holtz-Eakin et al., 1988; Śmiech and Papież, 2013). Moreover, research on VAR has made it possible to identify unobservable structural shocks and examine the dynamic effects of these shocks on observable data (Hamilton, 1994). In our study, VAR model is used to investigate the competition between local and non-local suppliers and their responses to outside shocks with no restrictions imposed.

3. Data We use one year (52 weeks from 26 December 2010 to 24 December 2011) of Nielsen ScanTrack weekly data of grape and cherry tomatoes from three major grocery chain stores in Honolulu, Hawaii. In the dataset, loose weight tomatoes are not included due to the inconsistency of stock-keeping units across purchases or stores. The tomatoes sold at other locations or farmers’ market are not counted in either. Annual household income of grocery store shoppers is found to be higher than for those who shop at direct outlets for food (Darby et al., International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

2006). The same study reported, that the amount consumers spend on produce does not differentiate by retail outlet type, with buyers at a farmer’s market pay slightly more for the same type of product. Nevertheless, our study should still be able to represent the tomato markets in Honolulu, as the Ulupono Initiative (2011) gives an estimate that 60% of sales of tomatoes are made in supermarkets, suggesting supermarkets are the major channel of tomato sales. To help consumers identify local tomatoes, in the supermarkets, shelf tags such as ‘Hawaii Grown’, ‘Local’ and ‘Locally Grown’ are displayed. We also search the websites of the producers contained in the dataset to verify the origin of tomatoes as local or nonlocal. Price and quantity information are obtained according to the stock-keeping units on weekly basis. With the sales data, we construct a panel including six time series: local price, imported organic price, imported nonorganic price, local volume, imported organic volume and imported nonorganic volume, respectively. The price series are formed by calculating the average price weighted by volume of sales, including promotions that are recognized as price cuts and recorded in the data. Volume series are the sum of quantity sold in one week. The summary statistics are shown in the Table 1. Prices of local tomatoes are lower than those of imported organic tomatoes, though slightly higher than imported nonorganic ones. We tentatively attribute part of price difference to transportation costs and production costs, respectively. From the time series plots (Figure 1 and 2), we observe some seasonal changes and periodical price cuts for grape and cherry tomatoes, however, the patterns for the three categories are different. Local grape and cherry tomatoes’ prices are low frequency series compared to the imported ones. Also, fewer fluctuations are observed for local tomato sales in the dataset. Table 1. Descriptive statistics of grape and cherry tomato data set. Grape and cherry tomatoes

Price per pound

Weight (pounds)

Market share (weight)

Local Imported organic Imported nonorganic

6.520 6.942 6.472

722.6 212.1 1,498.0

29.52% 9.88% 60.61%

Local price Imported organic price

Imported nonorganic price

Price ($/lb)

4 0

Week

Figure 1. Price series for grape and cherry tomatoes.

International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

Local volume Imported organic volume

Imported nonorganic volume

4,000

Weight (lbs)

3,000

2,000

1000

0 0

Week

Figure 2. Volume series for grape and cherry tomatoes.

4. Models 4.1 Univariate characteristics of the price and volume series Before directly investigating variablesâ&#x20AC;&#x2122; influences on one another, the first task is to examine the time series properties of each of the four series and to confirm whether the quantity and price series are stationary. The tests are performed by using the most common approach, that is, the augmented Dickey-Fuller test (Dickey and Fuller, 1979, 1981). Johanson cointegration test is conducted to confirm that a first differenced VAR model is suitable. We also determined according to Schwarz information criterion that significant pricequantity interactions between local and imported tomatoes last at a maximum of 4 weeks. Details of the tests to validate our model are included in Supplementary results S1. 4.2 Vector Autoregressive-dynamic effects As the variables are nonstationary, but not cointegrated at levels, we investigate the dynamic relationships among the quantities and prices that are stationary with first difference. To describe the intertemporal correlations between the endogenous variables, we employ a panel VAR model. Economists advocate VAR model as the models provide a theory-free method to estimate dynamic economic relationships and work as an alternative to the â&#x20AC;&#x2DC;incredible identification restrictionsâ&#x20AC;&#x2122; in structural models. With no prior-modeling quantity-price relationship assumed in this study, we are able to estimate a model with least restrictions. The dynamic system of interest consists of price-oriented and quantity-oriented (inverse) demand equations, containing the six previously built series of local price, imported organic price, imported nonorganic price, local volume, imported organic volume and imported nonorganic volume. The reduced form VAR model would be able to evaluate the effects of price and quantity values in previous weeks on the current week, as specified in Equation 1 as follows:

International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

đ?&#x2018;&#x152;đ?&#x2018;&#x152;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; = đ??žđ??ž + â&#x2C6;&#x2018; đ?&#x2018;&#x161;đ?&#x2018;&#x161; đ?&#x2018;&#x2122;đ?&#x2018;&#x2122;=1 ÎŚđ?&#x2018;&#x2122;đ?&#x2018;&#x2122; đ?&#x2018;&#x152;đ?&#x2018;&#x152;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; + đ?&#x153;&#x20AC;đ?&#x153;&#x20AC;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;

Where

đ?&#x2018;&#x152;đ?&#x2018;&#x152;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; = [

(1)

đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x153;đ?&#x2018;&#x153;đ?&#x2018;&#x153;đ?&#x2018;&#x153; đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; ] đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x2030;đ?&#x2018;&#x153;đ?&#x2018;&#x153;đ?&#x2018;&#x153;đ?&#x2018;&#x153; đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;

a=location of production (local, imported organic, imported nonorganic) m=maximum lag numbers These equations can be estimated using ordinary least squares, while the contemporaneous effects of the endogenous variables are absorbed in the variance matrix of the errors (Hamilton, 1994). The lag is selected based on the underlying significance test. Coefficients beyond two lags are no longer significant, therefore, only variables up to two lags are included and the model results are presented in Table 2, with the columns presenting the results for each of the endogenous variables. Impulse Response Functions for each variable are shown in Supplementary results S1, which show consistent impact of shocks of various endogenous variables.

Our first finding, according to the VAR results, is fluctuations in the price sold in the market will correlate with sales. Evidenced by the significant coefficients of lagged volumes (one or two lags), as well as substantial responses of prices to volume shocks according to the impulse response functions, we find that prices of local and imported tomatoes are significantly affected by their inventory or stock shocks, which might be caused by seasonal changes in the past one or two weeks. A positive relationship shows that weak sales (cumulating consumer demand) in the current week will lead to price cuts or promotions in the following week. The explanation lies in the fact that a high inventory condition would be related to the timing of price promotions, which agrees with findings of Levin et al. (2009), Aguirregabiria (1999) and Pesendorfer (2002). However, we find no evidence of a significant impact of prices on sales for either local or imported tomatoes. Periodic price cuts have been discussed by Doyle (1983), where sellers in the non-durables market would cut their prices periodically to attract consumers who are uncertain about their preferences towards the product. In our study, we find that previous prices, as well as the duration since last promotion would affect the current prices, but this is only true for imported tomatoes. A promotion would immediately follow a higher price in the current week for imported organic tomatoes. For the imported nonorganic tomatoes, current price is positively correlated with previous prices. A possible explanation for this unexpected correlation would be the price rigidity or adjustment cost for prices (Slade, 1998, 1999). Local tomato prices are not serially correlated, showing fewer adjustments compared to imported tomatoes. Since local prices only respond to variations of local volumes, a possible explanation could be the smaller fluctuations in the sales of local tomatoes, which is a signal of stable supply. Last but not least, local market is relatively isolated as there are no observed dynamic adjustment processes between local vs imported organic or local vs imported nonorganic tomatoes. There are certain interactions between volumes and prices between imported organic and nonorganic markets, for examples, imported nonorganic price move in the same direction as imported organic price in the past week, and imported organic price one week before significantly affect current imported nonorganic volume. Both observations suggest imported organic tomatoes are substitutes for nonorganic tomatoes, although not vice versa. However, such interactions are missing between local and imported tomatoes. Our â&#x20AC;&#x2DC;Granger-causalityâ&#x20AC;&#x2122; tests confirm the above findings. This study tests dynamic specifications of the causality models. Regarding the use of dynamic lagged models to test for market influences, in economic terms the lagged effects in the model are likely to arise from sluggishness in price adjustment, delays in transportation and information transmission, cold storage inventory holdings, and the formation of expectations under price uncertainty (Elston et al., 1999). As for the Honolulu grape and cherry tomato market, a lag length of 4 is selected using weekly data, as this lag length accommodates the possible effects of commodity flows in both markets. Hence, the models were assessed for equal lag lengths in both markets for lengths of 4. For

International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

Local pricet-1 Local pricet-2 Imported organic pricet-1 Imported organic pricet-2 Imported nonorganic pricet-1 Imported nonorganic pricet-2 Local volumet-1 Local volumet-2 Imported nonorganic volumet-1 Imported nonorganic volumet-2 Imported organic volumet-1 Imported organic volumet-2 Constant Observations 1 ***

0.258 (0.222) 0.153 (0.242) -0.297 (0.312) -0.0321 (0.299) -0.204 (0.352) 0.0564 (0.359) 0.00259*** (0.000955) 0.00192* (0.00105) -0.000522 (0.000496) -0.000297 (0.000510) -0.00357 (0.00470) 0.000352 (0.00480) -0.170 (0.171) 49

-0.123 (0.127) 0.157 (0.139) -0.480*** (0.179) -0.123 (0.172) -0.0915 (0.203) -0.292 (0.206) -0.000657 (0.000549) 0.000522 (0.000605) -0.000106 (0.000285) -0.000447 (0.000293) 0.00746*** (0.00270) 0.00400 (0.00276) 0.0360 (0.0982) 49

-0.0302 (0.199) 0.00555 (0.218) 0.500* (0.280) 0.106 (0.269) -0.463 (0.317) 0.576* (0.323) 0.000116 (0.000859) 0.000373 (0.000946) -0.000305 (0.000446) 0.000954** (0.000459) 0.00472 (0.00423) -0.00257 (0.00432) -0.0173 (0.154) 49

-69.80 (55.13) -66.13 (60.17) -6.602 (77.48) -33.37 (74.37) 13.82 (87.63) 22.22 (89.23) -0.561** (0.238) -0.637** (0.262) 0.0212 (0.123) 0.122 (0.127) -0.193 (1.169) 0.185 (1.194) 47.20 (42.51) 49

-5.697 (9.472) -3.240 (10.34) 1.127 (13.31) -0.578 (12.78) 3.251 (15.05) 19.92 (15.33) -0.0127 (0.0408) -0.0234 (0.0450) -0.00569 (0.0212) 0.0250 (0.0218) -0.842*** (0.201) -0.370* (0.205) -2.843 (7.303) 49

(6) Imported nonorganic volumet

(5) Imported organic volumet

(4) Local volumet

(3) Imported nonorganic pricet

(2) Imported organic pricet

Variables

(1) Local pricet

Table 2. Vector Autoregressive estimates of the dynamics of prices and volumes of local and imported tomatoes.1

84.74 (144.3) 17.76 (157.5) -354.2* (202.8) 4.568 (194.7) 12.43 (229.4) -448.1* (233.6) 0.00275 (0.622) -0.315 (0.685) -0.279 (0.323) -0.760** (0.332) -3.007 (3.061) 3.622 (3.127) -0.753 (111.3) 49

P<0.01; ** P<0.05; * P<0.1; values between parentheses are standard errors.

the local market, causality is observed in one direction, which indicates that the influence of volume on that of tomato price is significant, whereas the hypothesis of price causality on sales is rejected. Moreover, there are no causal relationships between local and imported tomatoes, either organic or nonorganic. The local sales may not rely much on competitors’ pricing or volumes. The prices of local tomato sellers are only driven by their sales conditions, not subject to the price adjustment of imported tomato suppliers, suggesting that local tomatoes are significantly differentiated from imported counterparts and local tomato sellers do have market power of their products. Other features, such as brand recognition or ‘buy local’ incentive may have high weights on consumers’ behavior. Given that seasonality is less pronounced in Hawaii, and outside shocks which affect imported supply are not transferred to the local tomatoes, the volumes have relatively small fluctuations and prices of local tomatoes adjust less frequently compared to imported tomatoes. International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

5. Conclusions Competition of local food markets has not been studied empirically as much as other differentiated product markets for two major reasons: the difficulty of defining a ‘local’ market (Martinez et al., 2010), and the availability of data to test the empirical hypotheses based on existing demand models. Hawaii, as a state surrounded by the Pacific Ocean, provides a natural setting to overcome the first problem with its geographic isolation. The VAR model avoids the data requirement on the characteristics of the local food and market shares of each brand, as well as solving the endogeneity problems between prices and quantities. Built on the dynamic relationship in the tomatoes market, the proposed study discovered a distinguished pattern of interrelated demand systems. First, the demand for local and imported organic and nonorganic tomatoes are all quantity oriented, therefore, the inventory changes or supply shocks will feed into their respective price variations, but not vice versa. This finding is confirmed by the causality tests and impulse response functions and suggests that pricing strategies might not be effective in changing sales. Second, some evidence of substitutability of imported organic tomatoes towards imported nonorganic is found. However, imported tomatoes, either organic or nonorganic, are not interacting with local ones. Therefore, local tomatoes seem to be isolated from the competition of imported counterpart, and some market power as well as premiums are received from its ‘local’ feature. According to the theory of responsible consumption (Knight, 2004), quality, impact on environment and social responsibility as well as ‘green’ feature of local products can be recognized by the demand side, thus creating product differentiation and market power for local suppliers. Our findings show that consumer preferences could be the major support of local food system development. The resulting differentiation of local tomatoes from off-island competitors makes local farmers price makers and enhances their competitiveness, possibly profits. Consequently, strategies like advertising to increase the recognition of ‘locality’ of local tomatoes, could further increase the degree of differentiation and contributes to a sustainable competitive advantage of local tomatoes. The empirical results also have implications for the long-term stability of local tomato production. Although the market share of local grape and cherry tomatoes are not dominant, the local tomatoes are more resistant to outside shocks, which contributes to less frequent price adjustments and more stable market for local producers. Different types of tomatoes have different culinary values and functions, and this paper presents a case study of a particular agricultural product market. While we cannot generalize this finding to the entire tomatoes market or local food system, our finding suggests that it would be paramount to extend this study to other important local food items. Limited by the data, the study is not able to estimate a comprehensive demand system for local and imported food. Supplied with consumer data, e.g. household income, family size, age and education of household head, and residence location, etc., we would be able to derive price-cost margins as well as substitution pattern between local and imported tomatoes and get a better picture of competition in local market. Cost structure estimation for local farms would be viable with firm data to uncover the supply side conditions and sustainability behind the differential price and market shares of local food.

Acknowledgement This work was in part supported by the USDA National Institute of Food and Agriculture, Hatch project HAW01122-H, managed by the College of Tropical Agriculture and Human Resources, University of Hawaii at Manoa. Financial support from College of Business, Hawaii Pacific University is gratefully acknowledged. We would like to thank the reviewers for their constructive comments. We would also like to thank Prof. Ken Schoolland for his suggestions that greatly improved the manuscript.

International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2016.0139. Supplementary results S1.

References Adams, D.C and M.J. Salois. 2010. Local versus organic: a turn in consumer preferences and willingnessto-pay. Renewable Agriculture and Food Systems 25(04): 331-341. Aguirregabiria, V. 1999. The dynamics of markups and inventories in retailing firms. The Review of Economic Studies 66(2): 275-308. Amato, D.M. 2007. Going for the green. Chain Store Age 83(6): 42. Andreatta, S. and W. Wickliffe. 2002. Managing farmer and consumer expectations: a study of a North Carolina farmers market. Human Organization 61(2): 167-176. Awokuse, T.O. 2005. Exports, economic growth and causality in Korea. Applied Economics Letters 12(11): 693-696. Berry, S.T. 1994. Estimating discrete-choice models of product differentiation. The RAND Journal of Economics 25: 242-262. Berry, S.T., J. Levinsohn and A. Pakes. 2004. Differentiated products demand systems from a combination of micro and macro data: the new car market. Journal of Political Economy 112(1): 68-105. Bonini, S.M.J. and J.M. Oppenheim. 2008. Helping ‘green’ products grow. The McKinsey Quarterly 3(2): 1-8. Brooker, J.R. and D.B. Eastwood. 1989. Using state logos to increase purchases of selected food products. Journal of Food Distribution Research 20: 175-183. Brown, C. 2003. Consumers’ preferences for locally produced food: a study in southeast Missouri. American Journal of Alternative Agriculture 18(4): 213-224. Chamberlin, E.H. 1933. The theory of monopolistic competition: a re-orientation of the theory of value. Harvard University Press, Cambridge, MA, USA. Coley, D., M. Howard and M. Winter. 2009. Local food, food miles and carbon emissions: a comparison of farm shop and mass distribution approaches. Food Policy 34(2): 150-155. Darby, K., M.T. Batte, S. Ernst and B. Roe. 2006. Willingness to pay for locally produced foods: a customer intercept study of direct market and grocery store shoppers. Available at: http://tinyurl.com/yb8bary3. Davidson, K., M. Pan, W. Hu and D. Poerwanto. 2012. Consumers’ willingness to pay for aquaculture fish products vs Wild-caught seafood – a case study in Hawaii. Aquaculture Economics and Management 16(2): 136-154. Day, G.S. 1994. The capabilities of market-driven organizations. The Journal of Marketing 58: 37-52. De Crombrugghe, D., F.C. Palm and J.-P. Urbain. 1997. Statistical demand functions for food in the USA and the Netherlands. Journal of Applied Econometrics 12(5): 615-645. Deaton, A. and J. Muellbauer. 1980. An almost ideal demand system. The American economic review 70(3): 312-326. Dickey, D.A. and W.A. Fuller. 1979. Distribution of the estimators for autoregressive time series with a unit root. Journal of the American statistical association 74(366a): 427-431. Dickey, D.A. and W.A. Fuller. 1981. Likelihood ratio statistics for autoregressive time series with a unit root. Econometrica 49: 1057-1072. Doyle, C. 1983. Dynamic price discrimination, competitive markets and the matching process. The Warwick economics research paper series (TWERPS 229). Department of Economics, University of Warwick, Coventry, UK. Eastman, J.K., R.E. Goldsmith and L.R. Flynn. 1999. Status consumption in consumer behaviour: scale development and validation. Journal of Marketing Theory and Practice 7(3): 41-51. Eastwood, D.B. 1996. Using customer surveys to promote farmers’ markets: a case study. Journal of Food Distribution Research 27: 23-30.

International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

Eastwood, D.B, J.R. Brooker and M.D. Gray. 1999. Location and other market attributes affecting farmer’s market patronage: the case of Tennessee. Journal of Food Distribution Research 30: 63-72. Elston, J.A., J. D Hastie and D. Squires. 1999. Market linkages between the US and Japan: an application to the fisheries industry. Japan and the world economy 11(4): 517-530. Feenstra, G.W. 1997. Local food systems and sustainable communities. American journal of alternative agriculture 12(1): 28-36. Gabszewicz, J.J. and J.-F. Thisse. 1979. Price competition, quality and income disparities. Journal of Economic Theory 20(3): 340-359. Geslani, C., M.K. Loke, M. Barnes-Mauthe and P. Leung. 2015. Seafood purchasing preferences of Hawaii chefs: comparing actual purchase to stated preferences from conjoint choice experiment. Journal of International Food and Agribusiness Marketing 27(1): 50-63. Govindasamy, R., M. Zurbriggen, J. Italia, A. Adelaja, P. Nitzsche and R. van Vranken. 1998. Farmers markets: consumer trends, preferences, and characteristics. Parking 52(28.3): 16.0. Guthrie, J., A. Guthrie, R. Lawson and A. Cameron. 2006. Farmers’ markets: the small business counter – revolution in food production and retailing. British Food Journal 108(7): 560-573. Hamilton, J.D. 1994. Time series analysis. Vol. 2. Princeton university press, Princeton, NJ, USA. Hausman, J., G. Leonard and J.D. Zona. 1994. Competitive analysis with differentiated products. Annales d’Economie et de Statistique 34: 159-180. Holtz-Eakin, D., W. Newey and H.S. Rosen. 1988. Estimating vector autoregressions with panel data. Econometrica 56: 1371-1395. Hotelling, H. 1929. Stability in competition. The Economic Journal 39(153):41-57. Keahiolalo, K. 2013. Hedonic price analysis of fresh packaged tomatoes in the Honolulu market: an exploratory investigation. MA Capstone. Department of Economics, University of Hawaii, Manoa, Honolulu, HI, USA. Keeling-Bond, J.J., D. Thilmany and C. Bond. 2009. What influences consumer choice of fresh produce purchase location? Journal of Agricultural and Applied Economics 41(1): 61-74. Kezis, A.S., U.C. Toensmeyer, F.R. King, R.L. Jack and H.W. Kerr. 1984. Consumer acceptance and preference for direct marketing in the Northeast. Journal of Food Distribution Research 15(3). Knight, A. 2004. Sustainable consumption: the retailing paradox. Consumer Policy Review 14(4): 113-115. Krishnakumar, J., C. Chan-Halbrendt, P. Sullivan and K. Love. 2007. The willingness and likelihood of growers’ participation in fresh produce supply chain management: special focus on the Hawaiian avocado industry. IAMA 17th Annual World Forum and Symposium, Parma, Italy. Levin, Y, J. McGill and M. Nediak. 2009. Dynamic pricing in the presence of strategic consumers and oligopolistic competition. Management Science 55(1): 32-46. Loke, M.K, X. Xu and P. Leung. 2015. Estimating organic, local, and other price premiums in the Hawaii fluid milk market. Journal of dairy science 98(4): 2824-2830. Loke, M.K., X. Xu and P. Leung. 2016. Estimating local, organic, and other price premiums of shell eggs in Hawaii. Poultry science 95(5): 1050-1055. Low, S.A. and S.J. Vogel. 2011. Direct and intermediated marketing of local foods in the United States. USDA-ERS Economic Research Report 128. Available at: http://tinyurl.com/yd7pqy5q Martinez, S. 2010. Local food systems; concepts, impacts, and issues. Diane Publishing, Collingdale, PA, USA. National Agricultural Statistics Service. 2005. Statistics of Hawaii agriculture 2005. NASS, U.S. Dept. of Agriculture, Washington, DC. Available at: http://tinyurl.com/yat4ckfd. Nevo, A. and C. Wolfram. 2002. Why do manufacturers issue coupons? An empirical analysis of breakfast cereals. RAND Journal of Economics 33: 319-339. Pesendorfer, M. 2002. Retail sales: a study of pricing behavior in supermarkets. The Journal of Business 75(1): 33-66. Pinkse, J., M.E. Slade and C. Brett. 2002. Spatial price competition: a semiparametric approach. Econometrica 70(3): 1111-1153. Powell, L.M. and F.J. Chaloupka. 2009. Food prices and obesity: evidence and policy implications for taxes and subsidies. The Milbank Quarterly 87(1): 229-257.

International Food and Agribusiness Management Review

Fang et al.

Volume 21, Issue 1, 2018

Pretty, J.N., A.S. Ball, T. Lang and J.I.L. Morison. 2005. Farm costs and food miles: an assessment of the full cost of the UK weekly food basket. Food Policy 30(1):1-19. Salois, M.J. 2012. Obesity and diabetes, the built environment, and the ‘local’ food economy in the United States, 2007. Economics and Human Biology 10(1): 35-42. Salop, S.C. 1979. Monopolistic competition with outside goods. The Bell Journal of Economics 10: 141-156. Slade, M.E. 1998. Optimal pricing with costly adjustment: evidence from retail-grocery prices. The Review of Economic Studies 65(1): 87-107. Slade, M.E. 1999. Sticky prices in a dynamic oligopoly: an investigation of (s, S) thresholds. International Journal of Industrial Organization 17(4): 477-511. Slade, M.E. 2004. Market power and joint dominance in UK brewing. The Journal of Industrial Economics 52(1): 133-163. Śmiech, S. and M. Papież. 2013. Fossil fuel prices, exchange rate, and stock market: a dynamic causality analysis on the European market. Economics Letters 118(1): 199-202. Sobel, J. 1984. The timing of sales. The Review of Economic Studies 51(3): 353-368. Sullivan, P., C. Chan-Halbrendt and J. Krishnakumar. 2013. Are farmers’ market shoppers different from cross-shoppers? The case of Hawaiian avocado purchasers. Journal of Food Products Marketing 19(5): 363-375. Swenson, D. 2009. Investigating the potential economic impacts of local foods for southeast iowa. Leopold Center Pubs and Papers. 66. Department of Economics, Iowa State University, Ames, IA, USA. Available at: http://tinyurl.com/yblpz4bd. Ulupono Initiative. 2011. Local food market demand study. Available at: http://tinyurl.com/yc8892s4. Wang, Z. and D.A. Bessler. 2006. Price and quantity endogeneity in demand analysis: evidence from directed acyclic graphs. Agricultural Economics 34(1): 87-95. Warner, E.J. and R.B. Barsky. 1995. The timing and magnitude of retail store markdowns: evidence from weekends and holidays. The Quarterly Journal of Economics 110(2): 321-352. Wolf, M.M. 1997. A target consumer profile and positioning for promotion of the direct marketing of fresh produce: a case study. Journal of Food Distribution Research 28(3): 11-17. Wolf, M.M., A. Spittler and J. Ahern. 2005. A profile of farmers’ market consumers and the perceived advantages of produce sold at farmers markets. Journal of Food Distribution Research 36(1). Woodruff, R.B and S. Gardial. 1996. Know your customer: new approaches to understanding customer value and satisfaction. Wiley, Hoboken, NJ, USA. Xu, X., K. Keahiolalo, M.K. Loke and P. Leung. 2015a. Local premium or local discount: the case of packaged fresh tomatoes in Hawaii. Journal of Agricultural and Applied Economics 47(3): 345-358. Xu, X, M.K. Loke and P. Leung. 2015b. Is there a price premium for local food? The case of the fresh lettuce market in Hawaii. Agricultural and Resource Economics Review 44(1): 110-123. Zeithaml, V.A. 1988. Consumer perceptions of price, quality, and value: a means-end model and synthesis of evidence. The Journal of marketing 52: 2-22. Zepeda, L. and C. Leviten-Reid. 2004. Consumers’ views on local food. Journal of Food Distribution Research 35(3): 1-6. Zepeda, L. and J. Li. 2006. Who buys local food? Journal of Food Distribution Research 37(3): 385-394.

International Food and Agribusiness Management Review

100

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2017.0022

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Received: 20 February 2017 / Accepted: 29 September 2017

Efficient farming options for German apple growers under risk – a stochastic dominance approach RESEARCH ARTICLE Maren B.K. Röhrig a, Bernd Hardewegb, and Wolfgang Lentzc aResearch

Associate and bResearcher, Center for Business Management in Horticulture and Applied Research, Institute of Horticultural Production Systems, Leibniz University of Hanover, Herrenhäuser Straße 2, 30419 Hanover, Germany cProfessor,

Faculty of Agriculture/Environment/Chemistry, University of Applied Sciences Dresden (HTW Dresden), Pillnitzer Platz 2, 01326 Dresden, Germany

Abstract For a sustainable economic performance of apple production, the determination of efficient farming options considering production risk is crucial. Relying on a permanent crop, apple producers are less flexible to react upon disturbances. Based on data of 134 apple producers operating in the two main production areas in Germany, we compare and determine efficient production options. Furthermore, appropriate risk management instruments (RMIs) are identified using stochastic dominance criteria. In addition, we use Stochastic Efficiency with Respect to a Function to evaluate farming options for defined ranges of relative risk aversion. The results indicate that Red Prince is the most efficient variety in the north and subsidized hail insurance with frost irrigation is superior to frost irrigation as single RMI. In the south Braeburn should be chosen by rational decision makers, but the tested insurance solutions are not as efficient as the common practice of producing apple under hail nets. Keywords: crop insurance, risk perception, risk management, historical data approach JEL code: Q54, Q58, Q15 Corresponding author: roehrig@zbg.uni-hannover.de

101

Röhrig et al.

Volume 21, Issue 1, 2018

1. Introduction

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Apple production is a challenging business. Aspects of pests and diseases, changing market demands as well as weather conditions and volatile prices are predominant sources of risk, which have to be considered during the planning phase of an apple orchard (Catalá et al., 2013; Menapace et al., 2013). Due to the high complexity, this article focuses on the latter two issues of farm planning for apple production, as they represent two of the key risks (Ahmed and Serra, 2015). Farmer organizations have repeatedly pleaded for state subsidization of multi-peril insurance for fruit production. The debate was revived most recently by the occurrence of unusually strong late frosts in April 2017 inflicting severe damage on fruit and wine production especially in Southern Germany. The debate is often based on incidental information and – even when referring to risk – mostly ignores the nature of risk, which may be due to the complexity of the phenomenon and lacking data. This paper contributes to closing this gap for apple production. Therefore, farmers’ risk perception is compiled and distributions of the net present value (NPV) of different risk management strategies for apple production are calculated. The results indicate that a multi-peril insurance even with subsidies is dominated by other combinations of risk management instruments (RMIs). This explains the relatively low adoption of frost insurance in Southern Germany and provides evidence in favor of market-based solutions without government interference. For German apple growers, available risk management strategies to cope with weather-related risks are hail nets and frost irrigation. In the north of Germany subsidies for these risk management strategies are not available, whereas in the south hail nets are subsidized up to 50% by producer organizations (Dirksmeyer et al., 2014: 59-60). In addition, hail insurance that protects against revenue loss is available in both regions. For hail insurance, no governmental subsidy schemes exist (Bielza Diaz-Caneja et al., 2009). Even if producer organizations subsidize insurance policies, farmers, especially in the south of Germany, often decide not to participate in hail insurance, as high premium rates are common. Political programs to support apple growers in reducing risk require information on farmers’ risk behavior. According to the subjective expected utility framework, risk perception besides risk preference, is the main factor determining risk behavior. The former is the probability an individual associates with a particular uncertain situation and the likelihood to be susceptible to a specific event (e.g. Pennings et al., 2002). Knowledge of apple producers’ risk perceptions provides essential information for the development of political programs (Menapace et al., 2012). Deterministic crop budgets for an economic assessment of apple production systems in Germany are available (KTBL, 2010). However, no information is available on which risk management strategies are the most promising ones and whether new strategies, for instance combined frost-hail insurances, could provide appropriate instruments for apple growers in Germany. This article aims to evaluate combinations of different production systems (i.e. choice of variety and planting density) and RMIs according to their economic performance under different levels of apple growers’ relative risk aversion. The insights are also of political relevance as calculations of agricultural policy measures often rely only on cost-benefit analyses, based on weighted average values in terms of money and do not consider the effect of risk aversion. As a result, inappropriate conclusions are drawn when designing risk mitigation programs (Kaufman, 2014). For an in-depth risk analysis, a survey assessing perceived risk of German apple farmers in the two most important production regions (Altes Land, Lower Saxony and Lake Constance) has been conducted. After combining the data obtained from this survey with historical information stochastic dominance relations were applied in order to determine appropriate farming strategies.

