Volume 20 Issue 3 by IFAMA

International Food and Agribusiness Management Review

Official Journal of the International Food and Agribusiness Management Association

Volume 20 Issue 3 2017

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 Alessio Cavicchi, University of Macerata, Italy Hans De Steur, Ghent University, Belgium Klaus Frohberg, University of Bonn, Germany 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

Ajuruchukwu Obi, University of Fort Hare, South Africa Nick Vink, Stellenbosch University, South Africa 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 20 Issue 3, 2017

TABLE OF CONTENTS Research Articles 1.

Examining the strategy-performance link of Latin American businesses – a configurational approach Esteban R. Brenes, Luciano Ciravegna, Caleb A. Pichardo....................................................................................................................

External relationships and entrepreneurial orientation of tea manufacturing firms in Sri Lanka Aruni Wickramaratne, Akira Kiminami, Hironori Yagi.............................................................................................................................

p. 279

p. 293

Governance structures in smallholder pig value chains in Uganda: constraints and opportunities for upgrading Emily Ouma, Justus Ochieng, Michel Dione, Danilo Pezo.........................................................................................................

p. 307

CSR activities in the German poultry sector: differencing preference groups Henrike Luhmann and Ludwig Theuvsen .................................................................... p. 321

Pre-employment costs associated with H-2A agricultural workers and the effects of the ‘60-minute rule’ Fritz M. Roka, Skyler Simnitt, and Derek Farnsworth..

p. 335

Factors affecting cattle producers’ willingness to adopt an Escherichia coli O157:H7 vaccine: a probit analysis Brian J. Ochieng and Jill E. Hobbs....................

p. 347

6. 7.

Possibility of exporting halal-certificated food in Hokkaido, Japan: acceptance by Malaysian consumers Satoko Kubota, Hiroichi Kono, and Takuhiro Chiba ..........................................................................................................................

 2017 International Food and Agribusiness Management Association (IFAMA).

p. 365

Table of Contents

Volume 20 Issue 3, 2017

What do consumers think about farm animal welfare in modern agriculture? Attitudes and shopping behavior Heinke Heise and Ludwig Theuvsen ........................................................................................................................

Influence of product type and individuals’ perceptions on the geographic boundary for local products Madiha Zaffou, Alicia L. Rihn, Benjamin L. Campbell, Hayk Khachatryan, Omer Hoke ..................................................................................................

p. 379

p. 401

Case Study 10. Managing the pork supply chain through a cooperative: the case of Jinzhong Food Co. Ltd. Chen Ji, Fu Jia, and Jacques Trienekens .................................... p. 415

Industry Speaks 11. Management lessons learned in supply chain development: the experience of PICS bags in West and Central Africa Theodore Nouhoheflin, Jeanne Y. Coulibaly, Stephen D’Alessandro, Codjo C. Aitchédji, Maiyaki Damisa, Dieudonné Baributsa, and James Lowenberg-DeBoer ................ ............................................ ...................

p. 427

Table of Contents

Volume 20 Issue 3, 2017

iii

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2015.0163

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Received: 10 August 2015 / Accepted: 21 December 2016

Examining the strategy-performance link of Latin American businesses â&#x20AC;&#x201C; a configurational approach RESEARCH ARTICLE Esteban R. Brenes a, Luciano Ciravegnaa,b, and Caleb A. Pichardoa aProfessor,

Steve Aronson Chair of Strategy and Agribusiness, INCAE Business School, 2 km west of Procesa Nursery 1, La Garita, 960-4050, Alajuela, Costa Rica bAssociate

Professor, King's College, University of London, International Development Institute, Strand, WC2R 2LS London, United Kingdom

Abstract This study examines the sales performance of agribusinesses based in the humid tropics of Latin America. Its focus is on strategic positioning and specific functional areas, ranging from marketing to human resources and purchasing. This paper contributes to the literature through a novel methodological approach, a configurational angle, which allows for the identification of multiple causal explanations for consistently good sales performance. It also advances the research agenda on Latin American businesses by studying the strategic recipes of agribusinesses. The findings have both theoretical and managerial implications. Keywords: emerging markets, Latin America, performance, QCA, strategy JEL code: Q10 Corresponding author: esteban.brenes@incae.edu

ÂŠ 2017 Brenes et al.

279

Brenes et al.

Volume 20, Issue 3, 2017

1. Introduction

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Few topics have raised more debate in the field of management than the effects of strategy and strategy implementation on business performance in terms of sales and revenues (Grant, 2008). For example, whether a differentiation strategy has any effect on sales and profitability remains a heavily debated question and one with several potential answers (Campbell-Hunt, 2000; Dess and Davis, 1984; Kotha and Vadlamani, 1995; Pertusa et al., 2009; Robinson and Pearce, 1988). In addition to the debate over differentiation strategies, many other unaddressed questions remain regarding the effects of prioritizing certain strategic areas, such as marketing, human resources and supply chain management. Well-developed theoretical arguments justify investment in each functional area, often backed by empirical evidence. For example, scholars of operations management focus on the importance of the production process and product quality as determinants of performance, whereas scholars of marketing insist on the role of brands and advertising (Keller, 2009; Shah and Ward, 2003). Some of the most influential strategy scholars, such as Porter and Mintzberg, have used typologies or combinations of multiple, but not necessarily linear, causal factors that lead to competitive advantage (Mintzberg, 1979; Porter, 1985; Stonehouse and Snowdon, 2007). By doing so, they capture the complex nature of the relationships linking organizational structure, strategy and performance (Fiss, 2011). The problem is that the strategy-performance relationship has been studied mainly through linear causality methods, which do not allow for multiple paths leading to the same outcome and, thus, do not fit well with the configurational nature of strategy and strategic priorities (Öz, 2004). The diversity of theory-backed explanations for the strategic priorities of firms does not provide clear implications for practitioners and makes it difficult to test causality. This study contributes to the literature on the strategies of Latin American businesses by examining the strategic configurations of 51 firms through Qualitative Comparative Analysis (QCA), a method based on Boolean algebra, which allows for multiple causal explanations. It addresses the following research question: what strategic configurations lead to superior sales performance in the case of Latin American agribusinesses? Thus, this paper contributes to the study of strategy and performance through a different set of methodological lenses, building on the work of Fiss (2007, 2011) and responding to calls for methodological diversity in management research (Öz, 2004; Welch et al., 2010). Several authors point out that the environment in which a firm operates strongly affects the strategy-performance link (Ghemawat, 2005). The regulatory framework, growth and stability of markets circumscribe a firm’s set of strategic choices and determine which strategic configurations may lead to competitive advantage (Boehlje et al., 2005; Peng, 2002). Research on Latin American firms remains scarce, though there is some empirical evidence supporting the idea that diversification, vertical or horizontal integration, and flexibility are essential for firms to survive and perform well in such environments (Khanna and Palepu, 2013). This research study builds on the existing strategy literature by extending its analytical reach to firms based in Latin America, responding to calls for more evidence from this region (Nicholls-Nixon et al., 2011). It also focuses on agribusiness as opposed to industrial manufacturing, thus contributing to the body of knowledge by examining a rather under-studied sector, relevant to firms based in emerging and developed economies alike, as it is strongly linked to the food supply chain (Porter and Kramer, 2011).

2. Strategy and performance One of the most relevant questions in management and international business is whether and how strategy affects business performance (Acquaah and Yasai-Ardekani, 2008; Pertusa et al., 2009). Porter’s generic strategy argument continues to generate both criticism and support in the academic community (Chew, 2000; Jones and Butler, 1988; Porter, 1996; Reitsperger et al., 1993; Sashi and Stern, 1995; Stonehouse and Snowdon, 2007; Treacy and Wiersema, 1995). The issue may not be simply whether or not firms adopt a unique strategy, but how they try to accomplish that – for example, whether they position themselves International Food and Agribusiness Management Review

280

Brenes et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

as premium, commanding higher prices than competitors, or whether they opt for a low-cost strategy. Firms imitate successful competitors and their strategies, adapting them to their resources and capabilities (Deephouse, 1999). Superior performance may be linked to different strategic priorities as opposed to being the result of a single set of antecedents, and this link may help to explain the lack of clear empirical testing in this area (Fiss, 2007). The literatures on relational marketing and supply chain management point to the relevance of maintaining long-term, collaborative relationships with suppliers, as such relationships help firms control for quality and pursue innovation (Dyer and Singh, 1998; Grönroos, 1997). Scholars of strategy in emerging markets argue that diversifying activities, such as integrating vertically or horizontally, can help firms manage market failures and, hence, help achieve competitive advantage (Khanna and Palepu, 2013). Other scholars emphasize the importance of the final quality of the product as a driver of competitiveness (Dhamvithee et al., 2005; Shah and Ward, 2003). A more Schumpeterian, entrepreneurial perspective of firm strategy would maintain that innovation is at the core of competitive advantage (Dhamvithee et al., 2005; Zahra and Das, 1993). From a resource-based perspective, human capital is an important resource that may distinguish successful from unsuccessful firms; investing in the managerial team may contribute to a firm’s capacity to sustain its competitive advantage in entrepreneurial ways (Becker and Gerhart, 1996). Within the realm of firms that deploy a unique strategy, empirical evidence shows that environmental sustainability is linked to superior performance (Porter and Van der Linde, 1995). Another common differentiation strategy adopted by food producers is leveraging the country of origin of their products as an element of competitive advantage (Usunier, 2006). The literature covering the strategic priorities that may lead to performance is very extensive, precisely because it is split among different academic disciplines, such as marketing, human resource management, supply chain management, and strategy; thus, reviewing it is beyond the scope of this study. This section provides some examples to illustrate the theoretical and empirical foundations for the analysis in this study. Within the teaching and practitioners’ realm, the strategy-performance question becomes even more heated: disciples of marketing seek to illustrate that marketing is the main determinant of sales; strategists emphasize positioning and strategy execution as antecedents of performance; leadership gurus focus on the role of the executive team and the capacity to lead people; operation management experts look at internal processes; and scholars of the relational perspective illustrate the role of relationships with clients and suppliers. One of the fallacies of this debate is that it focuses on clear cause-effect relationships between, for example, emphasis on marketing and firm performance, ignoring the fact that in real life, companies may simultaneously deploy strategies that pursue multiple priorities, such as investing in marketing and improving relationships with suppliers (Fiss, 2011; Miller, 1996). Focusing on a narrow set of strategic priorities, or having one recipe for strategic success, may be consistent with the idea of pursuing a clear positioning in the market, but it can also lead to excessive emphasis on relevant areas at the expense of other, no less relevant areas (Miller, 1993). From the perspective of firms’ capabilities, companies often prioritize the capabilities deemed critical or strategic (Lumpkin and Dess, 1996). Prioritizing capabilities can simplify strategy execution and focusing on the core business, but it can also prevent firms from developing other capabilities and adapting to the market (Levinthal and March, 1993). For these reasons, firms tend to execute their strategy by emphasizing different sets of strategic priorities (Fiss, 2011). Firms combine strategic priorities to obtain a recipe, or configuration, which they deem suitable to achieving competitive advantage (Miller, 1996; Mintzberg, 1979). Different combinations of strategic priorities may lead equally to competitive advantage. These priorities are not necessarily linked to the outcome in a symmetric way, given that organizational features and strategic priorities form complex systems, where each variable is linked to the other and to the outcome through multiple causality links (Bowman and Ambrosini, 1997; Fiss, 2011). In other words, focusing on a given strategic priority, such as marketing, might lead to high International Food and Agribusiness Management Review

281

Brenes et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

performance only if combined with other strategic priorities. Perhaps having a high-price strategy results in superior performance only if coupled with other strategic priorities, such as investing in product quality and innovation. This possibility is captured by the principle of ‘equifinality,’ which allows for different causal paths leading to the same outcome – a principle that finds much empirical support in reality but does not fit well with the linear causation methods generally used to test strategy and international business theories (Fiss, 2007; Ragin, 2008). What if there were multiple strategic recipes to achieve superior performance? This study adopts a methodological approach that allows for multiple causal relationships – namely QCA – to identify the relationships between strategy and performance.

3. Methods and data This study examines strategy implementation by domestic and international Latin American value added agribusiness firms specializing in food production through different activities such as selection, cleaning, and/ or manufacturing process, which have higher capabilities and more- diverse strategic choices than commodity producers have (Garcia, 2005). The study is particularly relevant given the growing role of emerging markets as suppliers of food for the global market (Da Silva et al., 2009). This paper focuses on firms based in the humid tropics of Latin America, a region that extends from the south of Mexico to Northern Peru, so as to limit the effects that climate and geography have on agriculture (FAO, 2007) and because the region has become a leading global exporter of value added agricultural products, driven by innovative companies (Da Silva et al., 2009; Rosales and Kuwayama, 2012). By limiting our study to firms based in a specific region of Latin America, we also control for the effect of the external environment on strategy and performance – all firms are based in the same emerging market region (Hoskisson et al., 2000; Zahra et al., 2011). This empirical setting thus enables us to examine the strategy-performance link in an industry that, in spite of its global and local relevance, has yet to be thoroughly examined. Latin America’s humid tropics also differ strongly from the context of most strategy studies, which tend to be situated within developed economies (Khanna and Palepu, 2013). These differences entail that experimental and qualitative methods are more suitable than conventional methods, and particularly non-linear methods, such as QCA because they allow for multiple, non-symmetric, combinatorial explanations of the same phenomenon (Öz, 2004; Ragin, 2008). The framework presented here is the result of about 110 hours of open-ended conversations with the founders and senior executives of 17 agribusinesses from the same geographic region. The researchers visited their establishments several times to discuss both strategy formulation and strategy execution. This qualitative data collection, carried out as part of a broader study, helps to identify the strategic priorities of Latin American agribusinesses (Brenes et al., 2008, 2014). The lack of empirical evidence about strategy in the context of Latin American business and in the agribusiness sector makes this mixed method an appropriate one (Creswell, 2009; Nicholls-Nixon et al., 2011; Porter, 1985; Robson, 2011). From the information collected through qualitative interviews, the researchers created a list of strategies that Latin American agribusinesses consider important for achieving superior sales performance, including, among others, having higher-prices than competitors, investing in marketing skills and management team skills, and being innovative (Bowman and Ambrosini, 1997; Brenes et al., 2014). Building on this qualitative evidence and on Brenes et al. (2014), the researchers developed a survey questionnaire and began a second phase of data collection, which lasted from January 2013 to June 2014. The 250 targeted Latin American agribusinesses were first contacted via telephone and then given an online questionnaire. The researchers identified the target companies through Industrial Chambers, Commercial Guides and Agriculture Ministries and randomly contacted 250 of them. The response rate was 31.2%. This specific study focused on the 51 firms that achieved a minimum of 8% average annual sales growth for five consecutive years, which is well above the average gross domestic product annual growth for this region (Bárcena et al., 2014), while also keeping their annual operating profits constant or increasing. Firms that International Food and Agribusiness Management Review

282

Brenes et al.

Volume 20, Issue 3, 2017

met the 8% sales growth criterion but experienced a decline in operating profits were excluded from the sample. As inclusion criteria, the firms should not be part of a global multinational company, not listed on the stock market, and they had to sell their products within and beyond the domestic market. The focus was on these type of firms in order to limit the heterogeneity of agribusinesses. (Austin, 1992; FAO, 2007).

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Variables The questionnaires, which were corroborated with phone interviews when the answers were not clear, provided information about the strategic priorities in which the firms invested more resources (time, financial resources) as a means to achieve better performance. Fifty-one firms in this sample identified which among the strategic priorities they considered to be most important to their performance. They identified the following: geographic origin of the products as a marketing tool; proactive relationship with suppliers; emphasis on innovation capabilities; emphasis on sustainable environmental practices; good marketing skills; and good management team skills. Next, when asked to identify their areas of strength when compared to their closest competitors, the firms named three key areas: high-prices, quality of products, and vertical integration. The firms ranked strategic priorities and areas of strength versus closest competitors on a Likert Scale (1 to 7) which generated the data for the analysis. In the case of strategic priorities, the value of 1 means ‘not important at all’ and 7 represents ‘extremely important’ to achieve firms ‘success. For Areas of strategic strength, 1 means ‘much lower’ and 7 ‘much higher’ than closest competitors (Table 1). As a measure of performance we used the average annual sales growth for five consecutive years while also keeping their annual operating profits constant or increasing. The survey suggested the following options: 8, 13, 18, 23 and 25% average annual sales growth.

4. Fuzzy-set qualitative comparative analysis QCA examines the relationship between an outcome (sales performance in this case) and all possible combinations of causal conditions (strategic priorities and areas of strength in this case) (Rihoux and Ragin, 2009). The software Fuzzy-set qualitative comparative analysis (FsQCA) (University of Arizona, Tucson, AZ, USA) was used to examine the data. FsQCA is a program that uses multiple tools, such as Boolean algebra, combinatorial logic and fuzzy-set theory, to give different recipes of causal conditions that may be sufficient or necessary for a result to befall (Kent, 2008). A causal condition is considered as necessary if it must be Table 1. Abbreviations and variable meaning.1 Abbreviation

Meaning

focus_theorigin priceprod_r prodqual_r qualirelat_supp innovation_proc skillmanag_envi vertinteg_r marke_skills team_qual sales_growth

Geographic origin of the products as a marketing tool. Higher-prices relative to closest competitors. Products quality relative to closest competitors. Proactive relationship with suppliers. Emphasis on innovation capabilities. Emphasis on sustainable environmental practices. Vertical integration relative to closest competitors. Good marketing skills. Good management team skills. Average sales growth for past 5 years.

1 All

variables are on Likert scale, except ‘sales growth’. International Food and Agribusiness Management Review

283

Brenes et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

present for an outcome to be reached. On the other hand, a causal condition can be also defined as a sufficient if always conducts to a result of study by itself (Peréz, 2009; Ragin, 1987). FsQCA begins with a data matrix in which cases are either member or non-members of a category (Kent, 2008). In fuzzy-sets, the software records a value of 1 for full membership of a set, while zero means total non-membership. However, fuzzy-sets contain values that are not necessarily at the extremes of the continuum, and, for this reason, at least a third value is necessary to express a crossover point that defines the point of maximum ambiguity and a boundary for being in or out of a set, in other words, a third value is indicating a limit between necessary and sufficient conditions for the outcome to occur (Ragin, 2008). The FsQCA software distinguishes cases that are more in a set than others. With fuzzy-sets, a subset relationship exists when the membership scores in one set, such as a combination of causal conditions, are less than or equal to the membership scores in another set, such as the outcome. After the data matrix, causal conditions and one outcome may be selected to explain what set of conditions is necessary or sufficient for the outcome to occur. The truth table treats each single case as a configuration of the conditions selected (Elliott, 2013). FsQCA is based on the idea that combinations of causal conditions – not just one condition – are linked to the outcome. Several combinations of causal conditions may lead to the same outcome (Ragin, 2000; Verkuilen and Smithson, 2006). The paper models a relationship that is not necessarily symmetrical and reports conditions that are necessary but not always sufficient; and it uses measures of consistency and coverage to assess how well alternative causal recipes explain the outcome of interest (Kent, 2008; Schneider and Eggert, 2014). Both consistency and coverage are concepts that show a measure on every given model. Consistency shows the degree to which the cases share a given combination of conditions that is representative for the empirical evidence. Coverage shows the degree to which cases with the result of interest are explained by the final model (Peréz, 2009; Woodside and Zhang, 2011). This study treats consistency as higher than 0.80, while coverage must range between 0.2 and 0.6 (Ragin, 2006). For this paper, the outcome of interest is the firms’ sales performance in the last five years and different causal conditions that were grouped in two particular models defined by literature and confirmed through interviews. Whereby, we defined model one, which includes three causal conditions that are relative to closest competitors such as higher-price, better quality of their products and more vertical integrated; this model also includes four strategic priorities, these are proactive relationship with suppliers, innovation capabilities, marketing skills, and a good management team skills. We established a second model with a few differences from model one, first, it does not include vertical integration among causal conditions relative to closest competitors, and substituted two strategic priorities, better marketing and management skills, with two other, geographic origin of the products as a marketing tool and emphasis on sustainable environmental practices. These two models represent an effort of qualitative research to have different configurations in firms that consider the strategic priorities and areas of strength to reach a high performance that allow themselves to grow in a period of 5 consecutive years. It is allowed to see different recipes of causal conditions while variables may change for any reason the firms may consider, whence this reality is represented in two built scenarios. Calibration The first step to calibrating all the variables into fuzzy-set values is the selection of a fuzzy-set scale to get the fuzzy-set values, which are loaded into the FsQCA software. The first variable to be calibrated is sales growth (Table 2), where an increase of 20% represents full membership; the midpoint is 10%, and 5% is non-membership. On the other hand, causal conditions (Table 3 and Table 4) use the 1-7 Likert scale; in this

International Food and Agribusiness Management Review

284

Brenes et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Table 2. Calibration of outcome: average sales growth for the past 5 years. Values for outcome (sales growth)

Fuzzy-set values

0.25 0.23 0.18 0.13 0.08

0.99 0.98 0.92 0.71 0.23

Table 3. Calibration of variables: strategic priorities. Values for causal conditions1

Fuzzy-set values

1: not important at all 2: not very important 3: somehow important 4: does not matter 5: quite important 6: very important 7: extremely important

0.05 0.12 0.27 0.5 0.73 0.88 0.95

Factors that led to superior performance.

Table 4. Calibration of relative variables: areas of strength. Values for causal conditions1

Fuzzy-set values

1: much lower 2: lower 3: a little lower 4: same 5: a little higher 6: higher 7: much higher

0.05 0.12 0.27 0.5 0.73 0.88 0.95

Factors that led to superior performance relative to closest competitors’.

case, the membership score is given by three limits: 7 (full membership), 4 (midpoint) and 1 (no membership) (Basurto and Speer, 2012). Truth table for models 1 and 2 The truth tables show all logically possible combinations of causal conditions and each configuration’s empirical outcome (Ragin, 2008). The number of configurations is 2k (k represents all causal conditions considered for the outcome to occur). In this case, the value of 0 indicates that the condition is out of the set, and a value of 1 indicates that the condition is in the set (Table 5 and Table 6). Among 11 observations for model 1, all causal conditions are present. However, not all conditions are necessary for all cases: only proactive relationship with suppliers (qualirelat_supp) and better quality of products relative to the closest competitors’ (prodqual_r) are necessary conditions for the outcome to occur. In this case, there are no sufficient conditions because the membership values of the conditions are not lower than the membership of the outcome for all observations.

International Food and Agribusiness Management Review

285

Brenes et al.

Volume 20, Issue 3, 2017

priceprod_r

prodqual_!r

vertinteg_r

marke_skills

team_qual

qualirelat_supp

innovation_proc

number

sales_growth

1 1 1 0 1

1 1 1 1 1

0 1 1 1 1

1 0 0 1 1

0 0 1 1 1

1 1 1 1 1

0 1 1 1 1

1 1 1 2 11

1 1 1 1 1

raw consist. 0.974425 0.918418 0.892123 0.87686 0.849403

priceprod_r

prodqual_r

qualirelat_supp

innovation_proc

skillmanag_envi

number

sales_growth

Table 6. Truth table for model 2 with fuzzy-set values. focus_theorigin

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Table 5. Truth table for model 1 with fuzzy-set values.

1 1 1 0

0 1 0 1

1 1 1 1

0 1 1 1

1 13 2 2

1 1 1 1

raw consist. 0.984211 0.858652 0.854881 0.831451

Among 13 observations for model 2, all causal conditions are present. However, not all conditions are necessary for all cases: only better quality of products (prodqual_r), proactive relationship with suppliers (qualirelat_supp) and innovation capabilities (innovation_proc) are necessary conditions for the outcome to occur. In this case, there are no sufficient conditions because the membership values of conditions are not lower than the membership of the outcome for all observations. Findings of causal recipes The complex solution gives the most detailed solution since it makes no simplifying assumptions. It assumes that all configurations without cases (number=0) would have produced the lack of a result of interest (Peréz, 2009). There are three paths to see a causal condition in every configuration that makes an outcome occur: ‘YES’ means the condition is present; ‘NO’ means that the condition is absent; and ‘Indifferent’ means that the condition is impartial – that is, it is neither present nor absent.

5. Analysis Model 1 focuses on the strategic positioning of firms and the areas on which they focus. It examines: a highprice strategy; emphasis on the quality of the products; emphasis on the relationship with suppliers; emphasis on innovation capabilities; vertical integration; good skills of the management team; and good marketing skills. The results of model 1 show that three possible causal configurations lead to superior performance (Table 7). In each configuration, firms prioritize better product quality and proactive relationships with their suppliers. However, in the first and third cases, firms achieve superior performance through a highprice strategy, whereas in the second case, whether or not firms sell at higher prices than their competitors International Food and Agribusiness Management Review

286

Brenes et al.

Volume 20, Issue 3, 2017

Table 7. Configurations that led to superior sales good performance in model 1.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Configuration

Solution

High-prices relative to closest competitors Products quality relative to closest competitors’ Vertical integration relative to closest competitors’ Good marketing skills. Good management team skills Proactive relationship with suppliers Emphasis on Innovation capabilities. Consistency Raw coverage Unique coverage Overall solution consistency Overall solution coverage

First

Second

Third

YES YES YES NO Indifferent YES YES 0.863383 0.320933 0.028639 0.843006 0.669029

Indifferent YES YES YES YES YES YES 0.848399 0.617951 0.309123

YES YES NO YES NO YES NO 0.974425 0.224978 0.020372

is indifferent. Firms prioritize innovation capabilities in the first two configurations, but in the third, they achieve high performance through other strategic priorities. Firms with a high- price strategy emphasize either marketing or vertical integration. Firms for which a high-price strategy is indifferent to the outcome integrate vertically and simultaneously focus on marketing. The good management team skills is a strategic priority only in the second configuration, whereas it is indifferent for the first and not present for the third. Firms that may not have a high-price strategy prioritize product quality and collaborative relationships, like all other firms, and they also focus on good management skills, good marketing skills, innovation capabilities and vertical integration. However, firms that do have a high-prices strategy take different paths to achieve high performance. In configuration one, model 1 these firms focus on better product quality, innovation, relationships with suppliers and vertical integration. In configuration three, model 1, firms with a high-price strategy achieve high performance by focusing on product quality, marketing, and relationships with suppliers. Firms that do not necessarily implement a highprice strategy and are successful do prioritize the quality of their management team skills (configuration 2, model 1). Having a proactive relationship with suppliers and focusing on product quality are key priorities in all configurations of model 1. The configurations generated by model 1 are consistent with the arguments of scholars of strategy in emerging markets: they show that to achieve superior performance, firms have to prioritize multiple areas – in our case, product quality and relationships with suppliers – coupled with different configurations of others (Acquaah and Yasai-Ardekani, 2008; Pertusa et al., 2009). The results show that no firm achieves superior performance by prioritizing only one or two areas. Model 2 examines which combinations of the following strategic priorities lead to superior performance: high-price strategy; quality of the products, geographic origin of the products as marketing tool; emphasis on innovation capabilities; proactive relationships with suppliers; emphasis on sustainable environmental practices. Model 2 builds on model 1 by incorporating not only the critical areas identified in the pilot study and in previous publications (e.g. Brenes et al., 2014), but also two different and not necessarily exclusive ways of differentiating by positioning products as environmentally sustainable or with a strong emphasis on their geographic origin of their products. The analysis shows two possible configurations of these priorities that lead to the same outcome (Table 8). Only in the second configuration do firms command higher than average prices relative to closest competitors, which is considered here as a proxy for having a high-price strategy. Emphasis on product quality and innovation International Food and Agribusiness Management Review

287

Brenes et al.

Volume 20, Issue 3, 2017

Table 8. Configurations that led to superior sales good performance in model 2.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Configuration

Solution

High-prices relative to closest competitors Products quality relative to closest competitors’ Proactive relationship with suppliers. Emphasis on innovation capabilities Emphasis on sustainable environmental practices. Geographic origin of the products as a marketing tool Consistency Raw coverage Unique coverage Overall solution consistency Overall solution coverage

First

Second

NO YES YES YES Indifferent YES 0.859155 0.396221 0.041630 0.810003 0.688515

YES YES YES YES YES Indifferent 0.819678 0.646885 0.292294

capabilities also appears in both configurations. In both cases, the relationship with suppliers also proves to be causal antecedents for superior performance. The main difference between the two configurations of model 2 is that, in the first, firms emphasize the origin of their products, whereas in the second configuration, they emphasize their efforts in the field of environmental sustainability. Having a high-price strategy is key for firms that prioritize their environmental credentials, such as those of configuration 2, model 2. The results suggest that Latin American agribusinesses can achieve high performance by prioritizing the geographic origin or sustainability of their products. To do so, they focus on a better product quality and maintaining good-proactive relationships with the suppliers that provide inputs, as, presumably, the latter facilitate quality control across the value chain. They also prioritize innovation capabilities as part of the effort to introduce new products and processes to attract new consumers and maintain high quality standards.

6. Conclusions This study examines the link between strategy and performance through a configurational approach, building on previous work by Fiss (2007, 2011) and extending the research agenda to a different context – that of Latin American agribusinesses. The study shows that multiple configurations can lead firms to achieve superior performance, though they share many elements, such as focus on product quality and the relationship with suppliers. An important methodological contribution here is that using configurational method can enable uncovering multiple, parallel paths that firms use to be successful in different contexts, which conventional, linear regression analysis fails to reveal (Fiss, 2007; Öz, 2004). A high-price strategy tends to be linked to high-performing firms, supporting the continuing validity of the Porter´s generic strategy argument in different geographies and industries (Thornhjill and White, 2007). Nonetheless, in one of the three configurations of model 1, firms achieve the same outcome with or without a high-price strategy – again, proof of the fact that there may not be only one way to success and that strategy is best examined through configurational lenses (Fiss, 2011; Miller, 1993). All of the strategic configurations contain strategic priorities related to the final product, such as focusing on product quality, relationships with suppliers, marketing, and innovation. Relationship with suppliers, marketing, and innovation contribute to the quality and appeal of the final product. The first allows for better control over the inputs used; the second helps to communicate quality to consumers; and the last supports continuous improvement of product lines. The implications are that high-performing Latin American International Food and Agribusiness Management Review

288

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Brenes et al.