International Food and Agribusiness Management Review

102

Röhrig et al.

Volume 21, Issue 1, 2018

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

2. Literature review Up to now only a very limited number of studies deal with apple growers’ behavior towards risk in industrialized countries. Menapace et al. (2014) analyzed risk perception of apple farmers in Italy in the context of climate change hazards at province level. For a long-term perspective of twenty years, respondents believing in climate change stated significantly higher probabilities for suffering from weather and disease related effects, than non-believers, whereas perceptions for the short-term view did not differ significantly (Menapace et al., 2014). Results of their survey further indicate a strong effect of different heuristics in farmers’ decision making processes. These are mental simplifications which reduce the complexity within decision-making processes. A significant effect was observed for availability heuristics (use of experience from the past for future decisions), representativeness heuristics (alignment of unfamiliar events with familiar ones) and biased assimilation (preexisting attitudes that lead persons to acquire indications which support their opinion and to reject indications against it) (Menapace et al., 2012). For conventional and organic apple production in the Pacific Northwest of the U.S. Chen et al. (2007) compare the risk reduction properties of a multi-peril crop insurance covering yield shortfalls to an incomebased insurance, focusing on deviations of income as a product of yield and prices. Their analysis is based on historical price and yield data. In the context of government subsidies, they find, that an income-based insurance would be more cost efficient than multi-peril crop insurance. However, as the associated certainty equivalents (CE) reveal, the income-based insurance provides a lower welfare, with only one variety-dependent exception (Chen et al., 2007). The use of historical data as the single source for probability estimates is, however, not advisable for risk analysis that addresses an uncertain future. The predictions may be insufficient, because underlying circumstances might change over time. Therefore, an appropriate risk analysis should include subjective probability estimates as well (Hardaker et al., 2004: 62-63; Lien et al., 2011). The historical data approach (Hardaker et al., 2004: 80-82) allows one to combine historical data and the farmers’ subjective probability estimates to reproduce the correlation structure and therefore to account for stochastic dependencies (Hardaker et al., 2004: 168-169). Lien and Hardaker (2001) use this technique to evaluate the appropriateness of different subsidy schemes for the Norwegian agricultural sector with a utility-efficient programming model and Lien et al. (2011) apply it for the calculation of gross margins of a typical Norwegian lowland farm. For capturing risk, it is recommended to work with probability distributions (Hardaker, 2000; Lien, 2003). Clancy et al. (2012) use a stochastic budgeting model in their work. In comparison to deterministic models, this approach is more appropriate to consider various uncertainties, as for example volatile prices, yields, costs and weather conditions, all factors, which are simultaneously affecting revenues and profits in farmers’ reality. For all variables of interest, stochastic budgeting assigns probabilities to values, resulting in probability distributions (Clancy et al., 2012; Lien et al., 2007a). Ranking farming options according to their efficiency under consideration of the associated cumulative density functions (CDF) and underlying farmers’ risk attitudes may be achieved by applying Stochastic Dominance (SD) criteria, Stochastic Dominance with Respect to a Function (SDRF) or Stochastic Efficiency with Respect to a Function (SERF). Presuming a positive marginal utility, a ranking based on First Degree Stochastic Dominance (FSD) is appropriate, if CDFs do not cross. If the condition Fa(x)≤Fb(x) for all x with at least one strict inequality is met, farming option a dominates option b independently of the underlying risk attitude. However, if an intersection exists, Second Order SD (SSD) needs to be applied. It requires risk aversion for all values of x, which means that the associated utility function is positive with a decreasing x* x* slope. Under SSD option a dominates option b if ∫-∞ Fa(x)dx≤∫-∞ Fb(x)dx for all x* with at least one strict inequality (Hardaker et al., 2004: 147-150; Smidts, 1990: 125-126). Similar to SSD, where limits regarding risk attitude (r) are set as 0<r<∞, SDRF defines also positive, but more restrictive boundaries for risk attitude (r1<r<r2), which allows a stricter discrimination (Hardaker et al., 2004: 153). Harper et al. (2013) apply SDRF for an evaluation of apple varieties, namely Crimson Gala, Ginger Gold and Fuji, as well as training International Food and Agribusiness Management Review

103

RĂśhrig et al.

Volume 21, Issue 1, 2018

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

systems with respect to their associated net returns for an eight year harvest period. Data for the analysis were obtained during a ten-year field experiment in Pennsylvania (USA). They observe that higher net returns are afflicted with higher risk. SDRF analysis further indicates that growers, independent of their risk attitude, prefer Fuji as cultivar (Harper et al., 2013). For SERF-analysis values of a utility function are converted by the inverse utility function into CE for a given range of risk aversion coefficients. CEs have the advantage that they can be expressed in monetary terms. The CE is the sure payment which provides the same utility as a risky prospect (Hardaker et al., 2004: 153-155; Lien et al., 2007a). Similar to SDRF, SERF relies on a range of risk aversion coefficients, but with the additional assumption that parameters of risk aversion remain constant for varying levels of payoffs (Hardaker and Lien, 2010). Recently, Schenk et al. (2014) applied SD as well as SERF for assessing Australian farmersâ&#x20AC;&#x2122; decision making concerning crop-choice, focusing on five arable crops and pasture, given uncertain amounts of water supply. Similarly, Clancy et al. (2012) considered the previously mentioned methods for their evaluation of the economic efficiency regarding biomass crops in Ireland. Up to now investment decisions for apple growers in Germany have not been analyzed considering the main sources of risk, different risk management tools and alternative risk protection strategies. The objective of this study is therefore to determine the most efficient farming options by applying stochastic dominance criteria and SERF to data of net present values (NPVs) for investments in apple orchards.

3. Data and methods To obtain a stochastic ranking of farming options, the deterministic budget is extended in order to calculate cumulative distribution functions (CDFs) of the NPV for apple orchard investments. The NPV is calculated over 16 years of full bearing for a combination of one hectare of a certain variety and the respective risk management strategies by summing up the discounted net cash flows simulated for each year. The juvenile phase of the orchard in the initial three years after planting is considered as a deterministic component of the NPV. For risk ranking, stochastic dominance criteria are subsequently applied to the CDFs. 3.1 Survey sample Apple production on owner-operated farms in Germany is concentrated in two regions, the Altes Land, located in the north at the mouth of the river Elbe, and the Lake Constance area in the south near the Alps. As the distance between these areas amounts to 900 kilometers, climatic conditions are different. In the north especially late frosts lead to higher yield and quality reductions, whereas in the south hail events are more frequent and pose a major risk to fruit quality. During the winter season 2013/2014, the apple growers were first contacted by local extension and research stations. A number of 500 growers in each region received an invitation letter or a call for participation in the newsletter of producer organizations. Starting with 16 volunteers in the north and 3 in the south, a pyramid scheme was used to acquire further participants. In the end, data of 66 farmers from the north and 68 from the south were collected through two-hour face-to-face interviews. Besides information on farmersâ&#x20AC;&#x2122; risk perception, details on their risk attitude were obtained. 3.2 Elicitation of subjective probabilities For the elicitation of probabilities, the estimation of probabilities based on the experience technique was applied, as it only requires three values and in consequence, represents one of the simplest question frameworks (Hoag, 2010: 212-213). However, only estimations of yield under normal conditions were successfully elicited with this technique, resulting in a Program Evaluation and Review Technique (PERT) Distribution. In contrast, when focusing on losses and quality reductions due to weather related risks as well as prices, the pretest revealed, that farmers do not feel comfortable to assign a minimum, maximum and modal International Food and Agribusiness Management Review

104

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Röhrig et al.

Volume 21, Issue 1, 2018

value. As applied in the work of Menapace et al. (2013), the fixed value method was used in this study in order to assess distributions. To achieve a reduction in bias, farmers were asked to recall the frequency of occurrence for different events in the past 10 years, before they stated their estimates for the upcoming decade in both frameworks. A time interval of ten years was set, since longer time intervals might result in a lower willingness to participate and a decline in attention during the interview. After recapitulation of the past, farmers were asked to indicate their expectations for the upcoming production years by allocating ten years to given intervals of losses and prices, respectively. These absolute frequencies were converted into relative ones and the midpoints of the given intervals were used for further calculations. In order to evaluate the preventive effect of RMIs, apple growers were asked to give their estimates for all circumstances, i.e. under absence and existence of the RMIs. An example is given in Figure 1. 3.3 Parameter setting for Stochastic Efficiency with Respect to a Function SERF analysis requires the choice of a utility function, which is not a trivial task. With focus on terminal wealth constant relative risk aversion (CRRA) is recommended as it is unaffected by different levels of wealth. In contrast, constant absolute risk aversion is more convincing for transitory income, which is relatively small in relation to wealth (e.g. Hardaker and Lien, 2010). As apple farms in Germany are less diversified and wealth is predominantly determined through a long term success of the orchard, CRRA will be used in this study, represented by the following functional form of utility (Equation 1). U=

1 W (1–rr(WT)), WT > 0 1 – rr(W ) T

(1)

Where WT is total wealth rr(W ) is relative risk aversion related to total wealth T

Please state the number of years in which yield losses occurred due to hail by focusing on the last ten years: ____ No. of years Your expectation for the next ten years: How often will hail lead to the following losses (%). Please allocate ten years to the given intervals:

If your expectation is based on the existence of hail nets, which losses would you expect under the absence of hail nets?

Loss due to hail (%)

Absolute (estimation) 7

Relative (calculated)

0.7

1-4

5-9

10-19

20-29 30-39

Relative (calculated) 0.5

20-29 2

30-39

0.2

40-49

50-59

60-69 70-79

Absolute (estimation)

60-69 1

70-79

0.1

80-89

90-100

Figure 1. Application of the fixed value method related to the survey design. International Food and Agribusiness Management Review

105

0.5

RĂśhrig et al.

Volume 21, Issue 1, 2018

CRRA implies constant relative risk aversion and in consequence decreasing absolute risk aversion as the absolute amount of money for risk-investments increases with increasing wealth, whereas the relative proportion remains constant.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

As described in Lien et al. (2007b) the total wealth is assumed to be WT=W0+Ws. Where W0 is the nonstochastic wealth, equaling 45,000 â&#x201A;Ź per hectare, and Ws is the stochastic wealth of apple production. Even if the simulation is based on a one hectare level, which is relatively small compared to whole-farm wealth, annual gain from one hectare is seen as a permanent source of income. Upscaling of the area planted leads to a large portion of terminal wealth and thus, the relative risk aversion coefficient is assumed to be constant (cf. Hardaker et al., 2004: 112). The range of risk aversion coefficients of 0 to 3.00 is set according to the results of the risk attitude analysis. Risk attitude was elicited with a hypothetical, farm profit-framed Holt and Laury lottery, originally pioneered by Holt and Laury (2002). About half of apple growers exhibited risk aversion, characterized by risk aversion coefficients above zero (Table 1). In this paper we follow a normative approach in order to give advice how risk averse farmers should behave under uncertainty. Thus, only data of risk averse farmers are provided. In this context, 38% of risk averse apple growers can be described as nearly risk neutral with risk aversion coefficients close to zero, whereas the others indicated stronger tendencies to risk averse behavior. 3.4 Description of the model and calculation of key variables The risk model was developed in MS Excel (Microsoft Corporation, Redmond, WA, USA). Regarding the key variables of the model, parameters, which are substantial risky determinants of revenue, were set as stochastic ones (cf. Clancy et al., 2012). As reported by Bravin et al. (2009), yield and quality have an important impact on farm profit, whereas production costs are less important. Therefore, it was decided to treat production costs (except for those proportional to yield) as deterministic variables, which can be taken from the literature (KTBL, 2010), whereas yield and quality under non-hazardous conditions of production, as well as prices and weather related impacts, i.e. frosts, hail and sunburn, are considered as stochastic variables. In addition, the event of fire blight (Erwinia amylovora), a bacterial infection, is included as a stochastic variable in the simulation as a rare but severe event. After combining historical yield and price data with subjective probabilities, the Palisade add-in @Risk for Latin hypercube simulation (@Risk 6.0 Industrial Edition, Palisade Corporation, Ithaca, NY, USA), an advancement of the Monte Carlo simulation, is used to generate probability distributions for stochastic variables of interest. For the simulation, 5,000 iterations were performed. Figure 2 shows the program flow of the simulation model, which is described in detail, afterwards. The historical data approach was used for implementing stochastic dependencies between subjective risk perceptions of yield and price variables (Hardaker et al., 2004: 80-82). As the focus is on region-specific hazards rather than single farm simulations, the means of the subjective estimates were determined. Price data for years 1993-1996 were obtained from ZMP (1998: 77-79), for 1997-2001 from ZMP (2002: 68Table 1. Elicited relative risk aversion coefficients. Risk aversion coefficient

Absolute frequency

Relative frequency (%)

Level of risk attitude

0.155 0.470 0.815 1.265 2.000 n

25 15 14 3 8 65

38 23 22 5 12 100

Risk aversion level 1 Risk aversion level 2 Risk aversion level 3 Risk aversion level 4 Risk aversion level 5

International Food and Agribusiness Management Review

106

Röhrig et al.

Volume 21, Issue 1, 2018

Historical data (input data)

Subjective estimates (input data) Yield – normal conditions (Y no risk)

Historical yield (Yh)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Simulation activities 5000 iterations for years 4-20 (a)

Yield reduction: frosts (FYR)

Insurance parameters (QYL and QRYL)

Yield reduction: hail (HYR)

Revenue (R)

Yield reduction: fire blight (FBYR)

Yield specific costs

Quality – normal conditions (HQ no risk) Quality reduction: hail (HQR) -

Deterministic costs

Quality reduction: sunburn (SBQR) Market price (extra, class 1) (Pclassdk)

= Discounted cash flows (years 4-20)

Market price (class 2) (Pclassdi)

Historical prices (Ph)

Market price (processing fruit) (Pclassl) Farmer-to-consumer direct marketing (extra, class 1) (Pclassek)

Deterministic cash flows (years 0-3)

Farmer-to-consumer direct marketing (class 2) (Pclassei)

= Net present value

Figure 2. Schematic overview of the model’s input data and stochastic calculations. 69), for 2002-2006 from ZMP (2006: 25-27), and for 2007-2012 from AMI (2013: 74-76) and adjusted for inflation, using the consumer price index with 2010 as the base year, provided by the Federal Statistical Office (2014). Yield data were taken from the Landesbetrieb für Statistik und Kommunikationstechnologie Niedersachsen (2008, 2012). Equation 2 provides an example for the combination of historical prices with subjective estimates. Pvy = E[Psv] + {[Phvy – E[Phv]] / σ[Phv]} × σ[Psv] (2) Where E[Psv] is the expected value E of the subjective price estimation PS of the variety v Phvy is the historical price of variety v in year y E[Phv] is the expected value of the historical price Ph of the variety v σ[Phv] is the standard deviation of the historical price Ph of the variety v σ[Psv] is the standard deviation of the subjective price estimation PS of the variety v The following calculations of yield and qualities under the consideration of risky events are based on subjective estimates. Risk-adjusted yield (Yafter risk) and the percentage of high quality apples (HQ) of the actual year a are calculated as follows (Equations 3 and 4). Ya after risk = Ya no risk × (1 – FYRa) × (1 – HYRa) × (1 – FBYRa) (3) Where frosts (FYR), hail (HYR) and fire blight (FBYR) may lead to yield reductions.

International Food and Agribusiness Management Review

107

RĂśhrig et al.

Volume 21, Issue 1, 2018

HQa after risk = HQa no risk Ă&#x2014; (1 â&#x20AC;&#x201C; HQRa) Ă&#x2014; (1 â&#x20AC;&#x201C; SBQRa) (4)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Where HQR indicates quality reductions due to hail and SBQR due to sunburn. For the yield and quality reductions due to severe weather events, two additional assumptions were made. First, events of hail may lead to a considerable decrease in revenue if quality classes extra, one and two become apples for processing purposes. However, it was assumed that hail has a similar impact on both quality classes and thus only apples for processing purposes were distinguished from HQ. This simplification is justified by price adjustments. If adverse weather events lead to widespread damages, the price level of class two commonly equals those of classes extra and one. Thus, it was decided not to differentiate between hail damage of classes one and two separately, but instead reduce the percentage of both quality classes equally. As a consequence, the model realizes revenue reduction and accounts for higher market prices of class two simultaneously. Second, in reality frosts affect yield and quality of apples. Nevertheless, the pretest revealed that apple growers are more concerned about yield losses due to frosts. In consequence, questions addressing quality reductions due to frosts were not answered and therefore frost-related quality reductions are not considered in the model. Insurance indemnity payments are generally related to quantitative and qualitative losses. The actual yield after spring frost provides the basis for the calculation of quantitative yield loss (QYL) due to hail (Equation 5). In addition, quality-related yield loss (QRYL) is calculated after subtraction of frost and hail related yield loss (Equation 6). QYLa = Ya (1 â&#x20AC;&#x201C; FYRa) Ă&#x2014; HYRa (5) QRYLa = Ya (1 â&#x20AC;&#x201C; FYRa) Ă&#x2014; (1 â&#x20AC;&#x201C; HYRa) Ă&#x2014; HQa Ă&#x2014; HQRa Ă&#x2014; QQuota (6) Where QRYL indicates the amount of high quality apple HQa is the share that becomes processing fruit after hail (HQRa). This value is further multiplied with QQuota, which is determined as a loss ratio of 70% and represents a common rating, applied by an insurance company in Germany (Vereinigte Hagel, 2017). The sum of quantitative and qualitative losses due to hail, divided by expected yield after excluding the effect of losses from spring frosts, result in the total loss ratio from hail (TLRH). As the model is confined to the years of full harvest, the expected yield is calculated as a mean of production years four to twenty (Equation 7). â&#x2C6;&#x2019;1

1 đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; = [(đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; + đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; ) â&#x2C6;&#x2014; [ â&#x2C6;&#x2014; (â&#x2C6;&#x2018; đ?&#x2018;&#x152;đ?&#x2018;&#x152;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; â&#x2C6;&#x2014; (1 â&#x2C6;&#x2019; đ??šđ??šđ??šđ??šđ??šđ??šđ?&#x2018;&#x17D;đ?&#x2018;&#x17D; ))] ] â&#x2C6;&#x2014; 100 (7) 16 đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;=4

The total sum insured equals expected revenues as a mean of the production years four to twenty, considering the higher quality classes extra, one and two and their respective market prices. If the decision maker participates in frost insurance, indemnity payments for covering frost damages are subtracted. Multiplying the sum insured with the TLRH leads to the total amount of economic loss. A percentage of this economic loss represents deductibles, which are paid by the apple grower. For deductibles conventional calculations considering the TLRH of a single year were used (Vereinigte Hagel, 2017). After deductibles have been subtracted, the value of the economic loss equals the indemnity payment. For the calculation of the hail insurance premium, the basic insurance premium (IP) equals 10% of the sum insured in the north and 21% in the south. As the insurance premium has to be adjusted in order to consider the extent and the variation of overall damages, it is further multiplied with a correction factor. For the first nine years of full harvest, the factor equals 100% and for the following years it is determined on the basis of the average TLRH observed during the ten previous years.

International Food and Agribusiness Management Review

108

RĂśhrig et al.

Volume 21, Issue 1, 2018

The calculation of parameters regarding frost insurance follows the same procedure as explained for hail. The sum insured equals the calculated sum insured for hail. Only information of yield losses due to frosts (Equation 8) and the total loss ratio of frosts (TLRF) (Equation 9) are required for the calculation of the insurance premium and indemnity payments.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

QYF = Ya Ă&#x2014; FYRa (8) 20

â&#x2C6;&#x2019;1

đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x2021;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; = [đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; â&#x2C6;&#x2014; (â&#x2C6;&#x2018; đ?&#x2018;&#x152;đ?&#x2018;&#x152;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; â&#x2C6;&#x2014; 16â&#x2C6;&#x2019;1 ) ] â&#x2C6;&#x2014; 100

(9)

đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;=4

In line with existing frost insurance schemes deductibles are not calculated. Furthermore, information on quality reduction is not available and therefore not considered in this study. For frost insurance 7 and 2% represent the basic insurance premium levels for the north and south, respectively. As for hail, the premium is multiplied with an adjustment factor, which is set at 100% for the first nine years and relies on the average TLRF of the previous ten years, afterwards. Finally, the amount of annual indemnity-payments is obtained as TLRF is multiplied with the sum insured. Even if frost insurance as RMI has not been established for apple production in Germany, it is already available as a RMI in neighboring countries. Thus, it is suggested as an alternative to frost irrigation in the north and as a supplement in the south. Revenues are calculated as shown in Equation 10. đ?&#x2018;&#x2019;đ?&#x2018;&#x2019;

đ?&#x2018;&#x2013;đ?&#x2018;&#x2013;

đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; = â&#x2C6;&#x2018; â&#x2C6;&#x2018;[ ( đ?&#x2018;&#x152;đ?&#x2018;&#x152;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x; â&#x2C6;&#x2014; đ??ťđ??ťđ??ťđ??ťđ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x; ) â&#x2C6;&#x2014; (1 â&#x2C6;&#x2019; đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; ) â&#x2C6;&#x2014; đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x2014;đ?&#x2018;&#x2014;đ?&#x2018;&#x2014;đ?&#x2018;&#x2014; â&#x2C6;&#x2014; đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192; đ?&#x2018;?đ?&#x2018;?đ?&#x2018;?đ?&#x2018;?đ?&#x2018;?đ?&#x2018;? ] đ?&#x2018;?đ?&#x2018;?=đ?&#x2018;&#x2018;đ?&#x2018;&#x2018; đ?&#x2018;&#x2014;đ?&#x2018;&#x2014;=đ?&#x2018;&#x2DC;đ?&#x2018;&#x2DC;

+ [ đ?&#x2018;&#x152;đ?&#x2018;&#x152;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x; â&#x2C6;&#x2014; (1 â&#x2C6;&#x2019; đ??ťđ??ťđ??ťđ??ťđ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x; )] â&#x2C6;&#x2014; đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192; đ?&#x2018;&#x2122;đ?&#x2018;&#x2122;đ?&#x2018;&#x2122;đ?&#x2018;&#x2122; đ?&#x2018;&#x2019;đ?&#x2018;&#x2019;

đ?&#x2018;&#x2013;đ?&#x2018;&#x2013;

(10)

+ â&#x2C6;&#x2018; â&#x2C6;&#x2018;( đ?&#x2018;&#x152;đ?&#x2018;&#x152;(đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;â&#x2C6;&#x2019;1) đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x; â&#x2C6;&#x2014; đ??ťđ??ťđ??ťđ??ť(đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;â&#x2C6;&#x2019;1) đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;đ?&#x2018;&#x17D; đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x;đ?&#x2018;&#x; ) đ?&#x2018;?đ?&#x2018;?=đ?&#x2018;&#x2018;đ?&#x2018;&#x2018; đ?&#x2018;&#x2014;đ?&#x2018;&#x2014;=đ?&#x2018;&#x2DC;đ?&#x2018;&#x2DC;

â&#x2C6;&#x2014; đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;(đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;â&#x2C6;&#x2019;1) â&#x2C6;&#x2014; đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x201E;đ?&#x2018;&#x2014;đ?&#x2018;&#x2014;(đ?&#x2018;&#x17D;đ?&#x2018;&#x17D;â&#x2C6;&#x2019;1) â&#x2C6;&#x2014; (đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192;đ?&#x2018;&#x192; đ?&#x2018;?đ?&#x2018;?đ?&#x2018;?đ?&#x2018;?đ?&#x2018;?đ?&#x2018;? â&#x2C6;&#x2014; đ?&#x2018;&#x201C;đ?&#x2018;&#x201C;đ?&#x2018;?đ?&#x2018;?đ?&#x2018;?đ?&#x2018;? )

R indicates revenue, S the amount of apples stored and Qclassj the percentages of the classes extra and one (k) as well as class two (i) of higher quality apple. Pclasscj are the corresponding prices for qualities, referring to the two distribution channels c. Traditionally, apple growers can sell the fruits directly to consumers (market d) or via the wholesale market e. Furthermore, the variable Pclass l represents the price for processing quality (l). The price increase after storage is further considered by means of the factor f, which was calculated according to data provided by AMI (2014). Here, higher prices for stored apples are variety specific for the classes extra and one and are calculated as an increase of 4-12% of the price occurring in the actual year. In a last step direct, variable and fixed costs are subtracted from operating and non-operating (i.e. indemnities and subsidies) revenue. Costs associated with harvest and sorting of apples are considered as yield-dependent costs. Discounting with a rate of 4% and summation of the discounted cash flows leads to the NPV of the farming options. An overview of the cost calculation is provided in the supplementary material (Supplementary Table S1). The farming scenarios include varieties, which are common in the considered production areas. In the north of Germany, mainly the varieties Braeburn, Elstar Jonagored, and Red Prince are produced, whereas in the south Braeburn, Elstar, Jonagold, and Gala are the predominant ones, usually grown on M.9 rootstocks. These varieties provide the basis for the analysis of one hectare of certain farming options over a period of 16 years of full bearing capacity. To evaluate the effect of already existing as well as non-established risk management tools, the following scenarios for common varieties of each region are compared.

International Food and Agribusiness Management Review

109

Röhrig et al.

Volume 21, Issue 1, 2018

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

The first step determines the optimal planting density for each variety under standard risk management strategies. In the North, frost irrigation is usually installed, whereas in the south hail nets serve as a standard risk management strategy. The best options determined in the first step, are further analyzed in a second step. For the north frost irrigation, combined with a hail insurance is considered and for the south hail insurance (HI) is analyzed as an alternative for hail nets. The preferred options are further considered in a third and last step of the analysis, where the focus lies on a hypothetical set of combined frost-hail insurance (FI & HI) for both regions. As currently available subsidies cover up to 50% of installation costs and insurance premium costs, a twostep simulation, with and without subsidy payments is performed. For the subsidy schemes the following assumptions were made: in the south the material for hail nets, the installation costs of frost irrigation systems and insurance premiums are subsidized at a rate of 50%. In the north, the calculation is based on the sum insured multiplied by a distribution factor as laid down in the subsidy scheme of the producer organization. This distribution factor represents 1% of total apple sales and claim settlements, divided by the sum insured over all enterprises. Subsidies may not exceed the costs of the insurance premium and the sum insured may not be higher than 20,000 €/ha.

4. Results and discussion It is of particular interest how apple growers can protect themselves against farm risk in open field production. For this purpose, a stochastic budgeting model was developed, which considers weather-extremes as well as price risks for miscellaneous varieties of apple and planting densities. An example of the results of the simulation model is shown in Table 2, where fruit yields are measured in decitonnes (dt). The columns indicate the simulation results of the calculated mean, the standard deviation, the coefficient of variance, as well as the 5% and 95%-percentiles and the associated Minimum and Maximum. The total revenue is calculated based on the operating revenue, representing the revenue achieved through sales activities, plus insurance indemnity payments and subsidies. The sum insured is calculated by taking the expected yield achieved without extreme weather events multiplied by market prices. In the case of Red Prince, simulated with 3,300 trees per hectare, the sum insured exceeded the maximum of 20,000 €/ha. As a subsidized scenario is represented, the simulation is calculated under consideration of the maximum mentioned above. The monetary loss due to weather events is calculated by multiplying the sum insured with the total loss ratio (cf. Equation 7 for hail and Equation 9 for frost). For calculating the amount of hail related indemnity payments, deductibles which are determined by the total loss ratio have to be subtracted as described above. Costs of the insurance are set with a basic premium rate of 10%. After nine production years, the average total loss ratio determines an adapted premium rate, which represents a variable component in the model. Multiplying the sum insured with the distribution factor of 0.0424, which is a 4-year average stated by the producer organization (C. Greisiger, personal communication), average subsides of 848 € are calculated in the following scenario. Four region-specific varieties under common practice are part of the first analytic step. Figure 3 displays the CDFs of NPVs for one hectare of the respective option for the north. In general, higher planting densities of a variety clearly dominate the lower densities in the sense of FSD, but also show higher variation. Notably, the varieties of Elstar and Jonagored at a density of 1,800 trees per hectare can be considered as less efficient, since the probability to achieve a positive NPV is small. Comparing varieties planted at the same density, Braeburn and Red Prince dominate Elstar as well as Jonagored in the sense of FSD. In comparison to Red Prince, Braeburn indicates a steeper curve and is thus less risky. Furthermore, Red Prince at 2,500 trees per hectare dominates Elstar in terms of SSD, as their associated curves cross close to p=1.0. In contrast, the discrimination of the most efficient option among Braeburn and Red Prince at 3,300 trees per hectare is not possible with FSD and SSD. Later, SERF helps to achieve a clearer differentiation. On the basis of these results, Braeburn, Jonagored as well as Red Prince at 2,500 and 3,300 trees per hectare will be part of the further analysis. Results of the second analytic step for the north are given in Figures 4 to 6. International Food and Agribusiness Management Review

110

Röhrig et al.

Volume 21, Issue 1, 2018

Red Prince 3,300 HI sub 10% IP

Mean

Net present value (€/ha) Annuity (€/ha/a) Direct costs (€/ha) Variable costs (€/ha) Fix costs (€/ha) Total yield (dt/ha) Classes extra and one (sold per year) (dt/ha) Class two (sold per year) (dt/ha) Processing fruit (sold per year) (dt/ha) Percentage stored (%) Percentage of market sales (%) Market price classes extra and one (actual year) (€/dt) Wholesale market price classes extra and one (actual year) (€/dt) Operating revenue (€/ha) Total revenue (€/ha) Discounted profits (€/ha) Sum insured (€/ha) Indemnity payments hail (€/ha) Insurance premium hail (€/ha) Subsidies (€/ha)

109,688.96 20,416.93 8,071.11 1,502.31 11,304.78 1,107.72 4,493.12 735.20 4,023.32 426.07 575.40 123.15 433.03 134.40

Stdev

0.05 perc. 0.95 perc.

0.19 0.19 0.10 0.16 0.11 0.21 0.31

77,193.81 5,680.06 9,155.60 2,951.56 3,272.73 317.18 168.82

Min

Max

145,020.98 38,497.16 10,670.90 2,832.69 12,884.90 7,869.46 5,466.56 1,212.52 4,642.44 2,180.23 738.45 25.88 633.29 20.18

187,930.53 13,828.26 13,276.55 5,466.56 5,107.31 738.45 711.93

25.95

32.54

1.25

0.00

96.67

0.00

162.56

105.85

112.50

1.06

13.27

338.86

0.68

706.50

0.95 0.87

0.01 0.01

0.93 0.85

0.97 0.89

0.91 0.84

0.97 0.90

42.88

7.55

0.18

30.17

54.64

30.17

54.64

165.67

26.21

0.16

121.52

206.50

121.52

206.50

30,679.52 10,615.82 33,505.10 11,205.97 11,247.15 9,336.99 20,000.00 0.00 1,976.87 3,430.25

0.35 0.33 0.83 0.00 1.74

12,178.06 14,456.56 -4,542.63 20,000.00 0.00

48,064.85 2,161.86 52,252.22 3,197.35 27,184.53 -14,341.62 20,000.00 20,000.00 10,042.48 0.00

60,522.03 74,242.38 47,598.98 20,000.00 20,000.00

2,000.00

0.00

2,000.00

848.71

0.00

848.71

HI = hail insurance; IP = insurance premium; dt = decitonnes; CV = coefficient of variance; 0.05 perc = 5% percentile; 0.95 perc = 95% percentile.

1.0

Braeburn 1,800 Elstar 1,800 Jonagored 1,800 Red Prince 1,800 Braeburn 2,500 Elstar 2,500 Jonagored 2,500 Red Prince 2,500 Braeburn 3,300 Elstar 3,300 Jonagored 3,300 Red Prince 3,300

0.9 0.8 0.7 Probability (p)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Table 2. Red Prince with 3,300 trees/ha, frost irrigation and a subsidized hail insurance (north).1

0.6 0.5 0.4 0.3 0.2 0.1 0.0 -60,000

60,000

120,000

180,000

Net present value (€/ha)

Figure 3. Basic scenario with frost irrigation (north). International Food and Agribusiness Management Review

111

Röhrig et al.

Volume 21, Issue 1, 2018

1.0 0.9 0.8

Braeburn 3,300 Jonagored 3,300 Red Prince 2,500 Red Prince 3,300 Braeburn 3,300 HI sub 10% IP Jonagored 3,300 HI sub 10% IP Red Prince 2,500 HI sub 10% IP Red Prince 3,300 HI sub 10% IP

0.6 0.5 0.4 0.3 0.2 0.1 0.0

60,000

120,000

180,000

Net present value (€/ha)

Figure 4. Subsidized hail insurance (HI sub) at a 10% insurance premium-level (IP) with frost irrigation (north). 1.0 0.9 0.8

Braeburn 3,300 Jonagored 3,300 Red Prince 2,500 Red Prince 3,300 Braeburn 3,300 HI no sub 10% IP Jonagored 3,300 HI no sub 10% IP Red Prince 2,500 HI no sub 10% IP Red Prince 3,300 HI no sub 10% IP

0.7 Probability (p)

0.6 0.5 0.4 0.3 0.2 0.1 0.0

60,000

120,000

180,000

Net present value (€/ha)

Figure 5. Unsubsidized hail insurance (HI no sub) at a 10% insurance premium-level (IP) with frost irrigation (north). 160,000 Braeburn 3,300 Braeburn 3,300 HI sub 10% IP Braeburn 3,300 HI no sub 10% IP Jonagored 3,300 Jonagored 3,300 HI sub 10% IP Jonagored 3,300 HI no sub 10% IP Red Prince 2,500 Red Prince 2,500 HI sub 10% IP Red Prince 2,500 HI no sub 10% IP Red Prince 3,300 Red Prince 3,300 HI sub 10% IP Red Prince 3,300 HI no sub 10% IP

140,000 Certainty equivalent (€)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Probability (p)

0.7

120,000

100,000

80,000 0.00

0.50

1.00

1.50

2.00

2.50

3.00

Relative risk aversion coefficients

Figure 6. Stochastic Efficiency with Respect to a Function analysis with the basic scenario, subsidized hail insurance (HI sub) and unsubsidized hail insurance (HI no sub) at a 10% insurance premium-level (IP) (north). International Food and Agribusiness Management Review

112

Röhrig et al.

Volume 21, Issue 1, 2018

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

As indicated in Figure 4, insured crops of all varieties dominate the associated common production practices in terms of SSD. This effect arises from both, insurance and subsidies, where the latter amount to about 850 € per hectare on average. However, with respect to Jonagored and Red Prince at a planting density of 3,300 trees per hectare the associated curves cross more than once in their upper range and thus, the application of stochastic dominance criteria does not lead to a final ranking. Figure 5 presents the results of unsubsidized hail insurance. As can be seen, insurance policies without subsidies reduce risks as the associated curves become steeper in comparison to those of the basic scenario. Again, it is not possible to judge the performance of insurance according to FSD or SSD. The SERF analysis (Figure 6) reveals that for slightly risk averse decision makers, the basic production practice, with frost irrigation as the only RMI, is more appropriate, whereas for risk averse persons an unsubsidized hail insurance provides slightly higher CEs. Nonetheless, a combination of frost irrigation and subsidized hail insurance provides the most efficient risk management strategy, irrespective of the variety. The SERFanalysis further shows that in the basic scenario with frost irrigation, Red Prince provides the most efficient option in the north over a wide range of relative risk aversion (0≤rr≤3). These results are explained by the average yield-level of Red Prince exceeding those of Elstar and Braeburn. In addition, Red Prince attains higher net revenues than Jonagored due to slightly higher prices as well as lower variable costs. In contrast, an elevated price level causes higher revenues of Braeburn whereas yield is relatively low. As Braeburn is afflicted with smaller standard deviations in revenue as well as in direct and fixed costs, its CEs remain more stable across different risk aversion coefficients in comparison to Red Prince. The results further suggest that Jonagored is less efficient despite its high yield level, since it yields lower prices in the market. Furthermore, the standard deviation of the NPV for Jonagored is similar to the one obtained for Red Prince and likewise results in a considerable CE as well as utility reduction when risk aversion increases. The high yield-level of Jonagored as well as of Red Prince result in higher direct costs, as storage and harvesting costs are increased. These results indicate that rational apple growers in the north should combine frost irrigation with subsidized hail insurance. However, data shows that 30% of the apple growers, who are already members in a producer organization, do not participate in hail insurance. This observation can possibly be explained with the effect of reference dependence. As described by Bocquého et al. (2013), prospect maximizers might be risk averse for gains, but show risk-seeking behavior in a context of losses. They accept the possibility to suffer a high loss instead of paying a certain amount of insurance premium regularly (Bocquého et al., 2013). Furthermore, the results of SERF show that unsubsidized hail insurance only leads to a slight increase in efficiency. In consequence, it is not worthwhile for slightly risk averse growers to combine frost irrigation with unsubsidized hail insurance. Nevertheless, the reduction of standard deviation in NPV amounts to 4,520.75 €/ha on average for the highest planting densities. This leads to a slight increase in efficiency compared to the basic scenario, given a high risk aversion. Please note, that the unsubsidized hail insurance was simulated without a restriction of the sum insured (20,000 €/ha), which leads to higher indemnity payments as well as to a higher decrease in NPV standard deviations. The results of NPV-CDFs calculated for the south are illustrated in Figures 7 to 8. Apparently, apple growers in the north achieve higher revenues than in the south. These differences stem from deviations of yield estimates, which are variety specific and amount to 59 and 89 dt per hectare for Braeburn and Elstar, respectively. An explaination might originate from an overestimation of yield risks. As the results of Menapace et al. (2014) indicate, persons who experienced specific risks in the past, show a significant increase in their associated risk perception for future events. An additional question in our survey captures the influence of the two main weather related risks in the past ten years. In the north, 28.8% stated that the operating income of the enterprise was severely or more than severely affected due to hail, whereas in the south even 48.5% indicated a strong impact of hail. In contrast to yield risks, differences in price levels for market prices are quite small and between 5 to 10 € per dt, whereas variations of wholesale market prices up to 40 € per dt have presumably a higher effect on revenue.

International Food and Agribusiness Management Review

113

RĂśhrig et al.

Volume 21, Issue 1, 2018

1.0

Braeburn 2,500 Elstar 2,500 Gala 2,500 Jonagold 2,500 Braeburn 3,000 Elstar 3,000 Gala 3,000 Jonagold 3,000 Braeburn 3,500 Elstar 3,500 Gala 3,500 Jonagold 3,500

0.9 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 -60,000

60,000

120,000

180,000

Net present value (â&#x201A;Ź/ha)

Figure 7. Basic scenario with hail nets (south). 1.0 0.9 Braeburn 3,000 Braeburn 3,500 Gala 3,500 Jonagold 3,500 Braeburn 3,000 HI sub 21% IP Braeburn 3,500 HI sub 21% IP Gala 3,500 HI sub 21% IP Jonagold 3,500 HI sub 21%IP

0.8 0.7 Probability (p)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Probability (p)

0.8

0.6 0.5 0.4 0.3 0.2 0.1 0.0 -30.000

30.000

90.000

150.000

Net present value (â&#x201A;Ź/ha)

Figure 8. Subsidized hail insurance (HI sub) at a 21% insurance premium-level (IP) as an alternative for hail nets (south). In the basic scenario for the south, Braeburn Gala and Jonagold are more profitable than Elstar at equivalent planting densities and dominate the latter in the sense of FSD. Thus, results suggest that Elstar does not provide an efficient option in either area, as it shows a lower level of yield and therefore lower revenue. The reason for the low yield of Elstar may in part be explained by its high tendency for alternate bearing (Atay et al., 2013; Untiedt and Blanke, 2001). As a consequence, it is reasonable to suppose, that yield estimates of Elstar lie below of those of other varieties. Furthermore the basic scenario for the south shows that the CDFs for Jonagold indicate a higher risk, as their course is not as steep as the curves that represent the other varieties. Similar to the north, all varieties simulated at the highest planting density of 3,500 trees per hectare dominate the lower ones in sense of FSD. When focusing on the highest planting density, Jonagold is dominated by Braeburn in terms of SSD. However, no clear ranking according to FSD or SSD is observable when focusing on Braeburn at 3,000 trees per hectare as well as on Gala and Jonagold at 3,500 trees per hectare. Thus, Braeburn at 3,000 trees per hectare together with Braeburn, Gala and Jonagold at 3,500 trees per hectare are considered in the second part of the analysis. Figure 8 shows the results for hail insurance as an alternative choice to hail nets in the south. Despite subsidies covering 50% of the premium, hail insurance seems to provide no appropriate solution, since the variety specific comparison of the RMI reveals a decrease in efficiency. However, hail insurance combined with high density and profitable varieties as Braeburn or Jonagold at 3,500 trees per hectare appear as efficient as Braeburn at 3,000 trees or Gala at 3,500 trees per hectare. One may recognize that distances between the International Food and Agribusiness Management Review

114

Röhrig et al.

Volume 21, Issue 1, 2018

Figure 9 summarizes the results for the south in terms of CEs. Generally, subsidized hail insurance in the south is less efficient than growing the same variety under hail nets. Subsidized hail insurance can only compete with the basic scenario, if the apple grower is risk neutral to slightly risk averse and chooses varieties characterized by a high yield or an elevated price level. For risk neutral apple growers, subsidized insurance solutions of Braeburn and Jonagold at 3,500 trees per hectare are as efficient as the production under hail nets of Braeburn at 3,000 trees and Gala at 3,500 trees per hectare. However, as the SERF analysis reveals for Gala, the underlying risk attitude may be important. As Gala at 3,500 trees under hail net is still afflicted with lower risks it shows higher CEs for risk averse individuals, who should give priority to this option. This result makes evident that regardless of the variety specific decrease of standard deviation achieved by the participation in insurance, the general risk of a variety has to be taken into account. Furthermore, the results indicate that without subsidies, hail insurance would be clearly dominated by common practice with hail nets and thus provides no reasonable alternative. Therefore, subsidies appear as a certain and non-negligible source of revenue. The results further suggest that Braeburn is the most efficient variety due to its higher price level, even if the average yield of Braeburn is below that of Gala and Jonagold. Regarding the NPV’s standard deviations, Jonagold shows the highest and Gala the lowest risk. Compared to Jonagold, standard deviations of Gala regarding the operating revenue, as well as the direct, fixed and variable costs are smaller. Similarly, Braeburn shows a lower standard deviation with respect to the operating revenue. Consequently, and as the SERF analysis reveals, Braeburn is afflicted with lower risk and risk averse individuals should opt for Braeburn instead of choosing Jonagold. With respect to costs, Jonagold shows higher fixed and variable costs, as its high yield leads to higher labor costs for harvesting. 115,000

Certainty equivalent (€)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

basic scenario and subsidized hail insurance for Braeburn at a planting density of 3,500 trees per hectare are larger than for 3,000. This effect stems from an increase of expected yield, which is coupled with higher premium costs. Average values, obtained for a planting density of 3,500 trees per hectare, indicate that indemnity payments and subsidies do not meet revenue loss occurring without hail nets, which leads to a financial loss of about 632 €. Furthermore, direct costs increases up to 19.90%. With respect to unsubsidized hail insurance all basic scenarios dominate their analogs with unsubsidized hail insurance according to FSD.

Braeburn 3,000 Braeburn 3,000 HI sub 21% IP Braeburn 3,000 HI no sub 21% IP Braeburn 3,500 Braeburn 3,500 HI sub 21% IP Braeburn 3,500 HI no sub 21% IP Gala 3,500 Gala 3,500 HI sub 21% IP Gala 3,500 HI no sub 21% IP Jonagold 3,500 Jonagold 3,500 HI sub 21% IP Jonagold 3,500 HI no sub 21% IP

95,000

75,000

55,000

35,000 0.00

0.50

1.00

1.50

2.00

2.50

3.00

Relative risk aversion coefficients

Figure 9. Stochastic Efficiency with Respect to a Function analysis with the basic scenario, subsidized hail insurance (HI sub) and unsubsidized hail insurance (HI no sub) at a 21% insurance premium-level (IP) (south).

International Food and Agribusiness Management Review

115

Röhrig et al.

Volume 21, Issue 1, 2018

160,000 Braeburn 3,300 Braeburn 3,300 HI sub 10% IP Braeburn 3,300 FI sub 7% IP & HI sub 10% IP Jonagored 3,300 Jonagored 3,300 HI sub 10% IP Jonagored 3,300 FI sub 7% IP & HI sub 10% IP Red Prince 2,500 Red Prince 2,500 HI sub 10% IP Red Prince 2,500 FI sub 7% IP & HI sub 10% IP Red Prince 3,300 Red Prince 3,300 HI sub 10% IP Red Prince 3,300 FI sub 7% IP & HI sub 10% IP

Certainty equivalent (€)

140,000

120,000

100,000

80,000 0.00

0.50

1.00

1.50

2.00

2.50

3.00

Relative risk aversion coefficients

Figure 10. Stochastic Efficiency with Respect to a Function analysis with the basic scenario, subsidized hail insurance (HI sub) at a 10% insurance premium-level (IP) and combined frost-hail insurance (FI sub & HI sub) at a 7% IP for frost (north).

115,000 Braeburn 3,000 Braeburn 3,000 HI sub 21% IP Braeburn 3,000 FI sub 2% IP & HI sub 21% IP Braeburn 3,500 Braeburn 3,500 HI sub 21% IP Braeburn 3,500 FI sub 2% IP & HI sub 21% IP Gala 3,500 Gala 3,500 HI sub 21% IP Gala 3,500 FI sub 2% IP & HI sub 21% IP Jonagold 3,500 Jonagold 3,500 HI sub 21% IP Jonagold 3,500 FI sub 2% IP & HI sub 21% IP

95,000 Certainty equivalent (€)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

As mentioned before, RMIs for apple growers against weather related risks are rare. Thus, subsidized hailfrost insurance as a (so far) hypothetical alternative was implemented. The associated CEs are plotted together with subsidized hail insurance and the basic scenario in figures 10 and 11. Figure 10 presents the results for the north at a 7% insurance premium level for frost insurance. The insurance premiums were obtained by a comparison of average premium costs and indemnity payments.

75,000

55,000

35,000 0.00

0.50

1.00

1.50

2.00

2.50

3.00

Relative risk aversion coefficients

Figure 11. Stochastic Efficiency with Respect to a Function analysis with the basic scenario, subsidized hail insurance (HI sub) at a 21% insurance premium-level (IP) and combined frost-hail insurance (FI sub & HI sub) at a 2% IP for frost (south).

International Food and Agribusiness Management Review

116

Röhrig et al.

Volume 21, Issue 1, 2018

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

In the north, frost-hail insurance generally does not provide an efficient alternative to frost irrigation. The results suggest that subsidized, combined frost-hail insurance in the north is only attractive for very risk averse decision makers, even though the effect is variety-specific and depends on the planting density, as can be seen for Braeburn and Red Prince at 3,300 trees per hectare. Only extremely risk averse decision makers would obtain slightly higher CEs compared to the basic scenario of Red Prince at 2,500 trees per hectare and Jonagored at 3,300 trees per hectare. However, a disruption in the capability to supply regular customers with apples might jeopardize the business relationships in real life and apple growers are expected to prefer frost irrigation systems rather than insurance solutions. In the south the basic (and most commonly found) farming options remain the most efficient ones, compared to subsidized insurance solutions (Figure 11). A combined frost-hail insurance with a 2% insurance premium would lead to slightly higher incomes due to subsidies, whereas damages due to late frosts are marginal and indemnity payments for high density plantations amount to 300 € in average. The CEs of the combined hailfrost insurance are close to those of single hail insurance, although the associated values always lie above due to the assumption of higher subsidy payments. Thus, the added value of a frost-hail insurance is low. In 2017 frosts caused high damage in fruit and wine yards, located in the south of Germany. Coble and Barnett (2012) describe that ex post disaster assistance via direct payments in the United States reveal that these payments in contrast to insurance programs do not provide a support in sense of risk protection. In order to reduce the demand of ex post disaster payments, subsidies for insurance contracts clearly represent an incentive to increase the number of insurance contracts (Coble and Barnett 2012). Therefore, a subsidized multi-peril insurance could provide an appropriate solution to cope with damages due to frosts and hail. For example, a commercial multi-peril insurance is available for apple production in the Netherlands covering frost and hail damages in combination with other weather-related risks. This insurance receives a government subsidy up to 65% of the insurance premium (Berkhout et al., 2016). In contrast, our results indicate that the added value of a multiperil-insurance is low considering the estimated risk situation. Nevertheless, the effects of climate change may increase the occurrence of late frosts and multi-peril-insurances could become more relevant. Finally, a potential criticism regarding the use of the NPV as the stochastic investment criterion should be addressed. Using the NPV implies an aggregate evaluation of the total simulation results over the economic life of the apple orchard, which tends to level the effect of a catastrophic year that could have caused bankruptcy. This could lead to an underestimation of the true risk. In line with Clancy et al. (2012), it is assumed that each variety on a 1 ha basis only represents a rather small percentage of farming activities and farmers’ wealth, whose failure would not likely lead to insolvency of the enterprise. Also from a marketing perspective, apple growers are required to produce a certain mix of apple varieties to meet their customers’ demands, which precludes the recommendation of a single variety.

5. Conclusions Results of the present study reflect observed behavior in reality, where apple growers successfully apply available risk management strategies in their respective regions. In the north, Red Prince appears as the most efficient variety. Furthermore, subsidized hail insurance would provide benefits for risk averse farmers in general, whereas an unsubsidized hail insurance is only more efficient if the apple grower is highly risk averse. In the south, none of the considered RMI provides a more appropriate alternative to common practices of using hail nets when the same variety and planting density are considered. Even if recent events of frost damages in the south arousing thoughts of developing multiperil insurance programs, the results reveal that additional benefits under the present circumstances are low. As for the north, also in the South more efficient varieties could be identified. Braeburn is the most efficient variety and appropriate for slightly to highly risk averse individuals. Identifying efficient combinations of variety, planting density and RMI is only a first step, however, as diversification reduces farm income risk International Food and Agribusiness Management Review

117

Röhrig et al.

Volume 21, Issue 1, 2018

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

further and takes into account the customers’ requirement of a certain product mix. As a consequence, future work should utilize whole-farm risk programming to capture the interaction between varieties as well as constraints to the implementation of risk management strategies by considering additional requirements, such as pollination management and farm-specific restrictions As the present paper aims to discuss RMIs universally applicable to apple producing orchards, restrictions as farm debts cannot be generalized and thus are not discussed. Future work may include these topics and considering farm debt repayment activities during catastrophic years which could lead to bankruptcy. With respect to the hypothetical subsidized frost-hail insurance, variety specifications as well as the planting density have to be considered when interpreting the results. For high-yielding varieties in high density plantings, this insurance concept seems to be inappropriate. Otherwise, when considering very risk averse apple growers, using less intensive production systems, it may lead to a slight increase of efficiency compared to an ordinary frost irrigation. A subsidized, hypothetical multi-peril insurance covering frost and hail would lead to slightly higher net incomes than observed for the subsidized hail insurance alone in the south, due to its additional transfer payments. But nevertheless, the production of high yielding varieties catching good prices in high density plantations under hail nets still remains the most efficient option. Apple growers of other European countries however have access to multi-peril insurance policies, which are often subsidized by the government. To analyze the effect of different risk management concepts, future work should compare European multi-peril insurance concepts, which cover a broad extent of weather-related risks.

Acknowledgement Special thanks go to M. Büchele of the Kompetenzzentrum Obstbau-Bodensee (KOB) and M. Görgens of the Obstbauzentrum Jork, who supported us in making contacts to apple farmers. Funding of the research by the Federal Ministry of Education and Research is gratefully acknowledged.

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2017.0022. Table S1. Cost overview for Red Prince with 3,300 trees/ha, frost irrigation and subsidized hail insurance (north).

References Ahmed, O. and T. Serra. 2015. Economic analysis of the introduction of agricultural revenue insurance contracts in Spain using statistical copulas. Agricultural Economics 46: 69-79. AMI. 2013. AMI Marktbilanz Obst. 2013. AMI Agrarmarkt Informations-Gesellschaft mbH, Bonn, Germany. AMI. 2014. AMI Marktbilanz Obst. 2014. AMI Agrarmarkt Informations-Gesellschaft mbH, Bonn, Germany. Atay, A.N., F. Koyuncu and E. Atay. 2013. Relative susceptibility of selected apple cultivars to alternate bearing. Journal of Biological and Environmental Sciences 7(20): 81-86. Berkhout, P., M. Van Asseldonk, R.W. van der Meer, H.A.B. van der Meulen and H.J. Silvis. 2016. Evaluatie regeling brede weersverzekering. Wageningen, Wageningen Economic Research, Rapport 2016-070. Available at: http://tinyurl.com/y8mds3aj. Bielza Diaz-Caneja, M., C.G. Conte, F.J. Gallego Pinilla, J. Stroblmair, R. Catenaro and C. Dittmann. 2009. Risk management and agricultural insurance schemes in Europe. JRC Reference Reports, Institute for Protection and Security of Citizen, European Communities, Brussels, Belgium, pp. 1-32. Bocquého, G., F. Jacquet and A. Reynaud. 2013. Expected utility or prospect theory maximisers? Assessing farmers’ risk behaviour from field-experiment data. European Review of Agricultural Economics 41: 1-38.

International Food and Agribusiness Management Review

118

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Röhrig et al.

Volume 21, Issue 1, 2018

Bravin, E., A. Kilchenmann and M. Leumann. 2009. Six hypotheses for profitable apple production based on the economic work-package within the ISAFRUIT project. Journal of Horticultural Science and Biotechnology 7: 164-167. Catalá, L.P., G.A. Durand, A.M. Blanco and J.A. Bandoni. 2013. Mathematical model for strategic planning optimization in the pome fruit industry. Agricultural Systems 115: 63-71. Chen, X., H.H. Wang and L.D. Makus. 2007. Production risk and crop insurance effectiveness: organic versus conventional apples. Working Paper No. SCC-76. Economics and Management Risk in Agriculture and Natural Resources Gulf Shores, Alabama, AL, USA. Clancy, D., J.P. Breen, F. Thorne and M.Wallace. 2012. A stochastic analysis of the decision to produce biomass crops in Ireland. Biomass and Bioenergy 46: 353-365. Coble, K.H. and B.J. Barnett. 2012. Why do we subsidize crop insurance? American Journal of Agricultural Economics 95(2): 498-504. Dirksmeyer, W., H. Garming and K. Strohm. 2014. Horticulture report 2014. TI Braunschweig, Braunschweig, Braunschweig, Germany. Available at: http://tinyurl.com/y744hdrv. Federal Statistical Office. 2014. Gesamtwirtschaft & Umwelt – Verbraucherpreisindizes – Verbraucherpreise – Statistisches Bundesamt (Destatis). Available at: http://tinyurl.com/yawlx7nz. Hardaker, J.B. 2000. Some issues in dealing with risk in agriculture. Working paper No. 2000-3. Working Paper Series in Agricultural and Resource Economics, Graduate School of Agricultural and Resource Economics, University of New England, Armidale, Australia. Hardaker, J.B., R.B.M. Huirne, J.R. Anderson and G. Lien. 2004. Coping with risk in agriculture, 2nd ed. CAB International, Wallingford, UK. Hardaker, J.B. and G. Lien. 2010. Stochastic efficiency analysis with risk aversion bounds: a comment. Australian Journal of Agricultural and Resource Economics 54(3): 379-383. Harper, J.K., A.J. Jimenez-Maldonado, R.M. Crassweller and D.E. Smith. 2013. Impact of producer risk preferences on selection of fresh-market apple training systems. International Journal of Fruit Science 13(3): 265-273. Hoag, D.L. 2010. Applied risk management in agriculture. 1st ed. CRC Press-Taylor & Francis Group, Boca Raton, FL, USA. Holt, C.A, and S.K. Laury. 2002. Risk aversion and incentive effects. American Economic Review 92(5): 1644-1655. Kaufman, N. 2014.Why is risk aversion unaccounted for in environmental policy evaluations? Climatic Change 125: 127-135. KTBL. 2010. Obstbau. Betriebswirtschaftliche und produktionstechnische Kalkulationen, 4th ed. Kuratorium für Technik und Bauwesen in der Landwirtschaft e.V. (KTBL), Darmstadt, Germany. Landesbetrieb für Statistik und Kommunikationstechnologie Niedersachsen. 2008. Informationen für die Ernte- und Betriebsberichterstatter 8/2008 Available at: http://tinyurl.com/yafd8cba. Landesbetrieb für Statistik und Kommunikationstechnologie Niedersachsen. 2012. Informationen für die Ernte- und Betriebsberichterstatter 6/2012. Available at: http://tinyurl.com/y9na7xge. Lien, G. 2003. Assisting whole-farm decision-making through stochastic budgeting. Agricultural Systems 76 (2): 399-413. Lien, G., and J.B. Hardaker. 2001. Whole-farm planning under uncertainty: impacts of subsidy scheme and utility function on portfolio choice in Norwegian agriculture. European Review of Agricultural Economics 28(1): 17-36. Lien, G., J.B. Hardaker and O. Flaten. 2007a. Risk and economic sustainability of crop farming systems. Agricultural Systems 94: 541-552. Lien, G., J.B. Hardaker, M.A.P.M. Van Asseldonk and J.W. Richardson. 2011. Risk programming analysis with imperfect information. Annals of Operations Research 190(1): 311-323. Lien, G., S. Størdal, J.B. Hardaker and L.J. Asheim. 2007b. Risk aversion and optimal forest replanting: a stochastic efficiency study. European Journal of Operational Research 181(3): 1584-1592. Menapace, L., G. Colson and R. Raffaelli. 2012. Cognitive heuristics and farmers’ perceptions of risks related to climate change. Agricultural and Applied Economics Association’s 2012 AAEA Annual Meeting, Seattle, Washington, DC, USA, pp. 1-23. Available at: http://tinyurl.com/y9l4od5r. International Food and Agribusiness Management Review

119

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2017.0022 - Monday, January 29, 2018 11:31:54 AM - IP Address:98.232.238.224

Röhrig et al.

Volume 21, Issue 1, 2018

Menapace, L., G. Colson and R. Raffaelli. 2013. Risk aversion, subjective beliefs, and farmer risk management strategies. American Journal of Agricultural Economics 95(2): 384-389. Menapace, L., G. Colson and R. Raffaelli. 2014. Farmers’ climate change risk perceptions: an application of the exchangeability method. Available at: http://tinyurl.com/yar2f9cs. Pennings, J.M.E., B. Wansink and M.T.G. Meulenberg. 2002. A note on modeling consumer reactions to a crisis: The case of the mad cow disease. International Journal of Research in Marketing 19(1): 91-100. Schenk, J., P. Hellegers, M. van Asseldonk and B. Davidson. 2014. How do farmers react to varying water allocations? An assessment of how the attitude to risk affects farm incomes. Agricultural Water Management 136: 52-58. Smidts, A. 1990. Decision making under risk: a study of models and measurement procedures with special reference to the farmer’s marketing behaviour. PhD thesis, Wageningen University, Wageningen, the Nederlands. Untiedt, R. and M. Blanke. 2001. Effects of fruit thinning agents on apple tree canopy photosynthesis and dark respiration. Plant Growth Regulation 35: 1-9. Vereinigte Hagel. 2017. Vereinigte Hagel. Kölnische Hagel. Kundeninformation. Available at: http://tinyurl. com/yad9wbuu. ZMP. 1998. ZMP-Marktstatistik. Obst 1998. Absatzmengen und Verkaufserlöse der deutschen Erzeugermärkte. Zentrale Markt- und Preisberichtstelle für Erzeugnisse der Land-, Forst- und Ernährungswirtschaft GmbH, Bonn, Germany. ZMP. 2002. ZMP-Marktstatistik. Obst 2002. Absatzmengen und Verkaufserlöse der deutschen Erzeugermärkte. Zentrale Markt- und Preisberichtstelle für Erzeugnisse der Land-, Forst- und Ernährungswirtschaft GmbH, Bonn Germany. ZMP. 2006. ZMP-Marktstatistik. Obst 2006. Absatzmengen und Verkaufserlöse der deutschen Erzeugermärkte. Zentrale Markt- und Preisberichtstelle für Erzeugnisse der Land-, Forst- und Ernährungswirtschaft GmbH, Bonn, Germany.

International Food and Agribusiness Management Review

120

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2016.0166 Received: 14 October 2016 / Accepted: 15 September 2017

Lighting the flame of entrepreneurship among agribusiness students RESEARCH ARTICLE Lindsey M. Higginsa , Christiane Schroetera, and Carlyn Wrightb aAssociate

Professor, and bAgribusiness Student, Agribusiness Department, California Polytechnic State University, 1 Grand Ave, San Luis Obispo, California 93407, USA

Abstract Entrepreneurship and innovation play a key role in combating problems facing agribusinesses, including the need for water conservation, sustainable packaging, and environmental protection. These issues have led to an increasing demand for college graduates with technical skills and innovative ways of thinking. The objective of our research is to provide insight into character traits that signal entrepreneurial skills. We conducted a survey to examine entrepreneurial interests and perspectives among U.S. agribusiness students. A cluster analysis revealed that entrepreneurial-minded students were more likely to be male, consider themselves risk takers, and have parents directly engaged in production agriculture. Our results emphasize the importance for universities to build studentsâ&#x20AC;&#x2122; experiences through industry partnerships, where students can interact with entrepreneurial mentors and get hands-on knowledge through applied coursework and internships in entrepreneurship. In addition, our study aids industry managers to learn more about future employees and their perceptions of entrepreneurial activities. Keywords: agribusiness, women, entrepreneurship, innovation, critical thinking JEL code: A22, C83, J20, L26 Corresponding author: lhiggins@calpoly.edu

ÂŠ 2017 Higgins et al.