Volume 20, Issue 3, 2017

agribusinesses are highly focused on the final customer and the product. Model 1 also shows the importance of vertical integration. This is consistent with the argument that firms based in emerging markets integrate vertically to internalize market failures – for example the difficulty of finding certain inputs due to uncertainty in the regulatory and macroeconomic contexts, or the high costs of reaching suppliers due to infrastructural deficiencies (Khanna and Palepu, 2013). Our analysis also illustrates that firms may achieve high performance without emphasizing the quality of the managerial team, as long as they have a high-price strategy and focus on product quality, relationships with suppliers, innovation, and vertical integration. This finding shows that although investing in one’s management matters, it is not a necessary strategic priority to achieve superior performance through a high-price strategy. The effects of management skills on firm performance have been discussed by a large body of literature (Huselid, 1995). However, there continues to be scarce evidence that management training and emphasis on management skills have positive effects on the performance of small and medium enterprises, such as those examined hereby (Westhead and Storey, 1996). One of the explanations for our outcome is that management training has a higher opportunity cost for small and medium enterprises, as they have smaller management teams who often perform multiple functions (Marshall et al., 1995). Configuration 2 of model 1 shows that management skills are an antecedent of high performance for firms that do not necessarily have a differentiation strategy, yet focus simultaneously on innovation, product quality, vertical integration, relationship with suppliers, and marketing. Firms that are not vertically integrated and do not emphasize either innovation or marketing, on the other hand, achieve high performance without prioritizing management skills (configuration 3, model 1). With regard to the specific way in which firms project themselves to consumers, the companies in our sample – consistent with the literature on food producers – emphasize two strategic priorities: environmental sustainability and focus on the country of origin of products. These priorities were examined in model 2 (Table 8). Both configurations included either a focus on environmental sustainability or on the country of origin, suggesting that attempting to be different in multiple ways (e.g. country of origin and environment) may not be as effective as a simpler positioning. Interestingly, firms that prioritize environmental sustainability implement a high-price strategy, whereas firms emphasizing the geographic origin of their products do not. The implication could be that these firms attempt to be different by specifying where their products originate, but do not charge higher prices for this feature. The managerial implications of this study are that performance can be achieved through different strategic avenues in the agribusiness of Latin America, though it may be easier through a high-price strategy. That finding illustrates the importance of adopting strategic configurations that go beyond narrow definitions, instead incorporating elements of marketing, human resource management, supply chain management, and innovation. The finding also shows that there may be different ways to achieve a successful strategy, though they all involve strong emphasis on many other strategic priorities that support a company’s positioning, including focus on product quality, relationship with suppliers and innovation. The findings of this study illustrate that the complexity and multicausal nature of the link between strategy and performance is best examined through research methods that allow for equifinality (Ragin, 2008). This study suffers from the limitations of focusing on firms based only in the tropics of Latin America and operating in food production, which determined the sort of strategic priorities tested. Further research would be necessary to establish whether firms operating in other emerging markets and in other industries behave similarly. This study provides a first step towards a broader methodological approach towards the study of strategy and performance of Latin American firms.

International Food and Agribusiness Management Review

289

Brenes et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

References Acquaah, M. and M. Yasai-Ardekani. 2008. Does the implementation of a combination competitive strategy yield incremental performance benefits? A new perspective from a transition economy in Sub-Saharan Africa. Journal of Business Research 61: 346-354. Austin, J.E. 1992. Agroindustrial project analysis (2nd ed.). World Bank, Washington, WA, USA. Bárcena, A., A. Prado, J. Weller and R. Pérez. 2014. Estudio económico de América Latina y el Caribe. Available at: http://tinyurl.com/gs9bcut. Basurto, X. and J. Speer. 2012. Structuring the calibration of qualitative data as sets for qualitative comparative analysis. Field Methods 24: 155-174. Becker, B.E. and B. Gerhart. 1996. The impact of human resource management on organisational performance: progress and prospects. Academy of Management Journal 39: 779-801. Boehlje, M., A.W. Gray and J.D. Detre. 2005. Strategy development in a turbulent business climate: concepts and methods. International Food and Agribusiness Management Review 8: 21-40. Bowman, C. and V. Ambrosini. 1997. Perceptions of strategic priorities, consensus and firm performance. Journal of Management Studies 34: 241-258. Brenes, E.R., M. Mena and G. Molina. 2008. Key success factors for strategy implementation in Latin America. Journal of Business Research 61: 590-598. Brenes, E.R., D. Montoya and L. Ciravegna. 2014. Differentiation strategies in emerging markets: the case of Latin American agribusinesses. Journal of Business Research 67: 847-855. Campbell-Hunt, C. 2000. What have we learned about generic competitive strategy? A meta-analysis. Strategic Management Journal 21: 127-154. Chew, B. 2000. The geometry of competition. Internal Publication of the Monitor Company. New York, NY, USA. Available at: http://tinyurl.com/zldkw2t. Creswell, J.W. 2009. Research design: qualitative, quantitative and mixed approaches. Sage, Thousand Oaks, CA, USA. Da Silva, C., D. Baker, A, Sheperd, C. Jenane and S. Miranda-da-Cruz. 2009. Agro industries for development. Available at: http://tinyurl.com/jkscodj. Deephouse, D.L. 1999. To be different, or to be the same? It’s a question (and theory) of strategic balance. Strategic Management Journal 20: 147-166. Dess, G.G. and P.S. Davis. 1984. Porter’s (1980) generic strategies as determinants of strategic group membership and organizational performance. Academy of Management Journal 27: 467-488. Dhamvithee, P., S. Bhavani, A. Jangchud and P. Wuttijumnong. 2005. New product development in Thai agro-industry: explaining the rates of innovation and success in innovation. Agribusiness Management Review 8: 1-20. Dyer, J.H. and H. Singh. 1998. The relational view: cooperative strategy and sources of interorganizational competitive advantage. Academy of Management Review 23: 660-679. Elliott, T. 2013. Fuzzy set qualitative comparative analysis: part 2. Available at: http://tinyurl.com/jqslv64. Fiss, P.C. 2007. A set-theoretic approach to organizational configurations. Academy of Management Review 32: 1180-1198. Fiss, P.C. 2011. Building better causal theories: a fuzzy set approach to typologies in organization research. Academy of Management Journal 54: 393-420. Food and Agriculture Organization (FAO) 2007. Challenges of agribusiness and agro industries development. Available at: http://tinyurl.com/jrblt9j. Garcia, M. 2005. Agribusinesses in the new millennium. Comisión Económica para América Latina y el Caribe (CEPAL), Santiago, Chili. Ghemawat, P. 2005. Regional strategies for global leadership. Harvard Business Review 83: 98-108. Grant, R.M. 2008. Why strategy teaching should be theory based. Journal of Management Inquiry 17: 276-281. Grönroos, C. 1997. Value-driven relational marketing: from products to resources and competencies. Journal of Marketing Management 13: 407-419. Hoskisson, R.E., L. Eden, C.M. Lau and M. Wright. 2000. Strategy in emerging economies. Academy of Management Journal 43: 249-267. International Food and Agribusiness Management Review

290

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Brenes et al.

Volume 20, Issue 3, 2017

Huselid, M.A. 1995. The impact of human resource management practices on turnover, productivity, and corporate financial performance. Academy of Management Journal 38: 635-672. Jones, G.R. and J.E. Butler. 1988. Costs, revenue, and business level strategy. Academy of Management Review13: 202-213. Keller, K.L. 2009. Building strong brands in a modern marketing communications environment. Journal of Marketing Communications 15: 139-155. Kent, R. 2008. Using fsQCA: a brief guide and workshop for fuzzy-set qualitative comparative analysis. Available at: http://tinyurl.com/hr6yogc. Khanna, T. and K.G. Palepu. 2013. Winning in emerging markets: a road map for strategy and execution. Harvard Business Press, Boston, MA, USA. Kotha, S. and B.L. Vadlamani. 1995. Assessing generic strategies: an empirical investigation of two competing typologies in discrete manufacturing industries. Strategic Management Journal 16: 75-83. Levinthal, D.A. and J.G. March. 1993. The myopia of learning. Strategic Management Journal 14: 95-112. Lumpkin, G.T. and G.G. Dess. 1996. Clarifying the entrepreneurial orientation construction and linking it to performance. Academy of Management Review 21: 135-172. Marshall, J.N., N. Alderman, C. Wong and A. Thwaites. 1995. The impact of management training and development on small and medium-sized enterprises. International Small Business Journal 13: 73-90. Miller, D. 1993. The architecture of simplicity. Academy of Management Review 18: 116-138. Miller, D. 1996. Configurations revisited. Strategic Management Journal 17: 505-512. Mintzberg, H. 1979. The structuring of organizations: a synthesis of the research. Prentice-Hall, Englewood Cliffs, NJ, USA. Nicholls-Nixon, C.L., J.A. Davila, J. Sanchez and M.R. Pesquera. 2011. Latin America management research: review, synthesis, and extension. Journal of Management 37: 1178-1227. Öz, Ö. 2004. Using Boolean-and fuzzy-logic-based methods to analyze multiple case study evidence in management research. Journal of Management Inquiry 13: 166-179. Peng, M.W. 2002. Towards an institution-based view of business strategy. Asia Pacific Journal of Management 19: 251-267. Peréz, A. 2009. Instrucciones para utilizar fs/QCA (versión 2.0, 2007). Universidad de Salamanca. Available at: http://tinyurl.com/jkxzhq3. Pertusa, E.M., J.F. Molina and E. Claver. 2009. Competitive strategies and firm performance: a comparative analysis of pure, hybrid and ‘stuck-in-the-middle’ strategies in Spanish firms. British Journal of Management: 20: 508-523. Porter, M.E. 1985. Competitive advantage. Free Press, New York, NY, USA. Porter, M.E. 1996. What is strategy? Harvard Business School Press, Boston, MA, USA. Porter, M.E. and M.R. Kramer. 2011. Creating shared value. Harvard Business Review 89: 62-77. Porter, M.E. and C. Van der Linde. 1995. Toward a new conception of the environment-competitiveness relationship. The Journal of Economic Perspectives 9: 97-118. Ragin, C. 1987. The comparative method: moving beyond qualitative and quantitative methods. University of California, Berkeley, CA, USA. Ragin, C.C. 2000. Fuzzy-set social sciences. University of Chicago Press, Chicago, IL, USA. Ragin, C.C. 2006. Set relations in social research: evaluating their consistency and coverage. Political Analysis 14: 291-310. Ragin, C.C. 2008. Redesigning social inquiry fuzzy sets and beyond. University of Chicago Press, Chicago, IL, USA. Reitsperger, W.D., S.J. Daniel, S.B. Tallman and W.G. Chismar. 1993. Product quality and cost leadership: compatible strategies? Management International Review 33: 7-21. Rihoux, B. and C.C. Ragin. 2009. Configurational comparative methods: qualitative comparative analysis (QCA) and related techniques. Applied social research methods series, 51. Sage, Thousand Oaks, CA, USA. Robinson, R.B. and J.A. Pearce. 1988. Patterns of strategic behavior and their relationship to business- unit performance. Strategic Management Journal 9: 43-60. Robson, C. 2011. Real World Research. Wiley, Chichester, UK. International Food and Agribusiness Management Review

291

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0163 - Tuesday, May 09, 2017 8:27:06 PM - IP Address:24.21.169.207

Brenes et al.

Volume 20, Issue 3, 2017

Rosales, O. and M. Kuwayama. 2012. China y América Latina y el Caribe: Hacia una relación económica y comercial estratégica. Comisión Económica para América Latina y el Caribe (CEPAL), Santiago, Chili. Sashi, C.M. and L.W. Stern. 1995. Product differentiation and market performance in producer goods industries. Journal of Business Research 33: 115-127. Schneider, M. and A. Eggert. 2014. Embracing complex causality with the QCA method: an invitation. Journal of Business Market Management: 7: 312-328. Shah, R. and P.T. Ward. 2003. Lean manufacturing: context, practice bundles, and performance. Journal of Operations Management 21: 129-149. Stonehouse, G. and B. Snowdon. 2007. Competitive advantage revisited: Michael Porter on strategy and competitiveness. Journal of Management Inquiry 16: 256-273. Thornhill, S. and R. White. 2007. Strategic purity: a multi-industry evaluation timing and internationalization strategy. International Business Review 17: 473-487. Treacy, M. and F. Wiersema. 1995. The discipline of market leaders: choose your customers, narrow your focus, dominate your market. Addison-Wesley, Reading, MA, USA. Usunier, J.C. 2006. Relevance in business research: the case of country-of-origin research in marketing. European Management Review 3: 60-73. Verkuilen, J. and M. Smithson. 2006. Fuzzy set theory: applications in the social sciences (Vol. 147). Sage, Thousand Oaks, CA, USA. Welch, C., R. Piekkari, E. Plakoyiannaki and E. Paavilainen-Mäntymäki. 2010. Theorising from case studies: towards a pluralist future for international business research. Journal of International Business Studies 42: 740-762. Westhead, P. and D. Storey. 1996. Management training and small firm performance: why is the link so weak? International Small Business Journal 14: 13-24. Woodside, A. and M. Zhang. 2011. Identifying X-consumers using causal recipes: ‘whales’ and ‘jumbo shrimps’ casino gamblers. Journal of Gambling Studies 28: 13-26. Zahra, S.A., S.G. Abdelgawa and E.W. Tsang. 2011. Emerging multinationals venturing into developed economies: implications for learning, unlearning, and entrepreneurial capability. Journal of Management Inquiry 20: 323-330. Zahra, S.A. and S.R. Das. 1993. Innovation strategy and financial performance in manufacturing companies: an empirical study. Production and Operations Management 2: 15-37.

International Food and Agribusiness Management Review

292

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2015.0070

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

Received: 20 May 2015 / Accepted: 3 January 2017

External relationships and entrepreneurial orientation of tea manufacturing firms in Sri Lanka RESEARCH ARTICLE Aruni Wickramaratne a, Akira Kiminamib, and Hironori Yagic aLecturer,

Department of Agricultural Economics, Faculty of Agriculture, University of Ruhuna,Mapalana, Kamburupitiya, Matara, Sri Lanka

bProfessor

and cAssociate Professor, Department of Agricultural and Resource Economics, Graduate school of Agricultural and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo ku, Tokyo 113-8657, Japan

Abstract The purpose of this study is to examine the relation of external relationships and entrepreneurial infrastructure on the entrepreneurial orientation (EO) of tea manufacturing firms. An empirical survey was conducted with tea manufacturing firms located in a low grown area in Sri Lanka. Primary data were collected by administering a structured questionnaire at 109 tea factories. The findings indicated the importance of external relationships; specifically, relationships with supply chain partners and relationships with government facilitating institutions enhance the EO of tea manufacturing firms. However, relationships with other tea factories and educational and research institutions are not significant influences on firmsâ&#x20AC;&#x2122; EO in the context of the Sri Lankan tea industry. When considering the dimensions of EO, risk taking is influenced by external relationships and not innovativeness and proactiveness. Further, the entrepreneurial infrastructure provided by related institutions except other tea factories have positive links with EO. Such contributions are important to managers and policy makers to enhance the EO of tea manufacturing firms when facing competitiveness in the global market. Keywords: entrepreneurial infrastructure, entrepreneurial orientation, external relationships, Sri Lanka, tea manufacturing firms JEL code: M13 Corresponding author: aruni_dpw@yahoo.com

ÂŠ 2017 Wickramaratne et al.

293

Wickramaratne et al.

Volume 20, Issue 3, 2017

1. Introduction

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

Searching for new opportunities to increase firms’ competitiveness is necessary when facing fierce competition. The entrepreneurial literature argued that entrepreneurial firms display more innovative, risk taking, and proactive behavior than rival firms because of their entrepreneurial orientation (EO) (Covin and Slevin, 1991). This paper is to examine the role of external institutions related to improve the EO of firms and considers the case of the Sri Lankan tea industry. This study raises two guiding firm-level theories: the EO of the firm and the relational dimension of social capital theory. Entrepreneurial firms have been conceptualized as seizing the three main uniqueness characteristics of innovativeness, risk taking, and proactiveness (Covin and Slevin, 1991; Miller and Friesen, 1982). EO reflects the organizational process, methods, and styles that firms use to act entrepreneurially (Lumpkin and Dess, 1996). The relational dimension of social capital concerns the types of relationships that individuals have developed through a history of interactions. Further, relational capital takes into custody the level at which an entrepreneur actually obtains informational, physical, and emotional support in the business process (Liao and Welsch, 2005). Entrepreneurial firms which are able to establish strong relationships with external parties may develop competitive advantages over competitors that are unable or unwilling to develop such relationships (Dyer and Singh, 1998). Limited understanding exists about entrepreneurship in developing countries (Fairoz et al., 2010). Furthermore, studies on entrepreneurship in the Sri Lankan tea industry are rare and still in the growth stage. Hence, more in-depth studies emphasizing external relationships and entrepreneurship would assist in enhancing EO toward global competition in the tea industry. Empirical evidence exists on the independent effect of EO on performance (e.g. Zahra and Covin, 1995) and its contingent relationship with the external environment (e.g. Covin and Slevin, 1989). However, only few studies have examined the extent to which a firm’s embeddedness in inter-firm networks influences its EO (Simsek et al., 2003). Therefore, identifying how particular relationships enhance entrepreneurial behavior represents an important research agenda (Lee et al., 2001). According to Stam and Elfring (2008), studying the effect of different access to social capital on EO is worthwhile. This study seeks to fill this gap by examining the relationship of relational capital and EO dimensions of proactiveness, innovation, risk taking, and overall EO. This study investigated the degree of EO of tea manufacturing firms, the strength of their relationships with supply chain partners, other tea factories (OTF), government facilitating institutes, and educational and research institutes and the level of entrepreneurial infrastructure received from each institution. Therefore, specific objectives are: ■■ To examine the relationship between external relationships and EO and its dimensions as related to tea manufacturing firms. ■■ To understand the relationship between entrepreneurial infrastructure and the EO of tea manufacturing firms.

2. Theoretical background and hypotheses Sri Lankan tea industry The tea industry generates 60% of the country’s export agriculture revenue and is recognized as an important sector in Sri Lanka for reducing unemployment and poverty. Approximately 10% of the Sri Lankan population is employed directly and indirectly in the tea industry. The tea grown in Sri Lanka is classified into three different elevation zones: high grown (above 1,200 m), low grown (below 600 m), and mid grown (600-1,200 m). The principal production method in Sri Lankan tea factories is orthodox production, which accounts for more than 90% of the total production, whereas cut, tear, and curl (CTC) and green tea account for 7 and 1%, respectively. Additionally, instant tea and bio tea production accounts for less than 1%. More than 90% of the tea produced in Sri Lanka goes to the export market and more than 50% of tea exports are still in the International Food and Agribusiness Management Review

294

Wickramaratne et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

form of traditional bulk tea, with value-added tea exports accounting for approximately 40% of total tea exports. The Colombo Tea Auction is the main marketing channel use by tea manufacturers and accounted for approximately 95% of total tea sales. Private sales and direct sales are other marketing channels used by tea manufacturers. Sri Lanka is the one and only tea exporting country that firmly adheres to ISO 3720 standards (minimum product quality) for each kilogram of tea exported. Additionally, demand for the ‘Food Factory Concept’ and certification systems such as Hazard Analysis and Critical Control Point (HACCP), ISO 22000, and ISO 9001:2000, among others, is increasing (Sri Lanka Tea Board, 2012). To meet this demand, factory modernization or process automation is imperative. The following institutions are dedicated to improving the Sri Lankan tea industry: the Tea Research Institute (TRI), the Tea Small Holding Development Authority (TSHDA), the National Institute of Plantation Management (NIPM), and the Tea Board (TB). TRI is responsible for generating and transferring scientific knowledge and technologies appropriate for stakeholders to improve productivity and quality. Formal training programs are conducting by NIPM, adding to the accumulated knowledge available through work experience. TSHDA increases the productivity and quality of the tea smallholding sector by providing support services, thereby providing quality green leaf to manufacturers. As the apex authority of the tea industry, the TB is responsible for hygienic tea production and factory modernization. Compared with other tea growing countries around the world, some challenges that the Sri Lankan tea industry face today include the high cost of production, low field and factory productivity, and competition from other emerging producer countries that produce similar types of teas at much lower costs (Mohamed and Zoysa, 2006). Furthermore, a continual focus on orthodox and bulk tea could affect the country’s competitive position, as global consumption patterns increasingly turn toward more convenient types of this beverage. Accordingly, upgrading the EO of tea manufacturing firms is essential because doing so will reflect their innovativeness, proactiveness, and risk-taking qualities – predominant factors in mitigating the challenges arising in the competitive tea market. Entrepreneurial orientation Entrepreneurial firms strive to acquire competitive advantages by typically developing innovations and taking demanding risks (Miller and Friesen, 1982). The concept of entrepreneurship described at the organizational level is called EO (Covin and Slevin, 1991; Lumpkin and Dess, 1996). In line with prior research, EO is defined as the process, structure, and behavior of firms characterized by innovativeness, proactiveness, and risk taking (Covin and Slevin, 1989). Further, from the theoretical viewpoint, scholars have suggested that the dimensions of EO should be viewed as separate but allied constructs rather than as a single unifying characteristic (Lumpkin and Dess, 1996; Lyon et al., 2000). According to Naldi et al. (2007) firms can vary in their degrees of innovativeness, proactiveness, and risk taking, making them not equally entrepreneurial across all dimensions. Accordingly, this research employed both the one-dimensional and the multi-dimensional aspects of EO. The organizational imperative to introduce newness by adding value is described as an attribute of innovativeness. According to Lumpkin and Dess (1996), innovativeness reveals a tendency of a firm to actively seek new ideas, novelty, experimentation, and inspired solutions in pursuit of a competitive advantage. The concept of risk taking is well associated with entrepreneurship and can be described as the willingness of entrepreneurs to engage in calculated business-related risk (Brockhaus, 1980). Typically, EO firms display risk-taking behavior for which its risk-taking propensity can be inferred from its willingness to incur large resource commitments for uncertain businesses (Lee et al., 2001; Lumpkin and Dess, 1996) such as developing design teas. The tendency to take the initiative to compete aggressively with other firms is called proactiveness (Covin and Slevin, 1989). Proactiveness is an important organizational process because it entails a forward-looking perspective of the firm. Being a pioneer by anticipating and pursuing new opportunities and participating in emerging markets is a property of entrepreneurship (Lee et al., 2001). Therefore, EO may contribute to stronger performance by assisting a firm’s capability in identifying International Food and Agribusiness Management Review

295

Wickramaratne et al.

Volume 20, Issue 3, 2017

innovative opportunities with potentially large returns and target market segments, and by obtaining firstmover advantages (Lumpkin and Dess, 1996).

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

External relationships and entrepreneurial orientation External relationships perform a vital role in identifying entrepreneurial opportunities. This identification encompasses relationships with various entities, such as customers, suppliers, competitors, or research institutions. Kale et al. (2000) used the notion of relational capital to express the quality of a network. Relationships obviously matter to entrepreneurs; however, to identify how they function requires an understanding of social capital (Cope et al., 2007). Nahapiet and Ghoshal (1998: 244) viewed social capital as ‘the sum of the actual and potential resources embedded within, available through, and derived from the network of relationships possessed by an individual or social units.’ Further, they proposed that social capital is comprised of three dimensions: the structural dimension, the relational dimension, and the cognitive dimension. This study refers to the relational dimension of social capital. The nature of the relationships that develop between parties is referred to as the relational dimension. This dimension is manifested in strong versus weak relationships (Nahapiet and Ghoshal, 1998). Further, Tsai and Ghoshal (1998) argued that the relational dimension refers to firms’ direct relationships with others and the assets rooted in these relationships, such as trust. In addition to trust, the strength of a relationship is an important element within the relational dimension (Nahapiet and Ghoshal, 1998), and is address in this study. The strength of a relationship is reflected in the combination of characteristics, such as frequency, emotional intensity, intimacy, and reciprocal services (Granovetter, 1985). Cooke and Wills (1999) argued that relational capital facilitates the tacit exchange of information. Further, Lechner and Dowling (2003) emphasized that increasing relational capital can greatly enhance an enterprise’s opportunities. According to Gumusluoğlu and Ilsev (2009), firms develop a wide range of relationships with different parties, such as cooperating with universities and research institutions for technical assistance and consulting and with public and private organizations to receive financial and technical assistance for their innovative projects. Further, collaboration with universities and research institutes provides a means to develop technical knowledge (Santoro and Gopalakrishnan, 2000). According to Zimmerman and Zeitz (2002), external relationships facilitate access to valuable resources (e.g. information, knowledge, physical and emotional resources) that support business growth and survival. In this study, relationships with external institutions are categorized as those with supply chain partners, OTF, government facilitating institutes, and educational and research institutes. On the basis of previous research, this study proposes receiving either knowledge or resource-based support from external institutions to enable interaction with other EO factors. Entrepreneurial infrastructure This study also focuses on the issue of entrepreneurial infrastructure for the entrepreneurial process. Covin and Slevin (1991) confirmed that entrepreneurial activities need to utilize large quantities of resources. Therefore, resources obtain from outside organizations are defined as the entrepreneurial infrastructure. Furthermore, Gumusluoğlu and Ilsev (2009) argued that availability of slack resources improve firm’s innovative performance. According to Suzuki et al. (2002), firms seek better and more professional services and financial and institutional support. Therefore, the infrastructure needs to be constructed to strengthen the technical and business expertise to help firms take advantage of continually emerging venture opportunities. Further, government and policy makers need to remove barriers that prevent firms from acquiring the needed management resources by creating or facilitating public and private institutions that support the business.

International Food and Agribusiness Management Review

296

Wickramaratne et al.

Volume 20, Issue 3, 2017

Research hypotheses The following research model (Figure 1) is proposed on the basis of the aforementioned theoretical background and that considers the objective of the research. This study attempted to examine the role of external institutions in improving EO of tea manufacturing firms in Sri Lanka. Thus, the following hypotheses are proposed:

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

H1: Strong external relationships may enhance EO and its dimensions of tea manufacturing firms. H2: An entrepreneurial infrastructure has a positive influence on the EO of tea manufacturing firms. Figure l depicts the association between external relationships and EO. The entrepreneurial infrastructure, such as financial support, information sharing, education and training, innovation development, consultation, research and development support, and networking facilities, and except for suppliers, are provided by external institutions.

3. Materials and methods Sample and data collection In line with the research objectives, a tea manufacturing firm is considered to be the unit of analysis. The target population of the study was tea manufacturing firms located at low elevations with the highest contribution (60% in 2012) to total tea production of Sri Lanka, in contrast to high and middle elevations. In 2012, 425 tea factories were registered with the Sri Lankan Tea Board. First, this study selected four main districts â&#x20AC;&#x201C; Rathnapura, Galle, Matara, and Kalutara â&#x20AC;&#x201C; that make the highest contribution of low elevation production. A sample of 109 tea manufacturing firms were proportionately taken from each district depending on the number of tea manufacturing firms in each district and on the willingness of the owner or manager of the firm interviewed. Empirical data were obtained through purposive sampling, which allowed the sample to fulfill two criteria. First, the firm should be a private sector tea manufacturing firm. Since based on their management structure, Sri Lanka Tea Board has categorized tea manufacturing firms as private sector, plantation companies and state sector. Second, the firm should have been established for more than five years. To analyze above

Organization and managers characteristics (control variables) External relationships Supply chain partners (brokers and suppliers)

Entrepreneurial orientation

Other tea factories

Innovativeness

Government facilitating institutions (TSHDA and TB)

Proactiveness Risk taking

Education and research institutions (NIPM, TRI and University)

Figure 1. Model of associations between external relationships and the entrepreneurial orientation of the firm. TSHDA = Tea Small Holding Development Authority; TB = Tea Board; NIPM = National Institute of Plantation Management; TRI = Tea Research Institute. International Food and Agribusiness Management Review

297

Wickramaratne et al.

Volume 20, Issue 3, 2017

proposed relationships it is important to obtain data from well established firms. The respondents involved in this research are comprised of 109 owners or managers who had good awareness about past and present organizational practices.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

Instrument of measurement Primary data were collected during a survey using a pre-tested self-administered structured questionnaire (Supplementary Methods S1). Before data collection, we conducted a pilot study in which the owner/ managers of four tea factories verified the understandability and content validity of the survey instrument. This study defined external relationships through four variables, including relationships with supply chain partners, other tea factories, government facilitating institutions, and educational and research institutions. Green leaf suppliers and tea broker companies were categorized as supply chain partners. Government facilitating institutions have relationships with TSHDA and TB. The variable relationships with educational and research institutions consist of relationships with TRI, universities, and NIPM. Each relationship was operationalized as a compound of three measures: intensity, frequency, and reciprocal services (Granovetter, 1985). To measure the extent of the intensity, frequency, and reciprocal services of each institution, a Likert scale was used, where 1=no extent and 5=very high extent. This study adopted the three dimensions of EO – innovativeness, proactiveness, and risk taking – to measure the EO that numerous studies documented as a high level of reliability and validity (e.g. Knight, 1997; Kreiser et al., 2002). EO was measured using nine items, which were developed and tested for reliability by Covin and Slevin (1989) using a five-point Likert scale ranging from 1=strongly disagree to 5=strongly agree. This study used three items to measure innovativeness, three items to evaluate proactiveness, and three items to measure risk taking. To measure the level of entrepreneurial infrastructure received from each institutional category, the infrastructure used by Suzuki et al. (2002) was adapted, including financial support, information sharing, education and training, innovation development, research and development, consultation, and networking facilities. A Likert scale, where 1=very little extent and 5=very high extent, was used to measure the level of benefits received from each institutional category. Data analysis Regression analysis was used to test the hypothesized relationship in the research model. This analysis allows for an assessment of how well the dependent variable can be explained by the values of the independent variables. Hypothesis H1 was tested using hierarchical regression analysis, which allows for an examination of the set of independent variables and the dependent variable after controlling for the effects of other independent variables on the dependent variable. In stage 1, the control variables were entered as predictors of EO. Then, the main effect predictor variables of relationships (supply chain partners, OTF, government facilitating institutions, and education and research institutions) were entered. Before the proposed hypothesis was tested, regression analysis was performed to identify the control variables. Initially, organization and owners’/managers’ characteristics, including size, availability of green leaf, stage of the firm, having other businesses, owners’/managers’ education level, present age, training, and previous business experience were used as predictors of EO. Among them, the variables that significantly influenced EO and its dimensions were selected as controls in further analysis. Thereby, size, availability of green leaf, and owners’/managers’ present age were used as control variables to test hypothesis H1. Further, correlation analysis was used to test hypothesis H2 and, thereby, determine the relationship between entrepreneurial infrastructure and EO.

International Food and Agribusiness Management Review

298

Wickramaratne et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

As previously declared, the scales used to measure all of the variables discussed in this study appear in the literature and were used in several studies. Therefore, the validity and reliability of the scales is not a main issue in this study. Before proceeding to the main hypotheses, this study assess whether the questionnaire yielded reliable results across populations. To assess the reliability of the scale items, this study used Cronbach’s alpha, which is a widely used measure of internal consistency. Accordingly, the Cronbach’s alpha coefficients for the variables were higher than or approaching the recommended level in Nunnally (1978) and indicated better internal consistency.