121

Higgins et al.

Volume 21, Issue 1, 2018

1. Introduction A recent study found that more than 90% of executives believe that the long-term success of their businesses depends on their ability to come up with new ideas (Brooks, 2013). In order to make a profit, successful entrepreneurs must have the capacity to develop, organize, and manage a business venture, along with any of its risks. However, an integral part of entrepreneurship is the ability to be innovative (Deller and Conroy, 2016; Price et al., 2013). Entrepreneurs who are willing to assume the risks of taking on a new business venture must also be able to successfully implement new ideas to set themselves apart from competitors. Entrepreneurship and innovation play a key role in combatting problems facing agribusinesses, including the need for water conservation, sustainable packaging, and environmental protection (The Economist, 2016). According to United Nations (UN) estimates, the world population will increase to 9.7 billion people by 2050 (UN Department of Economic and Social Affairs, 2015). In order to keep up with the growing demand for food, the global agricultural industry must double its production and efficiency. While increasing production may be a solution to the need for food, it may also lead to negative externalities, such as resource scarcity. One of the biggest issues facing the global agricultural industry is water conservation and utilization. 60% of fresh water in America is used strictly for agriculture, although less than 10% of farms practice advanced on-farm water management, which includes moisture sensing tools and computerbased irrigation-scheduling tools (Zimmerman, 2014). In order to improve water usage in agriculture, the adoption of more effective irrigation systems that maximize the efficiency of the water used, while also minimizing waste, will become critical. As the populations of cities begins to grow there are more one to two-person households rather than larger suburban households, leading many companies to shy away from large, bulk packaging to smaller more compact packaging (Muratoglu, 2015). This shift has been motivated by taxes on plastic shopping bags and the promotion of compostable packaging in several countries such as Belgium and the United States (Chanprateep, 2010). While it may be difficult to completely eliminate these plastics given their ubiquitous use in food packaging, more food and beverage companies have been developing innovative solutions for bio-degradable packaging and bio-plastics made from crops such as corn or starch. By adopting innovative practices and products, innovative growers will be in a better position conquer potential threats to global agribusiness. Being innovative is an important quality for an agricultural entrepreneur, especially when the business faces strong competition and operates in a rapidly changing environment. Successful agribusinesses are those who adapt to changing environment to capture the opportunities from such disturbance and outperform those who do not adapt (Shadbolt and Olubode-Awosolab, 2016). It was once thought that entrepreneurial skills were innate, but now research has led to the conclusion that entrepreneurial education and exposure to entrepreneurial activities can help build a strong entrepreneurial skillset in entrepreneurship (Charney and Libecap, 2000; Souitaris et al., 2007). The demand for entrepreneurial education has increased globally, especially at the undergraduate level (Robinson and Josien, 2014). In recent years, the number of working Americans interested in pursuing a career in entrepreneurship has been on the rise. The Startup Activity Index, an early indicator of new entrepreneurship in the United States, registered another increase in 2016, after sharp increases two years in a row. New entrepreneurs who started businesses in order to pursue an opportunity rather than from necessity, reached 86.3%, which is more than 12% higher than in 2009 at the height of the Great Recession. Nearly 30% of all new entrepreneurs in the U.S. are first-generation immigrants, which is the highest level for just the second time in 20 years, climbing steadily from 13.3% in 1996 (Kauffman Foundation, 2017). Simultaneously, a much larger proportion of undergraduate students are attracted to the idea of creating a startup after graduation. College graduates are twice as likely to choose an entrepreneurial career path International Food and Agribusiness Management Review

122

Higgins et al.

Volume 21, Issue 1, 2018

compared to those with no high school education, and almost 50% more likely than high school graduates (Babson College, 2011). Universities and alternative online colleges are creating more entrepreneurial-focused curriculums, programs, and organizations for students interested in starting their own business (Schroeter and Higgins, 2016). Entrepreneurship is one of the fastest growing subjects in undergraduate curricula, responding to the pent-up demand from students, university administrators, and employers. Across the United States, entrepreneurship clubs, associations, internships, and even entrepreneurship-related majors at universities have sprung up (Kauffman Foundation, 2001). Many agribusiness programs have begun to respond to the demand for including concepts of innovation and entrepreneurship into their curriculum (e.g. Schroeter and Higgins, 2016). However, out of the approximately 40 universities in the U.S. with agribusiness programs, there are only a few that offer agricultural entrepreneurship programs or coursework. While it is important for agribusiness students to learn general business skills, there is also the need to acquire knowledge related to new opportunities and trends in the agricultural industry, given its dependence on limited resources like land and water. These issues will require college graduates with technical skills and the ability to come up with innovative ways of thinking. Given the increasing the demand for graduates who possess entrepreneurial skill sets required by agribusinesses, the objective of our research is to provide insight into the character traits among recruits who are likely to possess entrepreneurial skills. We aim at assessing opportunities that would encourage college student to pursue entrepreneurship. These findings are important for educators as well as industry managers, because there will be an increased pressure to incorporate concepts of innovation and entrepreneurship into college curricula. Understanding the entrepreneurial intent among college students will aid with developing tools and programs that expose students to the concepts of entrepreneurship.

2. Background 2.1 Characteristics, traits and attributes of entrepreneurs Prior research has helped shape our understanding of characteristics that are common of entrepreneurs (Sancho, 2010). Entrepreneurs tend to be risk-takers who push boundaries and enjoy being faced with challenges. Not only are entrepreneurs creative, but they also have the ability to communicate their ideas (Oosterbeek et al., 2010). These ideas represent the foundation on which entrepreneursâ&#x20AC;&#x2122; new products or services are built. Entrepreneurs tend to be driven by self-motivation and creativity, and the capacity to implement them (Knudson et al., 2004). Entrepreneurs are determined, persistent, and committed when it comes to their business ventures (Hand, 2010). Sometimes this commitment leads to entrepreneurs being deemed as selfish or self-focused individuals. Because of their intense motivation, innovators are considered goal-oriented people (Breugst et al., 2012). The ability to set and achieve a goal is seen as the most important trait for an entrepreneur to have; however, other traits and characteristics recognized among entrepreneurs are persistence, optimism, innovativeness, having a strong work ethic, and the ability to take initiative (Hand, 2010). While the characteristics and attributes of current entrepreneurs is well studied, little is known about these characteristic precursors in terms of undergraduate students pursuing an entrepreneurial career. In a study of university students, Ozaralli and Rivenburgh (2016) identified three over-arching themes in terms of potential influences that drive entrepreneurs: personality factors, social factors, and societal factors. The personality traits important for entrepreneurs are optimism, innovativeness, risk-taking, and competitiveness. In terms of social factors, the authors believe that constant exposure to new experiences and perspectives, like travelling or trying new things, boosts onesâ&#x20AC;&#x2122; creativity. Other social factors include entrepreneurial education and family exposure to entrepreneurship, while societal factors include perceptions about the economic and political climate.

International Food and Agribusiness Management Review

123

Higgins et al.

Volume 21, Issue 1, 2018

The majority of all college graduates are women; yet, they are less likely to pursue entrepreneurial paths (Anderson, 2016; Francis, 2007). Previous research suggests that women are less likely to become selfemployed or engage in other aspects of entrepreneurship. However, on a global scale, women represent one third of all professionals engaged in some form of entrepreneurship. In 2008, there were 10 billion firms owned by women, and those firms employed 13 million people. One major hypothesis for the increase in female entrepreneurs has been their frustration with the gender wage gap. While the wage ratio increased by 11% between 1980 and 1990, it only closed by an additional 5% from 1991 to 2005. Women who perceive they are not being equally valued in the workplace, are more likely to leave the traditional work environment and start their own business. Another important factor pertaining to women in entrepreneurship has to do with the development status of the country of residence. Female entrepreneurs in under-developed countries may face more problems entering the business world due to social beliefs about women in their country. Increasing the knowledge about female entrepreneurs and to females interested in pursuing an entrepreneurial path gives us insight into the modern-day business world, where both genders are more equally represented in the work force (Kobeissi, 2010). While one’s personality, gender, and nationality plays a big part in determining their likelihood to become entrepreneurs, another aspect is the impact of role models on potential entrepreneurs. Role models serve as an example for entrepreneurs to follow and imitate. Research shows that entrepreneurs with higher levels of education are more likely to have a role model than entrepreneurs who do not. Entrepreneurs’ role models tend to be close to home, including family members, close friends, and former employers or colleagues (Bosma et al., 2012). 2.2 Entrepreneurship and the undergraduate student Despite having innovative, risk-taking, and competitive skill sets, relatively few students in the U.S. anticipate becoming an entrepreneur (Ozaralli and Rivenburgh, 2016). An explanation for this may be the perceived risks of taking on a new business venture; American students tend to prefer the idea of a salaried job rather than investing in their own risky start-up (Ozaralli and Rivenburgh, 2016). In general, male students have a more positive perception of entrepreneurship prior to taking any entrepreneurial courses; however, evidence suggests that entrepreneurial education promotes and removes perceived barriers to entrepreneurship (Packam et al., 2010; Schroeter and Higgins, 2016). Students’ exposure to new experiences contributes to higher levels of creativity, which in turn leads to a higher chance of pursuing entrepreneurial activities. Universities are being pressured to produce new generations of workers who fit the workplace’s demands, and many are starting to implement curricula pertaining to entrepreneurship (Ollila and Williams-Middleton, 2011). Many entrepreneurial programs initially focused on teaching entrepreneurship, rather than creating entrepreneurs. In 2001, the Chalmers School of Entrepreneurship (CSE) in Sweden took a different approach to entrepreneurial education. The main focus of the CSE was not only to educate students on entrepreneurship, but also instill a ‘learn-by-doing’ attitude and organize them into groups to apply their understanding on entrepreneurship and participate in real-life ventures. Students were able to pitch an idea or product and collaborate with professors, business advisors, and alumni to put together portfolios for their ‘companies.’ Educators at CSE found that students who participated in their simulation were able to improve their current business skills and acquire new entrepreneurial skills (Ollila and Williams-Middleton, 2011). The majority of entrepreneurial education research revolves around the curricula and development of education, but it is also important to consider the effect of the teacher and how they present information to their students (Ruskovaara and Pihkala, 2014). Projects are the often used as the tool for teaching entrepreneurial skills. Group projects are a way for students to improve their ability to collaborate with others, exercise their problem-solving skills, as well as exposing them to both peer- and self-assessment throughout the project. Research found that in these kinds of scenarios, teachers move away from the traditional lecturer role and become more of a mentor for their students (Ruskovaara and Pihkala, 2014).

International Food and Agribusiness Management Review

124

Higgins et al.

Volume 21, Issue 1, 2018

While prior research has focused on entrepreneurial skillsets and perspectives of undergraduate students, little is known about agribusiness students, and the factors that contribute to lighting their entrepreneurial flame. Therefore, this research aims to isolate characteristics, perceptions, and entrepreneurial intent among agribusiness students.

3. Methodology A 25-question survey was developed in order to measure the entrepreneurial intent among agribusiness students. A large undergraduate agribusiness program (~600 students) in California was selected as the sampling frame. The survey was sent out electronically to current and recent graduates of the agribusiness program. The survey was open for two weeks, with two reminder emails sent during that time period. As an incentive for completing the survey, respondents were entered into a drawing for one of four $25 gift cards to Amazon.com. The survey included a list of 13 of personality traits characteristic of entrepreneurs including innovativeness, creativeness, and risk-taking ability (Hand, 2010; Knudson et al., 2004; Ozaralli and Rivenburgh, 2016). Respondents were asked to indicate to what extent they identify with each of those characteristics. They also were asked to indicate their top five strengths from Gallup’s StrengthsFinder test1. In order to get a better idea of students’ perceptions of entrepreneurship, respondents were asked to indicate their agreement with a series of 10 statements related to entrepreneurship and entrepreneurial education. A 5-point scale from ‘strongly agree’ to ‘strongly disagree’ was used. Statements included: ‘more schools are offering entrepreneurship programs than in the past’ and ‘entrepreneurs are more likely to be men’ and were formulated based on the extant literature (e.g. Caliendo and Kritikos, 2011; Noyes and Linder, 2015). Respondents were then asked to provide a self-assessment of their strengths for skills related to entrepreneurial activities (e.g. risk management, marketing), as well as how likely they were to pursue entrepreneurial activities and what influenced their interest in entrepreneurship. The motivation for including these questions in the survey instrument were to compare findings with those of Bosma et al. (2012) regarding the importance of role models influencing young entrepreneurs. Learning more about the antecedents and variables that influence respondents’ interest in entrepreneurship allows us to compare this sample of agribusiness students to those discovered more generally by Ozaralli and Rivenburgh (2016) in their study about influential factors of university students. The survey is attached in Supplementary Methods S1.

4. Results During the two-week period that the survey instrument was available, responses were gathered from 132 individuals. To meet the University’s Institutional Review Board requirements, responses to demographic survey questions were not required. Thus, this resulted in varying sample sizes by question, with response counts of 109 to 112 for demographic questions. The sample was distributed fairly equal among genders, with 48% males and 52% females (Table 1). Almost 70% of students surveyed were between the ages of 20 and 22, and nearly 80% of students were Caucasian, consistent with the demographics of the university. Although no incoming freshmen were surveyed, 31% of the respondents were juniors, 34% were seniors, and the remaining 36% were either sophomores, students going into their fifth year, or recent graduates. Roughly half of the students’ parents obtained at least a Bachelor’s Degree, and 57% of respondents came from a suburban hometown. The demographics of the sample are comparable to the agribusiness student population at the university of study, however it should be noted that the high percentage of Caucasian students many not be representative of agribusiness programs across the country.

The Clifton’s StrengthsFinder is a test that helps people determine dominant characteristics and strengths they possess related to the business world. The majority of students at this university complete the StrengthsFinder assessment during their first weeks on campus. 1

International Food and Agribusiness Management Review

125

Higgins et al.

Volume 21, Issue 1, 2018

Table 1. Demographics of survey respondents (where ‘n’ represents response counts). Gender (n=109) Male Female Age (n=110) 17-19 20-22 23-25 25+ Ethnicity (n=109) Asian/Pacific islander Caucasian Hispanic/Latino Native American Other Class level (n=110) Sophomore Junior Senior 5th year or above Alumni Parents’ highest level of education (n=112) Some high school High school/GED Some college Bachelor’s degree Master’s degree Adv. grad/Ph.D. Not sure Hometown (n=112) Rural Suburban Urban Ag. background (check all that apply, n=199) Farming family Ag. organization Hobby farms 1+ parents work in ag Prior work experience No ag. background Other

# 53 58

% 48 52

23 77 10 2

21 69 9 2

4 88 11 2 6

4 79 10 2 6

13 34 37 14 14

12 30 33 12 12

1 8 18 56 20 8 1

1 7 16 50 18 7 1

37 64 11

33 57 10

26 32 28 25 37 47 3

23 29 27 22 33 42 3

Students were asked to identify their strengths based on the results of their Clifton StrengthsFinder assessment. Interestingly enough, none of the respondents identified self-assurance as one of their strengths. This finding is consistent with the more than 6,000 students that have taken the assessment at the university of study; self-assurance is the least common strength, while achievement is the most common. It also suggests that our sample is representative of the university’s study body. Gallup describes people who possess the selfassurance trait as ‘confident in their ability to manage their own lives.’ They possess an inner compass that gives them confidence that ‘their decisions are right’ (Gallup Strengths Center, 2017). While responses to the remaining 33 traits were distributed fairly evenly, over 56% of respondents possess the achiever trait, International Food and Agribusiness Management Review

126

Higgins et al.

Volume 21, Issue 1, 2018

which is described as having strong work ethic and finds satisfaction in being productive. Another 27% of respondents share the competition trait, and 25% share the restorative trait, meaning they are good at identifying problems and finding solutions to them (Gallup Strengths Center, 2017). The top three strengths identified in our study confirm the key traits of entrepreneurs based on research by Hand (2010). Hand (2010) surveyed 257 current entrepreneurs who identified being self-confident and competitive as two major traits of successful entrepreneurs. 4.1 Gender differences Students were asked to indicate the extent to which each of the 13 key entrepreneurial characteristics describes them (Table 2). In general, respondents strongly identified themselves as being tenacious (87%), versatile (86%), competitive (87%), self-motivated (92%), and open-minded (88%). A majority of respondents (91%) also said they work well with others and are not afraid to ask for help as needed. Although none of the students said they possess the self-assurance trait from the Gallup StrengthsFinder, they still described themselves as being self-motivated. While there were no significant differences in the responses on personality characteristics based on class standing, we found that students who stated a higher interest in becoming entrepreneurs, thought of themselves as risk-takers (P-value of 0.000). Further, we found some interesting results when making comparisons by gender (Table 2). Male respondents, in general, are more confident that they possess entrepreneurial characteristics. Of the 13 characteristics used in the survey, female respondents were likely to possess creativity, self-motivation, open-mindedness, and not being afraid to ask for help (although none of those differences were statistically significant). On the other hand, male respondents were significantly more likely to indicate that they were risk-takers, innovative, willing to fail in order to learn, versatile, competitive, decisive, persuasive, and work well with others. All of these characteristics are important to becoming an entrepreneur. Among the seven topic areas, agribusiness students were most confident in their knowledge of economics and trends/issues in agribusiness (Table 3). Interestingly, compared to the personality characteristics question (Table 2), there was more gender parity with regard to their knowledge assessment. In fact, the only statistically significant difference was found in female respondents indicating they were more likely to have a good understanding of marketing products and services. These results suggest that while both genders may have Table 2. Personality characteristics by gender on a 5-point scale (where 5=describes me completely and 1=does not describe me at all). Characteristics

Male

Female

P-value

Signif.1

Works well with others Competitive Self-motivated Versatile Tenacious/persistent Open-minded Persuasive Willingness to fail to learn Not afraid to ask for help Innovative Risk-taker Decisive Creative

4.6346 4.6154 4.4808 4.4231 4.3077 4.2692 4.1923 4.1569 4.1154 3.9231 3.8462 3.8462 3.7885

4.3333 4.1404 4.6140 4.1754 4.2982 4.3333 3.7321 3.6491 4.2143 3.4211 3.3333 3.3333 3.9649

0.027 0.005 0.273 0.090 0.947 0.664 0.005 0.008 0.583 0.002 0.001 0.015 0.335

Significance levels denoted by ***, **, and * for the 0.001, 0.05, and 0.1 levels, respectively. International Food and Agribusiness Management Review

127

*** *

*** *** *** *** **

Higgins et al.

Volume 21, Issue 1, 2018

Table 3. Self-assessed knowledge levels by gender on a 5-point scale (where 5=strongly agree to have a good understanding of the topic and 1=do not have a good understanding). Economics Trends and issues in agribusiness Marketing products and services Legal issues related to business Innovation Entrepreneurship and what it means to be an entrepreneur Policies and laws pertaining to businesses 1

Male

Female

P-value

4.0000 3.9615 3.8850 3.7500 3.6923 3.6731 3.5577

3.7368 4.1786 4.1404 3.4561 3.4737 3.4386 3.4737

0.131 0.209 0.064 0.108 0.225 0.197 0.636

Signif.1

Significance levels denoted by * for the 0.1 level.

the same knowledge levels, female students are less likely to give themselves credit for certain personality traits and, as a result, may be less likely to see themselves as an entrepreneur. In order to determine how much emphasis students place on the importance of entrepreneurship and innovation in agriculture, and their perceptions of entrepreneurship, we asked respondents to indicate their level of agreement with different statements about entrepreneurship. After comparing the student responses and categorizing them based on class level, gender, background, and their likeliness to pursue future entrepreneurial activities, we discovered that perceptions of entrepreneurship and agribusiness were fairly homogeneous among the respondents (Table 4). However, male respondents were significantly more likely to agree that entrepreneurs are more likely to be men (P=0.009). Respondents were asked to rate their likelihood of pursuing entrepreneurial activities, followed by a question asking them to identify what influenced their interest, or disinterest, in becoming an entrepreneur. Nearly 70% of the respondents indicated they are likely or extremely likely to pursue an entrepreneurial career. Respondents who expressed their interest in entrepreneurship indicated that they know someone who is an entrepreneur (52%) or simply have a personal interest in becoming an entrepreneur (57%). Table 4. Level of agreement with statements related to entrepreneurship (where 5=strongly agree and 1=strongly disagree). Agreement statement Meeting the worldâ&#x20AC;&#x2122;s food demands will come as a result of innovation in agriculture There are many problems facing the global agribusiness industry Entrepreneurship and innovation is crucial for the agribusiness industry to continue to grow There will be more start-ups in the agribusiness industry in the near future Improving water usage in agriculture can be solved using technology we already have More schools are offering entrepreneurship programs than in the past Entrepreneurs are more likely to be young rather than old Students are more comfortable working with students of the same major Entrepreneurs are more likely to be men There are more entrepreneurs in industries like technology and medicine than in agribusiness 1

Male

Female

P-value

4.673 4.327

4.614 4.368

0.636 0.744

4.327 3.827

4.456 3.649

0.369 0.293

3.808 3.789 3.692 3.529 3.289

3.456 3.667 3.474 3.754 2.772

0.071 0.352 0.230 0.226 0.009

3.039

3.000

0.850

Significance levels denoted by *** and * for the 0.001 and 0.1 levels, respectively.

International Food and Agribusiness Management Review

128

Signif.1

***

Higgins et al.

Volume 21, Issue 1, 2018

We conducted independent sample t-tests based on students’ class level, their parents’ highest level of education, gender, where they grew up, and background in agriculture. While there were no major differences among these groups, we found that female students, in general, were less interested in pursuing an entrepreneurial career path. For the likelihood of pursuing entrepreneurial activities, we found that male respondents had a mean value of 1.98 while female respondents had a mean value of 2.37 (based on a scale where 5 was ‘extremely unlikely’ and 1 was ‘extremely likely’), a statistically significant difference (P=0.043). To further understand the differences between those that indicated a likelihood of pursuing an entrepreneurial career and those that did not, we conducted a k-means cluster analysis. The two identified clusters had statistically significant differences in terms of class level, ag background, gender, self-assessed strengths, and the likelihood of pursing an entrepreneurial career. The clusters had a 1.47 ratio of largest to smallest cluster. The larger of the two clusters consisted of students who are significantly more likely to pursue an entrepreneurial career. We may call this the ‘Entrepreneurial’ cluster. Students in the Entrepreneurial cluster were more likely to be male, consider themselves risk takers, and have parents directly engaged in production agriculture. Further, those in the Entrepreneurial cluster were closer to graduation. Cluster 2 consisted of students who were significantly less likely to pursue entrepreneurship. We may call this cluster ‘Administrative’, following the idea of Fairbrothers and Gorla (2001) that those opposite of entrepreneurs are more cautious and tend to focus on procedure. Administrative cluster students were more likely to be younger, female, and less likely to come from a production agriculture background. Consistent with previous research (Fairbrothers and Gorla, 2011; Stevenson and Gumpert, 1985), we find that Administrative cluster students are risk adverse. The results from our cluster analysis support our findings that showed that female students were less likely to strive for an entrepreneurial career path. In addition, the cluster analysis provides deeper insight into the profile of entrepreneurial students. Students were asked to identify their level of interest in learning more about entrepreneurship or building an entrepreneurial skillset, as well as how they would like to go about doing so (e.g. entrepreneurial classes, internships, clubs, and getting in contact with current entrepreneurs). We found that 67% of students were interested in developing an entrepreneurial skillset, 85% were interested in entrepreneurial specific courses, and 79% wanted to get in contact with current entrepreneurs for future work. Students who indicated that they were less likely to become an entrepreneur were, in turn, less interested about learning more about it. Contrary to male respondents, female students indicated that they were still interested in learning more about entrepreneurship regardless of their likelihood to become an entrepreneur.

5. Conclusions The purpose of this study was to examine the perceptions and influencers pertaining to agribusiness students’ interest in entrepreneurship and isolate differences between male and female students. We determined that a large proportion of students (70%) are interested in pursuing an entrepreneurial career path after graduation. Regardless of a student’s stated entrepreneurial intent, students still indicated interest in learning more about entrepreneurship (85% were interested in entrepreneurial specific courses). These students indicated they would be most interested in entrepreneurship-specific coursework and meeting current entrepreneurs to learn more about what makes them successful. Much like Bosma et al. (2012), who learned that a majority of the entrepreneurs had a role model that influenced their entrepreneurial interest, 52% of agribusiness students who indicated that they were likely to pursue an entrepreneurial career said they were influenced by a family member, friend, or coworker who is an entrepreneur. Further, those that ended up being in the entrepreneurial cluster were more likely to come from a background in production agriculture and, thus, may have been exposed to more family-owned farming operations, given that 97% of all U.S. farms are family-owned (United States Department of Agriculture – Economic Research Service, 2015). Consistent with Ozaralli and Rivenburgh (2016), having some form of a role model or current entrepreneur to look up to appears to be very influential to agribusiness students’ interest in pursuing an entrepreneurial career.

International Food and Agribusiness Management Review

129

Higgins et al.

Volume 21, Issue 1, 2018

Not only are agribusiness students interested in pursuing an entrepreneurial career, but most of these students possess the traits and characteristics of successful entrepreneurs. Previous literature suggests that successful entrepreneurs are often described as competitive, persistent, and innovative (Knudson et al., 2004). The results from our survey show that most of the agribusiness students described themselves as self-motivated, tenacious, and versatile, which are some of the most common traits of successful entrepreneurs. One of our more fascinating findings came from students’ StrengthsFinder results. We found that of the students who identified their StrengthsFinder results, 56% of them listed achiever as one of their strengths, followed by competition and restorative at only 27%, with the remaining traits present in anywhere from 2 to 18% of the sample. The large proportion of students possessing the achiever strength (meaning they have a strong work ethic and find satisfaction in being productive) suggests that agribusiness students possess many positive entrepreneurial traits. These students possess strong entrepreneurial traits, and they also have a good understanding of different aspects of business like economics and marketing. The combination of these personality traits and general business knowledge has led us to believe that students who are interested in pursuing an entrepreneurial career already possess the foundation of successful entrepreneurs and that there is an opportunity to further develop their entrepreneurial traits. While the majority of students indicated they are interested in pursuing an entrepreneurial career after their graduation, we found that in general, female students indicated that they would less likely to become entrepreneurs. We also found that when students were asked to indicate the extent that each of the 13 entrepreneurial personality characteristics describes them, women seemed to be less confident in their skills related to entrepreneurship. Caliendo and Kritikos (2011) discovered similar findings when studying business women in Germany, who were less likely to be self-employed or engage in entrepreneurial work than their male counterparts. Despite the lack of interest in pursuing an entrepreneurial career, we found that female students are still interested in learning more about entrepreneurship and building an entrepreneurial skillset. Our findings suggest that entrepreneurship education with activities to build an entrepreneurial skillset will aid female students to eliminate the perceived barriers into entrepreneurship. Furthermore, previous research points out that anybody who is willing to be mastery-oriented, i.e. able to learn and face challenges with an unknown outcome, is an entrepreneur. Thus, all what matters is the passion to make an impact in life (Fairbrothers and Gorla, 2011). Our research aims to encourage additional studies on entrepreneurship. One area of research would be to track students after graduation and follow them in their pursuit for an entrepreneurial career. It would also be useful to assess the possible precursors and influencers of students’ interest in entrepreneurship. More specifically, it would be valuable to learn more about females’ hesitations in pursing entrepreneurial careers, given their budding interest in developing an entrepreneurial skillset. In our study, we did not find a strong relationship between various demographic characteristics (ethnicity, class level, parents’ education, background in agriculture, etc.) and the students’ interest in entrepreneurship. In addition to demographics, there is a need for research that determines other antecedents that might influence students’ perceptions of, or interest in, entrepreneurship. This study of agribusiness students serves as a foundation for industry managers to learn more about future employees and their perceptions of entrepreneurial activities. In addition, our findings may help universities who are seeking to implement entrepreneurial education and promote students to venture out and pursue their entrepreneurial dreams. We found that a majority of students are interested in learning more about entrepreneurship. Given the increasing role of innovation and entrepreneurship in agriculture, universities may want to implement additional entrepreneurial education opportunities to meet the demands of students seeking to build an entrepreneurial skillset and pursue their own startup. Our results emphasize the importance for universities to build students’ experiences through learn by doing exposure to entrepreneurship such as industry partnerships, where students could interact with entrepreneurial mentors and get hands-on knowledge through applied coursework and internships in entrepreneurship.

International Food and Agribusiness Management Review

130

Higgins et al.

Volume 21, Issue 1, 2018

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2016.0166. Methods S1. Survey instrument.

References Anderson, N. 2016. Women break barriers in engineering and computer science at some top colleges. The Washington Post, September 16. Available at: http://tinyurl.com/y7lf9ba6. Babson College. 2011. U.S. entrepreneurship rates reverse trend, reach new heights. Available at: http:// tinyurl.com/ya7agkb5. Bosma, N., J. Hessels, V. Schutjens, M. Praag and I. Verheul. 2012. Entrepreneurship and role models. Journal of Economic Psychology 33: 410-424. Breugst, N., A. Domurath, H. Patzelt and A. Klaukien. 2012. Perceptions of entrepreneurial passion and employeesâ&#x20AC;&#x2122; commitment to entrepreneurial ventures. Entrepreneurship: theory and practice 36(1): 171-192. Brooks, C. 2013. Innovation: key to successful business. Business News Daily, September 23. Available at: http://www.businessnewsdaily.com/5167-innovation.html. Caliendo, M. and A. Kritikos. 2011. Searching for the entrepreneurial personality: new evidence and avenues for further research. Discussion paper No. 5790. Institute for the Study of Labor, Bonn, Germany. Chanprateep, S. 2010. Current trends in biodegradable polyhydroxyalkanoates. Journal of Bioscience and Bioengineering 110(6): 621-632. Charney, A., and G.D. Libecap. 2000. Impact of Entrepreneurship Education. Kauffman Center for Entrepreneurial Leadership, Kansas, MO, USA. Deller, S. and T. Conroy. 2016. Survival rates of rural businesses: what the evidence tells us. Choices 31(4): 1-5. Fairbrothers, G. and C. Gorla. 2011. Who is the administrator? Forbes, December 26. Available at: http:// tinyurl.com/ycn845p2. Francis, D. 2007. Why do women outnumber men in college? The National Bureau of Economic Research, Cambridge, MA, USA. Available at: http://tinyurl.com/cwv8vj8. Gallup Strengths Center. 2017. StrengthsFinder. Available at: https://www.gallupstrengthscenter.com. Hand, R.A. 2010. Entrepreneurial analysis: a study to identify traits and demographics of practicing entrepreneurs. Ph.D. Dissertation, Capella University, Minneapolis, MN, USA. Kauffman Foundation. 2001. The growth and advancement of entrepreneurship in higher education: an environmental scan of college initiatives. Ewing Marion Kauffman Foundation, Kansas City, MO, USA. Kauffman Foundation. 2017. Startup Activity swings upward for third consecutive year, annual Kauffman index reports. Available at: http://tinyurl.com/ycf5hvhe. Knudson, W., A. Wysocki, J. Champagne and H. Peterson. 2004. Entrepreneurship and innovation in the agri-food system. American Journal of Agricultural Economics 86(5): 1330-1336. Kobeissi, N. 2010. Gender factors and female entrepreneurship: international evidence and policy implications. Journal of International Entrepreneurship 8(1): 1-35. Muratoglu, S. 2015. 5 critical packaging trends for 2015. Packaging Digest, January 6. Available at: http:// tinyurl.com/y9m3l7d4. Noyes, E. and B. Linder. 2015. Developing undergraduate entrepreneurial capacity for social venture creation. Journal of Entrepreneurship Education 18(2): 113-124. Ollila, S. and K. Williams-Middleton. 2011. The venture creation approach: integrating entrepreneurial education and incubation at the university. International Journal of Entrepreneurship and Innovation Management 13(2): 161-178. Oosterbeek, H., M. van Praag and A. Ijsselstein. 2010. The impact of entrepreneurship education on entrepreneurship skills and motivation. European Economic Review 54(3): 442-454. Ozaralli, N. and N.K. Rivenburgh. 2016. Entrepreneurial intention: antecedents to entrepreneurial behavior in the U.S.A. and Turkey. Journal of Global Entrepreneurship Research 6(1): 1-32. International Food and Agribusiness Management Review

131

Higgins et al.