4. Results and discussion Profile of tea manufacturing firms This section describes certain characteristics of tea manufacturing firms used in this research as shown in Table 1. Among them, 62% of the firms that engaged only in manufacturing bought green leaf, whereas the others use both own and bought green leaf. In compliance with the national situation, the tea produced in these manufacturing firms is principally the orthodox type, and only 10% of firms also produce CTC tea. Table 1. Characteristics of tea manufacturing firms. Characteristics

Frequency

Category of tea manufacturing firm Both own leaf and bought leaf Bought leaf only Method of tea manufacturing Only orthodox Both orthodox and CTC1 Marketing channels Auction Direct export Direct sales through sales outlet (local) Through resellers (local) Quality certificates2 HACCP3 CQC and SLS4 ISO 22000 ISO 9001 Other Total number of employees (size) 30-149 (medium) ≥150 (large) Types of products Black tea in bulk form Special tea grades Unique tea grades Flavored teas Green tea Tea packets

Percentage

41 68

37.61 62.39

98 11

89.91 10.09

109 5 7 9

100.00 4.59 6.42 8.26

21 38 18 8 15

19.27 34.86 16.51 7.34 13.76

76 33

69.72 30.27

109 25 2 4 2 17

100.00 22.94 1.83 3.67 1.83 15.60

CTC = cut, tear and curl. Some factories have more than one certificate. 3 HACCP = Hazard Analysis and Critical Control Point. 4 Ceylon Quality Certificates and Sri Lanka Standards. 2

International Food and Agribusiness Management Review

299

Wickramaratne et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

Their main product is black tea in bulk, whereas 15.6 and 2% of factories produce tea packets and green tea, respectively, as other products. Similar to the national situation, all tea manufacturing firms utilize the tea auction as their main marketing channel. In addition, 5% engage in direct export and 6 and 8% of firms utilize direct sales and resellers, respectively, as the mode of disposal in the local market. Sri Lanka is proud of its quality tea, and 43% of tea manufacturing firms have achieved some type of quality standard. Moreover, 19 and 17% of manufacturing firms have HACCP or ISO 22000 food safety certifications, respectively. In addition, 15% of firms are in the process of acquiring quality standards. With respect to ownership, 43% of firms are sole proprietorships, 21% are partnerships, and 36% are limited liability companies. In terms of firm size, 70% of tea manufacturing firms are medium size in accordance with the definition of Small and medium-sized enterprises by the National Development Bank of Sri Lanka, indicating that they have between 30 and 149 employees. The figures imply that all of these tea manufacturing firms produce black tea in bulk form as their main product. Of the sampled firms, 50% have upgraded the quality of their tea grades and nearly 23% produce special tea grades such as silver tips, golden tips, special, and extra special tea grades. Two factories produce new tea grades – Jayachakra and Sun Pekoe – which are unique to them. A small minority of the firms produce tea packets, flavored tea, and green tea. External relationships and entrepreneurial orientation This section describes the empirical evidence on external relationships and EO, and the dimensions of EO of tea manufacturing firms in Sri Lanka. Table 2 provides the descriptive statistics of the variables used in the analysis. A two-stage hierarchical regression analysis was used to test hypothesis H1. In stage 1, the control variables (size, availability of green leaf, and owner/managers present age) were entered as predictors of EO. Next, the main effect predictor variables were entered and results were presented in Table 3. As model 1 illustrates, the coefficient for the relationship between supply chain partners’ relation and EO is positive and significant. Similarly, the coefficient for the relationship between the relation with government facilitating institutions and EO is positive and significant. These findings indicate that relationships with supply chain partners and government facilitating institutions have a significant influence on the EO of Table 2. Descriptive statistics of variables used for analysis.1 Variable

Mean

EO2 32.94 Innovativeness 9.27 Proactiveness 12.14 Risk taking 11.54 Relationship with supply chain partners 26.74 Relationship with OTF3 6.49 Relationship with government facilitating institutions 17.19 Relationship with education and research institutions 18.92 Size (no. of employees) 124.35 Availability of GL4 258.19 Present age of owner/manager 46.66

Standard deviation

Min

Max

Cronbach’s No. of alpha value items

6.890 2.990 2.481 2.949 2.675 1.956 3.560 4.186 96.772 14.464 11.334

16 3 6 3 20 3 7 9 40 238 25

45 15 15 15 30 12 27 31 500 277 85

0.833 0.720 0.783 0.682 0.780 0.752 0.806 0.834 n.a. n.a. n.a.

n.a. = not applicable. EO = entrepreneurial orientation. 3 OTF = other tea factories. 4 GL = Grean Leaf, this is the extent of tea cultivation in district/number of tea factories in each district. 2

International Food and Agribusiness Management Review

300

9 3 3 3 6 3 6 9 n.a. n.a. n.a.

Wickramaratne et al.

Volume 20, Issue 3, 2017

Table 3. Results of the analysis of external relationships and entrepreneurial orientation (EO).1,2

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

EO Model 1 Control variable Size (no. of employees) Availability of green leaf Owner/manager present age External relationships Supply chain partners OTF3 Government facilitating institutions Educational and research institutions F-Statistic R2 Adjusted R2 Change in R2a

Dimensions of EO Innovativeness Model 2

Proactiveness Model 3

0.048 (0.499) 0.134 (1.379) -0.038 (-0.369) 0.314*** (3.599) 0.317*** (3.570) 0.205** (2.177) -0.177** (-2.135) -0.174** (-2.055) -0.125 (-1.396) 0.265*** (3.220) -0.110 (-1.244) 0.379*** (3.681) 0.032 (0.282) 7.781*** 0.350 0.305 0.177

Risk taking Model 4 -0.008 (-0.073) 0.240** (2.564) -0.133 (-1.492)

0.163* (1.939) 0.209** (2.349) 0.280*** (3.159) 0.046 (0.507) -0.014 (-0.148) -0.291*** (-3.068) *** *** 0.283 (2.701) 0.347 (3.126) 0.306*** (2.774) 0.038 (0.330) 0.084 (0.669) -0.039 (-0.316) 7.033*** 4.678*** 4.833*** 0.328 0.245 0.251 0.281 0.193 0.199 0.095 0.161 0.163

Standardized coefficients are displayed in the table and t-values are in parenthesis. Significance level: *P<0.1; **P<0.05; ***P<0.01. 3 OTF = other tea factories. a Difference in adjusted R2 with and without external relationships in the model. 2

the firm. Further, the results of model 1 illustrate that relationships with OTF and educational and research institutions are not tend to influence on EO of tea manufacturing firms. Normally, firms tend to build relationships with external parties for knowledge acquisition and resource acquisition. According to Brüderl and Preisendörfer (1998), social networks are supposed to be vital because they open up entrepreneurial possibilities by providing useful and reliable information. Therefore, by acquiring relational capital, firms tend to receive informational, physical, and emotional support for the business process. Although networks may facilitate the performance of entrepreneurial firms, not all relations do so equally (Peng and Luo, 2000), as revealed by the findings of this study. This section would describe the relation of external relationships and EO of the tea manufacturing firms. As per model 1, the relationship with supply chain partners (coefficient=0.265, P<0.01) is positively and significantly related to the EO of tea manufacturing firms. Further, the results of models 2, 3, and 4 imply that the relationship with supply chain partners (the coefficients are 0.163, P<0.1; 0.209, P<0.05; and 0.280, P<0.01) is positively and significantly related to the firm’s innovativeness, proactiveness, and risk taking. Tea manufacturing firms have strong relationships with supply chain partners because they are tea brokers and green leaf suppliers. Tea manufacturers use the auction as their main marketing channel through tea brokers. Therefore, they are able to obtain detailed information from tea brokers’ companies through tea valuation reports, weekly and monthly average tea prices, and buyers’ special requirements. Having this information enhances the competitive behavior directed toward rival firms. This concept is in line with Hitt et al. (2001), who noted that information received from external relationships helps identify potential entrepreneurial opportunities. Further, tea broker companies have the propensity to provide financial support whenever tea manufacturing firms undertake risky decisions, such as committing resources to ventures with uncertain outcomes or borrowing heavily, because doing so tends to increase the risk-taking behavior of the firm. Additionally, tea brokers facilitate the discovery of opportunities related to the tea industry and, thereby, enabling tea manufacturing firms to enhance their EO. Another supply chain partner concern in this study is green leaf suppliers who are primarily small tea holder farmers that provide green leaf, the main raw material in the tea manufacturing process. To ensure a continuous supply of good quality green leaf, tea manufacturing firms tend to develop their suppliers by providing financial support, fertilizer, transport International Food and Agribusiness Management Review

301

Wickramaratne et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

facilities, technical expertise, and welfare facilities, ultimately enhancing the EO of manufacturing firms. As revealed by the results indicated in Table 3, the availability of green leaf is positively and significantly related to EO and its dimensions of tea factories. Dyer and Singh (1998) argued that the ability to select the correct partners and maintain relationships with them is required to gain a competitive advantage through inter-firm relationships. Therefore, maintaining better relationships with supply chain partners is important because doing so tends to enhance the EO and its dimensions of firms. As illustrated in model 4, the relationship with OTF (coefficients is -0.291, P<0.01) is negatively and significantly related to the risk taking of tea manufacturing firms. Additionally, as revealed by models 1, 2, and 3 (respectively, the coefficients are -0.110, P≥0.1; 0.046, P≥0.1; and -0.014, P≥0.1), the relationship with OTF does not significantly relate to the EO and dimensions of innovativeness and proactiveness. Results indicating that the relation with OTF not leads to risk-taking behavior of tea manufacturing firms. Because OTF are competitors, information that indirectly reaches them through such relationships may mislead and be distorted (Ingram and Roberts, 2000). In this context, tea manufacturing firms devote their resources to novel business such as producing designer teas individually. Furthermore, when considering the benefits received from OTF, Table 4 reveals that no benefit was significantly correlated with EO. Additionally, the findings of Nieto and Santamaria (2007) revealed that collaboration with a competitor negatively impacts the novelty of innovation. When considering the relationship with government facilitating institutions, the direct actions or inactions of governments influence the level of firm’s entrepreneurship, particularly as a result of the government’s regulations and policies. Government facilitating institutions related to the tea industry in Sri Lanka execute several programs that were geared to its development, such as financial support for tea factory modernization and process improvement, enhancement of the quality of manufactured tea and green leaf, and others. As evident from models 1 to 4, the relationship with government facilitating institutions (the coefficients are, respectively, 0.379, P<0.01; 0.283, P<0.01; 0.347, P<0.01; and 0.306, P<0.01) is positively and significantly related to EO and it dimensions of tea manufacturing firms’ innovativeness, proactiveness, and risk taking. Therefore, the relationship with government facilitating institutions has a significant influence on EO and its dimensions. This result is in line with the findings of Doloreux (2004), who stated that cooperation with the government has a significant influence on a firm’s innovation. The relationship between tea manufacturing firms and educational and research institutions typically occurs with TRI and NIPM. In addition, firms tend to also have relationships with universities. The results of models 1, 2, 3, and 4 illustrate (the coefficients are 0.032, P≥0.1; 0.038, P≥0.1; 0.084, P≥0.1; and -0.039, P≥0.1) that relationships with educational and research institutions are not significantly related to overall EO and its dimensions of the firm. Santoro and Gopalakrishnan (2000) argued that collaboration with universities and research institutions is a way to develop technical knowledge, an objective that a firm cannot accomplish singlehandedly. However, in this context, the findings show that the relationship with educational and research institutions does not significantly influence EO. Entrepreneurial infrastructure This section illustrates the level (mean values) of entrepreneurial infrastructure received by tea manufacturing firms from external institutions related to the tea industry in Sri Lanka. Correlation analysis was performed to identify the relationship between entrepreneurial infrastructures and a firm’s EO (Table 4). Tea manufacturing firms received significant financial support from tea brokers (mean value=3.99); however, this fact does not significantly correlate with EO because its intention is not specified. Yet, the objective of government financial support (mean value=1.64) is clearly defined as enabling upgrades of the production processes of tea factories. The results revealed that financial support received from government institutions has a positive and significant relationship with EO.

International Food and Agribusiness Management Review

302

Wickramaratne et al.

Volume 20, Issue 3, 2017

Table 4. Level of benefits received and the correlation between entrepreneurial orientation and types of benefits.1,2,3

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

Types of benefits

Benefit provider

Financial support Information sharing Education and training Innovation development support Consultation Research and development Networking facilities

Brokers

OTF4

Government facilitating ins.

Educational and research ins.

0.043 (3.99) 0.298** (4.02) n.a. 0.220* (3.28) n.a. 0.236* (2.20) n.a.

n.a. 0.155 (2.74) n.a. 0.047 (1.71) n.a. 0.049 (1.23) n.a.

0.407** (1.64) 0.158 (3.30) 0.216* (2.80) 0.306** (2.46) n.a. 0.193* (1.47) 0.321** (1.80)

n.a. n.a. 0.123 (3.16) 0.137 (2.72) 0.095 (3.00) 0.324** (2.05) 0.127 (1.59)

Mean values are in parenthesis. Significance level of correlation: *P<0.05; **P<0.01. 3 n.a. = not available. 4 OTF = other tea factories. 2

The results indicate that information shared by brokers has a positive and significant relationship with EO because the detailed information that they share helps improve forecasts of future demand and estimated customer preferences, as revealed by Uzzi (1997). Further, Shane and Venkataraman (2000) argued that social capital might make it possible to obtain information, which is an essential factor of entrepreneurial opportunities. However, information shared by OTF and government facilitating institutions is not significantly correlated with EO. Normally, education and training assists in enhancing the managerial and technical expertise of owners, managers, and other employees. This finding shows that education and training support received from government facilitating institutions is significantly correlated with EO. The results illustrate that the innovation development support received from brokers and government facilitating institutions has a positive and significant relationship with EO. Government facilitating institutions launched several programs to upgrade the input (green leaf) quality and ensure the quality of manufactured tea as regulatory bodies. Research and development support received from tea brokers, government facilitating institutions, and educational and research institutions have a significant relationship with the EO of the firm. Obviously, research and development support will accumulate technical and managerial expertise, which helps to enhance EO. Likewise, networking facilities initiated by government facilitating institutions are significantly correlated with EO because they assist in enhancing the accessibility of information and resources. Additionally, note that the entrepreneurial infrastructure provided by educational and research institutions are not significantly correlated with EO, except for research and development support in this context. This lack of correlation might be the result of the entrepreneurial infrastructure provided by institutions that is not in line with the requirements of tea manufacturing firms.

5. Conclusion This study addressed the external relationships and EO of tea manufacturing firms in low growing tea areas of Sri Lanka. The findings indicate the importance of external relationships; in particular, relationships with supply chain partners and government facilitating institutions enhance the EO of tea manufacturing firms. However, relationships with OTF and educational and research institutions do not significantly influence firmsâ&#x20AC;&#x2122; EO in the context of the Sri Lankan tea industry. This study reveals a positive and insignificant influence of relational capital on the EO of firms. Therefore, these findings are partially in line with previous research (e.g. Lee et al., 2001; Stam and Elering, 2008) that stated that social capital allows firms to enhance their EO. This phenomenon can be explained within the context of the Sri Lankan tea industry. At present, significant competition exists for green leaf among tea factories. Further, the findings indicate that the availability of International Food and Agribusiness Management Review

303

Wickramaratne et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

green leaf is a strong influence on EO and its dimensions of tea factories. According to Schmitz (1995), cooperation among firms depends on their perceived costs and benefits. Therefore, this study concludes that the relationship with supply chain partners and government facilitating institutions positively influence EO. In the Sri Lankan context, the relationship with competitors and educational research institutions does not significantly influence the EO of tea manufacturing firms. When considering the dimensions of EO, innovativeness, proactiveness, and risk-taking behavior are significantly influenced by the relationship with supply chain partners and government facilitating institutions. Furthermore, risk-taking behavior is negatively influenced by the relationship with OTF. Therefore, among the dimensions of EO, risk taking is highly influenced by external relationships. With respect to entrepreneurial infrastructure, the benefits provided by tea brokers and government facilitating institutions are more likely to have positive relationships with EO, whereas the facilities provided by educational and research institutions are less likely to be significantly correlated with EO. Therefore, governmental and institutional policy makers should consider the fundamentals of entrepreneurships and ensure that infrastructure requirements are created to strengthen firms’ technical and business expertise. The findings of this study show that relational capital provides external networks for the discovery of opportunities to test new product ideas and to attain resources, similar to the prior study of Lee et al. (2001). Therefore, that relationships between firms and external institutions can also improve a platform to enhance a firm’s EO under favorable industry context is concluded. Aloulou and Fayolle (2005) argued that strengthening entrepreneurship is essential to responding to a changing globalized environment. This study indicates that external relationships are associated with higher EO, except in the case of relationships with OTF where the association with risk taking is negative. Further, the entrepreneurial infrastructure is positively linked with EO. Such empirical data make important contributions to the existing literature, particularly in explaining the role of external relationships with EO and its dimensions under a developing economic context. Another contribution is that research studies on entrepreneurship related to the tea industry in Sri Lanka are still rare. Such contributions are important to managers and policy makers to enhance the EO of tea manufacturing firms in Sri Lanka with respect to facing competitiveness in the global market. This study has several limitations. It only examines the low country tea manufacturing firms; therefore, the results and recommendations are limited to within the industry context. Additional research that considers other industries would complement this study and enhance the generalizability of these findings. Further, the cross-sectional nature of this study may not have been the most appropriate approach because it fails to capture the dynamic interplay between relational capital and EO. Regarding the most advantageous external relationships, possibilities exist that institutions may want to have relationships with more entrepreneurial firms. Because this study did not analyze such trends, a longitudinal study may offer further remarkable insights. Further, it is important to consider additional variables such as manager’s competencies that touch on relation that emphasizes the impact external relations have on EO. Moreover, in this study, data were gathered from single source informants (owner/manager) in each firm. The underlying assumption behind this method is that such individuals are capable of providing opinions that reflect the company’s behavior. According to Lyon et al. (2000), research using single-source self-reports – when carefully performed – are suitable and necessary means of operationalizing key constructs. However, a multiple informant approach could be adopted in future research.

6. Managerial and policy implications The present study offers valuable insights for owner/managers and policy makers. The findings reveal that entrepreneurs can enhance their firms’ EO by acquiring relational capital. To enhance the degree of EO of tea manufacturing firms in Sri Lanka, developing effective institutional support to obtain financial, human, International Food and Agribusiness Management Review

304

Wickramaratne et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

and information resources for use when initiating new business opportunities to develop their external environment is necessary. Doing so enables firms to overcome the obstacles they face in the entrepreneurial process. Further, the findings imply that some entrepreneurial infrastructures provided by external parties are positively correlated with firms’ EO. Therefore, government and non-government sectors need to arrange their facilities in line with promoting the level of EO. At the same time, the findings reveal that external relationships with OTF have a negative influence on risk taking. Consequently, entrepreneurs need to fairly manage external relationships to minimize such effect. Research institutions and universities are capable of sharing information and developing the technical knowledge that is important to enhance firms’ EO strategies. However, the findings indicated that relationships with educational and research institutions are not significantly related to EO. Therefore, in the context of the Sri Lankan tea industry, the government needs to encourage research institutions and universities to assist tea manufacturing firms by focusing on their special requirements. Additionally, enhancing collaboration among government agencies and all stakeholders of the tea industry to enhance the productivity of tea lands and to ensure the quality of green leaf is needed because the availability of green leaf significantly influences the EO of firms and their dimensions.

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2015.0070. Methods S1. Questionnaire.

References Aloulou, W. and A. Fayolle. 2005. A conceptual approach of entrepreneurial orientation within small business context. Journal of Enterprising Culture 13: 21-45. Brockhaus, R.H. 1980. Risk taking propensity of entrepreneurs. Academy of Management Journal 23: 509-520. Brüderl, J. and P. Preisendörfer. 1998. Network support and the success of newly founded business. Small Business Economics 10: 213-225. Cooke, P. and D. Wills. 1999. Small firms, social capital and the enhancement of business performance through innovation programmes. Small Business Economics 13: 219-234. Cope, J., S. Jack and M.B. Rose. 2007. Social capital and entrepreneurship: an introduction. International Small Business Journal 25: 213-219. Covin, J.G. and D.P. Slevin. 1989. Strategic management of small firms in hostile and benign environments. Strategic Management Journal 10: 75-87. Covin, J.G. and D.P. Slevin. 1991. A conceptual model of entrepreneurship as firm behavior. Entrepreneurship Theory and Practice 16: 7-25. Doloreux, D. 2004. Regional networks of small and medium sized enterprises: evidence from the Metropolitan Area of Ottawa in Canada. European Planning Studies 12: 173-189. Dyer, J.H. and H. Singh. 1998. The relational view: cooperative strategy and sources of interorganizational competitive advantage. Academy of Management Review 23: 660-679. Fairoz, F.M., T. Hirobumi and Y. Tanaka. 2010. Entrepreneurial orientation and business performance of small and medium scale enterprises of Hambantota district Sri Lanka. Asian Social Science 6: 34-46. Granovetter, M.S. 1985. Economic action, social structure and embeddedness. American Journal of Sociology 91: 481-510. Gumusluoğlu, L. and A. Ilsev. 2009. Transformational leadership, creativity, and organizational innovation. Journal of Business Research 62: 461-473. Hitt, M.A., R.D. Ireland, S.M. Camp and D.L. Sexton. 2001. Strategic entrepreneurship: entrepreneurial strategies for wealth creation. Strategic Management Journal 22: 479-491. Ingram, P. and P.W. Roberts. 2000. Friendships among competitors in the Sydney hotel industry. American Journal of Sociology 106: 387-423. International Food and Agribusiness Management Review

305

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0070 - Tuesday, May 09, 2017 8:27:51 PM - IP Address:24.21.169.207

Wickramaratne et al.

Volume 20, Issue 3, 2017

Kale, P., H. Singh and H. Perlmutter. 2000. Learning and protection of proprietary assets in strategic alliances: building relational capital. Strategic Management Journal 21: 217-237. Knight, G.A. 1997. Cross-cultural reliability and validity of a scale to measure firm entrepreneurial orientation. Journal of Business Venturing 12: 213-225. Kreiser, P.M., L.D. Marino and K.M. Weaver. 2002. Assessing the psychometric properties of the entrepreneurial orientation scale: a multi-country analysis. Entrepreneurship Theory and Practice 26: 71-94. Lechner, C. and M. Dowling. 2003. Firm networks: external relationships as sources for the growth and competitiveness of entrepreneurial firms. Entrepreneurship and Regional Development 15: 1-26. Lee, C., K. Lee and J.M. Pennings. 2001. Internal capabilities, external networks, and firm performance: a study on technology-based ventures. Strategic Management Journal 22: 615-640. Liao, J. and H. Welsch. 2005. Roles of social capital in venture creation: key dimensions and research implications. Journal of Small Business Management 43: 345-362. Lumpkin, G.T. and G.G. Dess. 1996. Clarifying the entrepreneurial orientation construct and linking it to performance. Academy of Management Review 21: 135-172. Lyon, D.W., G.T. Lumpkin and G.G. Dess. 2000. Enhancing entrepreneurial orientation research: operationalizing and measuring a key strategic decision making process. Journal of Management 26: 1055-1085. Miller, D. and P.H. Friesen. 1982. Innovation in conservative and entrepreneurial firms: two models of strategic momentum. Strategic Management Journal 3: 1-25. Mohamed, M.T.Z. and A.K.N. Zoysa. 2006. Current status and future research focus of tea in Sri Lanka. Journal of Agricultural Sciences 2: 32-42. Nahapiet, J. and S. Ghoshal. 1998. Social capital, intellectual capital, and the organizational advantage. Academy of Management Review 23: 242-266. Naldi, L., M. Nordqvist, K. SjĂśberg and J. Wiklund. 2007. Entrepreneurial orientation, risk taking, and performance in family firms. Family Business Review 20: 33-47. Nieto, M.J. and L. SantamarĂa. 2007. The importance of diverse collaborative networks for the novelty of product innovation. Technovation 27: 367-377. Nunally, J. 1978. Psychometric (2nd Ed.). McGraw Hill, New York, NY, USA. Peng, M.Y. and Y. Luo. 2000. Managerial ties and firm performance in a transition economy: the nature of a micro-macro link. Academy of Management Journal 43: 486-501. Santoro, M.D. and S. Gopalakrishnan. 2000. The institutionalization of knowledge transfer activities within industry-university collaborative ventures. Journal of Engineering and Technology Management 17: 299-319. Schmitz, H. 1995. Collective efficiency: Growth path for small-scale industry. Journal of Development Studies 31: 529-566. Shane, S. and S. Venkataraman. 2000. The promise of entrepreneurship as a field of research. Academy of Management Review 25: 217-226. Simsek, Z., M.H. Lubatkin and S.W. Floyd. 2003. Interfirm networks and entrepreneurial behavior: a structural embeddedness perspective. Journal of Management 29: 427-442. Sri Lanka Tea Board. 2012. Annual Report of Sri Lanka Tea Board. Sri Lanka Tea Board, Colombo, Sri Lanka. Stam, W. and T. Elfring. 2008. Entrepreneurial orientation and new venture performance: the moderating role of intra and extra industry social capital. Academy of Management Journal 51: 97-111. Suzuki, A.K.S., S. Kim and Z. Bae. 2002. Entrepreneurship in Japan and Silicon Valley: a comparative study. Technovation 22: 595-606. Tsai, W. and S. Ghoshal. 1998. Social capital and value creation: the role of intra-firm networks. Academy of Management Journal 41: 464-476. Uzzi, B. 1997. Social structure and competition in inter-firm networks: the paradox of embeddedness. Administrative Science Quarterly 42: 35-67. Zahra, S.A. and J.G. Covin. 1995. Contextual influences on the corporate entrepreneurship-performance relationship: a longitudinal analysis. Journal of Business Venturing 10: 43-58. Zimmerman, M.A. and J.G. Zeitz. 2002. Beyond survival: Achieving new venture growth by building legitimacy. Academy of Management Review 27: 414-431. International Food and Agribusiness Management Review

306

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2014.0176

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

Received: 2 December 2014 / Accepted: 20 January 2017

Governance structures in smallholder pig value chains in Uganda: constraints and opportunities for upgrading RESEARCH ARTICLE Emily Ouma a, Justus Ochiengb, Michel Dionec, and Danilo Pezod aAgricultural

Economist, and cAnimal Health Scientist, International Livestock Research Institute, c/o Bioversity International P.O. Box 24384, Kampala, Uganda

bAgricultural

Economist, World Vegetable Center Eastern and Southern Africa, P.O Box 10, Duluti, Arusha, Tanzania

dConsultant,

Tropical Agriculture Research and Higher Education Centre (CATIE), 30501 Turrialba, Costa Rica

Abstract This paper analyses governance structures in Ugandaâ&#x20AC;&#x2122;s smallholder pig value chains by applying the New Institutional Economics framework. It utilises cross sectional and qualitative survey data from randomly selected pig value chain actors in 4 districts. A multinomial logit model is applied to assess the determinants of vertical integration among pig traders. The findings indicate that most relationships at the pig production node of the value chain are based on spot market governance structures supported by personal relationships and trust. Live pig traders are mostly vertically integrated. High integration levels of the pig traders are positively influenced by access to market information, value of investments in the value chain, and dedicated asset specificity in terms of backyard slaughter premises. Upgrading opportunities in the value chain in the form of value addition strategies, policy implementation and promotion of business models that link producer organisations to quality inputs and service suppliers through contractual arrangements are identified. Keywords: smallholder pig value chains, governance structures, transaction cost economics, vertical integration, upgrading, Uganda JEL code: D4, D23, L11, L14, Q13 Corresponding author: e.a.ouma@cgiar.org

ÂŠ 2017 Ouma et al.

307

Ouma et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

1. Introduction Agri-food value chain systems are being transformed with a shift from traditional sales and purchase channels to more coordinated linkages and value chain governance systems in developing countries. This is driven by increased demand from consumers for food safety and better food quality specifications. Value chain governance refers to the relationships among the buyers, sellers, service providers and regulatory institutions that operate within or influence the range of activities required to bring a product or service from inception to its end use (Kaplinsky and Morris, 2001). Governance is about power and the ability of actors to exert control at any point in the chain by setting or enforcing product or process parameters under which others in the chain operate. The key parameters of focus as indicated in value chain governance literature (e.g. Dolan and Humphrey, 2000; Gereffi et al., 2005) include; what is to be produced; including product design and specifications, how it is to be produced; in terms of the production processes including elements such as the technology to be used, quality control, labour standards and environmental standards, and how much and when it is to be produced; implying production scheduling and logistical arrangements. Several prototypes of governance structures are distinguished in the literature. There is consensus that these structures exist along a spectrum (Gereffi et al., 2005; Minot, 2011; Webber and Labaste, 2010). The spectrum has the spot market governance on one end and hierarchy or vertical integration on the other. Between the two governance poles are a variety of intermediary structures such as production and marketing contracts. Marketing contracts refer to pre-production agreements on the terms of sale for example quality, quantity and pricing, while production contracts refer to agreements that include pre-production assistance such as inputs, information and credit. Spot market governance involves transactions that require little or no formal cooperation between participants. It is characterised by many buyers and sellers leading to lack of repeated transactions, limited information sharing and significant flexibility and independence (Webber and Labaste, 2010). Market price provides powerful signals and is the principal coordinating mechanism. Commodities typically produced and traded in spot markets include staples. The hierarchical forms are characterised by full integration whereby all value chain activities are brought together within one legal entity or ownership. Horizontal and vertical coordination are also internalised in such structures. In the framework of transaction cost economics (TCE), such forms offer greater protection for specific investments and provide relatively efficient mechanisms for responding to change where coordinated adaptation is necessary (Shelanski and Klein, 1995). Full integration is common in commodities that are economically produced at scale and those that have high operational standards (Minot, 2011). Vertical coordination exists between actors with different functions while horizontal coordination exists among the actors who have the same function in the value chain. When specialised investments in the form of assets such as equipment and human resources are needed to carry out transactions, and when product or input markets are thin, central coordination of investment decisions and combined ownership become desirable. The shift from spot market to hierarchy along the spectrum entails a trade-off between price incentives and adaptive properties of the market, and the safeguards and central coordinating properties of the firm. Agri-food value chain governance is pivotal especially for small and medium enterprises (SMEs) for better integration and coordination of their activities in value chains and improved performance. The evolution and significance of value chain governance in promoting SME performance in agri-food chains has attracted considerable attention of agribusiness and policy analysts (e.g. Thorpe and Bennet, 2004; Trienekens and Willems, 2007). There are risks that as value chains develop and get more organised in order to meet the changing patterns in consumer demand, SMEs may be left out. An important challenge for the largest number of developing country SMEs is how to enter into these chains and become competitive. This is because they are often confronted with asymmetric power relationships that impact on the distribution of costs and benefits in the chain. This study analyses existing governance structures in smallholder pig value chains in Uganda and identifies inclusive models that could enhance integration and competitiveness of SMEs. It further examines the International Food and Agribusiness Management Review

308

Ouma et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

institutional environment under which the value chain operates and highlights opportunities for upgrading. Piggery is increasingly becoming an important sector dominated by SMEs in the country, yet there is limited evidence as regards its structure and functioning. The smallholder pig value chain in Uganda is a major source of livelihood for more than 1.1 million households (Uganda Bureau of Statistics, 2009), and contributes to national food security. The pigs are mainly kept in smallholder backyard systems, managed by women. Pigs are often preferred as they grow fast, produce 16-20 piglets per sow per annum, and are efficient transformers of various low value feed resources, mostly considered as ‘waste’, into high-value animal-source food, for sale or home consumption (Mutua et al., 2010). These livestock, normally considered as ‘living banks’, can quickly be sold for cash to meet household financial needs, such as school fees and inputs for crop cultivation. Demand for pork in Uganda is increasing rapidly and the annual per capita pork consumption, at 3.4 kg, is now the highest in East Africa (FAOSTAT, 2015). Fuelled by the increasing demand, the number of pigs in Uganda has increased from 0.2 to 3.2 million between 1980 and 2008. With a projected annual growth rate of 8% over the coming decade, the number of pigs in the country will come close to eight million by 2020 (Uganda Bureau of Statistics, 2009). Given its strategic location, there are opportunities for Uganda to export pig and pig products to neighbouring countries such as the Democratic Republic of Congo (DRC), Rwanda, and South Sudan where the demand for pork is also growing.

2. Theoretical and analytical framework The New Institutional Economics (NIE) framework, specifically TCE is applied to assess the rationale for governance structures and relations in the pig value chain. The basic premise of TCE is that transactions must be governed as well as designed and carried out, and that certain institutional arrangements affect governance better than others (Coase, 1998; Williamson, 1991). Transactions between firms are governed under conditions of bounded rationality and opportunism of the actors involved. Opportunism refers to unscrupulous behaviour of agents in an attempt to maximise their individual utility. Bounded rationality refers to the limited rationality of humans due to lack of relevant information and limited capacities. Williamson (2000) employs the bounded rationality assumption in TCE to suggest that all complex contracts are unavoidably incomplete, which leads to significant contracting issues in the face of opportunism. He suggests that contract incompleteness arises from two distinct mental bounds: cognitive and verbal limitations. The cognitive limitations prevent actors from generating all possible contingencies to include in the contract; while verbal limitations attenuate the contract’s content, because an idea cannot be included if it cannot be expressed in words. These limitations lead to incomplete contracts because actors can neither imagine all the possible contingencies that should go into the contract nor articulate them. In the presence of opportunism by agents, incomplete contracts lead to serious contractual difficulties, which can prevent exchanges from occurring. In smallholder farmer settings, limited capacities and lack of information lead to oral non-formal agreements based on trust between trading partners. Under these two conditions, transactions are characterised by hazards, thus measures must be taken to mitigate the losses arising from high transaction costs. TCE offers a set of normative rules on how trading partners choose governance arrangements that mitigate all forms of contractual hazards found between the partners in order to minimise transaction costs (Masten, 1993). There are key transactions cost-related constraints in the smallholder pig value chain in Uganda as in other agri-food value chains. These include high transaction costs due to high transportation costs and poor physical infrastructure; poor information flows, resulting in information asymmetry in buyer-seller relationships; low value final markets; weak legal and quality assurance systems; high costs of production inputs, high search costs, especially due to scarcity of pigs during subsequent periods after African swine fever (ASF) disease outbreaks (Minten et al., 2009). The institutions of principal focus in NIE are the institutional environment, that is ‘rules of the game’ and the institutions of governance referred to as ‘play of the game’ (Williamson, 1999). The institutional environment International Food and Agribusiness Management Review

309

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

Ouma et al.

Volume 20, Issue 3, 2017

refers to the formal rules, comprising policies, laws, regulations and property rights that create order in exchange (North, 1991). The institutions of governance refer to the use of governance structures such as spot market, hybrids or the hierarchical forms. Such structures consist of a collection of rules or institutions and constraints structuring the transactions between the various actors in the value chain. A fundamental notion of TCE is that market forces work to bring about an ‘efficient sort’ between transactions and governance structures, so that exchange relationships observed can be explained in terms of transaction costs. It also recognises that markets have difficulty dealing with some transactions because of asset specificity, bounded rationality, and opportunistic behaviour by the parties to the transaction. Williamson (1996, 1999) identifies key important properties of underlying transactions that influence governance structures. These include asset specificity, uncertainty about the future of the relationship, complexity of the transaction and frequency of transactions. Asset specificity refers to the transferability of assets (physical and human) that support a given transaction (Williamson, 1999). Transaction costs are affected by the extent to which specialised investments have to be made to carry out transactions. These specialised investments cannot be redeployed to alternative use without loss of productive value (Shelanski and Klein, 1995). Assets that are specific to a particular production relationship can empower those who own them because unique assets reduce the ease with which the asset owner may be replaced, should the production relationship deteriorate (Masten, 1993). This increases the dependence of other value chain actors in maintaining this production relationship. However, specific assets can also disempower the owner because they reduce the ease with which the asset owner can abandon the production relationship. Under such circumstances the owner may therefore seek stronger governance structures for protection. Bounded rationality further limits the capability of markets and simple contracts to handle asset specificity since all contingencies cannot be foreseen and contracted for by the parties. Spot market transactions offer no protection to asset owners against opportunism when transaction-specific assets are involved. Contracts and hierarchy forms of governance structures offer some protection for transactionspecific assets by tying the buyer and seller together for a specified period. When asset specificity, bounded rationality, and opportunism make contracting problems severe, vertical integration is preferred to ensure that the value of transaction specific assets is internalized. Uncertainties are unanticipated changes in the circumstances and behaviour surrounding a transaction. Aside from increasing information, contracting and ex post monitoring costs, uncertainty is the source of contractual disturbances to which costly adaptation may be required of two contracting parties (Williamson, 2008). Due to uncertainty, the exchange partners may find it difficult to write market contracts, leading to opportunism (Leiblein, 2003). As uncertainties increase, TCE hypothesizes that economic actors will respond by adopting coordination mechanisms and governance structures that offer increased control over unknowns. The probability of observing a more integrated governance structure depends positively on the amount or value of the transaction – specific assets involved in the transactions, the degree of uncertainty about the future of the relationship, the complexity of the transaction, and on the frequency of trade (Shelanski and Klein, 1995). Observed governance structures in the smallholder pig value chain are analysed using cross sectional and qualitative focus group discussion data. Specifically a multinomial logit (MNL) model is employed to examine the determinants of vertical integration observed in the pig trading node of the value chain by considering their transaction characteristics. Different levels of integration are observed in the live pig trading node, with traders carrying out multiple functions including logistical functions such as transportation and value addition roles such as pig slaughter and pork retail under single ownership. Other studies such as Mtimet and Baker (2013) also observe vertical integration among pig traders in Uganda. The traders are assumed to maximize their expected net returns from transactions, by selecting a given integration level that minimizes transaction costs. The expected net returns, E(π) from a given level of integration i can be represented by a latent variable:

International Food and Agribusiness Management Review

310

Ouma et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

E(π)i* = Zγi + μi (1) Although the E(π)i for the trader is not known to the analyst, Z is a vector of transaction related attributes, including asset specificity, uncertainty, other externalities and value chain actor characteristics that are observed. The relevance of these attributes on the levels of integration of pig traders is tested. The stochastic components and parameter estimates are represented by μ and γ, respectively. The behavioural model for the analyst is that a trader j, compares the expected net returns from transactions associated with different levels of vertical integration, and chooses alternative i from a finite set of alternatives I, with probability P(i) if the expected net return associated with i is greater than other alternatives. This is depicted in (2) as follows: P(i|Ij) =E(π)ij > E(π)kj = P(μkj–μij ) <(Zij–Zkj)

∀k ∈Ij = 1,…K; i≠k

(2)

The stochastic term, μ is assumed to be independent and identically distributed with a type 1 extreme value distribution, resulting in a MNL model: P(i|Ij) =

exp (Z ij μ ij ) ∑k exp (Z kj μ kj )

(3)

In the MNL, the vertical integration level for each trader is the dependent variable, and the transaction attributes and socio-demographic characteristics of the traders, the independent variables.

3. Data and description The data employed in this study are from pig value chain actor surveys that were conducted in 2013-2014 in Masaka, Kamuli, Mukono and Kampala districts, representing high pig supply and demand areas. Random samples of 376 pig farmers, 101 pig traders, 36 veterinary drug stockists and 36 livestock feed traders were used. The samples were drawn from lists of each value chain actor category prepared by local council authorities from 14 sub-counties in the survey districts. Survey tools were developed to capture information on characteristics of the actors, in particular the types of assets held, purchase of inputs and intermediate inputs in terms of cost, origin, supplier types, transaction characteristics, value addition, and marketing of outputs in terms of prices received, geographic destinations, buyer types, as well as quality attributes. The survey tools were administered by qualified and trained enumerators using in-person questionnaire interviews in the dialects of the value chain actors. Qualitative focus group discussion data from 1,400 randomly selected pig farmers were also used to complement the producer level data. In this paper, the value chain data is used to describe the overall pig value chain governance structures, while the pig trader data is used to model determinants of integration among pig traders by using a MNL model. The dependent variable considered in the MNL estimation is the level of integration of the live pig traders in the value chain. In this study, it is defined in terms of the number of value chain functions in which the trader is involved in the value chain. The traders involved in only one function are considered as low integrated while those involved in two to three functions as medium integrated. Traders involved in more than three functions are considered as highly integrated. Survey results indicate that 33% of the traders were highly integrated, while 35% and 32% were medium low integrated, respectively. The independent variables employed to explain determinants of integration levels among pig traders are based on the TCE literature (Shelanski and Klein, 1995; Trienekens, 2011; Williamson, 2000) and are presented in Table 1. 90% of the pig traders in the sample were men, with an average of 7 years of formal education, corresponding to primary level of education. The low integrated traders however, had an average of 5 years of education. In the sample, 50% of the traders had access to market information, of which 70% were vertically integrated compared to 30 and 40% of the low and medium integrated traders, respectively. Pig slaughter is carried out in unregulated backyard slaughter premises, that often lack appropriate waste disposal infrastructure and

International Food and Agribusiness Management Review

311

Ouma et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

Table 1. Definition of variables used in the multinomial logit model.1 Variables

Definition

Age Education Gender Experience Market information Trade association Assets (log)

Age in years 38.2 (9.3) No. of years of schooling 6.4 (3.3) 1=male; 0=female 0.91 (0.3) Years in pig business 9.6 (7.9) Access to market information on pork 0.5 (0.5) retail (1=yes; 0=otherwise) Member of a trade association 0.2 (0.4) (1=yes; 0=otherwise) Value in Uganda shillings of assets 12.3 (1.5) used in piggery business (log transformed) Access to a slaughter premise (1=yes; 0.6 (0.5) 0=otherwise) Trader’s pig market supply 0.7 (0.4) situation in last 5 years (1=increase; 0=otherwise)

Backyard slaughter Market perception 1

Overall sample (n=101)

Highly integrated (n=33)

Medium integrated (n=35)

Low integrated (n=33)

38.3 (10.1) 6.9 (3.1) 0.97 (0.2) 10.0 (8.9) 0.7 (0.5)

38.9 (10.4) 7.0 (3.2) 0.91(0.3) 8.5 (7.3) 0.4 (0.5)

37.6 (7.2) 5.1 (3.3) 0.91 (0.3) 10.3 (7.6) 0.3 (0.5)

0.2 (0.4)

12.8 (1.7)

12.2 (1.4)

11.8 (1.3)

0.7 (0.5)

0.5 (0.5)

0.4 (0.5)

0.6 (0.5)

0.8 (0.4)

0.7 (0.5)

Standard deviation in parentheses.

adequate water supply. The backyard slaughter premises are usually adjacent to pork retail outlets. In most cases, the traders collectively use such slaughter premises, often owned or rented by one of them. Most of the vertically integrated pig traders (70%) have access to such slaughter premises. The backyard slaughter premises for slaughtering pigs are less easily transferable for slaughtering other livestock species due to religious biases associated with pig consumption for religious groups such as Muslims and Seventh Day Adventists among others. The value of assets used in piggery is relatively similar across the 3 groups of traders in the sample.

4. Results and discussion Description of the pig value chain The pig value chain map is presented using GTZ’s ValueLinks approach (Springer-Heinze, 2007). Most relationships in the pig value chain are based on ‘arm’s length’ spot market governance structure (Figure 1). ■■ Pig production There are several smallholder farmers selling pigs for slaughter to several intermediary parties (traders and collectors) through uncoordinated spot-market transactions, based on oral agreements (Figure 1 and 2). This is usually supported by personal relationships and trust. The farmers sell on average 1-2 pigs at a time, twice a year to local traders or collectors working within larger traders’ business networks. The chain is long, with several intermediaries between the production and consumption nodes. There is limited availability of end market information on prices and preferred product attributes to producers. Asset specificity for individual pig farmers is generally low, given that the production assets associated with piggery can usually be deployed towards other livestock. There are generally no standards to adhere to and low barriers to entry. This system offers some advantages to the trading parties. Farmers are able to meet International Food and Agribusiness Management Review

312

Ouma et al.

Volume 20, Issue 3, 2017

Input supply & services

Pig production

Breeders (piglet producers)

Private/ government veterinarians & paravets

Transport

Slaughter Pig slaughter slabs

Pig fatteners (sale of pigs for slaughter)

In-country live pig traders

Live pig traders

Piglet producers (sale of weaners)

Brokers

Transporters

Processing

Wholesaling

Processors

Pork traders

Kampala pig slaughterhouse

Retailing

Consumption

Pork butchers

Pork joints

Supermarket

Restaurants

Processor retail shops

Individual households

Cross border live pig traders

Village breeding boars Feeds traders (shops/millers)

Research institutions (NALIRRI, Universities and CGIAR centres e.g. ILRI, CIAT) Finance service providers (BRAC, FINCA, SACCOs, VSLAs) NGOs (VEDCO, SNV, MADDO, etc) District local governments â&#x20AC;&#x201C; Production and marketing office

Ministry of Health â&#x20AC;&#x201C; Public health department

National Drug Authority MAAIF: NAGRC & DB, NADDEC, Livestock extension spot market

contracts (formal/informal) support organisations

hierarchical structure

value chain enablers

Figure 1. Pig value chain map. Direct to consumers

100

Neighbourhood butcher

Butcher in another town

Trader

Masaka

Kamuli

Figure 2. Pig sales outlets. International Food and Agribusiness Management Review

313

Mukono

Goma

Mukono TC

Kyampisi

Ntenjeru

Bugulumbya

Namwendwa

Kitayunjwa

Katwe-Butego

Nyendo-Ssenyange

Kimanya-Kyabakuza

Kabonera

Kyanamukaka

Kkingo

Proportion of producers (%)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

Veterinary supplies (Agro-vet shops)

Collection/ bulking

Ouma et al.

Volume 20, Issue 3, 2017

cash needs through spontaneous transactions and both transacting parties are not burdened by the governance costs of complex coordination structures. However, the disadvantage with such a coordination mechanism is that the transactions are not transparent and transaction costs are borne by individual players. Farmers receive prices imposed by the traders, based on visual estimates given that pig weights are not monitored to inform the price determination process.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

■■ Live pig trade Pig trading involves collection of pigs from individual pig farmers and bulking for sale to larger traders located in urban areas, neighbouring districts and to some extent neighbouring countries. The specificity of assets used in the live pig trade is low, thereby enabling the traders involved solely in live pig trade to easily switch from pig trading to other income generating activities during low pig sale seasons. The most important fixed investments owned by the pig traders interviewed include motorcycles for transporting pigs and mobile phones for obtaining market information and networking. The specificity of these assets is low insofar as they can be used for a range of trading activities. During low pig sales period, the traders switch from pig trading to transportation of people using the motorcycles. Results show that most of the pig traders are vertically integrated, performing several functions in the value chain under single ownership. Compared to pig farmers who are largely concentrated at the production node, live pig traders are also involved in the retail nodes of the value chain, operating pork butcheries and pork joints while also carrying out pig slaughter functions (Table 2). The multiplicity of functions by live pig traders was observed across the four survey districts (Figure 3). In these districts, the traders purchased pigs from producers and village brokers, provided transport services, slaughtered pigs and also retailed pork in pork joints and pork butcheries. Only 7-25% of the live pig traders were involved in sale of live pigs to other traders. Informal contracting arrangements between the traders and village brokers, who source pigs from farmers was common. Pig traders wielded a lot of power in the value chain, as they collaborate through informal groupings and collude in setting both producer price of pigs and pork retail price, making them rather powerful actors in the chain. 21% of the traders in the sample were members of such groupings. Such platforms were also used by the traders to share information on pig product prices and other market conditions. ■■ Pig slaughter There is absence of designated slaughter facilities for pigs, in both rural and urban areas in the districts, except for Kampala. This gap is filled by traders who invest in backyard slaughter premises that are not regulated and without ante-mortem inspection of pigs. The resulting pork is largely not inspected and Table 2. Traders engagement in other business activities along the value chain. Value chain activities

Proportion (%) of producers (n=376)

Proportion (%) of live pig traders (n=101)

Butchery Pork joint1 Live pig trading Pig slaughter Supply inputs (e.g. feeds) Pig production

1.1 0.8 0.5 0.5 0.8 100.0

95.1 62.4 100.0 60.4 0.0 22.7

1 Pork joints are common pork eateries in Uganda serving roasted or fried pork. Many people socialise and watch football matches in such joints.

International Food and Agribusiness Management Review

314

Ouma et al.

Volume 20, Issue 3, 2017

Kampala

Kamuli

Mukono

Masaka

Roast/fry

Fresh pork

Transport Sell to consumers

Buy Slaughter Live pigs

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

Sell to other traders

Transport Sell to other traders Buy 0

100

% of live pig traders

Figure 3. Functions of live pig traders, by district and product type. mainly sold through local retail outlets such as butcheries and pork joints (Ouma et al., 2015; Tatwangire, 2014), and there is no grading or standards of pork carcasses sold through such retail outlets. Furthermore, the hygiene of the backyard slaughter premises, and pork joints in addition to pork handling are often poor. The veterinary and public health departments in local governments have the responsibility of regulating such premises and enforcing standards to ensure that only quality and safe pork and other meat products are offered in the market. However, enforcement of such controls and standards in the value chain has been weak. The traders operating backyard slaughter premises tend to be vertically integrated in the value chain, sourcing for live pigs and also operating pork retail outlets within the proximity of slaughter facilities to minimise on transport cost for the carcasses. The slaughter premises are often adjacent to the retail outlets. Most of the live pig traders in the districts (more than 60%) are involved in backyard slaughter of pigs. As a consequence of significant pressure from religious groups, most of the backyard slaughter premises and pork joints are not visibly advertised (Worsley, 2013). There is only one officially recognised and regulated pig slaughterhouse located in the capital city in Kampala. Due to repeat transactions, informal and largely oral contractual relationships through a relational type of governance structure exist between the slaughterhouse and traders supplying pigs sourced from several districts for slaughter. â&#x2013; â&#x2013; Processing There are only two main pork processing firms in Uganda, mainly relying on the pig slaughterhouse in Kampala for supplies through formal contracts. Purchase of pigs from smallholder farmers by the firms is very seldom. The processors invest in processing and chilling equipment for their pork products. The pig slaughterhouse in Kampala is the main supplier of pork to these firms. The processors target high end consumers of their processed products. The high end consumers demand for high and consistent quality pork and pork products through the supermarket chains and retail outlets owned by the processors. The linkages between the pig slaughterhouse in Kampala and the pork processors on one hand and the processors and the supermarket chains on the other is formalised though written contracts.

International Food and Agribusiness Management Review

315

Ouma et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

■■ Input supply The value chain map shows input suppliers comprising livestock feed traders, and suppliers of veterinary products and services. Their relationship with the chain operators especially pig farmers is mainly on spot market basis. A number of policies are in place to promote animal production and productivity but these are either poorly implemented or lack a legal framework for implementation. Such policies include the National Animal Feeds Policy (2005) to govern the processing and sale of compounded animal feeds with the goal of developing the animal feeds industry (Ministry of Agriculture, Animal Resources and Fisheries, 2005), and the National Drug Policy to ensure availability, of essential, efficacious and cost-effective drugs though the National Drug Authority (Ministry of Health, 2002). The national feeds policy emphasizes the importance of the private sector in spearheading the supply of quality animal feeds. Nevertheless, constraints associated with poor quality feeds due to adulteration, and poor formulation, and inefficacious livestock drugs still abound. Uganda’s Draft Feed Bill which is the legal framework for implementation of the policy has been formulated but not yet approved and enacted. The Feed Bill is vital in providing a legal framework to guide feed compounders and traders, and regulators in the feed sector. Empirical results on determinants of vertical integration of pig traders in the value chain The parameter estimates of the choice of vertical integration levels used by pig traders are presented in Table 3. The base group for comparison is traders that are low integrated. Marginal effects are used to interpret the determinants of pig trader choices of level of vertical integration. The results show that the high integration level of pig traders is positively and significantly influenced by access to market information at the retail node, value of investments in the value chain, and dedicated asset specificity in terms of backyard slaughter premises. The marginal effect of market information variable is positive and statistically significant for high integration and negative for low integration, showing that if a live pig trader has access to market information, this increases the likelihood of being highly integrated. Market information can be exploited to reduce uncertainty. However, information tends to be more reliable for the retail node (downstream) and not the live pig trade node of the value chain (upstream) which is often characterised by supply uncertainties. This influences the traders to pursue higher integration levels in order to have increased control over supply uncertainties and protect value. Key sources of uncertainty in the pig value chain that may disrupt pig trade business include price fluctuation of pigs as a result of outbreaks of Table 3. Multinomial logit estimates of determinants of vertical integration among pig traders.1,2 Variable

Age Education Gender Experience Market information Trade association member Log of assets Backyard slaughter Market perception 1

Low integrated

Medium integrated

High integrated

Marginal effects

z-value

Marginal effects

z-value

Marginal effects

z-value

-0.001 -0.039 -0.210 0.001 -0.281 -0.033 0.000 -0.289 0.090

-0.07 -2.07** -0.75 0.09 -2.53*** -0.24 0.01 -2.42*** 0.72

0.010 0.000 0.250 -0.017 -0.127 0.225 -0.129 -0.147 -0.044

1.40 0.00 1.97** -1.73* -1.15 1.46 -2.85*** -1.25 -0.33

-0.009 0.039 -0.039 0.016 0.408 0.192 0.129 0.436 -0.046

-1.23 1.89* -0.14 1.80* 3.62*** 1.58 2.51** 3.89*** -0.36

Base group is low integrated traders. and *** denotes significant variables at 10, 5, and 1% levels, respectively.

2 *, **,

International Food and Agribusiness Management Review

316

Ouma et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

ASF disease. The infectious nature and high mortality rates associated with ASF, along with the absence of a vaccine and medical treatment has made the disease one of the most constraining factors (Dione et al., 2015). During outbreaks of the disease, price offer by pig traders to farmers are quite low due to panic sales by the latter, despite the largely stable retail pork price to consumers. Subsequent periods after ASF outbreaks are characterised by low pig supplies due to scarcity, thereby raising search costs for live pig traders. Studies such as Hennessy (1996) show the important role of information asymmetry in agro-food industry vertical integration. The positive and significant marginal effect of the asset variable for high integrated traders and negative effect for the medium integrated indicate that the traders who have a higher asset value in the value chain are more likely to be highly integrated and less likely to be medium integrated. This implies that traders with higher levels of investments in the value chain would prefer contractual relationships that provide sufficient security. This is even more so if such investments have higher levels of asset-specificity as described by Williamson (1979) and Joskow (1998). The backyard slaughter premises variable is an example of such an asset. Its marginal effect in the model is positive and significant for high integrated traders and negative for the low integrated traders, implying that traders operating slaughter premises are more likely to be highly integrated. The slaughter premises are located in close proximity to pork retail outlets such as butcheries and pork joints to minimise on transport related costs, thus representing some aspects of dedicated asset specificity. These findings are consistent with other studies on agri-food supply chains such as Trienekens and Willems (2007) that find governance structuring of South African grape supply chain towards chain coordination and vertical integration that lead to innovations that require investments in relationship-specific assets. A lot however still remains to be learnt regarding the role of asset specificity in vertical integration governance structures due to the dearth of empirical analysis as measurement issues remain a challenge. This concern is also raised by authors such as Joskow (2004). Marginal effects of socio-demographic variables such as education, experience and gender are not strongly significant in the model.

5. Conclusions The pig sector in Uganda is generally underdeveloped although it has high potential for growth, given the rising demand for pork domestically and in neighbouring countries such as South Sudan, Rwanda and the DRC. The smallholder pig production node is dominated by spot market and informal relationships. The implication of this is that pig producers are generally price takers and lack a common voice. In order to improve the value chain position of farmers, their value capture and reduce transaction costs associated with individual spot market transactions for inputs and sale of pigs, business models along the lines suggested by Cadilhon and Kobusingye (2014) could be promoted. Such models necessitate horizontal integration of producers into collectives in order to improve their bargaining power for better terms of trade and reduced transaction costs associated with transportation due to economies of scale. The farmer collectives can then be linked to different business development service providers to improve access to quality inputs and services including market information, through contractual arrangements. The results show minimal investments by private sector players for value addition and processing of pork and pig products, yet these are critical avenues for upgrading the value chain. The gap for processed pork products in the country is currently met through importation. Opportunities for such value addition exist but the cost of the investments including equipment, cold chain infrastructure that meets quality standards, and energy for processing tends to be high. This study reveals gaps in the pig slaughter node of the value chain, currently dominated by pig traders. Opportunities for investment in appropriate pig slaughter facilities exist, through private or public-private partnerships. Dedicated supply of pigs to such facilities can be in the form of contractual arrangements with pig farmer collectives. Government has a role in supporting the development of the pig value chains by providing a favourable business environment to incentivise private sector investments in the value chain. An additional and critical role of government in upgrading the value chain is the establishment of functional systems for implementation International Food and Agribusiness Management Review

317

Ouma et al.

Volume 20, Issue 3, 2017

of rules and regulation for quality input supply, pork quality assurance and standards as documented in other studies such as Lee et al. (2012).

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

Acknowledgements The authors are grateful to the Livestock and Fish program of the CGIAR, Irish Aid, and the European Commission and International Fund for Agricultural Development (IFAD) for funding support. This study is based on research conducted under the Uganda smallholder pig value chain development projects. We acknowledge the contribution of the field teams and the pig value chain actors who participated in the study. We extend our appreciation to the Managing Editor and 2 anonymous reviewers of this paper.

References Cadilhon, J.J. and L. Kobusingye. 2014. Adding value to aquaculture products: Kati Farms (Uganda) Ltd. International Food and Agribusiness Management Review 17(B): 133-137. Coase, R.H. 1998. The new institutional economics. American Economic Review 88: 72-74. Dione, M.M., J. Akol, K. Roesel, E. A. Ouma, J. Kungu, B. Wieland and D. Pezo. 2015. Risk factors for African swine fever in smallholder pig production systems in Uganda. Transboundary and Emerging Diseases. DOI: https://doi.org/10.1111/tbed.12452. Dolan, C. and J. Humphrey. 2000. Governance and trade in fresh vegetables: the impact of UK supermarkets on the African horticulture industry. Journal of Development Studies 37: 147-176. FAOSTAT. 2015. Database available at: http://www.fao.org/faostat/en/#data/CL. Gereffi, G., J. Humphrey and T. Sturgeon. 2005. The governance of global value chains. Review of International Political Economy 12: 78-104. Hennessy, D.A. 1996. Information asymmetry as a reason for food industry vertical integration. American Journal of Agricultural Economics 78: 1034-1043. Joskow, P.L. 2004. New institutional economics: a report card. Presidential address at the annual conference of International Society of New Institutional Economics, Budapest, Hungary. Available at: http:// economics.mit.edu/files/1171. Joskow, P.L. 1998. Asset specificity and vertical integration. In: The new Palgrave dictionary of economics and law, edited by P. Neuman. MacMillan, London, USA. Kaplinsky, R. and M. Morris. 2001. A handbook for value chain research. Available at: http://tinyurl.com/ p2n2gmf. Lee, J., G. Gereffi and J. Beauvais. 2012. Global value chains and agrifood standards: challenges and possibilities for smallholders in developing countries. Proceedings of the National Academy of Sciences of the USA 109: 12326-12331. Leiblein, M.J. 2003. The choice of organizational governance form and performance: predictions from transaction cost, resource-based and real options theories. Journal of Management 29: 937-961. Masten, S.E. 1993. Transaction costs, mistakes, and performance: assessing the importance of governance. Managerial and Decision Economics 14: 119-129. Ministry of Agriculture, Animal Resources and Fisheries. 2005. The national feeds policy. Entebbe, Uganda. Available at: http://tinyurl.com/h8ws6za. Ministry of Health. 2002. Uganda national drug policy. Kampala, Uganda. Available at: http://tinyurl.com/ htlb92u. Minot, N. 2011. Contract farming in Sub-Saharan Africa: opportunities and challenges. Available at: http:// tinyurl.com/z5dtt8h. Minten, B., L. Randrianarison and J. Swinnen. 2009. Global retail chains and poor farmers: evidence from Madagascar. World Development 37: 1728-1741. Mtimet, N. and D. Baker. 2013. The analysis of traders in a developing country value chain: pig traders in Uganda. Available at: http://tinyurl.com/gs4lzjd. Mutua, F.K., S. Arimi, W. Ogara, C. Dewey and E. Schelling. 2010. Farmer perceptions on indigenous pig farming in Kakamega district, western Kenya. Nordic Journal of African Studies 19: 43-57. International Food and Agribusiness Management Review

318

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

Ouma et al.

Volume 20, Issue 3, 2017

North, D.C. 1991. Institutions. Journal of Economic Perspectives 5: 97-112. Ouma, E.A., M.M. Dione, P. Lule, D. Pezo, K. Marshall, K. Roesel, L. Mayega, D. Kiryabwire, G. Nadiope and J. Jagwe. 2015. Smallholder pig value chain assessment in Uganda: Results from producer focus group discussions and key informant interviews. ILRI Project Report, IRLI, Nairobi, Kenya. Shelanski, H.A. and P.G. Klein. 1995. Empirical research in transaction cost economics: a review and assessment. Journal of Law, Economics and Organisation 11: 335-361. Springer-Heinze, A. 2007. ValueLinks manual: the methodology of value chain promotion. GTZ, Eschborn, Germany. Tatwangire, A. 2013. Uganda smallholder pigs value chain development: past trends, current status and likely future directions. ILRI, Nairobi, Kenya. Thorpe, A. and E. Bennett. 2004. Market-driven international fish supply chains: the case of Nile-perch from Africaâ&#x20AC;&#x2122;s Lake Victoria. International Food and Agribusiness Management Review 7(4): 40-57. Trienekens, J. 2011. Agricultural value chains in developing countries: a framework for analysis. International Food and Agribusiness Management Review 14(2): 51-77. Trienekens, J. and S. Willems. 2007. Innovation and governance in international food supply chains: the cases of Ghanaian pineapples and South African grapes. International Food and Agribusiness Management Review 10(4): 42-63. Uganda Bureau of Statistics. 2009. National livestock census report. Uganda Bureau of Statistics (UBOS), Kampala, Uganda. Webber, C.M., P. Labaste. 2010. Building competitiveness in africaâ&#x20AC;&#x2122;s agriculture: a guide to value chain concepts and applications. Agriculture and Rural Development. The World Bank, Washington, WA, USA. Williamson, O.E. 1979. Transaction cost economics: the governance of contractual relations. Journal of Law and Economics 22: 233-62. Williamson, O.E. 1991. Comparative economic organization: the analysis of discrete structural alternatives. Administrative Science Quarterly 36: 269-296. Williamson, O.E. 1996. The mechanism of governance. Oxford University Press, New York, NY, USA. Williamson, O.E. 1999. Strategy research: governance and competence perspectives. Strategic Management Journal 20: 1087-1108. Williamson, O.E. 2000. The new institutional economics: taking stock, looking ahead. Journal of Economic Literature 38: 595-613. Williamson, O.E. 2008. Outsourcing: transaction cost economics and supply chain management. Journal of Supply Chain Management 44: 5-16. Worsley, S. 2013. Report on the pig value chain impact pathways workshop held in June 2013 in Kampala. ILRI, Kampala, Uganda. Available at: http://tinyurl.com/jp6zvmu.