Volume 21, Issue 1, 2018

Packham, G., P. Jones, C. Miller, D. Pickernell and B. Thomas. 2010. Attitudes towards entrepreneurship education: a comparative analysis. Education and Training 52(8/9): 568-586. Price, D.P., M. Stoica and R.J. Boncella. 2013. The relationship between innovation, knowledge, and performance in family and non-family firms: an analysis of SMEs. Journal of Innovation and Entrepreneurship 2(1): 1-20. Robinson, P., and L. Josien. 2014. Entrepreneurial education: using ‘the challenge’ in theory and practice. Journal of Entrepreneurship Education 17(2): 172-185. Ruskovaara, E., and T. Pihkala. 2014. Entrepreneurship education in schools: empirical evidence on the teacher’s role. The Journal of Educational Research 108(3): 236-249. Sancho, F. 2010. Agricultural and rural entrepreneurship: concepts for modeling development. Communica January-July(5): 64-77. Schroeter, C. and L.M. Higgins. 2016. Learn by doing: a case study on enhancing students’ entrepreneurial skills. Western Economics Forum 15(1): 32-43. Shadbolt, N.M. and F. Olubode-Awosolab. 2016. Risk, resilience, and entrepreneurship. International Food and Agribusiness Management Review 19(2): 33-52. Souitaris, V., S. Zerbinati and A. Al-Laham. 2007. Do entrepreneurship programmes raise entrepreneurial intention of science and engineering students? The effect of learning, inspiration and resources. Journal of Business Venturing 22(4): 566-591. Stevenson, H. and D.E. Gumpert. 1985. The heart of entrepreneurship. Harvard Business Review 63(2): 89-92. The Economist. 2016. The future of agriculture. Available at: http://tinyurl.com/zhy8lbn. United Nations Department of Economic and Social Affairs. 2015. World population projected to reach 9.7 billion by 2050. Available at: http://tinyurl.com/hox2d2y. United States Department of Agriculture – Economic Research Service. 2015. Family farms are the focus of new agriculture census data. Available at: http://tinyurl.com/y8zrm8ws. Zimmerman, C. 2014. Four industry trends for 2015. AgWired. Available at: http://tinyurl.com/yatmh8wp.

International Food and Agribusiness Management Review

132

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2016.0095 Received: 10 May 2016 / Accepted: 18 March 2017

Supply chain re-engineering: a case study of the Tonghui Agricultural Cooperative in Inner Mongolia CASE STUDY Qianyu Zhu a, Cheryl J. Wachenheim b, Zhiyao Mac, and Cong Zhuc aAssociate

Professor and cGraduate Student, School of Agricultural Economics and Rural Development, Renmin University of China, 59 Zhongguancun St, Haidian Qu, Beijing, China P.R. bProfessor,

Department of Agribusiness and Applied Economics, North Dakota State University, 811 2nd Ave North, Fargo 58102, ND, USA

Abstract Benefits of cooperative organization in agriculture come from price advantages in procurement and marketing, cost reductions and efficiency gains from sharing of productive assets and processes, and improved access to and increased efficiency in using credit, logistics, and information. Efficacy of strategic activities designed to capture these advantages is investigated empirically in a case study of the Tonghui Agricultural Cooperative in Inner Mongolia, an autonomous region of China. Information from interviews, on-site visits, evaluation of cooperative, member and partner information, and participation in the advising process are used to evaluate the impact of efforts to re-engineer the supply chain for independent farmers through cooperative organization. Specific examples of marketing channel development and operation for Wallace melons and mutton represent implementation of strategic plans. The case also reviews the cooperativeâ&#x20AC;&#x2122;s credit system, designed increase access to and reduced cost of funds for members, use of alternative market venues, and horizontal expansion through tourism. Keywords: cooperative, China, supply chain, agricultural marketing, financing JEL code: P13, Q13, Q14 Corresponding authors: qyzhu2008@163.com; cheryl.wachenheim@ndsu.edu

ÂŠ 2017 Zhu et al.

133

Zhu et al.

Volume 21, Issue 1, 2018

Cooperative position and strategy are viewed here as the ability of cooperatives to successfully find or develop economic benefits for their farmer members in a modern, market-oriented food supply chain. – Hohler and Kuhl, 2014: 581

1. Introduction Cooperatives are economic and social organizations developed on the principles of voluntary membership, democracy, equality and mutual benefit to members. Since the 19th century, cooperatives in fields such as agriculture, industry, finance, insurance, medicine, transportation, science, and education have organized in more than 160 countries. The International Cooperative Alliance, established in London in 1895, is now one of the world’s largest non-government organizations with 292 member organizations in 95 countries (https:// ica.coop). Cooperatives are most widely used in agriculture where they have successfully been established to provide farmers and herdsmen production and marketing advantages not available to them individually. There is an array of literature aimed at assessing the financial performance of cooperatives (Hohler and Kuhl, 2014; Lerman and Parliament, 1990; Parliament et al., 1990; Sexton and Iskow, 1993; Soboh et al., 2009, 2011). However, the more interesting question is ‘why?’, that is, what are the underlying causes of failure or sustained success of cooperatives? A theoretical model seeking an explanation is based on the expected impact of various characteristics and behaviors of cooperatives on performance. Included are those factors that increase revenues such as increased unit price due to market power, quality improvements, or branding and those that decrease costs such as reduced cost of inputs due to market power, sourcing additional vendors, or reduced transactions costs and efficiency gains resulting from technology adoption, through shared knowledge and experience, and in logistics (USDA, 1980). Although there is no recipe for sustained success that can be applied generally, there are some relatively consistent findings within the literature about the source of advantages associated with cooperative behavior and these are consistent with this basic economic model. Holloway et al. (2000) and Roets (2004) concluded a cooperative organizational structure reduced costs, including transactions costs, in Africa. Abebaw and Haile (2013) identified efficiency gains from technology adoption among cooperative members in Ethiopia. Hohler and Kuhl (2014) identified sources of growth and success of farmer cooperatives in Europe to be consolidation and coordination, though vertical expansion and/or development of relationships within their marketing channel. They in particular noted the value of increased control over the quality and amount of product supplied that comes from coordination. Also telling is information from the literature about causes of failure of cooperatives. These causes can, in general, be categorized into those associated with membership, decision-making, leadership, financial difficulties, and lack of government support. Membership issues include insufficient membership, make-up of membership as dominated by larger operations, an unstable membership, and lack of understanding and education among members about the cooperative structure and its functioning (Garrido, 2007; Machethe, 1990; Ortmann and King, 2007; Sexton and Iskow, 1988). Democracy and efficiency of decision-making are related challenges. Van der Walt (2008) identified conflict between members as a contributing factor to cooperative failure. Ortmann and King (2007) discuss the control problem that results from a misalignment between member and management interests and the influence cost problem wherein members with diverse interests work to direct management so that resources are not allocated in the best interests of the organization as a whole. Bernard and Spielman (2009) point out that those with lower levels of education and wealth are less likely to participate in the cooperative and more likely to be excluded from the decision-making process. Financing adds another challenge. The limited capital compensation feature leads to difficulties in financing capital projects and other initiatives and meeting cash flow requirements. Sustained strong leadership is important to overcome these challenges (Ortmann and King, 2007; Van der Walt, 2008). If members are engaged and understand the benefits associated with working together, have International Food and Agribusiness Management Review

134

Zhu et al.

Volume 21, Issue 1, 2018

strong leadership that ensures they are provided with not only production, marketing and financial training, but that designed to help them understand cooperatives, they can be more effectively involved in the decision-making process (Ortmann and King, 2007; Machethe, 1990). Finally, is that this can best occur in a supportive external environment, including a proactive role by the government to support entrepreneurial activity such as the development of cooperatives (Ortmann and King, 2007). From the literature, we identified the following cooperative characteristics or behaviors with potential to support successful and sustained performance in a newly developed cooperative: member understanding of and participation in cooperative governance; strong leadership; an efficient democratic process for decisionmaking; a large base of members with medium and/or small-sized farms; a willingness to coordinate the marketing channel by integration, partnerships, agreements, or other coordinating activities; and the support of government in that an environment that supports entrepreneurial activities and cooperatives. A democratic governance structure should be established in agricultural cooperatives to prevent the decision-making power from being monopolized in the hands of a small number of members or by management (Zhu et al., 2015). Further, the cooperative must be both prepared and flexible as it faces new challenges and opportunities, including growth. 1.1 Cooperatives in China In China, rural reform in the late 1970s established a rural management system based on household contract management (Li and Ping, 2003). In this basic system, although they co-existed with large-scale farms employing relatively modern agricultural methods, small-scale farmers held little wealth, had minimal market power, were specialized in production, used low-technology methods, and did not have access to current market information. They sold products through intermediaries and the products sometimes changed hands several times before arriving at the end consumer, contributing to the relatively high farm to retail price margin. To help alleviate these challenges and improve the income of farmers, stabilize the price of agricultural production, and improve market efficiencies, China began to explore a wide range of cooperative economic organizations in the 1980s. Specialized farmer cooperatives have developed in China gradually since that time. The Law of the People’s Republic of China on Farmers’ Professional Cooperatives was approved in October 2006, and became effective on July 1, 2007. The law clarified the farmer cooperative’s principal status in the market and provided regulations for their internal management system and behavior, further accelerating their development. Six general trends typify agricultural cooperative development in China during the past three decades. First, the number has increased rapidly. By the end of October 2015, the number of farmer cooperatives registered with China’s National Administration for Industry and Commerce reached 1.48 million. Nearly 100 million farmers (farm households) are members of farmer cooperatives, accounting for an estimated 47% of the farm population (Ministry of Agriculture of China, 2015). Second, the extension of farmer cooperatives into consumer markets has grown. By the end of 2014, about 370,000 farmer cooperatives had obtained quality certifications; nearly 700,000 had registered trademarks; and 230,000 had set up wholesale or retail operations. Third, the profitability of farmer cooperatives has gradually increased. The total unified sale value of agricultural products by all types of farmer cooperatives in China reached more than 750 billion Chinese Renminbi (¥) (approximately $119 billion), with surplus earnings of nearly 100 billion ¥ (approximately $16 billion). Fourth, the organizational forms of agricultural cooperatives have continuously innovated, and current cooperatives include those dedicated to rural land shareholding, community shareholding, labor, tourism, property management, and consumers. General cooperatives are also growing in number. Fifth, farmer cooperatives have played a growing role in agricultural production systematization and reengineering, and optimization of the marketing channel for fresh and processed agricultural products. In recent years, many of the larger cooperatives have begun providing members with comprehensive services throughout the production and marketing processes including raw material procurement, new technology International Food and Agribusiness Management Review

135

Zhu et al.

Volume 21, Issue 1, 2018

adoption, information sharing, product grading and sorting, processing and packaging, transportation and brand management, as well as financial services (Kong et al., 2012). Many scholars have argued that cooperatives can help mitigate challenges faced by farmers with small landholdings such as lack of market power, market venues and technical services (Hu et al., 2015; Huang, 2011; Jia et al., 2012; Yang, et al., 2014). They have tended towards integration; horizontally, vertically, and both (Liao and Guo, 2015). Sixth, the role of cooperatives in leveraging the financial capacity of members is evolving and such is increasingly present as part of integrated agricultural cooperatives. This is important because there is considerable financial exclusion in rural areas (i.e. farmers have a difficult time securing funds from formal financial institutions). The World Bankâ&#x20AC;&#x2122;s Community Development Fund pilot project in rural Shaanxi Province of China provides some support to the idea that cooperative finance is an effective way to provide farmers with credit services (Zhu and Ma, 2013). Cooperative finance can reduce the prevalence of lending through usuries, open financing options, concentrate idle money, cultivate farmersâ&#x20AC;&#x2122; saving practices, and improve life for farmers (Kong and Jin, 2012). Despite the statistics and examples of the success of individual cooperatives in China, Hu et al. (2015) argue that there is considerable empirical evidence that most Chinese Farm Services Cooperatives are not organizations with viable operations under a cooperative structure. They looked at 45 cooperatives in 11 Chinese provinces over six years, selecting those that were established to demonstrate cooperative organization or had been recognized as success stories. They found many to be simply shells, organized for the purposes of receiving subsidies or for other similar reasons, and that others were operated as private firms with little or no benefit going to the farmer-members. Others had failed due to poor management or an unfavorable market. They concluded that government policy supporting the development of cooperatives has largely been ineffective. Hu et al. (2015), among others, call attention to some problems that cannot be ignored in the rapid development of agricultural cooperatives in China, mainly as follows. First, the number of cooperatives in China is relatively large, but the scale is generally small. For example, the average registered capital of cooperatives is close to that of individual industrial and commercial households, which is approximately 5% of that of the average private enterprise (Ministry of Agriculture of China, 2015). Second, the services provided by existing cooperatives are mainly concentrated in the production of agricultural products, and the role played in the sales link is still relatively limited. Many lack financial, insurance, social security and other services. Third, the internal operating structure of most cooperatives is not mature enough to ensure that smaller shareholders maintain their voice in the cooperative, and there is a lack of effective external supervision. Finally, the legal structure supporting the development of cooperatives is not yet sound. As noted, current cooperative development falls under the Farmersâ&#x20AC;&#x2122; Professional Cooperatives Act which lacks specific rules. Land management, the tax system and other key areas related to rural success and potential for cooperatives also suffer from a lack of effective legal guidance. 1.2 Problem statement Although evidence about the performance of cooperatives does not allow for drawing general conclusions, and is in fact sometimes conflicting, what is clear is that, with varying degrees of success, cooperatives have expanded their involvement in the marketing channel. Included are consolidated purchasing of inputs and sales of commodities, providing or facilitating financing, sorting of farm products, processing, logistics activities including storage and transportation, wholesaling and retailing, and branding. We look here at one cooperative in Inner Mongolia, China, Tonghui Agricultural Cooperative (TAC). Local farmers and herdsmen suffered from many of the same challenges that cooperatives are meant to address. Included are low education and literacy rates, lack of purchase and sales power and inadequate number of market venues, especially for sales, lack of availability of and ability to obtain financing, lack of market information, and poor logistics infrastructure. This paper covers the evolution, current situation, and future plans of TAC in an empirical application of the theory of cooperative establishment, investigating the performance of strategic and tactical decisions in improving income for farmer-members. Through extensive interviews

International Food and Agribusiness Management Review

136

Zhu et al.

Volume 21, Issue 1, 2018

with managers, consultants, farmers and herdsmen, and participants in the marketing channel1; on-site visits, and participation in the advising process, we investigate how TAC worked to re-engineer its supply chain for fresh produce, crops, and livestock, and the products of each, and how material, information, and fund flows are linked within a developing cooperative which offers production, supply, marketing, and facilitating functions. The agricultural production and marketing channels have been improved however measured (e.g. efficiency, coordination) and the income and welfare of farmers and herdsmen has increased. Risks associated with agricultural production and operation, as well as that faced by upstream and downstream partners and facilitating industries are reduced due to the emphasis on coordinating activities and integration. Finally, the gap in rural financial funds is reduced. This case study seeks to answer the following questions: 1. Specifically, how does TAC use the flow of information and capital, and its logistics system to integrate and coordinate the supply channel from the purchase and use of inputs to the end consumer to improve the welfare and income of local farmers? 2. Given that the development of many farmersâ&#x20AC;&#x2122; cooperatives in China has failed, what is different about the organization and strategic and tactical behavior of TAC that has led to success? 3. How has TAC impacted the rural economy? The remainder of this paper is organized as follows. Next is a discussion of the concept of engineering and optimizing the supply chain. TAC is then introduced and evaluated using the principles described for cooperative success. Detailed examples of the marketing channel development of two products are provided. Examples of horizontal expansion into rural tourism and of the industry facilitating function of credit are explored. The paper finishes with discussion and conclusions.

Contributing participants include the Chairman of the TAC Association, the TAC General Manager, the Chairman of ZhinongFumin Agriculture Co., Ltd., Deng County Agricultural Bank, Bank of China, Mengyin Village Bank, Dengkou County Rural Credit Cooperatives Credit Department. The authors visited TAC headquarters and facilities including the Wallace melon planting base, Dengkou sheep breeding base, and TAC warehouses and logistics facilities. TAC staff, farmers and herdsman including those responsive for the sheep breeding herd, and logistics staff were interviewed at length.

Upstream (agriculture) Research Machinery Capital

Midstream (industrial)

Crops Produce Livestock

Downstream (logistics and marketing) Storage Logistics Sales

Processing Formation

Optimization Loan guarantee

Regulatory (over-watch) Government

Cooperative

Financial institution Wholesale loans

Registration

Provide service Production

Consumption

Customers

Member

Marketing

Credit

Figure 1. Cooperative agricultural product supply chain. International Food and Agribusiness Management Review

137

Zhu et al.

Volume 21, Issue 1, 2018

2. The cooperative supply channel The supply chain of agricultural products runs through multiple stages that may include the supply of inputs, production, post-harvest preservation, processing, storage, wholesale, retail and consumption. Cooperatives can serve as an entity that brings together the many participants including farmers and herdsmen, researchers, government, the financial and agribusiness sectors, processers, logistics channel participants, and buyers including distributors, wholesalers, retailers, and the end consumer. The operation mechanism of cooperatives based on a supply chain of agricultural products serves to guide the discussion (Figure 1). The top shows the agricultural supply chain within which the cooperative operates and works to coordinate. The middle depicts government oversight of the cooperative and its guarantor role in member borrowing. Chinese cooperatives need to be registered with and are supervised by the State Administration for Industry and Commerce of the Peopleâ&#x20AC;&#x2122;s Republic of China. The rectangle at the bottom includes the supporting functions provided to members by the cooperative. Cooperatives provide their members with the services of production, supply and marketing, and provide financial support for the members directly or as a guarantor. Cooperatives also serve to integrate and coordinate, improving the efficiency of the marketing channel. Market relationships are motivated by opportunity. The cooperative works to help participants exploit potential benefits of a coordinated marketing channel (Table 1). The cooperative can help farmers and herdsmen achieve the benefits of scale economies and provide technical expertise, improve access to credit and information, reduce risk, and open new marketing channels. Participating enterprises in the supply chain include agricultural input suppliers and those who purchase primary agricultural products directly. Through a stable relationship with the cooperative, transactions costs can be reduced, forecasting improved, and profits increased. In addition, the relationship with the farmers through the cooperative allows the firms to provide technical guidance and other services that help farmers and herdsmen meet specific quality requirements. Financial institutions offer financing for the various participants in the supply chain. In rural China, financing efforts are hindered by inadequate information disclosure, high credit risk, high moral hazard, and high transactions costs and many farmers are unable to obtain operating loans from financial institutions. Coordinated activities allow for increased visibility on borrowers and reduced risks from additional guarantee mechanisms. Improved transparency also facilitates government oversight and the ability of the government to judge the need for and to distribute subsidies. This helps the government reach its objectives of improving the standard of living for farmers, promoting local economic development and enhancing its image. Consumers can also be beneficiaries. A coordinated marketing channel can improve quality monitoring and control, including that associated with ensuring food safety. In addition, supply chain management reduces the involvement of intermediate links and in other ways can reduce costs, which can reduce cost for consumers. Table 1. Potential benefits to supply chain participants with cooperative core. Participant Farmer/herdsman Enterprise Cooperative Financial institution Customer Government

Benefit Technical guidance and education on production and business; increased access to market information and credit, risk mitigation, reduced costs, improved revenues and profits. Improved forecasting, reduced production costs, increased profits, better control over product quality. Scale expansion, increased operating income, increased market influence. Alleviate challenges of asymmetric information and high transaction costs, develop the rural financial market, fulfill social responsibility, enhance image. Improved quality, quality and food safety assurance, reduced cost. Increased information, improved efficiency and equity in distributing subsidies, stimulate local economy, enhanced image.

International Food and Agribusiness Management Review

138

Zhu et al.

Volume 21, Issue 1, 2018

2.1 Optimization of the marketing channel After the development of the primary agricultural supply chain through integration and coordination, a cooperative can work to refine the system to capture further efficiencies and more effectively exploit market power. This second phase we call supply chain optimization. It usually includes one or more of the following: vertical extension, horizontal expansion and supply chain integration. Vertical extension of the supply chain can include extension into the farm-supply sectors including financing, into post-production sectors including processing, transport, storage and marketing of agricultural products; and the addition of intermediate links. In addition to traditional operating and capital inputs in the upstream sector, cooperatives may develop research programs, form technical education programs, and develop and enforce quality control standards. In the downstream channel, cooperatives can establish agricultural processing plants and cold storage, or develop a contractual relationship with enterprises that supply these value-adding services. Whether the tie is achieved through integration or coordination, processing fresh agricultural products can have a very important effect on the quality of the resultant products and the market price received for such. While large cooperatives can vertically expand to include sorting, grading, cleaning, processing, packaging, and inspection, the investment required to develop a transportation infrastructure can be prohibitive. As such, in rural China this extension is often through coordination with a third-party logistics company. Horizontal expansion of cooperative activities need not be limited to the addition of members and their agricultural products. Expansion can include an increase in the scale and technical level of value-adding processes (e.g. additional cold storage), but can also include activities such as eco-tourism, financing and insurance, and other secondary and tertiary industries, such as pick-your-own operations, eco-resorts, and farmhouse stays. In the process of supply chain integration, cooperatives will play a key role. They can act as intermediaries, providing services to their farmer members before, during and post production; act as the negotiating agent for purchasing farm inputs and selling agricultural products; and facilitate the development of value-adding processes.

3. Tonghui Agricultural Cooperative Inner Mongolian TAC is in Dengkou County, Bayannur City. Dengkou County covers an area of about 1,038,000 acres including approximately 260,000 acres of arable land. The county contains four towns and one SĂźme2. It has a population of approximately 130,000, of which 40% are rural residents. The area includes a complex variety of land forms, including mountains, deserts, plains and rivers. In the northern part of the county is Mount Langshan, which covers a rocky mountain area of about 239,000 acres. In the western part, there is the vast Ulan Buh Dessert whose surface is covered with dunes and psammophytes. Ulan Buh Dessert covers an area of about 703,000 acres. In the eastern part, there is the vast alluvial plain of the Yellow River. The plain is flat and fertile and covers an area of about 76,000 acres. Crossing channels that run through the plain make irrigation convenient. The Yellow River crosses the southern part, covering an area of 11,530 acres. The Yellow River brings rich water resources to the county. The environment is favorable for a variety of crop and livestock farming. The main grain crops are corn and wheat. Beef, sheep, and pigs are raised. Farming can be characterized as decentralized management by local farmers and herdsmen. Technology used and facilities are not standardized. There is little coordination in the supply chain. Each of these challenges has hindered the development of the regionâ&#x20AC;&#x2122;s agriculture. Since the Law of Farmer Cooperatives was implemented in 2007, leaders in Dengkou County have worked to guide the development of farmer cooperatives. By the end of 2015, farmer cooperatives registered in the 2

SĂźme refers to a community of minority people, generally larger than a village but smaller than a town.

International Food and Agribusiness Management Review

139

Zhu et al.

Volume 21, Issue 1, 2018

county by the Bureau of Agriculture and Animal Husbandry reached 153, including those of crop and livestock farmers, and their supporting industries. Approximately 80% of farmer cooperatives in the county are crop or livestock cooperatives and are small scale, providing only single service. As such, as noted about many Chinese cooperatives in general by Hu et al. (2015), existing cooperatives have had little effect on farmersâ&#x20AC;&#x2122; incomes and development of the agriculture marketing channel. TAC was established to help overcome these noted challenges facing the farmers and herdsmen. The cooperative has quickly helped farmers organize and coordinate internally and with market channel partners, improving the income of farmers and herdsman as well as that of the local economy. 3.1 Organizational Structure TAC was registered on March, 20, 2015 with registration capital of 10 million ÂĽ ($1,562,500) and inaugurated on May 20, 2015. The cooperative was initially organized by five individuals3, each of whom provided a capital investment of 2 million ÂĽ ($312,500). By September 2016, the cooperative claimed 471 memberhouseholds4. The charter specifies the cooperative principles of voluntary membership, and equal and democratic membership. Three bodies including the General meeting of members, Board of management and Board of supervisors are responsible for guiding the cooperative and oversight of management and decision making (Figure 2). There are four operating departments: production, marketing, finance and administration and eleven subdepartments. Division of labor and cooperation between different departments has been important to success. Each member has one vote regardless of their capital contributions or transactions, which promotes democratic decisionmaking. TAC actively engages members in decision-making at all levels. All-member meetings are held quarterly. TAC also invites its members to participate in forums and training designed to improve member understanding of the role and functioning of cooperatives, improve their production efficiency and gain 3 4

Ma Fei, Ren Guangyin, Wang Lu (Chairman of the Board of Management), Ma Shaoqi, and He Yongxiong (Chairman of the Board of Supervisors). Only one member of each household can be a member of the cooperative.

General meeting of members Board of supervisors

Finance

Board of management

Production

Marketing

Administration

Financial services

Purchasing

Marketing

Personnel

Accounting

Production

Sales

Administration

Storage

Electronic commerce

General logistics

Figure 2. Organization structure of Tonghui Agricultural Cooperative. International Food and Agribusiness Management Review

140

Zhu et al.

Volume 21, Issue 1, 2018

their input on means to do the same, and protect their rights and interests. Some examples of cooperativesponsored activities are shown (Table 2). To organization farmers together in the steps of production, marketing and logistics to effectively provide a broad range of services, TAC set the county as the management center entity. Branches are in towns and liaison service stations can be found in villages. These branches allow the cooperative to take advantage of the informal system of rural acquaintances so important in China to reduce transaction costs and manage risk. At present, TAC has nine grassroots branches and 27 service stations, providing service for more than 3,000 rural households (Table 3). Grass-roots branches play an important role in assurance, development, organization and information transmitting. They are also responsible for keeping track of transitions between members and the cooperative. To best serve the members, the cooperative selects appropriate shops in villages to work as liaison service stations. In village shops, members can exchange and disseminate information, and make purchases. Members can often take advantage of an offered discount by the village shop and those nearby. Membership requires an applicant have full capacity for civil conduct, be honest, have a good reputation, and be a farmer or herdsman with a demonstrated mastery of agricultural skills, all measured subjectively by the membership board. Members who are hardworking and honest are more likely to unite, improve production

Table 2. Participation of members of Tonghui Agricultural Cooperative. Activity

Date

Content

Participating households

Establishment ceremony National forum on industrialization and cooperative economy Grassroots branch trainings Nalin Xindi Minxing

May 2015 Dec. 2015

Ceremony, introduction to cooperatives Agricultural policy, Internet, project docking contracts, site visits

110 400

Jan. 2016 March 2016 June 2016

70 30 55

Experiment bureau

July 2016

Bianming

July 2016

Purchasing, sales, credit cooperation Operating and transferring land Trinity of cooperation, operation of cooperatives Cooperation in pensions and collective management of land Internet basics, Internet and traditional sales

67 70

Table 3. Grassroots branches of Tonghui Agricultural Cooperative. Name

Have Tonghui warehouse

Member households

Bianming Haigang Nalin Lake Minxing Experiment Bureau Xindi Dongdi Shajin SĂźme Wulanbahe Total

Yes Yes Yes Yes Yes No No No Yes

35 5 55 75 45 50 30 101 75 471

International Food and Agribusiness Management Review

141

Zhu et al.

Volume 21, Issue 1, 2018

efficiency and create a better life under the guidance of the cooperative. Each member is required to pay 100 ¥ ($16) for qualification shares. Wang Lu was voted as professional manager by members at the general meeting of members. He manages production and marketing. Staff were selected from among the membership to lead the departments of production, marketing, finance, and administration. There are currently 14 employees. All of them are between 25 and 50 years old and have at least a junior high school level education. Two hold graduate degrees and four hold certificates of completion from three-year vocational colleges. 3.2 Operation TAC provides and coordinates services for members and distributes proceeds5. Benefits to members originate from three sources, consistent with the foundational principles of cooperatives. First, production costs are reduced through increased market power for procurement of inputs and production efficiencies resulting from education and training. Second, sales revenues increase and market risk declines because of marketing efficiencies from economies of size and from cooperative storage, logistics, and marketing activities. Third, the cooperative increases access to and reduces the cost of credit for members. The cooperative has successfully earned profits, which are distributed to members based on their transactions volume with the cooperative. The cooperative also distributes subsidies to members on behalf of the government. ■■ Training and technical guidance TAC provides services, training and technical guidance for farmers and herdsmen to improve product quality and production efficiencies as well as to help members understand the production standards required for certifications in both crop and livestock production. Volunteer experts who understand and are experienced users of modern agricultural production techniques, including local experienced cooperative members, and active or retired government workers and professors, provide regular guidance and training for members and also work with farmers and herdsmen individually. For example, they will conduct farm soil testing and analysis to aid in the recommendation of seeds and fertilizers for individual fields. The Cooperative facilitates availability and understanding of market information. A Chinese farmer may be hard working and have gained productive skills honed from many years of experience, yet continue to lose money because he does not have access to and is not skilled at using market information. Acting independently, farmers often make production and marketing decisions based on market and production success and failures from the previous year (e.g. if the price of a product in the previous year was strong, they will increase production of this product). External expertise and that provided directly by the cooperative can guide members to adjust production scale and structure according to current and expected market conditions. ■■ Marketing cooperation Farmers who produce independently and in small scale are at a disadvantage in the market and lack negotiation capacity. TAC adds value for its members through reduced input costs, adding value to products internally through storage and processing, and increased prices and reduced risk by exploiting multiple marketing opportunities. Reduced input costs: TAC negotiates upstream purchases on behalf of its members. Presentation of cost savings from two key inputs serves as an example of this value. During its first sixteen months in operation, 5

An ideal contribution to this manuscript, as suggested by a reviewer, would be a quantified analysis of the value of the cooperative to individual farmer-members. Most farmers in the region are not well educated and for other reasons, their use of record-keeping, financial or otherwise, is minimal. In the absence of evidence about the financial impact for individual farmers we therefore provide some detail regarding revenue- and efficiency-enhancing and cost-reducing activities provided by the cooperative for membership groups. For example, cost-savings are estimated for a group of farmer-members purchasing their fertilizer and plastic sheeting through the cooperative.