International Food and Agribusiness Management Review

319

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2014.0176 - Tuesday, May 09, 2017 8:28:50 PM - IP Address:24.21.169.207

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2015.0156

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Received: 3 August 2016 / Accepted: 21 December 2016

CSR activities in the German poultry sector: differencing preference groups RESEARCH ARTICLE Henrike Luhmann

and Ludwig Theuvsenb

aResearcher

and bProfessor, Georg-August University of Goettingen, Department of Agricultural Economics and Rural Development, Chair of Management in Agribusiness, Platz der Goettinger Sieben 5, 37073 Goettingen, Germany

Abstract Corporate Social Responsibility (CSR) has long been an issue worldwide, and more and more industries are taking note. One important example is the poultry industry in Germany, which has become a focal point of public debate. Increasingly, consumers are demanding that firms take responsibility for their corporate actions. The goals of this study were, first, to analyze consumers’ preferences for poultry firms’ CSR commitment with the help of an adaptive conjoint analysis. Second, a cluster analysis is set out to define consumer groups according to their preferences. With regard to CSR, consumers in this study were most concerned with product quality, animal welfare and employee issues. TV, newspaper and product packaging are their most preferred information sources. Three clusters were identified. Firms should focus more on communicating what their firm’s commitment comprises and should adapt their CSR activities and firm communication methods to the preferences of different target groups. Keywords: ACA, agribusiness, cluster analysis, CSR, consumers’ preferences JEL codes: C1, Q13, M14 Corresponding author: henrike.luhmann@agr.uni-goettingen.de

321

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

1. Introduction Corporate social responsibility (CSR) has been an important topic worldwide for a long time and is receiving greater attention – both in research and in practice. From the scientific perspective, Bowen (1953), Davis (1960) and Carroll (1999) carried out path-breaking work in describing and developing the concept of CSR. As they define it, CSR denotes that a firm takes responsibility for its operations and thereby to meet the expectations of the society. From a consumer perspective, firms’ awareness of their social commitment is becoming increasingly essential. Consumers are focusing more and more on ecological and social aspects of supply chains and production processes when buying products (Hierholzer, 2010). But the mass media also plays an essential role in influencing public opinion and thus exerting pressure on companies for more transparency in their production and more social responsibility (Rommelspacher, 2012; Vanhonacker and Verbeke, 2014). Current research stresses the need for firms’ CSR activities to secure the legitimacy of their operations and their societal ‘license to operate’ (Luhmann and Theuvsen, 2016; Ross et al., 2015). Various industries are focusing more and more on CSR as they put social concerns on their agenda (Ranängen and Zobel, 2014). One industry that has come under high social pressure is the agribusiness sector. Livestock farming in particular has become a critical issue for public debate. There are a variety of reasons for this. Growing urbanization is accompanied by society’s growing estrangement from agricultural production (Albersmeier and Spiller, 2008; Böhm et al., 2009). Today’s agricultural value chains are characterized by high complexity that has caused information asymmetries for consumers. The non-communication of developments, especially in the meat industry, exacerbated by a number of food scandals, has created uncertainty for consumers and a widespread loss of trust in farmers and processors (Berk, 2012; Hierholzer, 2010; Spiller et al., 2010). As a result, there is a risk of growing resistance on the part of society to animalproducing companies. One industry particularly affected by these developments and a focal point of public concerns is the poultry sector. On the one hand, poultry production is a high-potential market. In 2015 the production volume in the EU-28 was 14.4 million tons of poultry products. Behind the USA and China, the EU is the third largest poultry producer in the world (Statista, 2016a). Germany is an emerging market in this regard. Consumption as well as production volume has been rising continuously since the 1990s. As a consequence, with a production volume of 1.51 million tons, Germany is now in fourth place in European poultry production – behind Poland (2.01 million tons), France (1.72 million tons) and the UK (1.69 million tons) (Federal Statistical Office, 2008; Statista, 2016b,c). Improvements in breeding, chicken husbandry and farm management have enhanced efficiency in production, leading to lower production costs than those involved in beef or pork production. The industry demonstrates a high potential for future competitiveness as well as the ability to meet future demand from the German market (BVDF, 2013; Spiller et al., 2010; Veauthier and Windhorst, 2011). The German market for poultry is characterized by oligopolistic structures and dominated mainly by large, vertically integrated enterprises. Four companies hold a market share of 75% of the total revenue (Four-firm concentration-ratio: € 4.2 billion) (Afz, 2012; Niemann, 2012). On the other hand, society is highly critical of poultry production. Protests against building new chicken barns and slaughterhouses or initiatives promoting higher animal welfare standards are accompanied by intense public and political debates. However, these developments in turn influence companies in the poultry sector to take ever greater responsibility for their actions (Berk, 2012; Niemann, 2012). As a result, companies in the poultry industry worldwide have become more sensitive to these issues and establish transparency campaigns or implement CSR strategies, through such means as focusing more on animal welfare, employee health and safety, supplier concerns, environmental issues and food safety (GNP, 2016; Heidemark GmbH, 2016; Keystone, 2016; PHW-Gruppe, 2016; Plukon Vertriebs GmbH, 2016). Despite the high relevance of the global poultry industry and the growing need to implement CSR strategies, scientific research in this area is still scarce. Amongst other things, there is a gap in research about consumers’ preferences of these commitments by poultry producing firms. Therefore, one central question of this study is this: what do consumers demand from firms taking social responsibility; that is, just what CSR commitment do they want International Food and Agribusiness Management Review

322

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

poultry firms to make? To analyze consumers’ preferences, we performed an Adaptive Conjoint Analysis (ACA) analyzing the CSR commitments favored by 250 consumers in Germany. Subsequently, a hierarchical cluster analysis is conducted to identify consumer groups with the help of their preferences regarding CSR commitments. Some initial management implications can be derived from the results. The remainder of this article is structured as follows: a literature review is presented in Section 2, followed by the study design and a statistical evaluation in Section 3. Section 4 describes the sample and the results of the survey. The study concludes with a discussion of the findings and some managerial implications and ideas for further research in Section 5.

2. Literature review CSR has the potential to help companies meet societal demands and criticism. It has been a research topic since the 1950s, when Bowen (1953), Davis (1960) as well as Carroll (1999) first described and defined the concept in the general management literature. But, despite several decades of research, there is no general definition of CSR or common understanding on what the concept covers (Dahlsrud, 2008). How companies perceive their responsibility towards society can vary considerably among sectors or within industries. More recently the concept has received growing attention by agricultural economists (e.g. Hartmann, 2011; Heyder and Theuvsen, 2012). Definitions concerning CSR in agribusiness widely parallel definitions from the common management literature on CSR (Luhmann and Theuvsen, 2016). Earlier research has focused on the effects of CSR and the determinants of a firm’s CSR commitment. Heyder and Theuvsen (2012) demonstrated that CSR efforts positively influence agribusiness firms’ reputation. Empirical results are different concerning their effects of CSR on firm profitability. Some studies have found positive effects (Heyder and Theuvsen, 2012; Rodriguez-Melo and Mansouri, 2011; Ross et al., 2015; Tang et al., 2012), whereas others have observed negative effects (Ingram and Frazier, 1980; Vance, 1975) or no effects at all (Peloza and Papania, 2008; Surroca et al., 2010). From a chain perspective, Hartmann (2011) identified various factors influencing an agribusiness firm’s CSR commitment. Especially economic considerations, such as improved competitiveness, are important reasons for implementing a CSR strategy. Another important determining factor is the size of a firm; the larger a firm, the more likely it is to engage in CSR. Likewise, firms with a strong brand are keener to implement a strong CSR commitment to protect that brand (Hartmann, 2011). These concerns particularly apply to the poultry sector, where some large companies with strong brands are already taking social responsibility or have implemented a CSR strategy by emphasizing issues like environmental protection, animal welfare, transparency in production processes and responsibility for their employees in order to counter public criticism challenging their reputations (i.a. PHW-Gruppe, 2016; Plukon Vertriebs GmbH, 2016). In this article, our understanding of CSR and the empirical analysis of consumer expectations follows Carroll’s (1991) pyramid model. According to this model, a company is said be socially responsible when it policies meet the four levels of economic, legal, ethical and philanthropic responsibility. Following Carroll (1991), a firm’s CSR is based on economic responsibility – establishing a strong market position, making profit or providing food products (Maignan, 2001; Maloni and Brown, 2006; Werner et al., 2009). Legal responsibility builds upon economic responsibility, as a firm acts within the law – by, for example, complying with regulations (Maignan, 2001; Maloni and Brown, 2006). Ethical responsibility applies to society’s norms and values. These requirements are expected by society. When a company does not comply with, it is sanctioned by society, resulting, among other things, in a deteriorating reputation. One important issue is animal welfare concerns for which societal demands surpass legal restrictions, such as not debeaking chicks, providing more freedom of movement than required and not shredding male laying hen chicks (ForsmanHugg et al., 2013; Maloni and Brown, 2006; Spiller et al., 2015). On the top of the CSR pyramid stands the philanthropic responsibility. Companies are seen as ‘good corporate citizens’ when they use their resources to support the local community. A philanthropic responsibility is desired but not required. Hence, if a firm defaults it is not sanctioned by the public (Carroll, 1991). Carroll’s (1979) pyramid model has served in International Food and Agribusiness Management Review

323

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

various studies as a basis for empirical research: in a study about CSR from a company’s perspective by Aupperle et al. (1985), the four responsibility levels from Carroll (1979) can be empirically confirmed. Research about CSR from a consumer perspective preferentially proves the general CSR perception of consumers. In most cases, studies analyze the influence of CSR on consumers’ buying behavior. Such studies do not follow Carroll’s (1991) pyramid model (cf. Berens et al., 2005; Brown and Dacin, 1997; Sen and Bhattacharya, 2001). In contrast, Maignan (2001) examined the four responsibility levels from the consumer perspective and developed items – or rather statements – based on Carroll’s research (1979). Maignan (2001) points out the various preferences for a firm’s CSR commitment depending on consumers’ country of origin. In a cross-country comparison of consumers from France, the USA and Germany, it was shown that, unlike people from the USA, consumers from Germany and France value commitments from the economic responsibility level less than the other three levels. German consumers rank the legal and ethical responsibilities of a firm as most important, followed by philanthropic responsibilities. Studies that focus on consumers and poultry production mainly examine individual aspects of a firms’ CSR. This research shows that consumers’ awareness focuses on food safety and risk perception (Badrie et al., 2006; Lobb et al., 2007; Stefani et al., 2008; Verbeke and Viaene, 2000) and therefore also on food scandals (Bánáti, 2011). For consumers, the use of antibiotics, diseases such as avian influenza and the use of growth hormones are the most important factors influencing their uncertainty about food safety (Goldberg and Roosen, 2008). One factor plays an important role in societal debates about poultry production as well as in consumers’ buying decisions: animal welfare (Burton et al., 2016). Consumers rate animal welfare in poultry production lower than in pork or beef production (Verbeke and Viaene, 2000). Shared values based on a firm’s CSR can help firms regain consumers’ trust in the food system. Additionally, information about production systems – depending on who receives the information and the recipients’ competence regarding it – is important for consumer’s trust (Arnot et al., in press). From the literature, it becomes obvious that consumers’ are mainly concerned about ethical matters, such as animal welfare or production issues (e.g. product quality and safety). The literature offers no information on consumers’ response to the philanthropic activities of poultry firms in their local community. In summary, studies that focus on CSR from a consumer perspective seldom consider Carroll’s (1979 and 1991) model of CSR. Research on consumers’ perceptions with relation to the poultry sector in particular fails to offer a comprehensive understanding of CSR. Literature in this area has focused more on individual aspects of CSR, such as animal welfare issues. Thus, consumers’ preferences for other CSR commitments on the part of poultry firms have not been investigated, indicating an important research gap.

3. Material and methods Study design In order to answer the research questions, we conducted an online questionnaire-based study in February 2013. We performed an ACA to evaluate consumers’ preferences regarding CSR activities and then conducted a cluster analysis to distinguish groups from a heterogeneous entity (Backhaus et al., 2008). The questionnaire comprised three sections: (1) demographic questions as income, age, housing situation and origin; (2) an ACA section comprising four steps to capture respondents’ preferences for CSR activities by poultry firms; and (3) questions about consumers’ knowledge about CSR and their attitudes toward and relation to agriculture or food production as well as their information behavior. The questionnaire contained mostly closed questions. A pretest was sent out first before the questionnaire was distributed to consumers all around Germany. 250 consumers completed the questionnaire. The demographic details – age, gender, education and origin – accord with the percentage distribution of the German population (Federal Statistical Office, 2012).

International Food and Agribusiness Management Review

324

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Statistical analysis The general premise of ACA methodology is that a subject contains a number of attributes and attribute levels. The four responsibility levels from Carroll’s (1991) pyramid comprise the attribute used to describe CSR: economic, legal, ethical and philanthropic responsibility. Through an extensive literature analysis, we explored 24 CSR commitments, which we then used in the form of statements to delineate each of the four responsibility levels. These CSR activities were operationalized as the attribute levels. For each CSR commitment, a part-worth utility was calculated, which shows respondents’ preferences for each commitment (Baier and Brusch, 2009). The online survey was designed using ACA 8.2 from Sawtooth Software (Orem, UT, USA). The ACA comprised four steps: in the first phase, respondents evaluated all CSR activities conducted by poultry companies for each responsibility level. In the second step, the importance of each attribute level (namely, the responsibility level) was captured. In the third step – the paired comparison – the program estimated the importance of each attribute, namely, CSR commitment (e.g. ‘Chicken meat has the quality I appreciate’). In the last step, the consistency of respondents’ answers was verified (Dietz, 2007; Harth, 2006; Reiners, 1996; Sawtooth Software, 2007). Before analyzing the sample results, the preference data from the ACA were estimated using the hierarchical Bayesian method in SSI Web 8.2 (Sawtooth Software Orem, UT, USA). Afterwards, they were merged with the survey results, and all analyses were made using IBM SPSS Statistics 23 (IBM, Armonk, NY, USA) (Baumgartner and Steiner, 2009; Sawtooth Software, 2006). The part-worth utilities for the CSR activities were estimated for each individual. To be comparable, the part-worth utilities were normalized using the zero – centered diffs method. The utilities were centered on zero and ranged in this study from -45.87 to 40.26. These part-worth utilities are significant based on metric scaling as long as one of the utilities for a specific CSR activity is twice the number of another. The utilities indicate how much the CSR commitments are preferred. Thus, a negative value does not mean that respondents would refuse a specific CSR activity but that they prefer others (Harth, 2006). Next, we conducted a hierarchical agglomerative cluster analysis to merge the respondents into homogenous groups. With regard to consumers’ preferences the most preferred aspects as well as the least preferred commitments of a firm’s CSR strategy (from each responsibility level) were used to separate the survey into groups; to this end, we included eight variables in the cluster analysis. At first groups with the lowest distance were combined using the single-linkage procedure; four outliers were identified, leaving 246 respondents in the sample. To identify the optimal cluster result, we used then the Ward method to combine respondents, which minimally increased the variance in the cluster group. The elbow criteria suggested three clusters. By using the k-means method, we determined the solution in six iterations. Finally, discriminant analysis confirmed that 98.8% of the cases had been classified correctly (Backhaus et al., 2008). To characterize the sample, descriptive statistics such as frequency distribution, crosstab, mean (µ) and standard deviation (±) were used (Backhaus et al., 2008). Sample description Of the 246 respondents to the survey, 124 were male, and 122 female. Five age groups are represented: 19.9% of the respondents are between 18 and 29, 17.5% are between 30 and 39, 24.8% are between 40 and 49, 21.1% are between 50 and 59 years old, and 16.7% are 60 or older. Of the respondents, 42.7% have a Certificate of Secondary Education and 30.1% a General Certificate of Secondary Education, while 27.2% finished school with an advanced technical college or university entrance qualification. These demographic details accord with the distribution of the population in Germany (Federal Statistical Office, 2012). Most consumers (52.1%) in the sample live in mid-sized cities of up to 50,000 residents; 41.5% of the respondents live in larger cities of between 50,000 and 1 million residents; whereas only 6.5% of interviewees live in a metropolis with more than 1 million inhabitants. More than half the respondents in the survey (58.9%) live with a partner and/or with children.

International Food and Agribusiness Management Review

325

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

Concerning meat consumption, 95.1% of the consumers in the survey eat meat1; 52% of them eat meat more than once a week, and another 30.1% say they eat meat exactly once a week. Pork is the preferred kind of meat (41.5%), with poultry in second place (37%). 14.2% of the interviewees say that they prefer beef the most.

4. Results

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Descriptive results Of the respondents, 40.7% state that they have no connection with agricultural production. When a meat scandal happens (e.g. antibiotic resistant germs in chicken meat), 21.1% of the respondents would trust a farmer as a source of information about the scandal. Most consumers (72%) would trust consumer protection organizations (e.g. the German organization Stiftung Warentest), but none agreed that politicians are trustworthy. If consumers want to know more about the CSR commitment of a chicken-producing firm, they prefer to watch a television documentary (61%), read a newspaper (59.3%) or read information on product packaging (45.9%). One central goal in this study was to measure consumers’ preferences for CSR commitment from poultry producing companies. Table 1 shows the part-worth utilities for each commitment according to the responsibility level. From the ACA, it is clear that on the economic responsibility level, the respondents most valued the quality of the chicken meat, whereas they least valued that long-term business plans are important for a strong market position and to increase profit. On the legal responsibility level, they considered most important the absence of prohibited substances, such as dioxin or antibiotics, in the meat and least important that a firm have a no tolerance policy regarding corruption. Concerning ethical responsibility, the most favored was increased animal welfare beyond applicable law and the least preferred that chickens are completely utilized. On the philanthropic responsibility level, the support of employees returning to work after illness or pregnancy was most preferred by the respondents, whereas they least favored firms’ support of culture and historical monuments. These eight CSR commitments – the most and least preferred commitments on each CSR level – were used to form the clusters that are described in the next section. Results of the cluster analysis Based on the descriptive results of the ACA, the cluster analysis distinguished three clusters. The results show that the groups differ in their preference structure and in some other characteristics. In the survey, we tested preferences for 24 CSR commitments – six from each CSR level. Below, the 12 most preferred commitments are shown for each cluster, and the clusters are described by their characteristics. Cluster 1 contains the 88 respondents (35.8% of the sample) whose preferences are mostly in the economic responsibility level (Figure 1). For this group, activities on the economic responsibility level (cf. CSR commitments no. 2, 3, 6, 10) fall into the top preference category, followed by three activities from the ethical (cf. CSR activity no. 5, 7, 11) and three from the philanthropic responsibility level (cf. CSR commitments no. 1, 4, 12). From the legal responsibility level, two CSR activities (cf. CSR commitment no. 8, 9) were among the 12 most preferred CSR measures in this cluster group. These respondents placed the most value on firms’ commitment to support employees returning to work after illness or pregnancy (philanthropic responsibility), chicken meat having the quality consumers appreciate and that animals are both kept and slaughtered in Germany (both economic responsibility).

The sample size for information about meat or meat consumption was n=234.

International Food and Agribusiness Management Review

326

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

Table 1. CSR commitments and part-worth utilities (n=246). 1 Responsibility level

CSR commitments

Part-worth utility2 µ

Legal responsibility

Ethical responsibility

Philanthropic responsibility

1 2

Chicken meat has the quality I appreciate Keeping animals and slaughtering poultry takes place in Germany Long-term rather than short-term employment Availability of apprenticeships and jobs to strengthen the German employment market Cheap prices for chicken meat through the use of modern manufacturing Long-term business plans for a strong market position and to increase profit No prohibited substances, such as dioxin or antibiotics, in the chicken meat Chicken health checks and meat inspections by a veterinarian Compliance with hygiene regulations Adherence to animal and environmental protection laws Chickens receive medications according to prescription No corruption Increased animal welfare beyond applicable law Short transfer distances from farm to slaughterhouse Fair working conditions and wages for all employees Chicken barn with a minimum impact on the environment Willingness to answer customers’ questions and to make full disclosure about all parts of the production process Preferably complete utilization of the chicken Support of employees returning to work after illness or pregnancy Assistance for employees in education and training Free child care for employees’ children during working hours Cooking classes in schools to provide responsible exposure to food Support of people in need in the local community Support of culture and historical monuments

40.26 21.73 9.08 2.86

27.95 32.41 28.42 28.75

-30.32 -43.62 17.89 9.33 4.70 1.13 0.09 -33.14 28.38 6.64 3.10 -7.76 -12.03

34.42 26.91 21.23 21.41 23.66 22.26 26.31 30.90 27.82 30.53 31.42 30.60 33.57

-18.32 28.67 24.65 1.14 -0.47 -8.11 -45.87

36.82 24.18 21.46 27.43 31.56 25.04 28.66

CSR = corporate social responsibility. µ = mean; ± = standard deviation. 1) Support for employees returning to work 2) Product quality

CSR commitments

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Economic responsibility

30.61

3) Keeping/slaughtering in Germany

25.46

4) Educational assistance for employees

25.29

5) Increased animal welfare

37.47

15.04

6) Long-term employment

12.63

7) Complete utilization of the chicken

12.61

8) No prohibited substances

11.77

9) Chicken health checks and meat inspection

6.34

10) Strengthening employment market

4.51

11) Short transport distances

4.50

12) Free child care for employees' children

3.43

Part-worth utilities

Figure 1. The 12 most preferred CSR activities by Cluster 1. CSR = corporate social responsibility. International Food and Agribusiness Management Review

327

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

All respondents were asked about their knowledge of CSR. There was no significant difference within or between the three groups, but the findings still give an overall impression of available knowledge. In comparison to the other groups, respondents in this cluster knew the least about CSR: Only 23.9% of the respondents knew what CSR is or can define it. To inform them about CSR, there are various opportunities for respondents to choose. Most of the consumers in this cluster said they would prefer to watch television documentaries about a firm’s commitment. There was a significant difference between the clusters concerning this source of information (0.009**). Their next preference was to read about CSR on product packaging. The second cluster group contains 77 respondents (31.3%) and can be characterized by their focus on commitments on the legal responsibility level (cf. CSR activity nos. 3, 7, 8, 10, 11). These were followed by three activities from the economic (cf. CSR activity nos. 1, 4, 12) and two each from the ethical (cf. CSR activity nos. 2, 9) and the philanthropic (cf. CSR activity nos. 5, 6) responsibility levels. Figure 2 shows the part-worth utilities of the 12 most preferred CSR activities within this group, to which 77 respondents (31.3%) belong. Overall, respondents belonging to Cluster 2 considered the most important CSR measures commitment to chicken meat having the quality consumers appreciate (economic responsibility), implementation of increased animal welfare beyond applicable law (ethical responsibility) and that no prohibited substances, such as dioxin or antibiotics, are in the chicken meat (legal responsibility). Concerning age there is no significant difference between Cluster 2 and the other groups, but 45.5% of the respondents in this group are middle-aged (40-59). There was a significant difference in gender between the 2 3

Nominal scales’ significances tested with Chi-square. Level of significance: P<0.05 significant*; P<0.01 very significant**; P<0.001 highly significant***.

1) Product quality

59.10

2) Increased animal welfare

36.64

3) No prohibited substances CSR commitments

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Overall, Cluster 1 contains the youngest consumers. 44.3% of this group are under 40, and 25% under 30. For most of the respondents in this group, the total available income per month for the whole household is less than € 2,000. Only 19.3% of respondents in this group live alone; 61.5% live with a partner and/or children. More than half of this group is male (54.5%); a significant difference between the clusters regarding gender was observed2 (0.004**3). Respondents in this cluster live mainly in mid-sized cities between 10,000 and 250,000 citizens (52.3%) and 17% have regular contact with some kind of agricultural production; 45.5% of them report that they have no connection with farming.

27.40

4) Keeping/slaughtering in Germany

21.99

5) Support for employees returning to work

19.08

6) Educational assistance for employees

18.98

7) Chicken health checks and meat inspection

15.08

8) Compliance with hygienic rules

8.99

9) Short transport distances

4.48

10) Adherence to animal and environmental protection laws

4.19

11) Medications according to prescription

1.81

12) Long-term employment -2.88 -10

Part-worth utilities

Figure 2. The 12 most preferred CSR activities by Cluster 2. CSR = corporate social responsibility. International Food and Agribusiness Management Review

328

Volume 20, Issue 3, 2017

three clusters (0.004**); much more than half of this cluster is female (64.9%). Respondents’ income tended to be higher in this cluster compared with the other clusters; 52% had a total household income between € 2,000 and € 3,999, and 10.4% of interviewees in this cluster declared a monthly income of € 4,000 or more. Respondents in this cluster people also mainly live with a partner and/or with children (61.1%) and predominantly in smaller cities of less than 50,000 inhabitants, and 35.1% of this cluster live in cities with fewer than 10,000 people. In this group, more people have a connection with agricultural production; only 33.8% stated that they have no contact with agriculture. Of respondents in this cluster, 35.1% said they have friends or family members who are connected with agricultural production or even live on a farm, and 33.8% of this cluster stated they have visited a farm, go to farm shops or spend their holidays on farms. Moreover, 18.2% of them are regularly in contact with agricultural production. All in all, respondents did not know much about the CSR concept. However, in Cluster 2, 32.5% of the respondents stated that they have an understanding of the concept. This is proportionally more than in all the other groups. They reported that they would not actively seek information about CSR. Most respondents in Cluster 2 would acquire information about CSR randomly, such as if they were to read something about it in the newspaper (0.032*). Their second preference would be watching a television documentary about CSR. Cluster 3 contains 81 respondents (32.9%). Respondents in this group preferred most CSR commitments from the economic responsibility level (cf. CSR activity nos. 2, 6, 7, 10), followed by three CSR commitments from the ethical (cf. CSR activity no. 1, 5, 9) and three from the philanthropic responsibility level (cf. CSR activity no. 3, 4, 11). Two commitments came from the legal responsibility level (cf. CSR activity no. 8, 12). Figure 3 shows the part-worth utilities of the 12 most preferred CSR commitments for this group. Consumers in this group most preferred that firms’ focus on increased animal welfare beyond applicable law (ethical responsibility), that chicken meat has the quality consumers appreciate (economic responsibility) and that companies’ focus on assistance for employees in education and training (philanthropic responsibility). This group contains only older people; 50.6% of the group is 50 and older; 22.5% of Cluster 3 is 60 and older. Of respondents in this cluster, 24.7% live alone, and 54.3% live with a partner and/or their children. The total income for more than half of this cluster (64.2%) lies between € 1,000 and € 2,999. Roughly a third

1) Increased animal welfare

35.02

2) Product quality

32.82

3) Educational assistance for employees CSR commitments

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Luhmann and Theuvsen

29.35

4) Support for employees returning to work

28.22

5) Fair working conditions and wages

19.28

6) Keeping/slaughtering in Germany

17.44

7) Long-term employment

16.60

8) No prohibited substances

15.49

9) Short transport distances

11.01

10) Strengthening employment market

10.23

11) Free child care for employees' children

9.53

12) Chicken health checks and meat inspection

7.11

Part-worth utilities

Figure 3. The 12 most preferred CSR activities by Cluster 3. CSR = corporate social responsibility. International Food and Agribusiness Management Review

329

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

(34.65%) of the respondents lives in larger cities of 250,000 inhabitants or more. In the third Cluster, 42% of respondents have some connection to agricultural production, but only 12.3% have regular contact as a farmer or through neighbors or going horseback riding.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Of respondents in Cluster 3, 70.4% know nothing about CSR and would most prefer to catch up on this topic by watching television documentaries (0.009**) or, secondly, by reading something randomly in a newspaper about it (0.032*).

5. Discussion and conclusions The first goal of this study was to analyze consumers’ preferences for poultry firms’ CSR commitment. The results show in general that consumers focus mainly on the product quality but also on animal welfare aspects. This confirms previous consumer studies which have stated that features of product quality, such as fat percentage, color and country of origin as well as animal welfare are important factors in consumers’ buying decisions (Burton et al., 2016; Goldberg and Roosen, 2008). Another strong emphasis is employeerelated issues, such as educational training or helping employees who are returning to work. These are new factors, which were not identified in earlier studies. Considering the results of the ACA (Table 1), it can be stated that consumers do not rank the commitments from each CSR level equally. They rank measures from the economic level (product quality) first; in second and third place, they put commitments from the philanthropic level (support of employees); and in fourth place, they again place an economic consideration (country of origin of the products). CSR activities that do not directly affect respondents (e.g. long-term business plans to gain strong market position and maximize long-term profit) are preferred less. This indicates a discrepancy between the various CSR aspects derived from Carroll’s (1991) framework and consumers’ preferences for a firm’s CSR obligation, but it is in line with Maignan’s (2001) results from German consumers and their CSR perception. These results also support De Quevedo-Puente et al. (2007), who determined that CSR obligations are the result of a process of exchange between society’s demands and a company’s reaction. This process does not reach a consensus and, therefore, varies for each country and society. The second goal of this study was to investigate different groups with the help of consumers’ preferences for various CSR commitments made by poultry firms. This is a new focus in consumer studies and has not been studied in earlier research. High standard deviations showed up in the descriptive analysis, supporting the assumption that there were different groups in the survey sample. Hence, a hierarchical cluster analysis was performed and three groups were distinguished. All three groups preferred specific CSR commitments that were important to them with different foci on the various responsibility levels distinguished by Carroll (1991). Cluster 1, containing the youngest respondents, has a low average income and most favors firms’ economic CSR commitment, especially what a firm does for its employees. The oldest respondents with a middle income are in Cluster 3. In its ranking of the CSR commitments, Cluster 2 differs from the other groups. By a wide margin, this group most prefers product quality commitments. Respondents in this group have the highest average income, are middle aged and especially value food quality and safety as well as animal welfare issues. This cluster contains mainly women. These findings parallel earlier research in such areas as consumer preferences for organic food and more animal welfare-friendly products. Beardworth et al. (2002), for instance, show that women are more aware of food quality issues and ethical concerns regarding production. Lueth (2005) confirms that social context highly influences preference for organic food, and income is especially suited for differentiating among the characteristics of various target groups. Hence, due to budget constraints, lower income groups often focus more on economic aspects than on process quality attributes such as organic or animal welfare-friendly. In contrast to earlier findings (cf. Lueth, 2005), Cluster 1 – the cluster with the lowest income – mostly focused on employee issues but also highly prefer issues regarding product quality and a firm’s commitment to animal welfare.

International Food and Agribusiness Management Review

330

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

Interesting implications can be derived from these findings for the design of CSR activities, not only in the poultry industry but also in other agribusiness subsectors. First, companies should tailor a communication strategy for their CSR commitments specifically for each target group. Furthermore, the results indicate that consumers focus mostly on a firm’s commitment but not exclusively on one responsibility level. If a company has a product for a specific target group differentiated by age or personal development status, this company could focus its CSR activities most on what this age or development group prefers. Nevertheless, Carroll (1991) showed that it is important to fulfill all four responsibility levels. Therefore, a company should address all the responsibility levels but can target its activities or give preference to a particular responsibility level depending on its target market. What is also important for companies is to communicate their obligation. The findings show that CSR awareness is rather low throughout the entire sample. There is no relationship between group characteristics and knowledge of CSR. Consumers’ knowledge influences how they perceive and search for information. In a broader sense consumers’ attitudes and buying behaviors can be influenced (Alba and Hutchinson, 1987). In Wigley’s (2008) study it can be shown how consumers’ knowledge about a firm’s CSR efforts their buying behavior. The greater consumers’ knowledge about a company’s obligation the more positive is respondents’ attitude towards CSR. Reports or communication on websites can help increase the CSR knowledge of consumers and, thus, contribute to a more positive response to CSR activities. These information channels should be chosen specifically for each target group. Studies’ limitations stem first of all from the sample itself. First, the sample is biased to a certain extent because only consumers who could be reached online could participate. To minimize this bias, quotas were set to ensure a percentage distribution with regard to gender, age, education and place of domicile similar to that of the German population as a whole. Second, the sample size is comparatively small as it contains only 250 respondents. Therefore, results should be viewed as initial insights into the research topic, and future research is needed. Additionally, another framework for analyzing consumers’ preferences could be used. For this study we adopted Carroll’s (1991) four responsibility levels as our theoretical framework. It is a well-known framework for analyzing CSR, but it was developed in the context of US society. Since Schoenheit (2006) and Backhaus-Maul and Braun (2007) explain that CSR differs in relation to its social context and there are great cultural differences between the Anglo-Saxon and the Continental European context (Hofstede, 1984), future studies should try to better take into account the cultural context of CSR. Future research should also address in greater detail whether consumers’ different preferences for CSR activities are sector-specific or whether distinctions between clusters are stable across various industry subsectors. Finally, it could be interesting to see to what extent the differences regarding preferences for CSR influence consumption patterns, consumers’ willingness-to-pay for product attributes and actual buying behavior.

References Alba, J.W. and J.W. Hutchinson. 1987. Dimensions of consumer expertise. Journal of Consumer Research 1: 411-454. Albersmeier, F. and A. Spiller. 2008. Supply Chain Reputation in der Fleischwirtschaft. Discussion paper (0811). Department of Agricultural Economics and Rural Development, University of Goettingen, Goettingen, Germany. Allgemeine fleischer zeitung (Afz). 2012. Die Top 150 der Fleischbranche. Allgemeine fleischerzeitung 38: 19-22. Arnot, C., Y. Vizzier-Thaxton and C.G. Scanes. In press. Values, trust and science-building trust in today’s food system in an era of radical transparency. Poultry Science. DOI: https://doi.org/10.3382/ps/pew168. Aupperle, K.E., A.B. Carroll and J.D. Hatfield. 1985. An empirical investigation of the relationship between corporate social responsibility and profitability. Academy of Management Journal 28: 446-463. Backhaus, K., B. Erichson, W. Plinke and R. Weiber. 2008. Multivariate Analysemethoden: Eine anwendungsorientierte Einführung. Springer, Berlin, Germany. Backhaus-Maul, H. and S. Braun. 2007. Gesellschaftlichen Engagement von Unternehmen in Deutschland. Konzeptionelle Überlegung und empirische Befunde. Rote Seiten – Stiftung und Sponsoring 10: 1-15.

International Food and Agribusiness Management Review

331

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

Badrie, N., A. Gobin, S. Dookeran and R. Duncan. 2006. Consumer awareness and perception to food safety hazards in Trinidad, West Indies. Food Control 17: 370-377. Baier, D. and M. Brusch. 2009. Erfassung von Kundenpräferenzen für Produkte und Dienstleistungen. In: Conjointanalyse – Methoden – Anwendung – Praxisbeispiele, edited by D. Baier and M. Brusch. Springer Verlag, Berlin, Germany, pp. 3-18. Bánáti, D. 2011. Consumer response to food scandals and scares. Trends in Food Science and Technology 22: 56-60. Baumgartner, B. and W.J. Steiner. 2009. Hierarchisch bayesianissche Methoden bei der Conjointanalyse. In: Conjointanalyse – Methoden – Anwendung – Praxisbeispiele, edited by D. Baier and M. Brusch. Springer Verlag, Berlin, Germany, pp. 147-159. Beardsworth, A., A. Bryman, T. Keil, J. Goode, C. Haslam and E. Lancashire. 2002. Women, men and food: the significance of gender for nutritional attitudes and choices. British Food Journal 104: 470-491. Berens, G., C.B.M. van Ries and G.H. van Bruggen. 2005. Corporate associations and consumer product responses: the moderating role of corporate brand dominance. Journal of Marketing 69: 35-48. Berk, J. 2012. Technik in der Geflügelhaltung. In: Jahrbuch Agrartechnik 2012, edited by L. Frerichs. Institut für mobile Maschinen und Nutzfahrzeuge, Braunschweig, Germany, pp. 1-9. Böhm, J., F. Albersmeier, A. Spiller and A. Zühlsdorf. 2009. Zukunftsfaktor gesellschaftliche Akzeptanz: Kommunikation mit der Öffentlichkeit – Mehr als Krisen-PR. In: Die Ernährungswirtschaft im Scheinwerferlicht der Öffentlichkeit, edited by J. Böhm, F. Albersmeier and A. Spiller. Eul Verlag, Lohmar, Germany, pp. 5-16. Bowen, H.R. 1953. Social responsibilities of the businessman. Harper and Row, New York, NY, USA. Brown, T.J. and P.A. Dacin. 1997. The company and the product: corporate associations and consumer product responses. Journal of Marketing Research 61: 68-84. Bundesverband der deutschen Fleischwarenindustrie (BVDF). 2013. Fleischverbrauch und Fleischverzehr je Kopf der Bevölkerung. Available at: http://tinyurl.com/gs3hhqs. Burton, E., J. Gatcliffe, H.M. O’Neill and D. Scholey. 2016. Sustainable poultry production in Europe. CABI, Oxfordshire, UK. Carroll, A.B. 1979. A three-dimensional conceptual model of corporate performance. The Academy of Management Review 4: 497-505. Carroll, A.B. 1991. The pyramid of corporate social responsibility: toward the moral management of organizational stakeholders. Business Horizons 34: 39-48. Carroll, A.B. 1999. Corporate social responsibility: evolution of a definitional construct. Business and Society 38: 268-295. Dahlsrud, A. 2008. How corporate social responsibility is defined: an analysis of 37 definitions. Corporate Social Responsibility and Environmental Management 15: 1-13. Davis, K. 1960. Can business afford to ignore social responsibilities? California Management Review 2: 70-76. De Quevedo-Puente, E., J.M. De la Fuente-Sabaté and J.B. Delgado-García. 2007. Corporate performance and corporate reputation: two interwoven perspectives. Corporate Reputation Review 10: 60-72. Dietz, W. 2007. Grundlagen der Conjoint-Analyse: Varianten, Vorgehensweise, Anwendungen. AV Akademikerverlag, Saarbruecken, Germany. Federal Statistical Office. 2008. Vom Erzeuger zum Verbraucher – Fleischversorgung in Deutschland Ausgabe 2008. Available at: http://tinyurl.com/ztltlkt. Federal Statistical Office. 2012. Statistisches Jahrbuch 2012. Available at: http://tinyurl.com/gtzcqmc. Forsman-Hugg, S., J.M. Katajajuuri, I. Riipi, J. Mäkelä, K. Järvelä and P. Timonen. 2013. Key CSR dimensions for the food chain. British Food Journal 115: 30-47. GNP. 2016: Social responsibility. Available at: http://tinyurl.com/hvmnsgu. Goldberg, I., and J. Roosen. 2008. Verbraucherbewertung der Qualität und Sicherheit von Geflügelfleisch. In: Zukunftsperspektiven der Fleischwirtschaft, edited by A. Spiller and B. Schulze. Universitätsverlag, Goettingen, Germany, pp. 341-352. Harth, M. 2006. Multikriterielle Bewertungsverfahren als Beitrag zur Entscheidungsfindung in der Landnutzungsplanung unter besonderer Berücksichtigung der Adaptiven Conjoint-Analyse und des Discrete Choice Experiments. PhD thesis, University of Halle-Wittenberg, Halle, Germany. International Food and Agribusiness Management Review

332

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

Hartmann, M. 2011. Corporate social responsibility in the food sector. European Review of Agricultural Economics 38: 297-324. Heidemark GmbH. 2016. Engagement. Available at: http://tinyurl.com/jqcdeyz. Heyder, M. and L. Theuvsen. 2012. Determinants and effects of corporate social responsibility in German agribusiness: a PLS model. Agribusiness 28: 400-428. Hierholzer, V. 2010. Nahrung nach Norm: Regulierung von Nahrungsmittelqualität in der Industrialisierung 1871-1914 – Kritische Studien zur Geschichtswissenschaft. Vandenhoeck and Ruprecht, Goettingen, Germany. Hofstede, G. 1984. Culture’s Consequences: International Differences in Work-Related Values. Sage, Newbury Park, CA, USA. Ingram, R.W. and K.B. Frazier. 1980. Environmental performance and corporate disclosure. Journal of Accounting Research 18: 614-622. Keystone. 2016: Corporate social responsibility. Available at: http://tinyurl.com/h43m7zw. Lobb, A.E., M. Mazzocchi and W.B. Traill. 2007. Modelling risk perception and trust in food safety information within the theory of planned behaviour. Food Quality and Preference 18: 384-395. Lueth, M. 2005. Zielgruppensegmente und Positionierungsstrategien für das Marketing von PremiumLebensmitteln. PhD thesis, University of Goettingen, Goettingen, Germany. Luhmann, H. and L. Theuvsen. 2016. corporate social responsibility in agribusiness: literature review and future research directions. Journal of Agricultural and Environmental Ethics 29: 673-696. Maignan, I. 2001. Consumer’s perceptions of corporate social responsibility: a cross-country analysis. Journal of Business Ethics 30: 57-72. Maloni, M.J. and M.E. Brown. 2006. Corporate social responsibility in the supply chain: an application in the food industry. Journal of Business Ethics 68: 35-52. Niemann, E. 2012. Die ‘Hähnchenblase’: Über die Krise der Geflügel-Agrarindustrie und eine erstarkende Bürgerbewegung. Der kritische Agrarbericht 2012: 122-125. Peloza, J. and L. Papania. 2008. The missing link between corporate social responsibility and financial performance: stakeholder salience and identification. Corporate Reputation Review 11: 169-181. PHW-Gruppe. 2016. Gesellschaftliche Verantwortung. Available at: http://tinyurl.com/jsdnk8o. Plukon Vertriebs GmbH. 2016. Nachhaltigkeit. Available at: http://www.stolle.de/nachhaltigkeit. Ranängen, H. and T. Zobel. 2014. Revisiting the ‘how’ of corporate social responsibility in extractive industries and forestry. Journal of Cleaner Production 84: 299-312. Reiners, W. 1996. Multiattributive Präferenzstrukturmodelle durch die Conjoint Analyse. Diskussion der Verfahrensmöglichkeit und Optimierung von Paarvergleichsaufgaben bei der adaptiven Conjoint Analyse. PhD thesis, University of Muenster, Muenster, Germany. Rodriguez-Melo, A. and S.A. Mansouri. 2011. Stakeholder engagement: defining strategic advantage for sustainable construction. Business Strategy and the Environment 20: 539-552. Rommelspacher, M. 2012. Corporate Social Responsibility aus Konsumentensicht. Gabler Verlag, Wiesbaden, Germany. Ross, R.B., V. Pandey and K.L. Ross. 2015. Sustainability and strategy in U.S. agri-food firms: an assessment of current practices. International Food and Agribusiness Management Review 18: 17-48. Sawtooth Software. 2006. The ACA/Hierarchical Bayes v3.0. Available at: http://tinyurl.com/jm96tnr. Sawtooth Software. 2007. The ACA/Web v6.0. Available at: http://tinyurl.com/jm96tnr. Schoenheit, I. 2006. Corporate Social Responsibility – Deutsche und internationale Perspektiven am Beispiel China. In: Corporate Social Responsibility – Verantwortung für nachhaltiges Wirtschaften in China, edited by I. Schoenheit, W. Iwand and R. Kopp. Beuth Verlag, Berlin, Germany, pp. 9-24. Sen, S. and C.B. Bhattacharya. 2001. Does doing good always lead to doing better? Consumer reactions to corporate social responsibility. Journal of Marketing Research 38: 225-243. Spiller, A., B. Schulze and A. Cordts. 2010. Was essen wir in zwanzig Jahren? DLG-Mitteilungen 1: 20-23. Spiller, A., M. Gauly, A. Balmann, J. Bauhus, R. Birner, W. Bokelmann, O. Christen, S. Entenmann, H. Grethe, U. Knierim, U. Latacz-Lohmann, J. Martinez, H. Nieberg, M. Qaim, F. Taube, B.-A. Tenhagen and P. Weingarten. 2015. Wege zu einer gesellschaftlich akzeptierten Nutztierhaltung. Berichte über Landwirtschaft-Zeitschrift für Agrarpolitik und Landwirtschaft, Sonderheft 221. International Food and Agribusiness Management Review

333

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0156 - Tuesday, May 09, 2017 8:29:14 PM - IP Address:24.21.169.207

Luhmann and Theuvsen

Volume 20, Issue 3, 2017

Statista. 2016a. Produktion der führenden Erzeugerländer und -regionen von Geflügelfleisch weltweit in den Jahren 2013 bis 2015. Available at: https://de.statistica.com. Statista. 2016b. Produktion von Geflügelfleisch in der Europäischen Union nach Ländern in den Jahren 2012 bis 2015. Available at: https://de.statista.com. Statista. 2016c. Produktion von Geflügelfleisch in Deutschland und in der Europäischen Union in den Jahren 1990 bis 2015. Available at: https://de.statista.com. Stefani, G., A. Cavicchi, D. Romano and A.E. Lobb. 2008. Determinants of intention to purchase chicken in Italy: the role of consumer risk perception and trust in different information sources. Agribusiness 24: 523-537. Surroca, J., J.A. Tribo and S. Waddock. 2010. Corporate responsibility and financial performance: the role of intangible resources. Strategic Management Journal 31: 463-490. Tang, Z., C.E. Hull and S. Rothenberg. 2012. How corporate social responsibility engagement strategy moderates the CSR-financial performance relationship. Journal of Management Studies 49: 1274-1303. Vance, S. 1975. Are socially responsible corporations good investment risks? Management Review 64: 18-24. Vanhonacker, F. and W. Verbeke. 2014. Public and consumer policies for higher welfare food products: challenges and opportunities. Journal of Agricultural and Environmental Ethics 27: 153-171. Veauthier, A. and H.-W. Windhorst. 2011. Die Wettbewerbsfähigkeit der Deutschen und Niedersächsischen Geflügelfleischerzeugung – Gegenwärtige Strukturen und Prognosen bis 2020. Vechtaer Druckerei und Verlag, Vechta, Germany. Verbeke, W.A., and J. Viaene. 2000. Ethical challenges for livestock production: meeting consumer concerns about meat safety and animal welfare. Journal of Agricultural and Environmental Ethics 12: 141-151. Werner, C., S. Janisch, U. Kuembet and M. Wicke. 2009. Comparative study of the quality of broiler and turkey meat. British Poultry Science 50: 318-324. Wigley, S. 2008. Gauging consumers’ responses to CSR activities: does increased awareness make cents? Public Relations Review 34: 306-308.

International Food and Agribusiness Management Review

334

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2016.0033

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

Received: 18 February 2016 / Accepted: 23 December 2016

Pre-employment costs associated with H-2A agricultural workers and the effects of the ‘60-minute rule’ RESEARCH ARTICLE Fritz M. Roka a, Skyler Simnittb, and Derek Farnsworthc aAssociate

Professor, University of Florida, Institute of Food and Agricultural Sciences,Southwest Research and Education Center, 2685 State Road 29 North, Immokalee, FL 34142, USA bGraduate

Student and cAssistant Professor, University of Florida, Institute of Food and Agricultural Sciences, Food and Resource Economics Department, P.O. Box 110240, McCarty B 1109, Gainesville, FL 32611, USA

Abstract Agricultural employers increasingly are turning to the foreign guest worker program, known as H-2A, as a means to secure a legal workforce. This paper outlines the procedural aspects and costs of recruiting and hiring H-2A workers. Cost data is from a 2014 survey of citrus harvesters and defines pre-employment costs as filing fees, advertising, surety bonds, travel, and housing. The pre-employment costs associated with guest workers are estimated to be nearly $ 2,000 per worker. The survey was motivated by the ‘60-minute rule’ imposed by the U.S. Department of Labor prior to the 2012-13 citrus harvesting season. Cost data were collected across two crop season, 2012-13 and 2013-14, to analyze the cost implications of the rule. We found that the 60-minute rule significantly increased filing fees. These fees, however, represent a very small share of total costs and overall pre-employment costs associated with the H-2A program did not significantly change. Keywords: farm labor, guest workers, citrus harvesting JEL code: Q12 Corresponding author: fmroka@ufl.edu

335

Roka et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

1. Introduction Fruit and vegetable growers in the United States depend on a significant number of workers to grow and harvest their crops. On average, 42% of variable production costs for U.S. fruit and vegetable producers are attributable to labor (Calvin and Martin, 2010). This observation is particularly relevant to Florida’s citrus industry, where every piece of fruit is harvested by hand (Roka, 2013). A shrinking supply of domestic workers and more aggressive enforcement of U.S. immigration laws have pushed Florida citrus growers to recruit an increasing number of foreign agricultural guest workers through the H-2A visa program (Taylor, 2010; Taylor and Charlton, 2012). During the 2014-15 season, more than 50% of Florida citrus crop was harvested by foreign guest workers hired through the H-2A program (M. Carlton, Director of Labor Relations, Florida Fruit and Vegetable Association, personal communication 2015). H-2A workers can be hired directly by growers or by licensed farm labor contractors (FLCs). FLCs are employers and historically have provided most of the labor to harvest Florida’s citrus production. Prior studies indicate that many FLC’s are either recent immigrants or at least share the language and cultural background with the workers they supervise (Huffman, 2005). As the H-2A program has expanded many FLCs have leveraged their knowledge of Mexican and other Central American labor markets to become employers of H-2A workers. This article provides an updated perspective on how citrus growers navigate the H-2A recruitment process and how they manage pre-employment expenses associated with recruiting and hiring foreign agricultural workers. We report and interpret the findings from a 2014 survey of licensed FLCs. The survey was motivated by a change in the U.S. Department of Labor’s (DOL) definition of ‘area of intended employment’ for which individual petitions are based. Prior to 2012, the ‘area of intended employment’ was not precisely defined within the H-2A program rules. In 2012 the DOL removed this ambiguity by defining the ‘area of intended employment’ to be no greater than a 60-minute commuting radius from the employer’s H-2A housing facility. The rule pertained only to employers who were FLCs. Growers who directly hire H-2A workers were not affected by the 60-minute rule. Many FLCs have been participating in the H-2A program since 2009 and they harvest citrus over a wide area of Florida. For the most part, these FLCs establish a single housing area, and while they attempt to find a geographically central housing location, in some cases their harvesting contracts force them to transport workers longer than a 60-minute commute. The 60-minute rule requires these FLCs to either cancel harvesting contracts with growers outside the 60-mile radius, or set-up separate housing facilities to keep all commuting distances to be less than 60 minutes. Since H-2A petitions are tied to a single housing location, expanding the number of housing facilities to a second or third location would force FLCs to submit additional petitions. FLCs assert that the 60-minute rule would significantly increase their costs of using the H-2A program. The objectives of the 2014 survey were two-fold. First, to document pre-employment costs associated with the H-2A program and second, to measure the extent to which these costs changed as a result of the 60-minute rule. Pre-employment costs were defined in the survey as expenses related to the H-2A application process, worker recruitment, round-trip travel between Florida and the workers’ home towns, and provision of housing facilities. Our initial hypothesis was that costs associated with the application process would increase with additional petitions. The costs associated with worker recruitment and in-bound/out-bound travel, however, would not be affected by the 60-minute rule unless a FLC chose to eliminate contracts outside of the 60-minute radius and thereby reduce the number of requested H-2A workers. Early in the survey design, we decided that questions regarding housing costs should relate only to the cost of acquiring ‘bed-spaces,’ and not the costs associated with managing the housing facilities during the harvest season. Our reasoning was based on comments from FLCs who argued that housing costs would increase because the 60-minute rule would force current housing to be underutilized and relatively more expensive properties to be acquired at new locations. Furthermore, management and operational costs associated with International Food and Agribusiness Management Review

336

Roka et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

H-2A workers during a contract period are likely to differ across different agricultural industries. Our goal was to define a set of costs that are likely to be similar among all agricultural specialty crops. Average preemployment costs per H-2A worker are estimated in 2012-13, the season prior to the 60-minute rule, and compared against similar costs during the 2013-14, the season when the new rule was implemented. Total boxes of oranges harvested by the survey participants are reported for each season to estimate the overall per box impact of H-2A pre-employment costs on harvesting costs. This article is organized as follows. First, we review the labor requirements for hand-harvested crops, with a focus on citrus production. We then provide a brief history of the H-2A visa program and the mechanics of implementing the H-2A program. Next, we explain our survey methodology and interpret the collected data. Lastly, we summarize the research and suggest additional avenues for investigation.

2. Need for farm labor The U.S. agricultural labor market underwent a significant transformation during the 20th century. In 1920, 30% of the U.S. population lived on farms. By 1990, less than 2% of U.S. citizens were involved directly in agriculture production (Dimitri et al., 2005). Rapid mechanization was largely responsible for this trend. Today, most agronomic crops in the U.S. are mechanized completely from planting through harvest. Horticultural crops have seen significant advances in mechanization as well, particularly crops which move fruit into processing channels such as paste tomatoes and tart cherries (Huffman, 2005). Machine harvest, however, can cause significant cosmetic and physical damage to fruits and vegetables thereby rendering them unsuitable for fresh market venues. Oranges and other citrus crops are important agricultural commodities grown in Florida with total sales exceeding $ 1.8 billion annually (Florida Department of Agriculture and Consumer Services, 2013). Approximately 95% of Florida’s sweet orange crop is processed into juice (National Agricultural Statistics Service, 2016), which in turn, supplies more than 80% of the U.S. domestic orange juice market (Wexler, 2014). Mechanical harvesting systems gained some momentum in the late 1990s when growers became concerned over whether a sufficient number of domestic workers would be available to harvest an expanding volume of sweet orange production. The number of acres mechanically harvested increased from 5,500 in 1999 to more than 36,000 acres in 2006. The discovery of citrus greening, also known as Huanglongbing (HLB), in 2005 and its subsequent spread across Florida curtailed mechanical harvesting efforts. By 2014 all mechanical harvesting efforts were suspended as growers sought to minimize any stress on HLB infected trees (Florida Department of Citrus, 2015). HLB has taken a significant toll on Florida citrus production. Growers harvested 242 million (90-pound) boxes of oranges during the season just prior to HLB’s discovery (2003-04). Eleven years later, the 201415 season produced less than 97 million boxes (National Agricultural Statistics Service, 2016). The decline in production resulted from both a 28% decrease in the number of bearing trees and lower per tree yields (Singerman and Useche, 2016). The Florida industry made a serious attempt to register an abscission compound known as CMNP, which if successful, could have facilitated further development of citrus mechanical harvesting in Florida. In theory, abscission would have dramatically lowered the overall force necessary remove fruit and in so doing lessen the harvesting stress to the tree. Unfortunately, the EPA abscission registration effort was not successful and abscission’s impact on tree stress could not be fully tested. As long as HLB remains a production threat and until more gentle mechanized harvesting equipment is developed, the only harvesting option for citrus growers will be hand-labor from seasonal agricultural workers.

International Food and Agribusiness Management Review

337

Roka et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

3. A foreign-born workforce on U.S. farms Across countries, as industrialization occurs and agricultural workers move into the manufacturing and service sectors in pursuit of higher wages, a country’s domestic agricultural labor supply decreases in elasticity (Taylor, 2010). Other factors drawing farm workers away are the perceived benefits of living in an urban setting such as better access to healthcare, educational opportunities, and more varied cultural amenities. Some growers have responded to this attrition of domestic workers by increasing wages, mechanizing, and growing different, more mechanically friendly crops. For other growers, and in particular Florida citrus growers, the least costly option has proved to be importing foreign workers. Hiring foreign guest workers to work seasonal agricultural jobs was not commonplace until the First World War when the U.S. government established a series of bilateral agreements with Mexico to encourage the seasonal migration of Mexican workers to U.S. farms (Martin, 2003). Martin (2003) refers to these agreements from 1917 to 1921 as the ‘first bracero program.’ The U.S. federal government implemented a second bracero program during the Second World War to address domestic farm labor shortages. Before the program was discontinued in 1964, 7.5 million contracts were signed. Recent surveys by the United States Department of Agriculture (USDA) indicate that there are slightly more than one million agricultural workers in the US, including full-time, part-time, and agricultural service workers (Economic Research Service, 2014). Many of these workers do not have legal status for employment in the U.S. More than 50% of the farm workers interviewed for the National Agricultural Worker Survey self-report that they are working in the U.S. without legal documentation (Employment and Training Administration, 2014). Data from the Social Security Administration suggest that a truer percentage of undocumented agricultural workers may be closer to 70% (Gunderson et al., 2009). In a 2014 Florida strawberry industry survey, half of the growers believe that illegal workers account for 90% of their seasonal workforce (Guan et al., 2015). The H-2A visa program provides U.S. agricultural employers with a legal avenue to hire foreign agricultural workers. The H-2 temporary work visa program has been in place since the Immigration and Nationality Act was passed in 1943. The program was separated into the H-2A and H-2B programs in 1986 with the passage of the Immigration Reform and Control Act. The latter visa category allows employers in the service and manufacturing sectors to recruit foreign workers, while the former is specifically for agricultural workers. The number of Florida H-2A positions certified by the DOL has been steadily growing since 2009 both in the actual number and as a percentage of total U.S. H-2A workers (Table 1). During fiscal year (FY) 2010, Florida agricultural employers hired 4,432 H-2A workers, or 5.6% of the total H-2A workers certified across the U.S. In FY 2015, Florida H-2A numbers had grown to nearly 18,000 and accounted for 12.8% of the total U.S. H-2A certified positions. The statistics on the number of Florida H-2A workers hired specifically to harvest citrus during FY 2013 and 2014 were not recorded, or at least summarized. In the two years prior to our survey, however, the percentage of citrus H-2A workers in Florida had increased from 71 to 84% of all the H-2A workers in Florida (Office of Foreign Labor Certification, 2014).

4. Mechanics of the H-2A visa program The H-2A application and recruitment process involves three federal, one state, and at least one local government agency. The federal agencies are Department of Labor, Homeland Security, and State Department. At the state level, there are the state workforce agency1 and the local department of health where worker housing is located. Between 60 and 75 days before the start of the season, a grower or FLC completes an application for foreign guest workers (Employment and Training Administration, 2016). The application packet is submitted initially 1

The Department of Economic Opportunity is Florida’s state work force agency.

International Food and Agribusiness Management Review

338

Roka et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

Table 1. Number of H-2A applications and certified positions by the U.S. Department of Labor across the United States (US) and in Florida (FL) between Fiscal Year (FY) 2010 and 2015.1 Florida crop year-citrus Certified H-2A applications US FL Certified H-2A positions US FL FL citrus (%) FL AEWR2 ($/hr)

FY2010

FY2011

FY2012

FY2013

FY2014

FY2015

2009-10

2010-11

2011-12

2012-13

2013-14

2014-15

6,988 51

7,000 54

7,845 78

8,118 108

9,405 N/A

N/A N/A

79,011 4,432 N/A 9.20

77,246 5,741 71 9.50

85,248 6,945 84 9.54

98,821 10,051 N/A 9.97

116,689 13,544 N/A 10.26

139,832 17,942 N/A 10.70

Data of FY2010 to FY2013 adapted from OFLC (2013); FY2014 from OFLC (2014); and FY2015 from OFLC (2015). N/A = data not available. 2 AEWR = adverse effect wage rate.

to the state workforce agency and includes the ‘job-order.’ The job-order (form ETA-790) specifies the number of foreign workers being requested as well as the start and end dates of the contract period. The job-order is designed to be the written contract between employer and worker, outlining all the job responsibilities of the worker and benefits to be paid by the employer. The job order is both crop and activity specific. The job order sets forth the duration of the contract period, the minimum weekly number of ‘offered’ hours, and the minimum hourly rate of pay. Typically, the federal-mandated Adverse Effect Wage Rate (AEWR)2 is the minimum hourly rate, unless a higher collective bargaining rate or local prevailing wage rate is already in place. The AEWR, which varies across states, is set to ensure that H-2A workers do not push down the wages of domestic workers performing similar jobs. The job-order guarantees that each worker will earn at least 75% of the minimum earnings set forth during the contract period (Office of Foreign Labor Certification, 2012). A foreign H-2A worker can be sent home prior to earning the 75% guarantee if he is replaced by a domestic worker. Otherwise, only an ‘act-of-God’ that prematurely destroys a crop will exempt an H-2A employer from the three-quarter guarantee.3 Finally, the job order lists all benefits accorded to the foreign guest worker such as free housing, free in-bound transportation, and if the worker completes the contract, free transportation back to the worker’s home town. Housing facilities must be inspected and certified that they meet local health and safety standards. If the housing units do not have kitchen facilities, the employer must cater three meals a day to all H-2A workers. Submission of the H-2A application package with the state workforce agency initiates a recruitment process for ‘domestic’ workers. Assuming that the number of job referrals from the state work force agency is not sufficient to cover the employer’s labor needs, the application is forwarded onto the DOL with form ETA9142A, an Application for Temporary Employment Certification. The employer-applicant continues to recruit for domestic workers by placing ads in media outlets in the ‘area of intended employment’ as well as in additional markets as directed by DOL (Office of Foreign Labor Certification, 2012). Active recruitment of domestic workers continues through the half-way point of the H-2A contract period. If domestic workers are hired prior to the contract’s start date, DOL reduces the requested number of foreign guest workers onefor-one. If domestic workers are hired after the contract’s start date, employers have the option of retaining foreign guest workers or sending them home. Employers are prohibited from hiring H-2A workers if they have laid off U.S. workers within 60 days of the date of need, unless the released U.S. workers were offered and rejected the agricultural job opportunities for which the H-2A workers were sought (Eisenbrey, 2009).

2 3

As of January 1, 2016 the AEWR in Florida was $ 10.70 per hour. The DOL must approve that a given event is an ‘act-of-God’.

International Food and Agribusiness Management Review

339

Roka et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

Once the labor certification process (forms ETA 790 and 9142A) are completed, a petition for non-immigrant worker visas (form I-129) is submitted to the Department of Homeland Securityâ&#x20AC;&#x2122;s Citizenship and Immigration Services (CIS). At this point, guest workers are recruited and CIS conducts background checks. With the list of recruited workers vetted, the Department of State issues individual H-2A visas. Employers are responsible for all expenses associated with the H-2A application and the recruitment process including filing and visa fees, any fees charged by third-party recruiters, bond expenses, and advertising costs to recruit domestic workers. The costs associated with the petition application, travel, and housing are identified as pre-employment costs. These costs, incurred as a result of the H-2A process, are not crop specific. Therefore, results of this study can be generalized to other agricultural employers seeking H-2A workers.