International Food and Agribusiness Management Review

142

Zhu et al.

Volume 21, Issue 1, 2018

taking advantage of negotiation power originating from size of purchase and exploiting multiple procurement venues, TAC reduced member cost of fertilizers and plastic film by more than 20,000¥ ($3,175) (Table 4). This may not seem like an important savings in the context of western agriculture, but it is important in China where the average farmer earned an annual net income of $1,175 in 2015 (Green Book on China’s rural economy released by the Chinese Academy of Social Sciences and quoted in http://tinyurl.com/yb57yeqv). Professional centralized storage: Recommend Cooperative organization also facilitates activities associated with increasing revenue. Fresh produce has a short shelf life and is difficult to preserve. Product loss due to spoilage is high for local farmers using self-storage because their facilities are primitive and improper storage techniques are widely employed. TAC built local warehouses for storage, including advanced preservation technology to extend the storage life for fresh produce, frozen meat, and other products. This has substantially reduced loss. The warehouses further facilitate cooperative growth through multiple functional usages. In the spring, agricultural inputs such as seed and fertilizer are stored in the warehouses. Warehouses are used to dry agricultural products in summer, and store them in the fall. Stored products can be used as collateral for bank loans throughout the winter season. Marketing diversification: there are few options for local farmers and herdsmen selling their products independently. To help its members overcome this challenge and thereby increase revenues and decrease risk, TAC worked to develop a diversified marketing system. The cooperative sells over 160 products including both commodities (e.g. grains, sunflowers, fruits, vegetables, whole chickens, eggs) and the products of those commodities (e.g. wheat, sunflower seeds). Channels employed include an electronic commerce platform, mobile application store, and sales through wholesale markets and chain stores, directly to independent retail stores, and through a store front with a home delivery option. China is the world’s largest e-commerce market (http://tinyurl.com/yb26p9fr) and TAC has exploited the value of the internet, including building Taobao6 and Wechat7 stores for the mobile internet customer, and cooperating with electronic commerce platforms such as Tootoo Commune (Figure 3)8. As a means to communicate to consumers, including those purchasing online, the attributes of food products otherwise considered commodities, the cooperative registered the Tonghui brand. Fresh products (e.g. pork, chicken, lamb) are assembled from independent members. The Cooperative then examines, processes, transports, and sells these products under the Tonghui brand. Use of e-commerce has been an important venue to expand brand awareness and gain market share. The cooperative also provides fresh produce for supermarkets directly including two supermarkets in Wuhai City, a nearby city with a population of over 500,000.

TaoBao (淘宝网; pinyin: Táobǎo Wǎng; literally interpreted as ‘searching for treasure website’) consists of a Consumer to Consumer Platform (Taobao Market place, similar to Ebay), Tmall.com (a business to consumer platform similar to Amazon), and eTao (an online shopping search engine). It was founded in 2003, has a gross income estimated to be more than 1 billion ¥ (approximately $156 million), and retains a majority market share in Mainland China (https://en.wikipedia.org/wiki/Taobao). 7 Wechat (微信; pinyin: ēixìn; literally interpreted as ‘micro message’), originally introduced as Weixin, was first released in January 2011 and is a one-interface mobile service center. It serves as a social networking site (similar to Facebook) and also facilitates text messaging, broadcasting, video conferencing, and sharing of locations, photographs and videos. It includes an embedded translation service, supports payments and money transfer, and supports city services, which allows for such activities as making doctor appointments, paying city fees, and arranging transportation (e.g. booking a taxi). Wechat also serves its approximately 600,000 active users as an e-commerce site. 8 TooToo Organic Farm (http://www.tootoo.cn/index-en-1.html) was established in 2008. The e-commerce site sells organic produce, fish and meat, and grain products; and a small selection of drinks, bakery items, and daily-use items such as paper towels.

Table 4. Materials purchased through Tonghui Agricultural Cooperative (TAC) (June 2015 to September 2016). Year

Product

Quantity

Market price ($)

TAC price ($)

Reduced cost ($)

2015 2016

Urea Plastic film Diamine Urea

100 kg 266 sheet 94.4 kg 239 kg

178 17 441 203

171 16 433 197

640 346 755 1,530

International Food and Agribusiness Management Review

143

Farms 32 30 38 78

Zhu et al.

Volume 21, Issue 1, 2018

Figure 3. Examples of Wechat store products of Tonghui Agricultural Cooperative. At the end of 2015, maintaining the principles of high quality, healthful products produced in an environmentally-sustaining manner, TAC entered the agricultural products sales market in Wuhai under their Home Life Company (HLC). HLC provides TAC with a business outlet as their traditional store front in a new agricultural products sales market. The outlet consists of 108 m2 of floor space. It was renovated with the ‘Tonghui Farm’ theme (Supplementary Figure S1). Nineteen employees are responsible for its operation, store management, product promotion, technical maintenance, and financial management. The cooperative uploads details about agricultural products for sale regularly onto its HLC consumer application. Consumers can select and purchase goods using this application. HLC collects and transfers the sales data to the cooperative weekly, and the staff screen, process and package the products. TAC continues to utilize third-party logistics. In this case, TAC partnered with the Wuhai Post Office. Every Friday, the cooperative sends food products to Wuhai outlets and, from there, the Wuhai Post Office sends the packaged foods directly to consumers. Targeted couponing: to increase sales, TAC has worked to develop high-income consumer markets. More than 400,000 high-quality customers use the Wuhai Mobile Company’s application. Many in this target market work in the government sector, at financial institutions, and at state-owned enterprises, have a higher net worth and/or income, and value product attributes including healthy foods and those produced using environmentally-sustainable methods. In March, 2016, TAC and Wuhai Mobile Company signed a business cooperation promotion agreement. Wuhai Mobile Company provides approximately 4,000 customers in TAC’s target market the opportunity to earn points towards a Tonghui Farm Products cash coupon. This coupon can be used directly for purchases from the Wuhai HLC. Pre-sales channels: the cooperative also carries out pre-sale business. Tonghui farm provides consumers in Wuhai with monthly, quarterly, half-year and full-year pre-sale services. Taking the full-year pre-sale services as an example, consumers sign a contract with Tonghui Farm to pay 3,000¥ ($476) annually, and receive products including rice, flour, and meat delivered to their home each Friday. The delivery volume specified in the contract is designed to feed a typical family for one week for 60¥ ($9.50) per week. This system provides some volume and price stability for the cooperative and helps promote the Tonghui brand.

International Food and Agribusiness Management Review

144

Zhu et al.

Volume 21, Issue 1, 2018

■■ Production efficiencies Since establishment in March 2015 through the end of the year, TAC operated at a loss as they had significant cost investments and were not yet operating at a break-even level. Yet, because TAC provided benefits to and improved the income of its local farmer members from the beginning, the cooperative was granted subsidies. In December 2015, at its National Agricultural Industrialization and Cooperation Economic Summit Forum, TAC was granted subsidy funds in the amount of 100,000¥ ($15,873) from the Dengkou County Agricultural Animal Husbandry Bureau. In March, 2016, TAC implemented its experimental project to explore the Wallace melon rejuvenation and seed breeding, having obtained an additional subsidy from the same source; this one for 160,000¥ ($25,397). The government not only encourages and fosters an environment that facilitates entrepreneurial ventures, but provides funds; funds which are critical to accelerate the growth of the cooperative. Since the second quarter of 2016, TAC has become profitable (Figure 4).

4. Supply chain development Dengkou County covers a vast geographic area. There is a diverse terrain, good soil, and clean air and water, bringing forward considerable potential to produce high quality agriculture products. At present, total farmed area approximates 128,500 acres and improved varieties are already used on most planted acres. Production agriculture has in fact been developing at a faster rate than the supply channel needed to successfully bring the products to market. Facilities used for agricultural production in 2013 had already grown to 1,450 acres. By this time, these facilities included 742 acres of solar greenhouses and 708 acres of plastic greenhouses, both of which support large scale off-season growth and availability of melons, vegetables and seedlings. Livestock farming has also been developing rapidly with inventories at 662,500 animals. And, many farmers have begun producing using methods that qualify the products as a Green Health Product which is analogous to an organic food certification in the U.S. Despite progress in production, the location of Dengkou County is remote to cities and the local rural infrastructure is not well developed. Poor transportation and low-capacity, primitive agricultural storage facilities greatly impeded local agricultural development. In addition, the county continued to struggle with inadequate financial services. To overcome logistics and storage challenges, TAC worked to integrate resources in planning, procurement, production, and sales and distribution to fit local conditions, including building partnerships with upstream and downstream market participants. For this case study, we consider two distinct products, the Wallace Honeydew Melon and Dengkou mutton, to investigate how the cooperative worked to coordinate the marketing channel, facilitate member access to the resources needed to successfully 600.000 500.000

Dollars

400.000 300.000 200.000 100.000 0 2015Q2

2015Q3

2015Q4

2016Q1

2016Q2

2016Q3 Revenues

Figure 4. Revenues and expenditures of Tonghui Agricultural Cooperative. International Food and Agribusiness Management Review

145

Costs

Zhu et al.

Volume 21, Issue 1, 2018

serve potential markets, and increase member income. We focus consideration on these two products in the manuscript because, as was noted by Reardon and Barrett (2000), cooperative structure may be especially important for produce and meat because cooperatives can provide training and resources to help facilitate production to grading and contract standards, serve to sort products accordingly and ensure quality. Further, both are highly perishable and therefore coordinated storage and logistics can produce considerable benefits to farmer-members. 4.1 Wallace Melon The production area of the Wallace honeydew melon9 is adjacent to the Ulan Buh Desert on an alluvial plane of the Yellow River. The deep and fertile soil contains abundant organic minerals. The climate is characterized by over 3,000 hours of annual sunshine, dry air, and a high temperature difference between day and night. Despite little precipitation, irrigation from the Yellow River is convenient. This growth environment results in the unique features of the Wallace Melon that differentiates it from the traditional honeydew melon including a thin skin, thick flesh, sweet juice, moderate crispness, rich fragrance and a high level of nutrition. Freshness of the Wallace Melon directly affects taste and it is generally not suitable for longer-term storage. As there was neither a mature preservation technology nor a well-developed logistics system in Dengkou, both the production and distribution activities were previously conducted by farmers independently. As there had been no efforts in market development, output was low. Most melons were sold within the county or offered as gifts to visiting friends and relatives. ■■ Recruiting producers Within months of its establishment, TAC began positioning the Wallace Melon as one of the most promising products in Dengkou County. Cooperative leadership worked to simultaneously build a market for and expand production of these melons. They developed a recruitment plan specifically aimed at farmers currently growing melons, especially those with experience growing Wallace Melons; and worked to reduce melon seed costs and simplify seed procurement for farmers by collectively purchasing seeds from the Local Agriculture and Animal Husbandry Bureau. Previously, farmers worked with the Bureau independently. The cooperative began to purchase seeds in bulk and subsequently sell them to farmers at the wholesale price. Planting area increased from 165 acres in 2013 to more than 824 acres in 2015. ■■ Improving and standardizing production (capability) The Cooperative hired professional melon planting experts to carry out field investigations, develop and put forward recommendations, and launch a school for melon farmers. Experts provided information and consulting on land selection, land preparation requirements, ridge preparation and fertilization, seed sowing, disease and insect control, topdressing, cultivating and weeding. As such they not only helped growers improve their production capacity and efficiency, but helped the cooperative standardize production management, improving the ability of members to meet qualifications such as those for Green Food Base, Good Agricultural Practices, and export registration base certifications. Volume of production was further encouraged in February, 2016 when the Department of Agriculture and Animal Husbandry of Inner Mongolia Autonomous Region and Agriculture and Animal Husbandry Bureau of Dengkou County offered grants of 200,000 ¥ ($32,000) and 130,000 ¥ ($21,000), respectively, to build greenhouses. This support facilitated cooperative development of new varieties and promoted fertilization, 9

Referred to as the ‘Wallace Melon’ in China, this is similar to the honeydew melon. It was bred to be slightly sweeter and has higher moisture content and a shorter shelf life. There are multiple historic accounts of the use of the term Wallace Melon Chinese: 华莱士; pinyin: Hualaishi), ranging from that commonly understood in Inner Mongolia, that Vice President Wallace visited the region and so enjoyed the melon that the people thereafter referred to it by his namesake, to the name originating from the donation by Mr. Wallace, who was also the founder of Pioneer Hybrid Seeds, of the melon seeds while visiting China in the 1940s.

International Food and Agribusiness Management Review

146

Zhu et al.

Volume 21, Issue 1, 2018

disinfection, pest control, harvesting and other steps by helping the cooperative and its members adopt advanced technologies. New generation Wallace greenhouses were built, each covering about 2.5 mus (0.41 acres) with an expected output of 2,500 kilograms per mu. ■■ Shelf-life extending technology Cooperative leaders understood that adopting the technology required to extend the storage period of the melons was the key to the successful development of external markets. However, there were no domestic storage technologies available that could match the specific requirements of the Wallace Melon. That changed shortly after Mr. Wang became acquainted with Mr. Liao, Director of the Taiwan Farmers Association, at an economic development forum. They discussed a preservation technology used in Taiwan that applies ozone sterilization, preservative and coating treatments, and advanced refrigerated storage techniques steps in the storage and preservation process. This technology is well suited for Wallace melons produced in Dengkou, and cooperation with the Taiwan Farmers Association provided TAC with the solution required. The introduction of this advanced preservation technology extended the shelf life of the Wallace Melon from three days to fifteen (Table 5). ■■ Market development TAC simultaneously worked on market development. Beijing was identified as having considerable potential as the large city nearest to Dengkou. Distance is important as production and particularly logistics costs of Wallace Melons are high. Because of the high logistics costs, a higher income market was targeted. Faculty at Peking and Renmin Universities in Beijing and several supporting industry partners helped to develop the market-entry strategy. The strategy identified the primary marketing channels for Wallace Melons to include an e-commerce platform and direct sales to supermarkets. In 2015, TAC signed an agreement with All-China Federation of Supply and Marketing Cooperatives (ACFSMC), enabling Wallace Melons to directly enter the e-commerce platform of the ACFSMC and be sold throughout China. This is an example of the willingness to work with other cooperatives identified as an International Cooperative Alliance principle (Ortmann and King, 2007). The admission to the e-commerce platform also represents the acquisition of the ‘quality endorsement’ from the ACFSMC, which provides further support to brand image. Supermarkets place orders directly and the cooperative schedules deliveries from TAC warehouses. The products are loaded in the early afternoon and scheduled to arrive in Beijing by 2 a.m. so they can be placed on shelves in all physical supermarkets prior to 6 a.m. The cooperative uses a third-party logistics company, and engages the technology noted previously to help guarantee the quality of the product when it arrives in Beijing. 4.2 Dengkou Mutton Livestock plays an important role in the economic development of Dengkou County. In 2013, the output value of animal husbandry reached $69 million, comprising nearly one-third of the total value of agricultural output. In recent years, the number of standardized, large-scale livestock farms has increased steadily. A Table 5. Preservation technology and equipment cost of fresh agricultural products. Product

Non-refrigerated shelf life

Cold storage shelf life

Refrigeration equipment type

Equipment cost ($)

Honeydew Wallace melon Dengkou mutton

5 days 3 days 5 days

3 months 15 days 1 year (frozen)

OBBH3-45M OBBH3-45M OBBL2-250L

87,302 87,302 103,175

International Food and Agribusiness Management Review

147

Zhu et al.

Volume 21, Issue 1, 2018

majority of the livestock raised in this region are sheep10. Shepherds can provide a high-quality product at a relatively low cost, and Dengkou County had secured a place in the domestic fresh product market. However, over time, lack of organization, antiquated production facilities and processes, and lack of a brand contributed to the declining price for live sheep for local producers and, regardless of herd size, herdsmen have suffered losses. Because of these struggles and the regional importance of sheep, TAC selected the industry as a strategic priority. The cooperative employed methods to cut cost and improve market access like those developed for the Wallace Melon including recruiting additional producers, inviting experienced members and other professionals to offer training, consolidating procurement and marketing activities, and adoption of improved logistics and storage technologies. The most influential contribution was the role of the Tonghui warehouses. In a similar strategy to that employed for the Wallace Melon, to attract sheep farmers to join the cooperative, TAC negotiated reduced prices for raw materials and invited experts to provide technical expertise and teach short courses. On the supply-side a different strategy would need to be employed. Approximately 70% of the meat produced in the county is exported. The cooperative needed to consider development of marketing channels for export and worked under the reality that the world market dictates price and increases substantially the risk faced by herdsmen. At this time, sales models such as contract farming, electronic business platforms, and directly arranged links between farmers and supermarkets were not mature and not used by local herdsmen. To extend the viable sales window to help mitigate risk associated with price volatility, the cooperative decided to establish additional town- and village-level warehouses and equip them with freezers to store meat. Although this strategy allowed member-herdsmen to reduce the effects of external market volatility, there was a need to do more. As sales of Dengkou mutton had declined in recent years, capacity use of small-scale local processing factories was relatively low, and some were nearly idle. Even the larger-scale facilities faced a sharp reduction in use. Cooperative leadership proposed that the warehouse enterprises be combined with existing local fresh-processing capability, and the idea gained support among the members. Tonghui Warehouse began its expansion in September 2015 with a focus on integrating the fresh-processing facilities. The cost of integrating existing facilities was far less than the establishment of new facilities. The warehouses provided the cooperative the flexibility to carry out secondary processing steps based on storage capacity and forecasted orders. Their use would also serve to bring additional jobs to the area. Tonghui quickly built or updated seven warehouses, each with a supply radius of about 25 miles. The largest one covers 32,290 square feet and the smallest is just 4,300 square feet. They upgraded existing processing equipment, and introduced meat roll and packaging equipment and quick-freezing warehouse technology, expanding their ability to offer both primary and processed mutton. The option to contract out the processing step to large-scale meat processing factories was retained. ■■ Brand development Currently, some domestic fresh mutton brands stand out among others, such as Xinliguole, Donglaishun, and Tianmashengtai. Foreign brands such as New Zealand Sheep have also become popular in recent years. In addition to managing processing and storage, the cooperative was faced with the challenge of creating a brand image that would stand out in a field that had become increasingly crowded with branded product. During their market analysis, the cooperative discovered that most mutton sold under existing brands came from large sheep farms rather than from individual households, the latter being the case in Dengkou County. The local way of raising sheep is through open-range grazing as it is a long-held local belief that sheep free to graze on open land result in a more natural and delicious product. The cooperative chose to build on their production practices rather than adapt them to the confined facilities used to support most other brands. They developed a ‘free-ranging, naturally healthy’ brand image.

Hand-grabbing mutton is a rich tradition among the residents of Inner Mongolia. Local residents believe that the local grasslands grazed by sheep and cattle contain five herbs necessary for complete seasoning so that mutton can be cooked without seasoning. The mutton is simply cut into pieces and boiled in salt water.

International Food and Agribusiness Management Review

148

Zhu et al.

Volume 21, Issue 1, 2018

Challenges for Tonghui Cooperative were small animal numbers, a lack of market power, and cultural and other differences between members within the cooperative that made it difficult to coordinate as required for brand establishment and development. Chairman Wang summed up the outcome of discussion among cooperative members in his statement, ‘Our limited funds determine we cannot promote our brand by advertisement over a short period. We can only establish our own brand to be low-cost, and then improve our product step by step, accumulating brand influence over time. The market will embrace us as long as we have good products’. The cooperative registered the Tonghui Cooperative brand and started simply by upgrading the packing machines so every product delivered from cooperative facilities had the Tonghui Warehouse label. Overtime, the strategy has been relatively successful. The Tonghui brand has captured an increasing share of the market, having processed and marketed over half of locally-raised sheep by the end of 2016. To take advantage of spillover effects associated with introducing the brand, the cooperative entered other surrounding markets including Baotou, the largest city in Inner Mongolia and home to more than 1.7 million residents. They also actively employed e-commerce channels including Taobao, Wechat and a supply and marketing business platform for cooperatives. Aided by e-commerce, the Tonghui brand has begun to gain recognition around the country. Through the establishment of Tonghui Warehouse, longitudinally extending the supply chain, brand development, and actively pursuing multiple additional marketing venues, TAC established a coordinated mutton supply chain extending from the herdsman through the storage, processing, and logistics and transportation steps, and concluding with high-potential marketing outlets.

5. Horizontal expansion through tourism With the objectives of promoting their brand, taking advantage of additional place-dependent opportunities and enhancing economic development, the cooperative decided to explore development of a rural tourism industry. Dengkou has many prestigious human and natural landscapes, such as Sanshenggong scenic spots, Agui temple, Nalin Lake, and Hateng Taohai Nature Reserve. In December 2015, TAC invited the Beijing Tourism Association to examine Dengkou’ tourism resources, help plan tourism routes, and establish specific strategies for individual farm households. In China, the Mongolian are called ‘the nation on horseback’. Now, the Han, China’s ethnic majority, are the majority among Inner Mongolia residents, but the personality characteristics of the local people have persevered (honest, forthright and resilient). The literally-interpreted ‘eating hunks of meat and drinking wine in big bowls’ is a signature of the culture. To embrace this and to further the interest of outsiders in the life and culture of Inner Mongolia, TAC adopted a rural tourism slogan ‘drink and eat as you like’. The hand-grabbing mutton, slow cooked over an open fire, would be an important part of the strategy. Distinctive ethnic characteristics and magnificent regional natural scenery are the natural advantages for the rural tourism at Dengkou County (Figure 5). TAC divided the tourism product structure into three layers. The first layer product is experiencing the local customs. During the inspection period with the Vice President of the Beijing Tourism Association, Mr. Wei, and based on member input, the tourism route was designed. The main tourist attractions include the Sanshenggong, Demonstration picking Garden, Ten-kilometer Ecological Park, and sunflower and fruit farms. The second layer is a homestay at a ‘Happy Farm House’. Guests will eat farm cuisine, sleep in the farm house, partake in farm work, enjoy the farm scenery, participate in activities such as riding horses, and purchase branded farm products. Participating member households need to renovate their houses, register with the Business Administration Department, and pass a home inspection. TAC is cooperating with the county government to carry out infrastructure maintenance around the scenic areas and ensure there are procedures to ensure that accommodation facilities are maintained and the food and environment are safe for guests. TAC will participate in organizing trips to show the local culture and will provide a variety of farm services. The third layer is the addition of tourism-specific activities. For example, TAC will use internal funds and sponsored fundraising efforts to rent a sightseeing helicopter. Visitors will be able to enjoy desert flight and deepen their understanding of the natural scenery, local conditions and customs of Inner Mongolia. Other International Food and Agribusiness Management Review

149

Zhu et al.

Volume 21, Issue 1, 2018

Figure 5. Examples of Tonghui Agricultural Cooperative rural tourism sights. excursions will also be added based on additional market research, guest input and the existing activities and capabilities of participating member households (e.g. sand dune rides in the desert).

6. Supply chain finance As elsewhere, agricultural production in Dengkou is dependent on weather and other external factors. And, there is considerable volatility in prices. Because of the risk, relatively low profitability, lack of liquidity, and high service costs, local farmers experience a considerable level of financial exclusion (i.e. demand for commercial financing far exceeds supply). To help cover the credit gap, and supported by new legislation, TAC introduced supply chain financing to their mix of member services. The cooperative was authorized to begin offering credit in April, 2015, and has successfully done so, providing members with improved credit access and reduced rates. The cooperative can provide loans for members with its own funds. It can also cooperate with financial institutions such as banks to facilitate loans to members with external funds. It does so by serving as a guarantor on behalf of its members. 6.1 The Jinfurong integrated management system An important initial step in developing a credit system was information management. The cooperative worked with Peking (Beijing) Fumin Commune to introduce the Jinfurong Integrated Management System (JIMS). This system integrates member financial information such as family income and credit activities, property, and member non-financial information such as family membersâ&#x20AC;&#x2122; status and reputation; membership details; cooperative service, financial, and procurement management; and cooperative credit activities. Joining and departing members, records and accounting of purchases between the cooperative and members, daily expense reimbursement, and period (e.g. monthly) statements for the Cooperative are processed through the system. The system allows for standardized management of all member personal, financial and operation data. It allows the cooperative to establish and maintain financial criteria and greatly facilitates transparency International Food and Agribusiness Management Review

150

Zhu et al.

Volume 21, Issue 1, 2018

within the organization and with downstream, upstream and facilitating partners. Use of the system allows TAC to quantitatively assess member creditworthiness, provide fair lending quotas, and reduce credit risk for members and for the cooperative. 6.2 Use of cooperative funds to provide internal financing through trade credit Agricultural production factors require financing, and local farmers generally require external credit to provide this financing. Lending from formal financial institutions not only has a complex procedure and approval process, but also generally carries a high interest rate. To overcome this challenge, the cooperative made use of investment contributions from members and founders to, according to individual credit assessments, issue loans to members by the form of joint guarantee of three households. The joint guarantee system is commonly employed in China. A joint guarantee group responsibility can greatly reduce risk associated with loan non-repayment. Tonghui Cooperative’s loan origination in practice is relatively flexible. It works on the assumption that credit needs vary from farmer to farmer. Loans usually consist of short-term funds and are used to provide capital for activities such as building or repairing a greenhouse or planting a crop. Within the Cooperative’s supply chain, if members using contract marketing require funds to purchase capital equipment, the cooperative can provide a loan by trade credit. Regardless of the amount or details of the loan, the Cooperative plays the role of fund circulator within in the supply chain. In their first year of operation, Tonghui Cooperative issued loans of approximately 200,000 ¥ (about $32,000) using internal funding. A typical loan is 20,000 to 30,000 ¥ ($3,200 to $4,800) at a rate approximating 9.7%, which is 2% lower than that members could obtain directly from formal financial institutions if they were successful in obtaining such a loan. Borrowing from the cooperative also reduces pressure on the member because payments are amortized rather than the balance being due at one time. 6.3 External financing function Tonghui Cooperative is still in its early stages of development. Its financial credit strength therefore is limited. To better meet member needs during this time of initial growth, the cooperative explored other means to increase member access to credit. The result was establishment of relationships with local banks including the Bank of China, Agricultural Bank of China, Rural Credit Cooperatives and rural banks. The cooperative prepares the joint guarantee agreement and supervises the use of loans taken by its members from these financial institutions. Amortized loans are used to reduce risk for the bank. From August to October 2015, Tonghui Cooperative cooperated with Mengyin Rural Bank issuing 2,300,000 ¥ ($36,500) in loan funds for use by 48 households. In February 2016, Tonghui Cooperative worked with the Bank of China and Agricultural Bank of China to offer loans to more 100 households. Under the agreement, Agricultural Bank of China and Mengyin Rural Bank offer policy-based lending with low interest rates of between 6.35 and 6.9%. Generally, the Bank of China and Rural Credit Cooperatives provide commercial lending, with annual interest rates of between 7.2 and 11%, which are higher than policy-based lending but still lower than the market average. Potential loan volume is large. Bank of China has approved issuance of 20 million ¥ ($3.17 million) in loan funds. The Bank of China and the Rural Credit Cooperative require a joint guarantee of three households and Mengyin of five households (Table 6). Loan terms are generally three to six months. Tonghui Cooperative first investigates the credit worthiness of the potential borrower including consideration of their personal history, accounts, and operation and operational history, loan amount, and intended use of funds (Figure 6). After selecting suitable borrowers, the cooperative develops loan teams comprised of three (or five) borrowers placed in a joint guarantee on a voluntary basis. The information is packaged and sent to external financial institutions. The bank will approve loans quickly, issuing a Funong card for every International Food and Agribusiness Management Review

151

Zhu et al.

Volume 21, Issue 1, 2018

Table 6. Member loan details. Lender

Rate

Amount ($)

Form

Nature

3,175 to 4,762

Three household guarantee and cooperative recommendation Cooperative recommendation and guarantee Five household guarantee and cooperative recommendation Three household guarantee and cooperative recommendation

Commercial

Bank of China

7.2%

Agricultural Bank of China Mengyin Rural Bank

6.35% 7,937

Rural Credit Cooperative

6.9% 11%

3,175 to 4,762 7,937

Government support loan Commercial Government support loan

farmer with a limit set to match the loan amount. The Funong card is similar to a debit card with the account defined at the level of the loan limit. Approved farmers bring certificates and other evidence to complete the loan process, and receive their Funong card. 70% of the loan amount on the card cannot be directly withdrawn, but is transferred to the cooperative and is available to the farmer to purchase chemical fertilizer, seeds, and other operating items from the cooperative. 30% is provided for direct use by members for additional operating expenses and for resources that cannot be obtained through the cooperative or that farmers do not choose to obtain through the cooperative. Farmers pay off the loans in installments. If overdue, the other members of their joint household guarantee must shoulder the responsibility. A farmerâ&#x20AC;&#x2122;s product storage in a Tonghui warehouse can also be used to pay off the debts

Credit check Lending group and recommendation Approval (issue credit card) Draw money from branch office of cooperative Collective account

Cash

Unified purchase

Use for personal producing

Installment repayment Success

Overdue Group joint liability and debt repayment

Figure 6. Loan operating process.

International Food and Agribusiness Management Review

152

Zhu et al.

Volume 21, Issue 1, 2018

6.4 Look to the future of credit operations In the future, TAC plans for JIMS and the Tonghui Warehouse to play more important roles. The JIMS will improve the quality and timeliness of information and make it more efficient and convenient for enterprises to obtain loans. The Tonghui Warehouse, as the logistics base for storage, will provide collateral on behalf of farmers as well as their upstream and downstream partners (Figure 7). When upstream enterprises provide production inputs (e.g. seed, machinery and equipment) to cooperative members, they also face cash flow requirements to cover their costs during production and after sale to the cooperative and to its members that they cannot meet internally. In the proposed system, this upstream partner can gain approval for their own external financing needs when supported by credit information available in the JIMS, and using as (partial) collateral, materials stored in Tonghui Warehouse. This collateral guarantee by the cooperative can make it easier and less expensive for the upstream partner to obtain external financing. Downstream enterprises will benefit from the same system. Furthermore, commitment by the downstream partner to purchase raw or processed commodities reduces risk for the external lender, further increasing the value of lending to cooperative members and marketing channel partners. The Jinfurong system is the facilitating mechanism. It provides information including inventories, values, and historic use of raw materials provided by upstream firms and the production plans of cooperative members interpreted in terms of intended purchase and sale orders. A full understanding of the likelihood of success of the proposed system depends on the readerâ&#x20AC;&#x2122;s understanding of the culture of agriculture and of business in China. It is not necessarily intuitive under the western culture of financing and business to business relationships. For example, an important value offered by the cooperative organization discussed at length in this case is that resulting from its market power on both ends of its marketing channel, procurement and sales. This remains important, but its value originates less from a new negotiating position and more from easing transactions between the users, independent farmers now represented as one entity, and solidifying relationships with upstream and downstream partners. To understand why the described system brings value, it is necessary to view the relationships between the cooperative and its marketing channel partners as true partnerships. For example, for the relationship with the cooperative to provide a guarantee that is of value to external financial institutions, the commitment by all parties must be made prior to production. Specifically, for an external financial institution to accept the cooperative as a financial guarantor for the input supplier, the supplier must have a commitment from the cooperative and

Loans supporting upstream partners

Financial institution

Credit information

Jinfurong system

Member information Upstream partner

Warehouse

Loans supporting downstream partners

Collateral information Downstream partner

Cooperative

Supermarket, end-consumer

Figure 7. Future planning of finance in the cooperative supply chain.

International Food and Agribusiness Management Review

153

Zhu et al.

Volume 21, Issue 1, 2018

its members to purchase its materials. The same is true for the relationship between the cooperative and downstream firms using the relationship to acquire financing to fund its cash flow requirements. Tonghui Cooperative is well on its way to successfully building capacity for farmers and herdsman by continuing to develop financing coordination that combines the resources and needs of the Cooperative and its warehouses, upstream and downstream enterprises, financial institutions and farmers and herdsmen. This model not only is a breakthrough in solving challenges of resourcing financing needs for all parties, but also forms a close collaboration of risk- and benefit-sharing by guiding business cooperation among the cooperative, farmers, enterprises, and financial institutions.

7. Discussion and conclusions TAC has quickly and successfully designed, and continues to improve, its agricultural supply chain by adding value to commodities, market expansion, increasing efficiency of logistics, and improving information and capital flow (Figure 8). The cooperative facilitates efficiency throughout the supply chain through unified purchase and sale and partnerships with buyers and suppliers. Increased access to information helps the cooperatives, its members, and its partners in the marketing channel to better plan and reduces transactions costs and risk. Use of the JIMS provides information about members and up- and down-stream partners to help the cooperative and external financiers assess credit worthiness, facilitates the provision of financial services, and allows the cooperative to assess the drivers of its financial performance. Through the exploitation and coordination of resources, TAC has effectively developed a value-added marketing channel that provides income and mitigates risk for its producer members and its marketing channel partners. It has also promoted rural development in the county. We conclude the paper by revisiting the three questions posed early in the paper.

Service supply (training, technology, information, finance)

Supply of seeds, agricultural machinery, pesticides and fertilizers

Plant cultivation animal breeding

Crop/livestock growth/harvest

Storage processing logistics

Direct-sales stores Direct-order sales

Consumers

Happy farmhouse

Sightseeing, picking garden, desert flight, eco-tour, family travel Backward integration

Forward integration

Figure 8. Optimized agricultural supply chain.

International Food and Agribusiness Management Review

154

Horizontal integration

E-business platform

Zhu et al.

Volume 21, Issue 1, 2018

1. How does Tonghui Agricultural Cooperative use the flow of information and capital, and its logistics system to integrate and coordinate the supply channel from the purchase and use of inputs to the end consumer to improve the welfare and income of local farmers? TAC is representative of a new generation of cooperatives in China. This type of cooperative introduces the principle of proportionality through the option of equity stock. These cooperatives remain true to the basic principles of cooperatives: member-owned, controlled and benefited. Members benefit through purchasing through the cooperative at favorable prices in large part due to negotiation power not available to them independently; lower rate, easier-to-access financing; education and technical services; and sale of individually-produced agricultural products at favorable prices with less risk. As the owners, members have the option to inject capital into the cooperative for additional earnings opportunities. Members receive these advantages without giving up their democratic right to participate in the decision-making process. Through the cooperative members of the General Assembly, members democratically elect the Board of Management and the Board of Supervisors to guide strategic and tactical development. Professional leadership and fulltime management in sales, production, finance and administration provide the business astuteness required to maintain and grow the cooperative in the evolving production and marketing environments. Cooperative earnings are returned to the members in proportion to the volume of trade through patronage. 2. Given that the development of many farmersâ&#x20AC;&#x2122; cooperatives in China have failed, what is different about the organization and strategic and tactical behavior of Tonghui Agricultural Cooperative that has led to success? The necessary but not sufficient keys to the success of TAC are leadership and entrepreneurship. Mr. Wang Lu, President of Tonghui Cooperative, and the other initial cooperative investors played a key role in the initial establishment and subsequent development of the cooperative, the formation and optimization of the agricultural supply chain, and the formal inclusion of marketing channel and supporting industry partnerships. They have continued to grow the cooperative in breadth and depth through their ambitious pursuit of alternatives for members to grow their operations, income, and technical expertise and their ability to effectively communicate with and include in decision-making, not only the farmer and herdsman members, but partners throughout the marketing channel and in supporting industries and the government sector. The leadership has also been proactive in consulting with university and business experts on plans and in adapting for use supporting development tools such as the JIMS. For example, the TAC President used his personal connections to overcome an important challenge facing the cooperative, transporting less-thanfull-truckload shipments of produce and other products over long distances. Produce, for example, is both bulky and heavy. And, because it is perishable and rather fragile, it is important that the product be moved as soon as possible out of the field. The challenge facing TAC was that third-party logistics firms usually require transport vehicles to reach near full capacity before shipment, a standard TAC does not yet regularly reach. Mr. Wang negotiated with the logistics company China Post for flexible, partial load transportation. The management team also worked with an entrepreneur from Taiwan to adopt a storage technology that extended the shelf life of TAC products allowing for additional time to make fuller-capacity loads. For his leadership, Wang Lu, elected by the cooperative members and by the Inner Mongolia Autonomous Region, won the 2016 â&#x20AC;&#x2DC;National Youth Agriculture Rich Leader Awardâ&#x20AC;&#x2122;. Earlier, we identified cooperative characteristics or behaviors with potential to support successful and sustained performance in a developing cooperative in addition to strong leadership to include: member understanding of and participation in cooperative governance; an efficient democratic process for decisionmaking; a large base of members with medium and/or small-sized farms; a willingness to coordinate the marketing channel by integration, partnerships, agreements, or other coordinating activities; and the support of government in that an environment exists that supports entrepreneurial activities and cooperatives. And, finally, that a cooperative must be both prepared and flexible as it faces new challenges and opportunities, including growth. The evolution of TAC has benefited from these defining characteristics and strategies.

International Food and Agribusiness Management Review

155

Zhu et al.

Volume 21, Issue 1, 2018

TAC has made an important part of its mission the education of its members. This includes, but it not limited to, the technical aspects of production and the business of procurement, sales and credit. They also educate their members, formally through all-member and branch seminars and informally during day to day interactions, about the role of the cooperative and cooperative governance. Through the General Membership Body, which elects the Board of Management and the Board of Supervisors, each member has a vote in the decision-making process. The system works because it allows for members to have a voice but allows for the efficiency in decision-making and flexibility in management that is so important as TAC faces new challenges and opportunities, including growth. While less formal evidence, it is worth noting that the President is from the region and the management team and staff are members, ensuring that members have ready access to their leadership. As such, and because regional farms are generally small, individual family operations, there is little concern about members with larger operations dominating the decision-making. For at least the intermediate term, relative homogeneity in farm size and wealth, or lack thereof, is equaling among the members and potential members. Where the cooperative really excels is the criteria of a willingness to coordinate the marketing channel by integration, partnerships, agreements, or other coordinating activities. The case study is filled with examples of the efforts by TAC to exploit opportunities for marketing channel development. This is important because the primary motivation for membership in TAC is to benefit from services and opportunities it offers. To date, the cooperative has extended its partnerships and coordination to every link of the marketing channel, and the supporting industries. TAC developed and is multiplying the cooperative brand. It has diversified the marketing channel and the production opportunities for its members. For example, members raising farm products have begun the process of renovating for home-stay tourism; others have become involved in further processing of their agricultural products and those of their peer members. Increasing the breadth of production of individual member families and of cooperative offerings has allowed TAC to combine the advantages of specialized economies of scale and multi-scope economies. Finally, the cooperative is blessed with strong support by government both nationally and locally. They have received financial support to add process infrastructure and local support to facilitate development of the tourism industry. Cooperative management worked to obtain the support of local government entities, both financial and through partnerships, to develop value such as through the development of tourism sites and establishment of the rules accompanying home-stay visits. 3. What is the impact of Tonghui Agricultural Cooperative organization and growth on the rural economy? Well-run cooperatives will show growing positive externalities as they evolve. Based on our global understanding of cooperatives as demonstrated through the literature and confirmed in our case analysis of TAC, cooperatives in general provide the most value in peasant industries with perishable products sold in a commodity market, and with particular logistics requirements; products with relatively large investment requirements for production and processing technology; and in communities where farmers and herdsmen have relatively low levels of market power, education, or literacy. A close-knit community based on cooperation and acceptance like that found in Dengkou County and other Inner Mongolian regions guided by strong leaders will find an easier road to success. Although many cooperatives were originally a combination of vulnerable individuals, the growth of cooperatives and their well-functioning societies as is the case with TAC can provide strong externalities. TAC has more than 160 agricultural products with quality certifications, has established a product tracking system working with the government that not only provides brand value but mitigates the chance of food safety instances and, if they occur, will help mitigate their breadth and impact. By way of partnerships with procurement and marketing vendors, the cooperative can reduce the steps associated with moving products from the farm to the end consumer and help stabilize regional prices. Their internal credit and development and use of the JIMS provides information that reduces transactions costs and risks for financial institutions, both increasing availability and reducing risk of financing for all parties. International Food and Agribusiness Management Review

156

Zhu et al.

Volume 21, Issue 1, 2018

Offering regular education and training activities on the cooperative business structure and operation, production, and new technology, among other topics, and providing timely and readily available information to members, the cooperative can work to gradually cultivate its farmer-members into more efficient operators with market astuteness and the capacity to grow and accept additional risk. Farmers can diversify their operations through production, vertical expansion (e.g. processing), or horizontal expansion (e.g. home stays). The cooperative itself employs staff, increasing the job base and providing employment that allows new college graduates the opportunity to return to their home region. Further, the series of production and marketing activities added in the marketing channel such as the expansion of warehouses creates jobs. And, the cooperative provides a platform for discussion and exchange among scattered farmers through the regular meetings of members of grass-roots bureaus, which arouses the enthusiasm of farmers and improves the organization and cohesion of rural and farmers, thus enhancing rural harmony and stability. 7.1 Opportunities and challenges ahead TAC provides comprehensive services for members, brings benefits to members and distributes proceeds to members. Benefits to members originate from three sources. First, production costs are reduced through increased market power for procurement of inputs and production efficiencies resulting from education and training. Second, sales revenues increase and market risk declines because of marketing efficiencies from economies of size and from cooperative storage, logistics, and marketing activities. Third, the cooperative increases access to and reduces the cost of credit for members. The cooperative also has successfully earned profits, which are distributed to members based on their transactions volume with the cooperative. Less than two years after establishment, TAC has been able to engineer a supply chain that successfully joins the products of farmer and herdsmen members with external partners throughout the marketing channel. It extends the theoretical value of cooperative organization to include the cultural realities of interactions between marketing channel participants in China. The variety of fresh and processed products offered by the cooperative continues to grow. Sound connections with industry and academic partners, a creative spirit, hard work, and drive have facilitated growth of the size and breadth of activities of the cooperative. Carefully engineered marketing channels have been established for many products. Tonghui Cooperative coordinates upstream and downstream enterprises and carries out quasi-integrated management throughout the marketing channel. It provides comprehensive services including facilitating cooperation in production, procurement and marketing and credit; improving member welfare. TAC is ready for the next steps. One opportunity is participating in the public information platform planned for Dengkou County to help in brand-building efforts. The platform will allow for the recording and public availability of detailed information about production and subsequent logistics, thus establishing complete product traceability. To protect the interests of consumers, the information platform is being developed by the government per the legislated standards. The continuing challenge for TAC in the coming years will be to identify additional opportunities and work to solve challenges that arise.

Acknowledgements The research is funded by the Science Research Foundation of Renmin University of China (15XNB025).

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2016.0095. Figure S1. Direct-sale store of Tonghui Agricultural Cooperative.

International Food and Agribusiness Management Review

157

Zhu et al.

Volume 21, Issue 1, 2018

References Abebaw, D. and M.G. Haile. 2013. The impact of cooperatives in agricultural technology adoption: empirical evidence from Ethiopia. Food Policy 38: 82-91. Bernard, T. and D.J. Spielman. 2009. Reaching the rural poor through rural producer organizations? A study of agricultural marketing cooperatives in Ethiopia. Food Policy 34(1): 60-69. Garrido, S. 2007. Why did most cooperatives fail? Spanish agricultural cooperation in the early twentieth century. Rural History 18(2): 183-200. Hohler, J. and R. Kuhl. 2014. Position and performance of farmer cooperatives in the food supply chain of the EU-27. Annals of Public and Cooperative Economics 85(4): 579-595. Holloway, G., C. Nicholson, C. Delgado, S. Staal and S. Ehui. 2000. Agroindustrialization through institutional innovation: transaction costs, cooperatives and milk-market development in the east-African highlands. Agricultural Economics 23(3): 279-288. Hu, Z., Q.F. Zhang and J.A. Donaldson. 2015. Understanding the failure of farmers’ specialized cooperatives in China. IPP Review. Available at: http://tinyurl.com/y8zuwlzw. Huang, P.C.C. 2011. China’s new-age small farms and their vertical integration: agribusiness or coops? Modern China 37(2): 107-134. Jia, X., J. Huang and Z. Xu. 2012. Marketing of farmer professional cooperatives in the wave of transformed agrofood market in China. China Economic Review 23: 665-674. Kong, X. and H. Jin. 2012. A research study on agricultural cooperatives of foreign countries: emergence, mechanisms and experiences. China Agriculture Press, Beijing, China, pp. 379-381. (in Chinese). Kong, X., B. Shi and Z. Zhong. 2012. Research on the operating mechanisms and social benefits of farmers’ cooperatives in China: based on surveys from one hundred cooperatives and one thousand households. China Agriculture Press, Beijing,China, pp. 62-79. (in Chinese). Lerman, Z. and C. Parliament. 1990. Comparative performance of cooperatives and investor-owned firms in US food industries. Agribusiness 6(6): 527-540. Li, J. and D. Ping. 2003. Rural land tenure reforms in China: issues, regulations and prospects for additional reform. Land Reform 3: 59-72. Liao, A. and X. Guo. 2015. The logic and tendency of the evolvement of China’s agricultural organization system: an analysis framework of industrial chain integration. China Rural Economy 2: 13-21. (in Chinese). Machethe, C.L. 1990. Factors contributing to poor performance of agricultural co-operatives in less developed areas. Agrekon 29(4): 305-309. Ministry of Agriculture of China. 2015. 1.47 million cooperatives cover forty percent of national farmer households. Available at: http://tinyurl.com/y85kpd2m. Ortmann, G.F. and R.P. King. 2007. Agricultural cooperatives II: can they facilitate access of small-scale farmers in South Africa to input and product markets? Agrekon 46(2): 219-244. Parliament, C., Z. Lerman and J. Fulton. 1990. Performance of cooperatives and investor-owned firms in the dairy industry. Journal of Agricultural Cooperation 5: 1-16. Roets, M. 2004. From folklore to feasibility: commercialisation of South Africa’s indigenous goats. Unpublished Ph.D. Thesis, Department of Agricultural Economics, Extension and Rural Development, University of Pretoria, Pretoria, South Africa. Sexton, R. and J. Iskow. 1988. Factors critical to the success or failure of emerging agricultural cooperatives. Giannini Foundation Information Series No. 88-3, Division of Agriculture and Natural Resources, University of California, California, CA, USA. Sexton, R. and J. Iskow. 1993. What do we know about the economic efficiency of cooperatives? An evaluative survey. Journal of Agricultural Cooperation 8: 15-27. Soboh, R.A., A. Oude Lansink, G. Giesen, G. and G. van Dijk. 2009. Performance measurement of the agricultural marketing cooperatives: the gap between theory and practice. Applied Economic Perspectives and Policy 31(3): 446-469. Soboh, R., A. Oude Lansink, and G. Van Dijk. 2011. Efficiency of cooperatives and investor owned firms revisited. Journal of Agricultural Economics 63(1): 142-157. International Food and Agribusiness Management Review

158

Zhu et al.

Volume 21, Issue 1, 2018

United States Department of Agriculture (USDA). 1980. Cooperative benefits and limitations: farmer cooperatives in the United States. Cooperative information report 1 section 3. Available at: http:// tinyurl.com/ya6vkgdv. Van der Walt, L. 2008. Collective entrepreneurship as a means for sustainable community development: a cooperative case study in South Africa. Forum Empresarial 13(2): 3-20. Yang, H., C. Leeuwis, R. Lie and Y. Song. 2014. The landscape of farmer cooperatives in China: functions and diversity in a changing environment. Rural China: an International Journal of History and Social Sciences 11: 119-155. Zhu, Q., X. Luo and J. Ma. 2015. Organizational costs, exclusive resources, and management of rural mutual fundâ&#x20AC;&#x2122;s initiators. Chinese Rural Economy 12: 49-62. (in Chinese). Zhu, Q. and J. Ma. 2013. Institutional arrangement of participation and self-organizing and the effective operation of CDF: taking the World Bankâ&#x20AC;&#x2122;s Baishui CDF projects in Shaanxi province as an example. China Rural Survey 4: 42-59. (in Chinese).

International Food and Agribusiness Management Review

159

OPEN ACCESS International Food and Agribusiness Management Review Volume 21 Issue 1, 2018; DOI: 10.22434/IFAMR2017.0053 Received: 15 June 2017 / Accepted: 21 September 2017

Growth strategies for a commercial farm: the AgroPastoril Campanelli case study CASE STUDY Roberto Fava Scarea, Allan Wayne Grayb, Rodrigo Lourenço Farinha c, Erin Chelsea Fullertond, and Marcos Fava Nevesa aProfessor,

and cUndergraduate Student, School of Economics, Business and Accounting of Ribeirão Preto (FEARP), University of São Paulo, Av. Bandeirantes, 3900, 14020-590, Ribeirão Preto, São Paulo, Brazil bProfessor,

and dMSc Student, University of Purdue, Agricultural Economics, 403 West State Street, West Lafayette 47907, Indiana, USA

Abstract At 2017, Victor Campanelli remembered the history of his family’s business – the AgroPastoril Campanelli farm. His thoughts were back to the challenges that his father had faced, including the investments on an integrated production system that contributed to the family’s business growth. The production of cattle, sugarcane, and corn are all part of this integration system that has helped the Campanelli family to develop a successful business model. Currently, AgroPastoril Campanelli has three main production activities: beef, sugarcane, and corn. With these activities, a sustainable circle was created to guarantee more efficiency in economic, environmental and social terms. As Victor Campanelli says, ‘there is a perfect synergy between sugarcane, moisture, corn, and cattle. Not just because of the better land and labor use, but also because of the inputs economy and the productivity improvement, both in agriculture and livestock’. However, the present time requires a new investment cycle, leading the whole family to consider new market opportunities, especially regarding farmer’s new demands. What should we do and where should we invest? Should we improve efficiency or open new areas? Should we invest in new agriculture enterprises or start new business outside agricultural market? Those and other crucial questions were posed by Victor Campanelli thinking on potential growth strategies for AgroPastoril Campanelli. Keywords: agribusiness, verticalization, strategy, farm, cattle, corn, sugarcane JEL code: Q12 Corresponding author: rlfarinha@usp.br

161

Scare et al.

Volume 21, Issue 1, 2018

1. Introduction It is a sunny morning in February 2017 as Victor Campanelli, co-owner and executive director for AgroPastoril Campanelli Farm, drives to his office in the city of Bebedouro, State of São Paulo. His company has three core business: the meat, sugarcane and corn production and has been well known due the high investments in technology and in an integrated production system. Victor drives and contemplates the future of his family’s business, he cannot help but remember the history of his family’s farm. His thoughts go back to the challenges his grandfather and father had faced, not the least of which included migrating to a new country and making investments in the coffee production business. He also thinks of the success his family achieved in the past as a result of facing these challenges, all of which contributed to the growth of the family business into the successful company it is today. The family had maintained a tradition of not only hard work and success, but also of movement and innovation, adapting to the quickly-changing agricultural environment. The environment and markets AgroPastoril Campanelli operates in are constantly changing. Producers cannot afford to remain stagnant if they want to remain at the top of the industry and maintain a profitable operation. The need for new investments leads the Campanelli family to constantly consider new market opportunities that can help the farm continue to improve efficiency and profitability. Victor Campanelli believes there is always room for improvement in their current operations. He ponders the questions: ‘what should we pursue next and where should we invest? What direction should we head to keep advancing and avoid remaining stagnant at our current position? Should we find a new business model? Or should we invest further in our current integration model?’ He sees diversification as a necessary path, but also as a challenge. In order for the operation to remain successful, all of the activities must move and advance together. The frequent addition of new family stakeholders can also create management obstacles to work around each year. AgroPastoril Campanelli has been successful in the past, but how must they proceed in order to remain at the top?

2. History of AgroPastoril Campanelli Paschoal Campanelli, Victor’s great grandfather, left Italy in the beginning of the twentieth century to buy some small coffee plantations in the rural region of Olimpia, located in western São Paulo, Brazil. Within a few years, Paschoal Campanelli had already become a successful rancher and coffee grower. Throughout the years, the existence of cheap land opened the opportunity to expand production, and also begin production of other products, such as oranges and cattle. In a town close to Olimpia, named Bebedouro, the AgroPastoril Campanelli SA company was founded in 1982 by all of the Paschoal Campanelli sons to honor the legacy of their father. Along with the official company, they also formed the enterprise board and the stakeholders’ council. Paschoal Campanelli is the grandfather of the generation of Campanelli’s directors, which includes Victor Campanelli as one of the owners and group directors. After founding AgroPastoril Campanelli in 1982, the company invested in the expansion of coffee plants and in orange tree farms, acquiring more farmland close to the first tract purchased by Senior Campanelli upon arriving in Brazil. While the company produced cattle at this time, it was a secondary production activity. In 1997, Campanelli family began to experience production problems with the rise of citrus canker, also known as greening disease. This crisis affected and damaged many crops throughout Brazil and Florida, leading AgroPastoril Campanelli to develop a plan of abandoning orange production and switching to production of sugarcane. In 2001, the Campanelli family set their plan into action and removed the first orange orchards, beginning with the less productive orchards. They continued to remove orchards through 2006, when the firm completely left the orange business. The removed orchards were gradually substituted by sugarcane fields as AgroPastoril Campanelli continued to increase sugarcane production, having all 8,000 hectares devoted to sugarcane by 2006. Livestock production was still a secondary production activity at International Food and Agribusiness Management Review

162

Scare et al.

Volume 21, Issue 1, 2018

this time. However, Victor Campanelli, who started getting more involved in the family business in that time, began to see synergies existed between sugarcane and cattle production and decided to invest more in cattle production, moving forward. Today, the company’s decisions are made by a board of directors composed by Victor, his grandfather and other family members. All the company’s strategic decisions are made by the board, however some of the directors have specific roles in the business. Victor’s grandfather is responsible for managing the calf and cattle that AgroPastoril Campanelli buys all year and Victor is the Executive Director of the company. Victor has a major in Business Administration by Fundação Armando Alvares Penteado and a postgraduate degree by Escola Superior de Agricultura ‘Luis de Queiroz’. Victor Campanelli business background transformed the family business and made him the person responsible for all the recent changes and innovation in AgroPastoril Campanelli. In 2016, the company had an average of 315 employees and approximately US $65 million in revenues. In observing the past of AgroPastoril Campanelli, two strong themes arise: tradition and movement. The generations work to preserve the values on which Paschoal Campanelli founded the company. They are pioneers in adopting the best practices and management and stay attuned to the market and environments in which they operate. They are willing to change and adjust their production activities or methods in order to stay profitable, operate at the highest level of efficiency possibly, and remain a top competitor in their industry.

3. The industry at large 3.1 The beef industry In 2016, the world production of beef and veal was over 60 million metric tons Carcass-Weight Equivalent (CWE) and the world consumption of beef and veal was over 58 million metric tons CWE. Brazil produced over 9 million metric tons CWE of beef and veal in 2016, making up 15.4% of world production, and consumed approximately 7.5 million metric tons of beef and veal (USDA, 2017b). Today, Brazil has made its way to one of the top three beef producers in the world, along with the United States and the European Union. Brazil exports around 10-12% of the world’s exported beef (Gerber et al., 2015). In 2015, Brazilian exports of beef totaled over US $5.9 million. The largest proportion of these beef exports went to Hong Kong (US $1.09 billion), followed by Egypt (US $661 million), Russia (US $579 million), Venezuela (US $534 million), and China (US $476.6 million) (ABIEC, 2015). The cow/calf operations are the first stage of the value chain for cattle production. The cattle are then sold to be raised either on pasture or in a confinement. Feedlots must also purchase or manufacture feed for their operation. Cattle from confinements and pasture are then sold to the slaughterhouses, which process the meat to be sold to end-users through retailers. Feedlot operations are usually large, vertically integrated, and fully mechanized. The meat value chain can be seen in Figure 1. The number of feedlots is rapidly growing in

International market

Inputs industry

Cattle production

Slaughterhouses

National market retailers

AgroPastoril Campanelli Business Scope AgroPastoril Campanelli produces cattle and components for the cattle feed, such as fibers, grains and micronutrients.

Figure 1. Meat value chain. International Food and Agribusiness Management Review

163

Consumers

Scare et al.

Volume 21, Issue 1, 2018

South America (Gerber et al., 2015). In contrast, economies of scale in beef production has lead the industry to experience large amounts of consolidation for slaughterhouses, characterizing this stage of the supply chain as highly concentrated. Three of the primary companies that purchase cattle for slaughter in Brazil are JBS, Minerva, and BRF. The small number of slaughterhouses in Brazil translates into relatively low bargaining power for the farmers supplying the cattle, leading to low, fluctuating profit margins for feedlots. Low profit margins create great risk for feedlot owners. Because of the high risk, a key to success in this stage of the value chain is maintaining efficiency and productivity in production. Efficiency of cattle herds (time in which cows are made into edible food product) is driven by: quality of feed, animal performance (weight gain), and breeding stock in the herd (Gerber et al., 2015). Obtaining high levels of efficiency in fattening cattle will aid feedlot owners in managing any risks associated with the sale of their cattle to slaughterhouses. Researchers expect production capacity to increase, particularly in low production areas, because of improvements in managing animal-crop and forest-pasture systems (Knoll et al., 2017). Victor Campanelli believes the beef market is in the process of transitioning. The current price of beef is relatively cheap for Brazilian consumers. However, the price of beef is expected to sharply increase in the coming years, leading to a shift in consumption to other proteins. Due to these changes that are occurring or anticipated in the near future, cattle producers will have to devote more focus to the quality of their animals and their fattening efficiency. It is possible that cattle producers may also be able to mitigate risk by seeking to strengthen partnerships with slaughterhouses in which they do business. Obtaining formal and complete contracts may aid feedlot owners in lowering the potential for high transaction costs or market failure. 3.2 The sugarcane industry More than 380 billion metric tons of sugar were produced in 2016. There were 55 million metric tons exported globally, of which Brazil accounted for 24 million metric tons, over 40%. Brazil is the largest sugarcane producer and exporter. The other major exporters following Brazil are Australia, the Philippines, Mexico, and Guatemala. Brazil has a natural competitive advantage for low cost production of sugarcane, largely due to the tropical climate and low to non-existent occurrence of natural disasters. Brazil is also the second largest producer of ethanol, which is often manufactured from sugarcane. Production in Brazil is expected to reach 47 million metric tons over the next five years, with exports predicted to reach as high as 34 million metric tons (Dâ&#x20AC;&#x2122;Costa, 2016). Sugarcane producers, such as the Campanelli family, are one of the early stages in the supply chain for sugar production, purchasing seedlings, fertilizers, and machinery from agricultural input suppliers. Producers can sell their sugarcane to manufacturers for processing of sugar for consumption, for use in bakery and other food products, for energy, or for ethanol. The manufacturing stage of the industry is defined as having relatively low concentration, particularly in Brazil. Low concentration can generally lend to high bargaining power for producers. However, location of the manufacturer is of importance in the production of sugarcane, because sugarcane must be crushed within 24 hours of harvest. In addition, the level of consolidation is expected to increase in the coming years, following the general trend within the agricultural industry of commercialization. Production of sugarcane is generally capital intensive, producing naturally high barriers to entry. Overall, sugarcane production is classified as a mature market and is expected to change little over the next five years (Dâ&#x20AC;&#x2122;Costa, 2016). Some countries have high levels of regulation regarding sugar production to protect local producers. Brazil, however, has operated in an environment with little to no regulation for the past 10-15 years. Many of the barriers for other countries, such as the United States and those in the European Union, are expected to decrease. This will likely spur an increase in international trade. The decrease in barriers and increase in

International Food and Agribusiness Management Review

164

Scare et al.

Volume 21, Issue 1, 2018

trade will have effects on the global price of sugarcane, as well as opportunities for sugarcane producers, in the coming years (D’Costa, 2016). In the past five years, the prices for sugarcane have experience high levels of volatility. Brazil’s large presence and impact on this market has been a driving force of the high price volatility. The price of sugar is directly related to ethanol prices. As energy prices increased in the years between 2010 and 2015, many producers devoted a large number of stocks toward ethanol production, decreasing the stocks of sugar available for consumption, and thus causing a large increase in the price of sugar. Following this, sugarcane production was encouraged, and an increase of 2 million metric tons produced in Brazil caused a large reduction in the price of sugar. Prices in 2016 rebounded from the decreases, and overall prices are expected to gradually increase over the coming years, leading to increased revenues in the sugar industry (D’Costa, 2016). 3.3 The corn industry The U.S. is the largest world exporter of corn; however, exports account for only 15% of production in the U.S. Global corn prices, therefore, are based largely on domestic trade in the U.S. and rely on weather within the corn belt region of the country. World producers wait to see the amount of corn that will be produced in the U.S. each year to make decisions on the size of their own crop (USDA, 2017c). For 2015-16, 963 million metric tons of corn were produced, and consumption totaled 981 million metric tons. The amount of corn traded worldwide peaked in 2013-14 at 130 million metric tons (USDA, 2017a). South America provides the primary source of competition for the U.S. (USDA, 2017c). Brazil exported 35,382 million metric tons and produced 67,000 million metric tons of corn in 2015/16. Both Brazil and Argentina had expected to have record crops for 2016/17, causing world production to largely increase from 2015/16 (USDA, 2017a). The reliance on U.S. crop size and prices creates the potential for great fluctuations in the global price, which can cause difficulties for corn producers, particularly in countries outside of the U.S.