5. Survey methodology The objectives of this study were to quantify the pre-employment costs associated with the H-2A program and determine the extent to which these costs were affected by the DOLâ&#x20AC;&#x2122;s 60-minute rule imposed before the start of the 2013 citrus harvest season. Cost data were collected from FLCs who harvested citrus and employed H-2A workers during the 2012-13 and 2013-14 seasons. The Florida Department of Economic Opportunity provided the names and addresses of 75 FLCs which comprised a complete list of contractors who were both citrus harvesters and H-2A petitioners in 2013. Surveys were sent to everyone on the list. The surveys asked questions regarding the FLCâ&#x20AC;&#x2122;s individual petitions including the requested number of workers, contract period, and the number of citrus boxes they were contracted to harvest. In addition, the FLCs were asked to provide cost data on (1) the petition application process; (2) in- and outbound travel costs for H-2A workers; and (3) rental or annual ownership costs pertaining to the housing facilities. Petition application costs included agency filing fees, advertising, bond, and visa costs, as well as any fees paid to consultants to help with petition filing or worker recruitment. Cost data were requested for 2012-13 and 2013-14, the season prior to and after the 60-minute rule was instituted. All costs are reported on a per worker basis. Worker costs are aggregated across categories and multiplied by the total number of H-2A workers to determine total pre-employment costs for each petition and employer. Likewise, the H-2A pre-employment cost per box of citrus is determined by dividing the total H-2A preemployment cost by the total number of boxes harvested by the total number of H-2A workers represented in the survey.

6. Survey results Fourteen of the 75 surveys were returned of which nine surveys (12% response rate) were complete with all requested cost information for both harvest seasons. While we would have liked a higher response rate, the employers who completed the survey hired a considerable number of H-2A workers. Collectively, these nine employers hired 2,518 H-2A workers to harvest more than 26 million boxes of citrus in the 2012-13 season, and 2,610 H-2A workers to harvest more than 21 million boxes during the 2013-14 season. For each season, the surveyed H-2A harvesting companies accounted for more than 17% of the total Florida citrus harvest. The employers who completed the survey hired 25% of the total number of H-2A workers in Florida during the 2012-2013 season, and 19% of the total number of H-2A workers in Florida during the 2013-14 season (Table 2). Eight of the nine employers had to establish additional housing facilities for their 2013-14 petitions to be in compliance with the 60-minute rule. Consequently, the number of individual petitions increased from 13 in 2012-13 to 22 in 2013-14 (Table 2). Between the 2012-13 and 2013-14 seasons, the overall average number of H-2A workers per employer increased by 10, from 280 to 290. The average number of H-2A worker per petition, however, declined from 194 to 119 between the 2012-13 and 2013-14 seasons. According to a twotailed two-sample t-test with heteroscedasticity, there is a statistically significant difference in the number of

International Food and Agribusiness Management Review

340

Roka et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

Table 2. Summary of surveyed citrus harvesters who hired H-2A workers during the 2012-13 and 2013-14 harvest seasons (adapted from National Agricultural Statistics Service, 2016). Employers (citrus harvesters) Individual petitions Duration of job order (months) H-2A workers hired, total Percentage of Florida H-2A workers1 H-2A workers hired, per petition H-2A workers hired, per employer Citrus boxes harvested by surveyed harvesters Total citrus boxes harvested across state Percentage of citrus harvest by sampled employers 1

2012-13

2013-14

9 13 7 (Nov-May) 2,518 25.1 194 280 26,914,877 156,230,000 17.2

9 22 8 (Nov-June) 2,610 19.3 119 290 21,781,993 124,030,000 17.6

See Table 1 for Florida H-2A totals.

workers per petition between years.4 The duration for most job-orders was between seven and eight months corresponding to the November through June Florida citrus harvesting season (Table 2). Table 3 summarizes the average pre-employment costs associated with petition filing, recruitment, travel, and housing of H-2A workers for each worker. Petition and recruitment costs include government agency fees, advertising, bonds, consultants, and worker visas. The DOL and CIS charge flat fees to process H-2A petitions. The DOL charges $ 100 per application plus $ 10 per worker up to a $ 1000 maximum application fee. The CIS charges $ 325 to process one petition for non-immigrant worker visas (form I-129). This cost structure, along with the time and effort required to file a petition, encourages employers to group as many workers onto a single petition as possible. One employer filed a petition for just 20 workers. Thus, the employer’s filing costs were $ 31.25 per worker. Another FLC petitioned for 390 workers and his filing fees were only $ 3.40 per worker. There is a statistically significant difference in the average DOL-CIS registration fees per worker between years for the sampled petitions (Table 3). The 2013-2014 season had fewer workers per petition and thus higher fees per worker. The sampled employers paid an average of $ 6.42 per worker in registration fees in the 2012-2013 season, and $ 8.97 in the 2013-2014 season. H-2A employers are required to actively recruit domestic workers by advertising job notices in the media markets defined as the ‘area of intended employment.’ The most common method of advertising is through newspapers. Advertising costs vary widely depending on whether the newspaper services an urban or a rural region, and whether the particular newspaper publishes a Sunday edition. Advertising is a lump-sum payment per petition. In general, as the number of workers increase per petition, one would expect the advertising cost per worker to decrease. However, despite observing more petitions with fewer workers per petition in the 2013-2014 season, the average advertising costs per worker decreased. Advertising costs averaged $ 10.76 per worker in the 2012-2013 season and decreased to $ 9.64 per worker in the 2013-2014 season. There is not a statistically significant difference in the average advertising cost per worker between years for the sampled petitions (Table 3). Some employers faced increases in advertising costs while others saw dramatic decreases. For example, an employer paid $ 5,867 for advertising two petitions in the 2012-2013 season, but only $ 1,711 for two similar petitions in the following season. There are several potential explanations for reduction in average advertising costs. One hypothesis is that new petitions resulting from the 60-minute rule tended to be located in more rural locations with relatively cheaper advertising costs. A second reason could be that employers are learning how to more cost-efficiently 4

The test reported a P-value of 0.0578 rejecting the null hypothesis with α=0.10 that the average number of H-2A workers per petition in the 2012-13 season was equal to the 2013-14 season. The reported H-2A workers per petition in the 2012-13 and 2013-14 seasons had corresponding standard deviations of 110.8 and 102.0, respectively. Satterthwaite’s degrees of freedom for the test were 23.64.

International Food and Agribusiness Management Review

341

Roka et al.

Volume 20, Issue 3, 2017

Table 3. H-2A pre-employment costs itemized with average and ranges summarized from surveys of H-2A employers during the 2012-13 and 2013-14 citrus harvesting seasons.1 2012-13

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

Average ($) Cost per hired H-2A worker DOL/CIS fees5 6.42 Advertising 10.76 Bond 9.49 Consultant – agent filing 25.00 Consultant – worker recruitment 94.66 Visa 190.00 Travel 407.57 Subtotal (no housing) 773.90 Housing 1,322.823 Average ($/H-2A) 2,066.72

% change2

2013-14 Range low-high

Average ($)

Range low-high

4-19 5-50 3-14 10-26 26-110 flat fee 186-662

8.97 9.64 12.06 25.00 95.79 190.00 463.48 804.94 1,165.363 1,970.30

3-31 2-78 3-38 15-29 25-120 flat fee 273-750

684-1,5664 1,494-2,317

415-1,4613 1,371-2,355

+40* –10 +27* – +1 – +14 +4 –12 –5

Indicates a statistically significant difference (with α=0.05) between years. two-sample t-test with heteroscedasticity tests for equal means between years. 3 Average cost for an 8-month contract (the median reported contract length). 4 Range includes contracts varying from 2.5 to 10 months long. 5 DOL = U.S. Department of Labor; CIS = Department of Homeland Security’s Citizenship and Immigration Services. 2 A two-tailed

manage their advertising costs as they gain more experience with the program. Third, the fall in average advertising costs could be attributable to normal variation or even misreporting. Overall, the survey responses reflect wide differences in advertising costs among employers, likely driven by the variance in advertising costs due to both geography and ad placement timing. FLCs are required to purchase a bond with each H-2A application. The bond, which grower-employers do not have to purchase, ensures that all financial obligations owed to the H-2A workers are fully met by the FLC. While the amount of the bond for a given petition increases by the number of workers being requested, the overall cost of an individual bond depends greatly on the asset level and prior employment history of the FLC petitioner. The average cost per H-2A worker to acquire a bond in the 2012-2013 season was $ 9.49 and increased to $ 12.06 per worker in the 2013-2014 season. There is a statistically significant difference in the reported bond costs per worker between years for the sampled petitions (Table 3). These results suggest that, on average, bond costs increase with the number of petitions. The increase in bond costs is logical given the legal requirements and hiring practices observed in the sample. FLCs must obtain a surety bond of $ 5,000 for petitions with fewer than 25 employees, $ 10,000 for 25 to 49 employees, $ 20,000 for 50 to 74 employees, $ 50,000 for 75 to 99 employees, and $ 75,000 for 100 or more employees. Given that many of the observed petitions requested over 200 workers in the 2012-13 season, splitting these petitions up resulted in a significant increase in bonding requirements. For example, a single 200-worker petition would only require a $ 75,000 surety bond, whereas a 150-worker petition accompanied by a 50 worker petition would require the same $ 75,000 bond plus another $ 20,000 bond. Conversely, it is possible that splitting petitions up could result in reduced bond costs due to the nonlinear bond requirement structure. For example, an 80-worker petition would require a single $ 50,000 bond, but two 40-worker petitions would require two $ 10,000 bonds. H-2A consultants are an important resource for many employers. Generally, consultants perform two types of services: (1) prepare and file the petition(s); and (2) handle in-country worker recruitment. Among the nine sampled employers, six paid consultants to provide one or both of those services. Other employers chose to International Food and Agribusiness Management Review

342

Roka et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

perform these tasks in-house and did not report a separate cost for petition filing and in-country recruitment. From those employers who paid consultants to prepare and file the H-2A petition, a typical cost was $ 25 per worker. The cost for in-country recruitment averaged $ 94.66 and $ 95.79 per worker for 2012-2013 and 2013-2014 seasons, respectively. There is not a statistically significant difference in the average consulting costs per worker between years for the sampled petitions (Table 3). These average costs are added to the costs reported by those employers who handled petition filing and in-country recruitment in-house in order to obtain a consistent estimate of the average total cost across the entire survey. The cost of an H-2A visa is set by the U.S. Department of State and was a flat fee of $ 190 per person for both 2012 and 2013. Travel costs consist of round-trip expenses for bringing H-2A workers from their respective home towns to the employer’s housing facility in Florida. Since all the guest workers came from Mexico, travel expenses include bus fare, hotels if necessary, and a per diem food allowance. Travel costs per worker represent nearly 20% of total pre-employment costs in the 2012-2013 season and increase to more than 23% of total pre-employment costs in the 2013-2014 season. Despite an increase in average travel costs, there is not a statistically significant difference between seasons (Table 3). In each year the reported range of travel costs varied significantly reflecting the longer distances some groups of workers had to travel and the need for hotel rooms during extended trips. H-2A employers can either own or lease housing units. In either case, employers provided data that reflected the cost of securing the necessary number of housing units as required per each H-2A petition. The average cost of housing is reported for an 8-month contract, the median reported H-2A contract length. In the 20122013 and 2013-14 seasons, the cost to rent housing accounted for 64 and 59% of total pre-employment costs, respectively. In the 2013-2014 season, the average monthly cost of housing an H-2A worker decreased 12% per worker from the previous year (Table 3). While there is not a statistically significant difference in the average housing cost per worker between years for the sampled petitions, the decrease in cost was not expected. Instead of building or purchasing new housing units, FLCs complied with the 60-minute rule by leasing the additional housing facilities. The fall in average housing costs is primarily driven by one observation where an employer switched from owned to leased housing and reported considerable savings as a result. Further, there was considerable variance in the reported housing costs. This variance may reflect that some employers do not accurately compute the full cost of providing housing, or that some employers provide housing at a much lower cost than other employers. As with travel and advertising costs, housing costs may differ significantly between employers depending on location, timing, and amenities. As stated previously, the cost of managing housing during the season was not included in our survey. While we felt justified in our reasons to exclude these costs, this does mean our analysis excludes the additional administrative costs from managing multiple housing sites such as the costs of employing separate management and maintenance personnel for each location. In total, the reported cost associated with recruiting, transporting, and housing one H-2A worker from Mexico for eight months in the 2012-2013 season averaged $ 2,067. The cost in the 2013-2014 season decreased by 5% to $ 1,970 per worker.5 Total reported costs were not significantly different between seasons (Table 3). Housing costs alone make up 60% of the total pre-employment costs. When housing costs are excluded, the remaining items increase costs by $ 31 per worker from 2012-13 to 2013-14. This increase, however, was not statistically significant. The only items that increased between years and were statistically significant were petition fees and bond costs. The combined increase in these items, however, amounted to only about $ 5 per worker. The H-2A visa cost only accounted for nearly 10% of the total pre-employment costs and did not change between years. Between 7 and 9% of these costs were spread among consultants and advertising, which also did not significantly change between years.

This decrease in total pre-employment cost per worker is primarily attributable to the reported decrease in housing costs between years; once again, this decrease in housing costs is driven by a single sampled employer converting from owned to leased housing and reporting reduced housing costs as a result.

International Food and Agribusiness Management Review

343

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

Roka et al.

Volume 20, Issue 3, 2017

Table 4 presents the average per box cost of the pre-employment H-2A process. This estimate is derived by dividing the aggregate sum of pre-employment costs by the total boxes harvested by the H-2A workers from the nine surveyed companies. The cost of the pre-employment H-2A program was 19-cents per (90-pound) box (Table 4). This estimate represents 9% of the total cost to harvest citrus in Florida during the 2012-13 season (Muraro, 2012). H-2A pre-employment costs increased to nearly 24-cents per box in the 2013-14 season despite a slightly lower per worker pre-employment cost (Table 3). The significant factor driving the increase in H-2A pre-employment unit costs was the decline in boxes harvested by H-2A workers during the 2013-14 season. Mainly due to the negative effects from citrus greening, 92 more H-2A workers (Table 2) harvested 5 million fewer boxes during 2013-14 than what was harvested in 2012-13.

7. Conclusions Despite rapid mechanization in the U.S. agricultural sector, labor-intensive farm positions continue to be an integral part of the fruit and vegetable economy (Calvin and Martin, 2010). Agricultural producers and consumers rely on seasonal and migrant workers to grow and harvest fresh produce each year in high-value industries where mechanized harvesting is not yet a feasible alternative. Florida’s citrus industry is especially reliant on manual labor and the prospect of mechanical harvesting has been put on hold as a result of the citrus greening disease (Roka, 2013). The number of legal domestic workers willing to do agricultural work continues to shrink, leading more employers to consider the H-2A visa program as a source of qualified labor (Taylor and Charlton, 2012). Currently, more than half of Florida's citrus crop is harvested by H-2A workers and most are hired using FLCs as intermediaries (M. Carlton, personal communication 2015). H-2A employers, however, are required to pay all petition and visa processing fees, and provide free housing and transportation both to and from a workers home town as well as free transportation during the season between the housing location and the various worksites. Further, H-2A employers must pay a higher minimum wage known as the AEWR. Across the country the AEWR varies between $ 10.59 and $ 13.80 (National Agricultural Statistics Service, 2015). In Florida the AEWR was $ 10.70 as of January 1, 2016. U.S. agricultural employers who hire H-2A workers are required to offer the same benefits to any domestic worker applying for the same job. The first objective of this paper was to quantify a list of pre-employment expenses incurred by employers of H-2A workers, which is frequently referenced but has not yet been precisely determined. This analysis estimates these costs per worker and per box using survey data from a number of H-2A employers who collectively manage between 20 and 25% of Florida’s H-2A workers, who in turn harvest approximately 17.5% of the Florida citrus harvest. Overall, we found a FLC who recruits and hires foreign agricultural guest workers through the H-2A program must spend close to $ 2,000 per worker before the first piece of fruit is picked. These costs cover H-2A petition filing, agency fees, worker recruitment, round-trip travel expenses from the workers’ respective home towns, and to acquire suitable housing space during the contract period. Given how we defined pre-employment costs, we argue that these costs would be similar for H-2A employers outside of citrus. Table 4. Cost per box of H-2A pre-employment costs. H-2A workers hired (Table 2) Pre-employment costs (Table 3) Total pre-employment costs Total citrus boxes harvested (Table 2) Average

2012-13

2013-14

2,518 $ 2,067 $ 5,204,706 26,914,877 $ 0.19 per box

2,610 $ 1,970 $ 5,141,700 21,781.993 $ 0.24 per box

International Food and Agribusiness Management Review

344

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

Roka et al.

Volume 20, Issue 3, 2017

A second objective of this paper was to analyze how H-2A pre-employment costs changed between the 2012-13 and 2013-14 seasons as a result from implementing the DOL mandated 60-minute rule. Toward that goal we found mixed results. We found there to be a statistically significant decrease in the number of H-2A workers per petition between the 2012-2013 and 2013-2014 seasons. Similarly, we observed a statistically significant increase in the reported DOL-CIS fees and bond costs per worker. We did not, however, observe a statistically significant increase in advertising, consulting, travel, housing, and total costs between seasons (Table 3). While the 60-minute rule increased some costs associated with hiring H-2A workers, the increase in these categories were a relatively small component of total pre-employment costs. Travel and housing costs, on the other hand, accounted for more than 80% of the total pre-employment cost, and these costs did not significantly change between seasons. The decrease in housing costs between years appears to contradict the conventional wisdom that the 60-minute rule would increase housing costs by requiring employers to arrange additional housing facilities, though it is difficult to confidently interpret this result given the large variance in reported housing costs between years and employers. Further, the sampled employers did not construct new housing to adhere to the rule, but rather leased the additional housing. The robustness of the 60-minute rule analysis was compromised by the design of the survey in which we chose not to include in-season managerial costs with respect to housing and transportation. While this decision was driven to collect cost data which would be applicable to operations other than citrus, by not including in-season housing management and transportation costs we could not fully explore the effect the 60-minute rule had on H-2A costs. It is not clear a priori how the inclusion of these costs would affect the net change in total H-2A pre-employment costs. On one hand, managing multiple housing locations could force higher costs with respect to employing more individuals in the management of each location. On the other hand, disbursing a given number of H-2A workers across the landscape may shorten daily trips to harvesting sites, thereby reducing fuel costs and worker time spent simply in transit. Limitations aside, we were able to capture cost information about a number of fixed costs associated with an H-2A petition. The H-2A program is becoming an increasingly important source of labor for U.S. specialty crop farms and in particular for the Florida citrus industry. By providing cost information about the H-2A program, we seek to mitigate some of the risk of participation for agricultural employers who are unfamiliar with the H-2A program. Furthermore, as Florida citrus growers seek to minimize production costs amidst the spread of citrus greening, information about harvesting costs and the increasing use of foreign guest workers could assist them in planning for future production as well as investing in future technologies such as mechanical harvesting.

Acknowledgements This project was supported in part with funding from the USDA Specialty Crop State Block Grants through the Florida Department of Agriculture and Consumer Services. The authors would to thank the Florida Fruit and Vegetable Growers Association for logistical support in carrying out the survey and for initial consultations in the survey design. The authors would also like to thank three anonymous reviewers and the IFAMR Managing Editor for the very constructive comments and suggestions that greatly improved this paper.

References Calvin, L. and P. Martin. 2010. Labor-intensive U.S. fruit and vegetable industry competes in a global market. Economic Research Service, United States Department of Agriculture, Washington, WA, USA. Available at: http://tinyurl.com/h2w9f57. Dimitri, C., A. Effland and N. Conklin. 2005. The 20th century transformation of U.S. agriculture and farm policy. Economic Research Service, United States Department of Agriculture, Washington, WA, USA. Available at: http://tinyurl.com/zrobddd. Economic Research Service. 2014. Farm Labor. United States Department of Agriculture, Washington, WA, USA. Available at: http://tinyurl.com/jzmd4qh. International Food and Agribusiness Management Review

345

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0033 - Tuesday, May 09, 2017 8:29:58 PM - IP Address:24.21.169.207

Roka et al.

Volume 20, Issue 3, 2017

Eisenbrey, R. 2009. The H-2 visa programs real need for reform. In: Getting immigration right, edited by D. Coates and P. Siavelis. Potamac Books Inc., Washington, WA, USA, pp. 201-218. Employment and Training Administration. 2014. The National Agricultural Workers Survey. United States Department of Labor, Washington, WA, USA. Available at: http://tinyurl.com/jqa5kqd. Employment and Training Administration. 2016. H-2A Temporary Agricultural Program. United States Department of Labor, Washington, WA, USA. Available at: http://tinyurl.com/h27hk7h. Florida Department of Agriculture and Consumer Services. 2013. Florida agriculture by the numbers 2013. Available at: http://tinyurl.com/z4ug7nl. Florida Department of Citrus. 2015. Citrus Mechanic Harvesting. Available at: http://tinyurl.com/h55day4. Guan, Z., F. Wu, F. Roka and A. Whidden. 2015. Agricultural labor and immigration reform. Choices 30: 1-9. Gunderson, M.A, A. Wysocki and J.A. Sterns. 2009. Postharvest Handling, Elsevier, Cambridge, MA, USA, pp. 129-152. Huffman, W.E. 2005. Trends, adjustments, and demographics, and income of agricultural workers. Applied Economic Perspectives and Policy 27: 351-360. Martin, P.L. 2003. Promise unfulfilled: unions, immigration, and the farm workers. Cornell University Press, Ithaca, NY, USA. Muraro, R.P. 2012. Estimated average picking, roadsiding and hauling charges for Florida citrus, 2011-12 Season. Citrus Research and Education Center, University of Florida, Institute of Food and Agricultural Sciences, Gainesville, FL, USA. Available at: http://tinyurl.com/zrqjcwh. National Agricultural Statistics Service. 2015. Farm labor. United States Department of Agriculture, Washington, WA, USA. Available at: http://tinyurl.com/j4cmjfs. National Agricultural Statistics Service. 2016. Florida citrus statistics 2014-2015. United States Department of Agriculture, Washington, WA, USA. Available at: http://tinyurl.com/hw29qgh. Office of Foreign Labor Certification (OFLC). 2012. Employer guide to participation in H-2A temporary agricultural program’ Employment and Training Administration, United States Department of Labor, Washington, WA, USA. Available at: http://tinyurl.com/z9gjypv. Office of Foreign Labor Certification (OFLC). 2013. Annual report October 1, 2012 – September 30, 2013. Employment and Training Administration, United States Department of Labor, Washington, WA, USA. Office of Foreign Labor Certification (OFLC). 2014. H-2A temporary agricultural labor certification program – Selected statistics. Employment and Training Administration, United States Department of Labor, Washington, WA, USA. Available at: http://tinyurl.com/jfwsmfk. Office of Foreign Labor Certification (OFLC). 2015. Adverse effect wage rate (AEWR) chart 2010-2015. Employment and Training Administration, United States Department of Labor, Washington, WA, USA. Available at: http://tinyurl.com/goc757g. Roka, F. 2013. Outlook on Agricultural Labor in Florida, 2012-13. Food and Resource Economics Department, University of Florida, Institute of Food and Agricultural Sciences, Gainesville, FL, USA. Availabl at: http://tinyurl.com/hsw22hl. Singerman, A. and P. Useche. 2016. Impact of citrus greening on citrus operations in Florida. Citrus Research and Education Center, University of Florida, Institute of Food and Agricultural Sciences, Gainesville, FL, USA. Available at: http://tinyurl.com/hc4q62k. Taylor, E.J. 2010. Agricultural labor and migration policy. Annual Review of Resource Economics 2: 369-393. Taylor, E.J and D. Charlton. 2012. The end of farm labor abundance. Applied Economic Perspectives and Policy 34: 587-598. Wexler, A. 2014. Future sours for orange crop. Barron’s. Available at: http://tinyurl.com/h5ku9zk.

International Food and Agribusiness Management Review

346

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2016.0177

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Received: 11 July 2016 / Accepted: 1 November 2016

Factors affecting cattle producers’ willingness to adopt an Escherichia coli O157:H7 vaccine: a probit analysis RESEARCH ARTICLE Brian J. Ochieng’a and Jill E. Hobbs

aMSc.

graduate, bProfessor, Department of Agricultural and Resource Economics, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK S7N 5A8, Canada

Abstract E. coli O157:H7 bacteria – a major cause of foodborne illness – occur naturally in the intestine of cattle but do not affect the health or productivity of the animal. A cattle vaccine that significantly reduces the risk of E. coli contamination was developed and commercialized in Canada and internationally, however, adoption by cattle producers remained extremely low. Utilizing data from a survey of cow-calf producers in western Canada, this paper examines the factors affecting cattle producers’ willingness to adopt the E. coli vaccine. Education, prior awareness of the vaccine, perception of who holds primary responsibility for E. coli risk reduction, and a producer’s external (versus internal) locus of control with respect to their ability to mitigate E. coli risks within the production environment are significant determinants of willingness to adopt. Adoption incentives are also evaluated, including policy interventions, market/supply chain incentives, production protocol, and producer reputation incentives. The analysis provides lessons for the development and commercialization of vaccines and other food safety intervention strategies that yield societal and supply chain benefits beyond the individual adopter. Keywords: binary probit, food safety, risk mitigation, technology adoption, locus of control, beef JEL code: Q13, Q16, Q18 Corresponding author: jill.hobbs@usask.ca

347

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

1. Introduction The recognition that the health of humans, animals and the environment are closely linked, so-called ‘One Health’, is garnering increasing attention across policy, scientific and industry contexts. The relationship between food safety, water quality and livestock production has provided an opportunity for innovations that mitigate human health risks through interventions at the livestock production stage. An example is the prevalence of Escherichia coli (E. coli) O157:H7 in the intestine of cattle, creating the potential for foodborne illnesses from E. coli O157:H7 contamination at meat processing plants or contamination of irrigation or drinking water supplies by run-off from livestock operations. Recognition of this risk factor spurred scientific research into pre-and post-harvest interventions to reduce the risks of E. coli O157:H7 contamination in food supply chains, resulting in the development and commercialization of a cattle vaccine. E. coli O157:H7 is one of five pathogens identified as the primary causes of foodborne illness in the U.S. (alongside Campylobacter, Salmonella, Listeria monocytogenes, and other strains of E. coli), which together account for an estimated US$7.7 billion in annual costs of foodborne illness1 in the U.S. (Scharff, 2012). E. coli O157:H7 (henceforth, E. coli) has been associated with a number of high profile incidents of food (and water) borne illnesses internationally. Examples include contamination at the XL Foods beef packing plant in western Canada in 2012 that resulted in the largest beef recall in Canadian history, contamination of organic beansprouts in Germany in 2011 resulting in over 30 deaths and thousands of illnesses and severe disruption to fresh produce supply chains before the source of the contamination was determined, a nationwide recall of bagged spinach in the United States in 2006, and contamination of municipal drinking water supplies in Walkerton, Ontario (Canada) in 2000 resulting in an estimated 2,300 illnesses and 7 deaths (Ochieng’ and Hobbs, 2016). In all of these cases, the source of E. coli contamination was eventually traced to cattle: through cross-contamination of fecal matter from hides to meat in the case of the Canadian beef packing plant, as a result of contaminated irrigation water in the case of the organic beansprouts in Germany and in the U.S. spinach case, and as result of run-off from a nearby livestock farming operation in the case of the contamination of municipal water supplies in Ontario. Cattle are asymptomatic carriers of the E. coli bacteria, which occur naturally in their intestine. A number of post-harvest interventions exist to reduce the risk of E. coli contamination in the food system, including (where permitted), irradiation, organic acid rinses and steam pasteurization. Nevertheless, it is recognized that pre-harvest interventions may also be an effective means by which to reduce the risk of E. coli contamination both within food supply chains and in a broader environmental context. One such pre-harvest intervention is a cattle vaccine that reduces the prevalence of E. coli in the intestine, therefore reducing the risk of cross-contamination during food processing or through environmental factors (water). An E. coli vaccine (Econiche) became available in a number of countries in 2008, beginning with Canada and the U.S., followed by Australia in 2011 and was available on a special certificate system across some Member States of the European Union the same year (Ochieng’ and Hobbs, 2016). Despite the existence of the vaccine, and scientific evidence of its effectiveness (discussed in the next section), levels of adoption by cattle producers were remarkably low, estimated to be in the range of 5% in Canada (Grier and Schmidt, 2009) By February 2015 the vaccine was no longer commercially available after the Canadian company which manufactured the vaccine (Bioniche) sold its animal health division to the French company Vetoquinol, who ceased production of Econiche (NFAN, 2015). As a commercial endeavor, the vaccine suffered from a classic externality problem: the cattle producer incurred the costs of vaccination, yet the direct benefits flowed elsewhere within the supply chain, to downstream meat packers/processors and retailers, and to consumers, or more broadly within the environment, such as users of municipal or irrigation water supplies. The additional costs of vaccination, together with the lack of a strong market incentive

1 Estimated annual costs of foodborne illness include direct health-related costs as well as a variety of indirect costs such as quality of life effects, lost productivity, foregone income.

International Food and Agribusiness Management Review

348

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

(premiums) for vaccinated cattle, have been identified as significant barriers to adoption of this technology (Ochieng’ and Hobbs, 2016; Tonsor and Schroeder, 2015). The vaccine, and the challenges associated with its adoption, provides a timely opportunity to examine the factors affecting the willingness of livestock producers to adopt technologies with beneficial human health impacts – or ‘socially beneficial’ technologies. This paper examines these questions within the context of Canadian cattle producers’ willingness to adopt the E. coli vaccine while it was still available for use. Using data from a survey of cow-calf producers in western Canada the paper uses a probit model to evaluate the factors affecting willingness to adopt. The degree of awareness of the vaccine among cattle producers and their primary sources of information about the vaccine further enriches our understanding of the adoption environment. Actions are often driven by perceptions, and in this case, producers’ perceptions with respect to who holds primary responsibility for reducing the risk of E. coli contamination, as well as who benefits from the reduced risks can influence willingness to adopt. The relative strength of a set of market, policy, and reputational incentives for adoption are also examined. The remainder of the paper is organized as follows: the next section provides further context for the focus on E. coli and documents the development of the E. coli vaccine as a pre-harvest intervention strategy. The Methods section outlines the survey data collection process and the methodological approach (binary probit analysis). Insights from survey data pertaining to producer awareness of E. coli problems and the E. coli vaccine, perceptions over the distribution of responsibility for E. coli risk reduction, the benefits that flow from risk reduction, and barriers to adoption of the vaccine are presented in the section ‘The adoption environment: descriptive statistics’ and set the scene for the probit regression analysis discussed in the Results section. The paper concludes with a discussion of policy and managerial implications, together with suggestions for further research.