4. AgroPastoril Campanelli AgroPastoril Campanelli is well known in the meat production market due the excellence in operations and pioneering in the integrated production of cattle, sugarcane and corn. These crops and livestock are produced in the company’s farms. There are four primary clusters of farms that are used for their production activities. The Santa Rosa Farm, where the feedlots are, is located in the city Altair, near Victor’s office, and has 4,500 hectares used for sugarcane, corn, and cattle. It is also the location of a newly built premix factory. Figure 2 shows the location of AgroPastoril Campanelli farms, Figure 3 shows the Santa Rosa Farm and Table 1 shows the area and production activities of each farm. 4.1 The integrated business model The production of cattle, sugarcane, and corn are all part of an integration process that has helped the Campanelli family to develop a sustainable business model. Currently, AgroPastoril Campanelli has three main production activities: beef, sugarcane, and corn. The core business is producing livestock, producing approximately 60,000 head of cattle per year and generating 59% of the company’s revenue. Two types of cattle are purchased. One type is yearlings that are around 280 kilograms. These cows, around 50% of Campanelli’s production, are shipped to the pasture for backgrounding and stay there until they reach around 400 kg. The cows are then shipped to the confinements. AgroPastoril Campanelli capacity for backgrounding depends on yearly rainfall in their pastures, but is generally around 25,000-30,000 heads of cattle. The other cattle are purchased at around 340 kg and shipped straight to the confinements. Cows stay in the confinements an average of 115-120 days and gain around 180 kg before being sold to the slaughterhouses. Cattle are purchased or sold for slaughter on a weekly basis. Table 2 presents AgroPastoril Campanelli’s purchase and selling of cattle throughout the years. The company’s confinements have the capacity of fattening 30,000 cattle heads at the same time. Approximately 74% of the beef produced by AgroPastoril International Food and Agribusiness Management Review

165

Scare et al.

Volume 21, Issue 1, 2018

Figure 2. Agro Pastoril Campanelli farms location.

Figure 3. Santa Rosa Farm, Campanelli feedlot located in Altair. Campanelli in 2014 went to Europe as part of the Hilton quota. The Hilton quota allows certain countries a specified amount of high quality beef (fresh, chilled, or frozen) to be exported to the European Union and then sold at a premium price. The quota consists of 58,100 metric tons of high quality beef from the United States, Brazil, Paraguay, Uruguay, Argentina, Australia, New Zealand, and Canada. Campanelli does not decide the amount shipped for the Hilton quote, however. The cows along with all their documentation are shipped to the slaughterhouses, and the slaughterhouses determine what will be shipped for the Hilton quota based on characteristics such as age and weight. AgroPastoril Campanelli devotes 8,000 hectares to sugarcane production, producing over 630,000 metric tons of sugarcane per year and comprising 36% of the companyâ&#x20AC;&#x2122;s revenue. To provide more fiber for the 36,000 confined cattle per year, the Campanelli group decided to use the sugarcane straw left in the soil after harvesting the sugarcane crop. About half of the total straw left in the soil after harvesting the sugarcane International Food and Agribusiness Management Review

166

Scare et al.

Volume 21, Issue 1, 2018

Primavera Santo Antônio São João Lago Azul Santa Rosa Conquista São Francisco São Luis Alvorada Iracema São Geraldo São José Santa Inácia Ipé Santa Maria São Pedro Santa Cruz São José (Alvorado) Total area (ha)

– 1,807.59 1,440.37 1,394.59 858.06 – 504.26 510.70 505.64 452.75 458.40 395.29 396.76 – 290.05 195.78 70.18 72.60 9,353.02

2,674.24 – – – 72.32 1,082.94 – 20.18 18.96 41.66 4.21 8.87 11.69 454.22 9.79 – – – 4,399.08

– – – – – – – – – – – – – – – – 2.65 – 2.65

4.79 12.15 – – 7.77 1.10 0.48 5.21 1.99 9.13 7.08 3.39 4.27 2.09 3.17 0.62 – – 63.24

– – – – 31.16 – – – – – – – – – – – – – 31.16

– – 20.02 – 94.92 – – – 13.39 – – – – – – 5.01 – – 133.34

209.52 267.75 352.39 154.21 220.79 164.15 107.62 70.49 44.80 47.36 72.07 130.09 105.26 16.99 24.12 37.59 22.14 – 2,047.34

Total area (ha)

Forrest

Tifton

Feedlots

Headquarters

Rubber

Farms

Pasture

Sugarcane

Table 1. Production activities areas (ha) for Campanelli farms.

2,888.55 2,087.50 1,812.78 1,548.80 1,285.02 1,248.19 612.35 606.58 584.78 550.90 541.76 537.64 517.98 473.30 327.13 239.00 94.96 72.60 16,029.82

Table 2. Purchase and selling of Agro Pastoril Campanelli’s livestock.1 Year

Cattle heads Cattle heads Total value Cost per cow Cattle heads Total in the year purchased of purchases (in Reais) sales per revenue per year (in Reais) year (in Reais)

Revenue per cow (in Reais)

2015 2014 2013 2012 2011 2010 2009 2008 2007

27,343 25,515 20,524 19,692 13,506 11,290 10,354 11,211 11,011

2,893 2,660 2,213 2,084 2,272 1,907 1,656 1,891 1,369

18,540 43,326 39,348 32,853 25,369 15,037 13,889 10,809 8,637

38,179,915 71,406,688 52,313,530 41,526,778 32,916,303 17,079,977 14,383,191 9,871,948 6,539,845

2,059 1,648 1,330 1,264 1,298 1,136 1,036 913 757

11,287 40,778 36,639 29,945 17,995 17,697 9,757 10,508 10,378

32,654,896 108,457,715 81,075,062 62,392,370 40,885,772 33,755,761 16,160,987 19,870,252 14,212,064

1 USD=3.17027 BRL, calculated on the basis of the exchange rate on October 18, 2017.

crop is used to feed the confined cattle, providing a great source of fiber, or is sold as biomass to sugarcane industries. The industry then burns the biomass in boilers to generate electricity. The straw is more efficient than the sugarcane bagasse normally used by the industry Sugarcane is a permanent crop and after some years it becomes unproductive and must be removed to make room to another crop. After a crop year the sugarcane can be replanted in the same area. Through constant research and testing, the company has increased the average age of their sugarcane crops before replanting International Food and Agribusiness Management Review

167

Scare et al.

Volume 21, Issue 1, 2018

to 9 years, compared to an average age of around 5 years for other sugarcane producers in Brazil. One of the biggest problems in sugarcane production, in Victor’s opinion, is soil compaction. Thus, Campanelli has invested in automatic machinery and methods to reduce the impacts of soil compaction. This has aided the company in increasing their yields per sugarcane crop. The crop areas where sugarcane must be removed, after its average life, are called ‘reform areas’. In the region of AgroPastoril Campanelli the common crop used by the sugarcane farms in their reform areas are peanuts. Then, these farms sell their peanuts production to industries, such as Santa Helena, a major candy producer in the region. However, AgroPastoril Campanelli decided that the best use for the land in its reform areas would be to produce corn, a major component for the feed to its cattle. Each year, approximately 3,000 of the hectares from unproductive sugarcane are used to produce corn, helping to reform the land before planting new sugarcane seedlings. The corn seeds are planted directly in the soil were the sugarcane straw was left, using the highest modern technology concerning planters and high precision agriculture instruments, such as the use of the Trimble FMX (FmX Integrated Display, Trimble, Sunnyvale, CA, USA), a GPS that has a sensor which controls seeds population and a sensor used to block the fertilizer application. The crop corn is also made with high quality, through harvesters with automatic pilots, generating more productivity gains. The amount of corn demanded to fatten cattle is quite large, around 30,000 metric tons. Around 21,000 metric tons of corn were cropped last year in the company’s farms to meet the cattle feed demand. The other 9,000 metric tons required by the confinement to fatten cattle are provided by industry ‘spinoffs’, such as cornstarch and Maltose Sugar. These choices have good nutritional value and are available in the region where the confinements are located with a reasonable price, making the market a good choice for fulfilling the remaining demand for cattle feed in the confinements. The state of São Paulo is a place where the feedlots have a higher cost of corn than those from Mato Grosso and Goiás. So, the production of corn to feed the cattle, rather than buying it, is a viable option for AgroPastoril Campanelli. However, the company monitors the price of the corn in the market, selling their corn when the price is high and buying corn when the price is low. All the corn is stored in silos in the AgroPastoril Campanelli farms. To maintain good standards of corn and sugarcane production, Victor Campanelli also worries about the quality of the soil in his farms. The soil in Brazil contains very few nutrients, making fertilizer a vital component of crop production. For Campanelli, the large amounts of cattle manure from their confinements created a solution for their fertilizer demand within their crop production. Additionally, most fertilizer in Brazil is imported from other countries, creating a relatively high price for farmers. Campanelli decided to invest in producing organic fertilizers using their cattle manure and produces around 50,000 metric tons of this organic fertilizer to use in the sugarcane fields. The production provides 70% of the total fertilizer needed for their sugarcane plantations. In all of the Campanelli farms, which cover approximately 15,000 hectares of land in total, the integration process is constant. The animals are slaughtered with an average weight of over 500 kg. The company is also one of the major moisture silage producers in the country, producing 17,000 metric tons per year. The AgroPastoril Campanelli’s integration process depicted in Figure 4 is one of the most successful business models of farm integration. A sustainable circle was created to guarantee more efficiency in economic, environmental and social terms. As Victor Campanelli says, ‘there is a perfect synergy between sugarcane, moisture, corn, and cattle. Not just because of the better land and labor use, but also because of the inputs economy and the productivity improvement, both in agriculture and livestock.’

International Food and Agribusiness Management Review

168

Scare et al.

Volume 21, Issue 1, 2018

Corn is used in the rotation

Storage of snaplage for cattle feed

Sugarcane straw

Use of sugarcane straw for cattle feed

Sugarcane fields

Use of fertilizer in sugarcane fields

Production of organic mineral fertilizer from cattle manure

Figure 4. AgroPastoril Campanelli integrated business model. 4.2 AgroPastoril Campanelli’s success Despite all the activities on AgroPastoril Campanelli’s farms existing under the same umbrella, the proceeds from one activity does not fund the other activities. Victor Campanelli ensures that each activity is individually competitive and profitable, otherwise it would be more beneficial for them to buy corn or fertilizer from the market rather than producing internally. Therefore, the corn produced by the company is ‘sold’ to the confinements in accordance with market prices, and the confinement ‘sells’ the fertilizer for the production of sugarcane and corn at the market price. This allows the Campanelli family to observe and evaluate the success of each activity separately, yet not neglecting the synergy between each activity. The successful integration model developed by the family has been recognized by multiple awards. One was the Nelson Pineda Award, received in 2012, which recognized AgroPastoril Campanelli as having the best cattle confinement practices in Brazil. One of the great advantages for the company is its performance in the futures market, specifically in the Bolsa de Valores, Mercadorias e Futuros de São Paulo (BM&F – São Paulo Stock Exchange), a major stock market exchange in Brazil. Cattle confinements are one of AgroPastoril Campanelli’s central activities. This high fixed cost business platform makes it essential for Campanelli to monetize their investments. The company uses BM&F to hedge the steer and hedge the inputs that are used in confinement. Another award was offered by Idea Group in 2013 for the best sugarcane plantation practices. It is becoming increasingly difficult in today’s economy to pursue profitability in agriculture and livestock, thus Victor seeks to use their integration model to maintain profitability and sustainable development. With the integration model, AgroPastoril Campanelli has reduced exposure to the risk of price fluctuations and holdup problems in obtaining necessary inputs. As mentioned previously, more than 60% of the energy needed for fattening their cattle and 70% of the necessary fertilizer are internally produced. However, it is important that the company continues to monitor the integration process and the transaction costs regarding each activity in order to ensure each activity is still contributing to the overall profitability and efficiency of the company. Table 3 shows AgroPastoril Campanelli income statement.

International Food and Agribusiness Management Review

169

Scare et al.

Volume 21, Issue 1, 2018

Table 3. AgroPastoril Campanelli income statement (2014-2015). Income statement (in thousand Reais)

2015

2014

Revenues from product sales Other Revenues Deductions, taxes and other expenses Net operating income Cost of goods sold Cattle Sugarcane and grain crop Operating expenses Depreciation Gross profit Operating revenues (expenses) General, administrative and commercial Financial expenses Financial receipts Interest on own capital Taxes Other revenues (expenses) Operating income (loss) before corporate holdings Net income (loss) before income taxes and social contributions Income tax and social contribution current Net income (loss)

195,405 97 -5,591 189,911 -174,106 -122,878 -36,647 -5,425 -9,156 15,805 -15,180 6,883 -3,748 -1,163 -4,120 -531 -1,060 625 625 -2,226 -1,601

151,945 53 -4,336 147,662 -133,306 -86,597 -27,188 -8,409 -11,112 14,356 -12,438 6,090 -2,468 -204 -3,500 -411 -174 1,918 1,918 -2,055 -138

While the company still has a profitable operation, changes within agriculture, including rising costs of inputs and down-stream consolidation, are providing challenges for AgroPastoril Campanelli. What actions can Victor take now to successfully face these challenges and improve internal operations?

5. Current context for AgroPastoril Campanelli Production today is no longer for amateurs and adventurers. In the Campanelli groupâ&#x20AC;&#x2122;s view, agriculture and livestock should be seen as a business. They understand it is essential to apply the best management concepts and practices to achieve consistent and lasting results. By constantly evaluating the agricultural markets, staying attuned to new developments and production technology, and seeking information from university research, AgroPastoril Campanelli has consistently stayed a couple years ahead of the market in their adoption of technology and production methods. This has allowed them to increase efficiency and profitability of their operation, maintain their integration model, and develop a trusted name among other producers in the industry. The company continually uses effective management procedures throughout all areas of the company. Victor identifies the company as having two primary pillars: technology and synergy. The goal of the family is to constantly improve their operations based on development of these two pillars. They certainly face some difficulties with their integration model and the extensive vertical integration, but overall it has provided to be crucial in their overall success. The integration model often allows them to hedge against some of the fluctuations in prices for commodities, such as with corn and fertilizer. AgroPastoril Campanelli has obtained a high level of efficiency in their production in the past; however, they have experienced a trend of increasing costs in recent years. As mentioned previously, the activities all operate independently from one another, so it is important to understand the current state of each production activity and how they are affecting each other. The Campanelliâ&#x20AC;&#x2122;s have not set up their integration model International Food and Agribusiness Management Review

170

Scare et al.

Volume 21, Issue 1, 2018

and then allowed it to go unchecked. They constantly evaluate the activities and how they are relating to the market. If market prices are more profitable for them, then they will discontinue production of that particular activity until it becomes more economical for them to begin again. Additionally, they ensure all activities advance together. It is of no benefit for them to focus primarily on advancing one area of production if they do not advance the other areas as well. 5.1 Current state of production activities AgroPastoril Campanelli strives to use the best managerial practices in all areas of productions. Within the confinements, feed purchases are made through simulations considering a variety of possible ingredients and suppliers. The choice is based on the goal of creating the greatest amount of economic efficiency possible. These simulations are made regularly to evaluate the alternatives. The company tries to avoid changing feed components throughout the year, but also tries to take all opportunities for advances in the market. Within in the cattle production industry, a major key to success is maintaining efficiency in feed distribution. AgroPastoril Campanelli transitioned their partnership for feed monitoring software to a new company three years ago because of problems they were experiencing in efficiency levels. All feed distribution is carried out and managed through automatic, computerized software that can be seen in Figures 5 and 6. The equipment allows the seed wagon to easily determine the exact volume that must be distributed in each picket. This

Figure 5. Ration control software and hardware.

Figure 6. Nutrition monitoring software. International Food and Agribusiness Management Review

171

Scare et al.

Volume 21, Issue 1, 2018

system also offers a strong managerial platform because it permits the managers to analyze how much feed was offered for each lot and if any problems exist with the ingredients. The system also opens the possibility of precisely identifying the cost, and thus the profitability, of each lot. The consolidation of slaughterhouses presents challenges for any producer. Within Sao Paulo, consolidation has occurred less than in other Brazilian states, leaving a greater number of small slaughterhouses. Because of the size of AgroPastoril Campanelli’s cattle production, however, they find it safer and more beneficial for them to deal with the large slaughterhouses and companies, even though it can be challenging. As mentioned previously, approximately 70% of Campanelli’s fertilizer demand is met through internal production of an organic fertilizer using cattle manure. This fertilizer does not contain enough nitrogen, so they purchase more nitrogen fertilizer from the market, and it is mixed with the organic fertilizer. Because of the high costs of fertilizer in Brazil, the production of organic fertilizer is always much cheaper than purchasing from the market. Victor notes, ‘The breakeven points are not even close.’ Precision agriculture and drones are used in fertilizer application, monitoring to make sure the proper amount of nutrients is applied throughout the crop. One specific device is called ‘Green Seeker’ (Greenseeker System, Trimble, Sunnyvale, CA, USA),which analyzes the sugarcane plant to evaluate its need for nitrogen fertilizer. A bio-digester was also tested in one of the farms in order to generate electric energy through cattle manure. However, considering the huge cattle manure production that is collected by the soil confinements, the bio-digester could not clean all the manure to keep it free from soil and other impurities. Additionally, the large fluctuation of energy prices made it risky and unprofitable for AgroPastoril Campanelli to operate the bio-digester. Thus, manure is currently only used to produce the organic fertilizer. Sugarcane production for the company is fully automatized and monitored closely to ensure the highest level of efficiency. Special tractors with wide-set wheels and run on auto-pilot are used to help alleviate soil compaction. They also use conductive electricity to identify the best locations to collect soil samples for testing. Regarding corn production, the corn seeds are planted directly in the soil were the sugarcane straw was left, using the highest modern technology concerning planters and high precision agriculture instruments. One tool they use is the Trimble FMX, a GPS containing a sensor which controls seed population, and a sensor used to block the fertilizer application. To prevent erosion and increase productivity AgroPastoril Campanelli also apply the no-till planting method in the production of the corn and sugarcane. No-till planting is a method of growing crops without disturbing the soil through tillage, as seen in Figures 7 and 8. This type of management increases the amount of water in the soil, increases the soil organic matter and also reduces the cost before planting the crops. Approximately 15

Figure 7. No-till planting for sugarcane. International Food and Agribusiness Management Review

172

Scare et al.

Volume 21, Issue 1, 2018

Figure 8. No-till planting for corn. metric tons of straw remain in the soil for planting sugarcane. This activity is important for the sustainability of the farmâ&#x20AC;&#x2122;s production model, considering the severe water restriction in the region, as seen in Figure 9. 5.2 Premix factory A recent investment and current focus for AgroPastoril Campanelli is the development of a factory to produce nutrient premix for their cattle. The group used to purchase a micro and macronutrient mix from the market to supplement their cattle feed. They realized companies in this industry were experiencing good profit margins and that AgroPastoril Campanelli possessed the facilities and capacity to produce their own mix of nutrients not only for their own production, but for other producers as well. They began construction of the plant in 2016. They are currently producing the premix now with the brand name of Tecnobeef, but are currently selling only to partners in order to test their product in the market. When they begin full production, only 10% will be used for their own cattle, the other 90% will be sold to the market. Their capacity for production is 100 metric tons/day. Profit margins for this activity are expected to be 10% (a 40% markup on cost). One of the largest resources in beginning production of this nutrient mix is the trusted and respected name of Campanelli in the marketplace. Within this market, the price farmers are willing to pay represents trust in the supplier and trust in the quality and the actual composition of the product being provided, particularly for custom products. Campanelli will provide two types of products: custom products,

Figure 9. Severe water restriction in the region of Agro Pastoril Campanelli farms. International Food and Agribusiness Management Review

173

Scare et al.

Volume 21, Issue 1, 2018

which will be ‘just-in-time’ inventory, and a basic product lines. Future expenses they will encounter include addition of salespeople and marketing to promote the product. The initial investment in the premix facility has been relatively low (approximately R$3 million1) compared to other investments, such as the energy facility (approximately R$75 million). 5.3 Challenges Right now, Campanelli has experienced big inflation in some costs and expenses. The primary increases in expenses have been for labor, energy, and external fertilizer purchases. Farms in Brazil are paying higher wages right now than other types of jobs, causing labor to become an increasing percent of Campanelli’s overall expenses. Energy prices have fluctuated greatly in the past few years, but right now have been increasing as a proportion of their costs, such as for diesel and electricity. Campanelli must purchase some of their fertilizer externally, and these prices have been increasing. Campanelli has also dealt with the overall downturn of the Brazilian economy in recent years. Campanelli states the integration helps alleviate the negative impacts of expense increases and price fluctuations in the market. Campanelli feels that a current challenge for the company is to continue to diversify the company, yet keep all current activities advancing. Additionally, the company has new stakeholders every year because of the size of the family. Incorporating the views of all the members and working together to manage the farm can create challenges.

6. What comes next for AgroPastoril Campanelli? The integration process on AgroPastoril Campanelli’s farms has proven to be a key component of Campanelli’s success. They continue to pursue innovative methods, adopt new technology, and seek for ways to improve efficiency and thus the profitability of their operation. With their most recent investments and ventures, including the production of energy (which has been shut-down for the current time), their use of precision agriculture technology, and the most recent decision to produce a premix of nutrients to add to cattle feed, Campanelli now wonders what opportunities await the operation to continue innovating and improve their current level of efficiency and productivity. Campanelli constantly evaluates the make-or-buy decisions of the integrated activities and acts accordingly. Most often, it is more efficient and profitable for them to continue internal production. Currently, Campanelli views three potential paths to pursue. AgroPastoril Campanelli can seek ways to further coordinate the supply chain network for beef or sugarcane production, look for a new method or technology to adopt in production, or they can consider expanding horizontally in the market for nutrient premixes. 6.1 Coordination in supply chain network Within any supply chain network, the producers involved in each stage of the value chain face the typical make-or-buy decision. The solution to this decision or dilemma often lies in the analysis of transaction costs and evaluation of the overall cost to buy versus the cost to make a material or component. If there is great potential for market failure in transactions, hold-up problems, and a large amount of opportunity cost or forgone profit potential in using the market, then it may be more beneficial for a company to vertically integrate or coordinate along the supply chain. Potential benefits of vertical integration include, but are not limited to: greater control over production quality, intellectual property and data; elimination of hold-up problems; and avoidance of information asymmetries. With the integration on Campanelli’s farm, the business currently plays a part in many steps of the value chain for both beef and sugarcane production. They decided that producing corn and using bales of sugarcane 1

1 USD=3.17027 BRL, calculated on the basis of the exchange rate on October 18, 2017.

International Food and Agribusiness Management Review

174

Scare et al.

Volume 21, Issue 1, 2018

straw for cattle feed, and producing their own micronutrient premix would cost less than using the market, and thus increase their profitability as an operation. With the high risk involved in owning feedlots and the low and fluctuating profit margins in selling cattle to slaughterhouses, feedlot owners must work to increase efficiency and productivity as much as possible to maximize profits in selling their cattle for slaughter. An issue that Campanelli must continue to evaluate is whether it is still profitable to continue producing the majority of their feed components, or if they would benefit more from switching back to the market. Campanelli is not considering any downstream integration at this time. Within the meat market it is very difficult to enter into the processing stage of the value chain. There are many decisions and opportunities to consider when processing meat, so thorough knowledge of the markets and sound decision-making are necessary skills that create natural barriers to entry for this stage of the value chain in meat production. They are beginning to move downstream in the supply of cattle feed with their production of a nutrient premix. Downstream integration is an option that can be evaluated in the future as Campanelli continues to develop their company. Issues they would need to consider if they wanted to integrate downstream include: level of competition in the specific market, size of the market, potential problems that could arise in the future due to market conditions (product price or cost of production), and the size of the profit margin. 6.2 New technology or production methods Next steps for both corn and sugarcane are biotech. The development of a GMO Corn is what allowed them to start planting corn in sugarcane reform areas. Biotech will help them to continue to increase production in a sustainable manner, preserving their land and soil, but increasing efficiency and profitability. Cattle is a little more difficult because Brazil has a lot of regulations regarding products than can be used on cattle, such as no products containing traces of DNA. There are nutrients that allow the cow to improve protein absorption, thus decreasing the amounts of protein that need to be used in the feed rations, but Campanelli must work around the regulations and find ways to increase their efficiency despite meeting regulatory requirements. Each year within cattle nutrition, though, there are new developments in products and methods. Also, Campanelli is evaluating a new way of applying nitrogen to sugarcane. All sugarcane producers apply the nitrogen to the soil. However, current research is showing that applying it to the leaf of the plant is more effective and can increase efficiency up to four times more than using it in the soil. Further research and development will have to be conducted in order to determine the best process and technology to use in order to use this method of applying nitrogen to their crops. 6.3 Horizontal expansion for nutrient mix Possible future options would be for them to enter the market of providing premix for other livestock and animals. The procedures and the facilities would be the same as the nutrient premix they supply for cattle. Campanellis states that it is actually easier to produce premix for smaller livestock, such as poultry. Their reasoning for entering the market with the company Tecnobeef and producing for livestock first is because of their trusted name in the market for beef production. Success within this market could lead to opening a path to provide nutrient premix for pet and other livestock feed. 6.4 Growth through expansion There is always opportunity for Campanelli to expand in the size of their production. Currently, Campanelli states the company is operating close to or at capacity of their facilities. In order to grow more in size of production, it would be necessary to invest in more mills, mill trucks, factories and land. Could this be a profitable option for Campanelli that would also open other opportunities for growth?

International Food and Agribusiness Management Review

175

Scare et al.

Volume 21, Issue 1, 2018

7. Future success for the Campanelli family Victor arrives at his office and proceeds inside to make a cup of coffee. As he sips his coffee, he looks to a portrait of his grandfather on the wall. He thinks of the current state of the agricultural industry and some of the challenges the company is currently facing. He feels affirmed in his position that AgroPastoril Campanelli needs to consider the available investment opportunities and keep moving forward. He knows they need to continue to improve upon internal operations to remain a profitable and successful supplier of cattle, sugarcane, and now their Tecnobeef nutrient mix. He contemplates the opportunities available for the business to pursue and realizes the necessity of a family meeting to discuss these opportunities and the future direction of the business. It is easy for companies to settle into a comfortable state when they reach a certain level of success, but Campanelli knows there are always to improve and to move forward. The company has a culture of innovation and constant improvement. No matter what the future may bring, Victor is confident that moving forward with these items in mind, and developing upon their pillars of synergy and technology, Campanelli will continue to remain at the top of their industry. He believes in the strength of his familyâ&#x20AC;&#x2122;s business and management capabilities to be successful and continue to improve their operations, remaining one of the most successful commercial farming operations in Brazil.

8. Questions for discussion 1. Many markets in agriculture are undergoing consolidation through large mergers and acquisitions. Does this market trend create any problems and/or opportunities for AgroPastoril Campanelli? 2. AgroPastoril Campanelli takes special care to continually monitor the industry and analyze how market prices compare to the price at which they can produce their own products. Are there any changes occurring in the industry currently that will make it more or less profitable to continue their integration model? Which products in particular? 3. The case discusses different opportunities for AgroPastoril Campanelli to pursue and direct their focus, including expanding their premix production to other animal markets, investing in new production methods or technology, vertically integrating in one of their production activities, or expanding the size of their production by investing in new facilities. Which of these opportunities is the most attractive for AgroPastoril Campanelli? Why? What factors need to be considered in making this decision?

9. Theoretical notes and teaching support This study is focused on issues involving a farm that achieved a strong level of integration on its activities and is deciding the next steps of the business considering the changes in the industry and economic landscape. The case can be used to discuss business growth strategies and value chain integration, using the Brazilian business environment as a background for the discussion. For this reason, this case is highly recommended for courses of strategy, marketing, agribusiness and farm management and can be used in both undergraduate and postgraduate courses. Also, this case can be used by farmers and in marketing and sales courses to capital goods industry or agricultural inputs industry. In courses directed to farmers, the students can put themselves in the place of the decision maker and discuss growth strategies, agro-industrial systems and farm management. The marketing and sales courses directed to capital goods and agricultural inputs industry should discuss key account management strategies, considering the challenges that AgroPastoril Campanelli faces, and business strategy having them as client.

International Food and Agribusiness Management Review

176

Scare et al.

Volume 21, Issue 1, 2018

The study is primarily focused on courses in the Brazilian students, however it can be also discussed in international courses. The case study session can follow the suggested approach: ■■ Brief explanation about AgroPastoril Campanelli and the Brazilian agribusiness. ■■ Questions to introduce the case discussion and challenges that the AgroPastoril Campanelli faces. ■■ AgroPastoril Campanelli history since the beginning and changes in its agricultural production through the years. ■■ Contextualization of the industry and business that the company is involved. This topic includes: – the beef industry; – the sugarcane industry; – the corn industry. ■■ The integrated business model of AgroPastoril Campanelli. ■■ Current context for AgroPastoril Campanelli. This topic includes: – the current state of production activities; – the premix factory; – the current business challenges. ■■ Alternatives for AgroPastoril Campanelli business. ■■ Main challenges of the case and conclusion.

References Associação Brasileira das Indústrias Exportadoras de Carnes (ABIEC). 2015. Brazilian beef exports. Available at: http://www.abiec.com.br/Exportacoes.aspx. D’Costa, V. 2016. IBISWorld Industry Report C1115-GL. Global sugar manufacturing. Available at: http:// tinyurl.com/yadgtod6. Gerber, P.J., A. Mottet, C.I. Opio, A. Falcucci and F. Teillard. 2015. Environmental impacts of beef production: review of challenges and perspectives of durability. Meat Science 109: 2-12. Knoll, S., C.S.S. Marques, J. Liu, F. Zhong, A.D. Padula and J.O.J. Barcellos. 2017. The Sino-Brazilian beef supply chain: mapping and risk detection. British Food Journal 119(1): 164-180. USDA Foreign Agricultural Service. 2017a. Grain: world markets and trade. Available at: http://tinyurl. com/yblk8ysm. USDA Foreign Agricultural Service. 2017b. Production, supply, and demand. Available at: http://tinyurl. com/zwdoyyu. USDA Economic Research Service. 2017c. Corn and other feed grains: trade. Available at: http://tinyurl. com/yblk8ysm.

International Food and Agribusiness Management Review

177