2. E. coli O157:H7 outbreaks and the development of a vaccine The Centers for Disease Control and Prevention tracked 390 E. coli O157:H7 outbreaks in the U.S. between 2003 and 2012 finding that these outbreaks resulted in 4,928 reported illnesses, 1,272 hospitalizations and 33 deaths (Heiman et al., 2015). Contaminated food was the source of the majority of these illnesses (65% of outbreaks or 255 incidents). Beef (primarily ground beef but also steak) accounted for approximately 21% of outbreaks (Heiman et al., 2015). Estimates of the costs of illness from E. coli infections include the direct costs of seeking medical attention, which can include hospitalization and, in severe cases, long-term medical care as a result of renal failure from hemolytic uremic syndrome, as well as the indirect costs of illness or premature death, including lost productivity and foregone earnings. In a Canadian study, Sockett et al. (2014) estimate that around 22,344 cases of primary E. coli infections occur in Canada annually. They estimate the mean annual cost of primary infections to be $26.9 million, of which 31% is accounted for by direct medical costs while 69% ($18.6 million) arises from lost productivity. Premature deaths accounted for the majority of the costs from lost productivity at $15.1 million. In addition, Sockett et al. (2014) estimate the annual costs of ongoing longterm illness at another $377.2 million per year, resulting in estimated mean costs of E. coli infections in Canada of $403.9 million annually. Additional costs of an E. coli contamination, not reflected in these numbers, include the costs to the food industry of product recalls, damage to industry reputation, market access restrictions and loss of consumer confidence (Loader and Hobbs, 1999; Teisl and Roe, 2010). The 2012 E. coli contamination at XL Foods Inc. in western Canada, for example, led to the largest beef recall in Canadian history and created short-run disruptions in export market access for Canadian beef products until the source of the contamination was identified. It also created severe disruption to domestic beef supply chains with the temporary closure of a major packing plant (the XL Foods plant represented 35% of Canada’s beef processing capacity), reflected in sharply lower cattle prices until the plant eventually reopened and markets stabilized (Lewis et al., 2013). International Food and Agribusiness Management Review

349

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

A number of potential interventions exist at both the pre-harvest and post-harvest stages of beef production as a means of addressing the E. coli pathogen. Pre-harvest interventions range from the use of vaccines, microbial probiotics designed to exclude or reduce microbial pathogens within the digestive tract of livestock, to biosecurity and herd management measures to reduce risks within the production environment. At the slaughtering and processing stage, interventions to reduce the risks of E. coli contamination within processing plants range from food safety management practices such as the use of Hazard Analysis And Critical Control Points plans, to technological measures such as electron-beam processing, irradiation, ethylene gas processing and steam pasteurization (Teisl et al., 2001). A wide body of scientific evidence, over a considerable period of time, has pointed to the potential for an E. coli vaccine to significantly reduce the risk of E. coli contamination within the beef supply chain (Ochieng’ and Hobbs, 2016). In early studies using stochastic simulation models, Jordan et al. (1999) and Signorini and Tarabla (2002) argue that vaccination is a key pre-harvest intervention strategy, with other post-slaughter measures, such as control of storage temperatures and the use of hide-wash cabinets in slaughterhouses remaining relevant complementary risk reduction measures. Smith et al. (2013) find vaccination to be highly effective at reducing both colonization and shedding, reducing the risk of environmental transmission of E. coli O157 within commercial farm operations. Withee et al. (2009) determine the maximum cost per unit at which a hypothetical E. coli vaccination would still be a cost-effective intervention for preventing E. coli illnesses in humans, finding that vaccinating the entire U.S. herd at a cost of between $2.29 and $9.14/unit would be a cost-effective intervention. The cost-effectiveness of interventions to reduce human illnesses varies with vaccine efficacy, and the authors suggest that vaccinating only a portion of the herd would be cost-effective for vaccines that are less effective or most costly. The recognition that cattle are a primary source of E. coli problems in both the food system and through environmental contamination of water supplies with cattle feces, along with the potential for a vaccine to reduce the prevalence of cattle shedding E. coli, led to research at the University of British Columbia and the Vaccine and Infectious Diseases Organization at the University of Saskatchewan in Canada to develop a vaccine. Building upon this research, a vaccine (Econiche) was licensed by the Canadian company Bioniche Life Sciences Inc., with early trials showing that vaccinated cattle were 92% less likely to be colonized by E. coli (Smith et al., 2009). The vaccine was authorized for use by the Canadian Food Inspection Agency (CFIA) as a three dose regimen for cow-calf operations or feedlots. Typically beef/dairy cattle received two doses of the vaccine within the first 12 months of life, with a recommendation that animals over 1 year receive one dose on an annual basis. The vaccine cost approximately CAD2 $3 per dose and in Canada was available through veterinarians (Grier and Schmidt, 2009). Despite the apparent effectiveness of the vaccine, adoption by cattle producers in regions where the vaccine was licensed for use was very low. A major impediment to adoption was the fact that cattle are unaffected by the presence of E. coli, which has no impact on the health of the animal or its productivity. Thus, cattle producers may perceive that the benefits of vaccination primarily flow elsewhere: to downstream actors in the beef supply chain, to consumers, to municipal water authorities, and so on. Other issues may have impacted the commercial uptake of the E. coli vaccine, including the cost of vaccination, the extent to which the required changes fit within existing practices and the scale of changes needed, lack of knowledge, the availability of veterinary advice, skepticism over the effectiveness of the intervention, financial constraints, human capital constraints, expected impacts on reputation, export market orientation, as well as anticipated government regulation and liability (Cobanoglu, 2012; Ellis-Iverson et al., 2010; Jayasinghe-Mudalige and Henson, 2006; Ochieng’ and Hobbs, 2016). A moral hazard risk may also be present if cattle producers anticipate that vaccinating their cattle will result in reduced efforts to mitigate food safety hazards on the part of other firms downstream in the supply chain (Marette et al., 2012; Ochieng’ and Hobbs, 2016). 21

CAD (Canadian dollar) = 0.76 USD; calculated on the basis of the exchange rate on December 12, 2016

International Food and Agribusiness Management Review

350

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

In the absence of incentives in the form of market premiums for vaccinated cattle, vaccination becoming a pre-requisite for securing a buyer, or the risk of direct penalties (market access, reputational or liability penalties) from being identified as the source of an E. coli contamination event, there appears to have existed little direct commercial incentive for cattle producers to vaccinate their cattle. Tonsor and Schroeder (2015) discuss the additional costs of vaccination and the absence of market premiums for vaccinated cattle as major impediments to widespread adoption of the vaccine by U.S. cattle producers, while Ochieng’ and Hobbs (2016) find market incentives (price premiums for vaccinated cattle or buyers requiring vaccination), along with measures to offset the costs of vaccination, to be attractive adoption incentives for Canadian cattle producers.

3. Methods Given the costs to the agriculture and food sector and to society as a whole from E. coli O157:H7 and the apparent promise of cattle vaccination as a pre-harvest intervention strategy, a deeper understanding of the factors affecting cattle producers’ willingness to adopt this type of vaccine is warranted. Using data from a survey of western Canadian cow-calf producers, a clearer picture of the adoption environment is provided, including producers’ experience with and awareness of E. coli and of the vaccine, their perceptions of who benefits from the vaccine within the supply chain and who bears the primary responsibility for reducing E. coli risks. A probit analysis is used to identify the significant factors influencing willingness to adopt the vaccine, including producer characteristics, attitudinal variables, and a set of market, policy, production protocol and reputational incentives for adoption. This section describes the data collection process and provides details of the probit model specification. Data were gathered through an online survey of western Canadian cow-calf producers drawn from an animal health/producer database managed by Ipsos Agriculture and Animal Health3. Cow-calf producers in the three prairie provinces of Alberta, Saskatchewan and Manitoba were surveyed in July 2014. These three provinces represent the largest share of the beef cow population (82%) in Canada at 40.7, 29.8 and 11.6% respectively (CanFax, 2014). A token incentive payment of CAD $20 was provided to participants who completed the survey. Pre-screening questions on location (province) of operation and number of beef cows ensured the survey sample matched the distribution of cow-calf production across the three prairies provinces and was comprised of producers with varying sizes of operations. Before proceeding with the survey, respondents were asked to confirm whether they were the person with overall or joint responsibility for the animal health management practices in use in their cow-calf operation. The survey gathered data on several aspects of the cattle farming operation, attitudes to and knowledge about E. coli risks and interventions including vaccination protocols, familiarity of cattle producers with E. coli and with the vaccine, attitudes towards E. coli risks, use of management practices/interventions within the cow-calf operation to mitigate E. coli risks, the extent to which producers exhibited an internal versus external locus of control with respect to E. coli risks within the production environment, barriers to adoption of the vaccine, and demographic information4. A more detailed discussion of the insights from the data is provided in the section ‘The adoption environment: descriptive statistics’. In analyzing producers’ willingness to adopt the vaccine, a binary probit model is developed. Probit model specification Survey respondents were asked to indicate their willingness to adopt an E. coli vaccine in response to an adoption incentive (discussed below), with three possible responses: yes, no, and don’t know/unsure. A standard probit model specification is used (Supplementary Methods S1 for details).

3 The survey received research ethics approval from the University of Saskatchewan Behavioural Research Ethics Board (BEH # 14-136) on June 27, 2014. 4 A copy of the survey instrument is available from the authors upon request.

International Food and Agribusiness Management Review

351

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Adesina and Baidu-Forson (1995) note the need for technology adoption studies to expand the range of explanatory variables beyond broad socio-economic, demographic and institutional determinants by including farmers’ subjective perceptions of the characteristics of new agricultural technologies. In this study, willingness to adopt is assumed to be a function of a set of producer and farm characteristics, attitudes, experience and the presence of a set of adoption incentives. The probit model estimation is specified as: Willingness to adopti = β0 + β1 Genderi + β2 Agei + β3 Educationi + β4 Experiencei + β5 Continuityi + β6 Sales/revenuei + β7 Livelihoodi + β8 Retaini + β9 Sizei + β10 Locationi + β11 Awarenessi + β12 Responsibilityi + β13 Benefitsi + β14 External locus of controli + β15 Internal locus of controli + β16 Individual best-worst scoresi + εi

(1)

where i represents the individual and εi: is assumed to be a random error with a zero mean and finite variance. The explanatory variables are described in Table 1 and include socio-demographic and farm-level characteristics to control for the effect of gender, age, level of education, number of years as a principal decision maker (experience), expected continuity as a cow-calf producer (continuity), the percentage of sales/revenues derived from cow-calf operations (revenue), dependence on cow-calf operations (livelihood), retaining ownership of cattle through the feedlot stage of production (retain), size of beef cow herd (size), and location (Alberta and Manitoba, with Saskatchewan omitted as a reference variable). For the most part, these are control variables with no a priori expectations as to the effect on willingness to adopt. Three exceptions are for livelihood (dependence on cow-calf operations), revenue, and retained ownership, where there is a weak expectation that producers with greater reliance on the cow-calf operation for their livelihood, or for whom a greater percentage of revenues is derived from cow-calf operations, or who retain ownership through the feedlot stage of production, might be more risk averse with respect to supply chain disruptions, liability or reputational risks from an E. coli outbreak that was traced to their cattle, and therefore have a higher willingness to adopt the vaccine. Other explanatory variables capture psychographic, attitudinal and knowledge variables. These include a variable capturing producer awareness of the E. coli vaccine technology and a set of dummy variables controlling for producers’ perceptions with respect to who bears primary responsibility for E. coli risk reduction between cow-calf producers, feedlots, packers, retailers, consumers, and regulators5. It is expected that producers who believe cow-calf producers bear the primary responsibility will be more willing to adopt the vaccine, or conversely those who believe primary responsibility lies elsewhere in the supply chain will be less likely to adopt. Using the same set of categories, respondents were asked who they perceived to be the primary beneficiaries of an E. coli vaccine. A priori we expect respondents who perceive cow-calf producers to be the primary beneficiaries will be more likely to express a willingness to adopt. This variable is specified as a dummy variable, with 1 where respondents indicated cow-calf producers as the primary beneficiary, and 0 otherwise. The survey assessed the extent to which respondents exhibited an internal versus an external locus of control with respect to their ability to control E. coli risks in their cow-calf operation. It is hypothesized that producers with an internal locus of control (meaning that they believe they can take concrete actions to reduce. E. coli risks within their farm environment) will have a higher willingness to adopt the vaccine, while those who believe that E. coli risks are largely exogenous to their actions (an external locus of control) are hypothesized to be less likely to adopt. Measured on a 5 point Likert scale from ‘Completely disagree’ to ‘Completely agree’ three survey questions reflected an internal locus of control: ‘I feel in control of potential E. coli contamination due to my existing management practices/interventions’; ‘Whether or not I’m successful in 5 Perception

of primary responsibility for reducing E. coli risks is comprised of a set of separate dummy variables for each of the categories given the difficulty of interpreting an aggregated coefficient. To avoid problems of collinearity, ‘cow-calf producers’ serves as the reference category and is omitted from the estimation. The fifth category dummy variable (retailers) did not receive any responses, hence its omission from the estimated model.

International Food and Agribusiness Management Review

352

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

Table 1. Variable descriptions.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Variable Dependent variable Willingness to adopt Explanatory variables Gender Age Education

Description Willingness of cow-calf producer to adopt an E. coli vaccine, coded 1 if yes, 0 otherwise.

Coded as 1 if male and 0 if otherwise. Actual age of respondent in years. Education level coded as 1, for respondents with an education level at or above college and 0 otherwise. Experience Years as principal decision-maker in the cow-calf operation. Mid-points of range: less than 4 years, 5-20 years, 21-35 years, 35 years and over. Continuity Years respondent plans to continue as a cow-calf producer. Mid-points of range: <1 year, 1-5 years, 6-10 years, 10-25 years, >25 years. Sales/revenues Percentage of revenues derived from cow-calf operations. Treated as a categorical variable with 1 representing 0-24%, 2 representing 25-49%, 3 representing 50-79%, and 4 representing 80-100%. Livelihood Importance of cow-calf operations to livelihood, coded as 1 if very important and essential and 0 if otherwise. Retain ownership Retained ownership of cattle/calves during the feeding/finishing process, coded 1 if ‘yes’ and 0 if ‘no’. Size of operation Number of beef cows in the cow-calf operation. Location Province/territory of operation with Saskatchewan as the reference (omitted) variable. Two dummy variables coded 1 if Alberta, 0 otherwise; and 1 if Manitoba, 0 otherwise. Awareness Awareness of E. coli vaccine, coded 1 if the respondent has heard of an E. coli vaccine and 0 otherwise. Responsibility Perception of who bears primary responsibility for reducing E. coli incidences, with cow-calf producers as the reference (omitted) category. A set of dummy variables each coded 1 if feedlots, 0 otherwise; 1 if packers/processors, 0 otherwise; 1 if retailers, 0 otherwise; 1 if consumers, 0 otherwise, and 1 if regulators and 0 otherwise. Benefits Perception of who would benefit the most from an E. coli cattle vaccine, coded as 1 if cow-calf producers and 0 otherwise. External locus of control External locus of control with respect to E. coli risks (average of Likert scale responses to three external locus of control questions): ‘My success as a cow-calf operator depends mostly on luck’; ‘It is not advisable for me to plan too far ahead by enhancing my current management practices because E. coli incidences are such that they cannot be fully prevented’; ‘To a great extent E. coli incidences on my cow-calf operation are determined by factors beyond my control’. Internal locus of control Internal locus of control with respect to E. coli risks (average of Likert scale responses to three internal locus of control questions): ‘ I feel in control of potential E. coli contamination due to my existing management practices/interventions’; ‘Whether or not I’m successful in mitigating/controlling E. coli depends mostly on my own ability’; ‘To a great extent E. coli incidences on my cow-calf operation are determined by the management practices I have in place’. Individual BWS scores Respondent’s best-worst scaling (BWS) scores for each of 13 incentives (see Table 2).

mitigating/controlling E. coli depends mostly on my own ability’; ‘To a great extent E. coli incidences on my cow-calf operation are determined by the management practices I have in place’. Three questions reflected an external locus of control: ‘My success as a cow-calf operator depends mostly on luck’; ‘It is not advisable for me to plan too far ahead by enhancing my current management practices because E. coli incidences are such that they cannot be fully prevented’; ‘To a great extent E. coli incidences on my cow-calf operation International Food and Agribusiness Management Review

353

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

are determined by factors beyond my control’. Responses to the three internal and external locus of control questions were averaged to give an internal and an external locus of control score for each respondent, each of which represented a separate variable in the analysis. A priori expectations suggest that, if significant, the internal locus of control variable will be positive and the external locus of control variable negative. The final set of explanatory variables capture cattle producers’ responses to a set of potential vaccine adoption incentives. The survey featured a set of best-worst scaling (BWS) questions in which respondents were asked to indicate the incentive which would be most likely to influence (best) and least likely to influence (worst) their adoption of an E. coli vaccine from a series of repeated choice sets featuring different combinations of thirteen adoption incentives. Each incentive appeared four times across the repeated choice sets, with each choice set featuring four incentives. A detailed discussion of the design and analysis of the BWS experiment is beyond the scope of this paper and is available in Ochieng’ and Hobbs (2016)6. Table 2 lists the thirteen adoption incentives, which are grouped into government (policy) intervention incentives, market/supply chain incentives, production protocol incentives, and producer reputation incentives. For each survey respondent, the number of times an incentive was selected as ‘worst’ in the choice sets is subtracted from the number of times it is selected as ‘best’ to provide a BWS value. As each of the 13 incentives appeared four times through the repeated choice tasks, the individual level scales for each of the incentives range from +4 to -4. For example, a score of +2 would be obtained if a survey respondent chose 6 The

analysis presented in Ochieng’ and Hobbs (2016) focuses on the Best-Worst Scaling experiment, which produces a preference ranking of the 13 adoption incentives and uses a latent class cluster analysis to identify heterogeneity among producers with respect to their responses to the adoption incentives. The present paper embeds the adoption incentives within a broader analysis of the factors affecting producers’ willingness to adopt the vaccine.

Table 2. Adoption incentives. Government intervention incentives • Government recommending use of E. coli vaccine for cattle. • Subsidy to compensate the costs of my adoption of the vaccine is available through a government vaccination program. Market/supply chain incentives • Premiums for E. coli vaccinated cattle are available through various programs (branded beef program) within the supply chain. • Attraction of a new set of buyers for my vaccinated cattle. • My buyer requiring use of vaccine as part of the production protocol as a condition for accepting my calves/ cattle. • Feedlots providing an assurance that they will give my cattle a booster of the E. coli vaccine to maintain the immunity of my cattle. • Through vaccination, my farm is less exposed to the effects of E. coli outbreaks, such as beef recalls and supply disruptions at packing plants. Production protocol incentives • Recommendation from my veterinarian to use the E. coli vaccine in my operations. • I can include an E. coli vaccination in my existing vaccination routine. • Duration of immunity for my calves/cattle is greater than six months. Producer reputation incentives • Beef products from my calves/cattle can be traced back to my farm/operations. • My reputation for a cattle producer is at risk because of higher consumer expectations concerning food safety. Other • My neighbors (other cattle producers) are adopting the E. coli vaccine. International Food and Agribusiness Management Review

354

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

an incentive as most desirable three times and least desirable once. Respondents’ individual BWS scores for each of the thirteen incentives are included as separate explanatory variables in the probit regression model.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Following the BWS choice sets, respondents were asked ‘Would you consider adopting an E. coli vaccine if presented with incentives such as some of those appearing above?’ Responses to this question provide data for the dependent variable in the probit regression: 145 (71%) responded ‘yes’, 8 ‘no’, and 50 ‘unsure/ don’t know’. For the purposes of the binary probit model, the dependent variable is coded 1 for ‘Yes’ and 0 for ‘No/unsure’.

4. The adoption environment: descriptive statistics A final sample of 203 survey responses were received, with the distribution of the sample closely corresponding to the cow-calf population across the three provinces (46.7% of respondents from Alberta, 35% from Saskatchewan, and 17.7% from Manitoba).7 The mean herd size was 210, ranging from 2 to 11,000. The mean age of respondents was 54, ranging from 29 to 77. Just over one third (34.5%) of the sample had high school level education, with another third (35.5%) having completed trade school/college and around 28% holding a university degree. This sample consists of relatively well-educated respondents which may reflect the mode of recruitment (producer animal health database) and the online nature of the survey. The sample is also comprised of relatively experienced cattle producers, with just over one third (35.5%) having over 35 years of experience as the principal decision-maker in the cow-calf operation, and another third (33.5%) having between 21-35 years of experience, while 27% of respondents had 5 to 20 years of experience. The survey data paint a rich picture of the vaccine adoption environment for this set of survey respondents, providing a useful backdrop to the discussion of the probit regression results that follows. Cow-calf producers expressed a fairly high degree of awareness of E. coli, with 76% of survey respondents indicating that they were either very familiar or somewhat familiar with E. coli 0157:H7. Only 5% of respondents were currently using the E. coli vaccine and a further 7% had previously used the vaccine but were not doing so at the time of the survey. Just under half of the respondents (46%) indicated that they had heard of the vaccine but had not used it. The remaining 42% of respondents had never heard of the vaccine. Use (either currently or previously) and familiarity with the vaccine were both associated with a higher stated willingness to adopt the vaccine in response to an adoption incentive. All of the producers currently using the vaccine and 86% of respondents who had previously used the vaccine indicated a positive willingness to adopt given the right incentive, while 75% of those who had previously heard of but never used the vaccine and 62% of those who had not previously heard of the vaccine indicated a willingness to adopt if the right incentive was in place. Primary sources of information on E. coli included producer associations (37%), followed by veterinarians (29%), then government agencies such as provincial government ministries (10%). About 11% of respondents indicated having no source of information regarding E. coli. A producer’s veterinarian was typically the first point of contact regarding E. coli for 44% of respondents, followed by government information services (18%), the Internet (17%), producer associations (13%), other cow-calf producers (7%), with only a couple of respondents (1%) indicating consultants. Asked who bears primary responsibility for reducing the risk of E. coli problems in the beef supply chain, the majority (56%) of respondents identified beef processors/packers. Only 15% of respondents indicated cow-calf producers, another 15% indicated feedlots, 8% said consumers and only 5% identified regulators such as the CFIA. Nobody indicated that the primary responsibility lay with food retailers. Conversely, survey respondents were asked who would benefit the most from a vaccine to reduce colonization and shedding of E. coli by cattle. The most popular response (41%) was consumers, followed by processors/packers (26%) 7 Respondents could exit the survey at any time with incomplete surveys not included in the final sample size, therefore, the final sample includes only fully completed survey responses. Information on the number of incomplete surveys or non-responses is not available.

International Food and Agribusiness Management Review

355

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

and cow-calf producers (21%), with the other potential agents (feedlots, retailers, municipal water security agencies) receiving relatively few responses. Responses to these questions – specifically whether or not cow-calf producers hold primary responsibility and are the primary beneficiaries – provide data for two of the explanatory variables in the probit regression. Figure 1 summarizes responses to the locus of control questions and shows that cow-calf producers within the survey sample on average tended towards an internal locus of control with respect to their perceived ability to control E. coli risks in their farm operation, although a larger proportion were neutral with respect to the external locus of control measures. The extent to which an E. coli vaccine fits within the producer’s existing vaccination protocol is expected to influence the likelihood of adoption. The vast majority (87%) of respondents vaccinate their calves, with 49% indicating that they did so twice a year, and 39% one a year. The majority of respondents (77%) typically carry out the vaccination themselves (or an employee does the vaccination), with only about 5.4% using their veterinarian. This suggests that an E. coli vaccine would need to be reasonably accessible and easy to use to remain consistent with typical vaccination protocols. 27% of respondents indicated they spend between $1 and 4.99 per calf on vaccinations per year, while 30% spend CAD $5 to 6.99 per calf, and 28% of respondents spend more than $7 per calf. At $3 per dose, the E. coli vaccine is a non-trivial addition to these costs. Finally, the extent to which various factors deter respondents from adopting the vaccine was explored (yes/ no responses). Focused on barriers to adoption, this set of survey questions differs from the BWS adoption

External locus of control

60%

Internal locus of control

56%

50% 44% 39%

40% 30%

26%

20% 14%

10% 0%

10% 6%

6% 2%

Completely Somewhat disagree disagree

Neutral

Somewhat Completely agree agree

Figure 1. Locus of control for E. coli risks: summary of responses (n=203). ‘Internal locus of control’ measured as average of responses to three Likert scale questions: ‘I feel in control of potential E. coli contamination due to my existing management practices/interventions’; ‘Whether or not I’m successful in mitigating/ controlling E. coli depends mostly on my own ability’; ‘To a great extent E. coli incidences on my cow-calf operation are determined by the management practices I have in place’. ‘External locus of control’ measured as average of Likert scale responses to three questions: ‘My success as a cow-calf operator depends mostly on luck’; ‘It is not advisable for me to plan too far ahead by enhancing my current management practices because E. coli incidences are such that they cannot be fully prevented’; ‘To a great extent E. coli incidences on my cow-calf operation are determined by factors beyond my control’. International Food and Agribusiness Management Review

356

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

A number of insights arise from this overview of the adoption environment for the E. coli vaccine and provide context for the probit analysis that follows. Although survey respondents exhibited a fairly high degree of awareness of E. coli and almost half of respondents had heard of the E. coli cattle vaccine, only around 5% were using the vaccine which is consistent with previous assessments of vaccine usage rates (Grier and Schmidt, 2009). A potential disconnect is evident between perceptions of who bears primary responsibility for risk reduction within the beef supply chain and who is the primary beneficiary of the vaccine. Few (15%) of producers believed that cow-calf producers bear primary responsibility for risk reduction or were the primary beneficiaries of an E. coli vaccine (21%), suggesting that commercial incentives to adopt the vaccine were weak. There are many potential barriers to adoption, but uncertainty over benefits and over efficacy (in a whole supply chain sense) appear to have resonated particularly strongly across the survey sample. We turn now to a formal consideration of the factors influencing willingness to adopt with the results of the probit analysis. 90% 80%

76.8% 71.4%

70%

68.5%

60% 49.3%

50% 41.4%

38.9%

40% 30%

26.1%

24.1%

20% 10%

n io pt fa

ar ot

od fo os

g in ov pr

ou tv

ou tp

ay -o ff/

dd i

tio

im n io pt

dd o ne a

in cc va li co

E. n ti en

stm ve

s er uy

tb in nt ra ar

w 't on sd ue

iss ty fe sa

od Fo

fe ee m ld ou w n tio

do p ra he

he t w to as

ty in ta er nc U

od fo g in

nc ha en n

ha st ue ss ri

he ot to n

ve gi y rit io

pr er at

co er ov y

sil ea ot en

ar n op tio

ad ith w

d te ia oc

ss sa ng e le

e m

ol oc ot pr n io at in cc va y m to de d ee sn ng e ha fc so Lo t

Ch al

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

incentives outlined in Table 2 and adds further richness to our understanding of both the incentives and barriers to adoption. Uncertainty over the market benefits of adopting the vaccine and over the efficacy of the vaccine as an intervention measure featured strongly in these responses, as shown in Figure 2.

Figure 2. Barriers to adopting an E. coli vaccine (percentage of respondents) (n=203). Percentage of respondents answering ‘Yes’ to the question: ‘Please indicate whether the following issues deter you from adopting an E. coli vaccine’.

International Food and Agribusiness Management Review

357

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

5. Probit regression results

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Cow-calf producers were asked whether they would consider adopting the E. coli vaccine if presented with the right incentives such as those appearing in Table 2. Responses to this question form the dependent variable in the probit analysis. The full probit model was run on all 32 variables specified in Tables 1 and 2. All estimations were conducted using STATA 2013 (StataCorp LP, College Station, TX, USA). Parameter estimates for the full model are provided in Supplementary Table S1. The results reveal that 20 of these variables are significant determinants of willingness to adopt, including a producer’s level of education, prior awareness of the vaccine, and perceptions of who bears primary responsibility for risk reduction, which is consistent with the assertions of Adesina and Baidu-Forson (1995). An external locus of control with respect to the perceived ability to control E. coli risks was also significant, as were all 13 adoption incentives as captured by the BWS scores. Other producer characteristics, such as gender and age, as well as the location (province) of the cow-calf operation, herd size, retained ownership, the relative importance of the cow-calf business to the producer’s livelihood, and the length of time the producer anticipating continuing in the business, were not statistically significant8. Following the estimation of the full probit model inclusive of all variables, a stepwise assessment of model fit allowed insignificant variables to be dropped one-by-one, with the model re-estimated sequentially as each variable was omitted. The resulting reduced model dampens the effects of the insignificant variables and produces more robust marginal effects estimations. Parameter estimates from the reduced model estimations are presented in Table 3. For purposes of interpretation, the marginal effects of these parameter estimates are needed. The marginal effect of an independent variable measures the impact of a change in an independent variable (e.g. Xi) on the expected change in the dependent variable (e.g. Y) 9. Thus, a marginal effect shows the change in probability when the explanatory or independent variable increases by one unit. Alternatively, for a dummy variable, marginal effects measure discrete change or how predicted probabilities change as the binary independent variable changes from 0 to 1. The marginal effects from the reduced model parameter estimates are presented in Table 4. The estimated marginal effect for education shows that the change from less than college education to at or above college increases the probability of willingness to adopt an E. coli vaccine by 20.3%. It is possible that producers with higher levels of education have a better understanding and appreciation of technologies to reduce food safety risks. Unsurprisingly, prior awareness of the E. coli vaccine had a positive and significant (P≤0.05) effect on stated willingness to adopt the vaccine. The estimated marginal effects show that on average, prior awareness increased willingness to adopt by 16.1%. The largest marginal effects include several of the ‘primary responsibility’ variables. Recall that these variables measure the effect on willingness to adopt of a producer’s perception that the primary responsibility for reducing risks of E. coli problems within the beef supply chain lies with the indicated party rather than with cow-calf producers (the reference category). The negative and significant marginal effect for each of these four variables is consistent with a priori expectations, which suggest that cow-calf producers are less likely to be willing to adopt the vaccine if they believe the primary responsibility lies elsewhere in the supply chain. The perception that primary responsibility lay with feedlots, consumers, and regulators were all highly 8 The

Chi-square value shows the goodness of fit of the data to the model. A value of 51.43 indicates the choice of the explanatory variables in the binary probit model explained the variations in adoption decisions fairly well. An ordered probit model was also estimated, where the dependent variable took the order of the three categories of ‘yes’, ‘don’t know/unsure’ and ‘no’ responses. Results of the ordered probit model, which are similar in nature, are not reported here in the interests of brevity but are available in Ochieng’ (2015). 9 Calculation of the marginal effects was as follows: Y = ɸ (β +β χ +β χ +…+β χ ), thus δY/δχ = β Ø (β +β χ +β χ +…+β χ ). 0 1 1 2 2 n n i i 0 1 1 2 2 n n

International Food and Agribusiness Management Review

358

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Table 3. Binary probit reduced model – parameter estimates.1 Variable

Coefficient2

Std. err.

P-value

Level of education Awareness of technology primary responsibility for E. coli risk reduction – feedlots Primary Responsibility E. coli risk reduction – processor/packer Primary Responsibility E. coli risk reduction – consumers Primary Responsibility E. coli risk reduction – regulators External locus of control Beef products can be traced to my farm Premiums available through branded programs Recommendation from my veterinarian Subsidy to compensate cost of adoption available Government recommending use of vaccine My reputation as producer is at risk My buyer requiring use of vaccine Feedlots providing assurance for a booster Duration of immunity greater than 6 months Through vaccination I can reduce supply disruptions My neighbors adopting vaccine Attraction of a new set of buyers I can include vaccine in an existing vaccination routine Constant N Pseudo R-sq.

0.639*** 0.522** -1.391*** -0.807** -1.497*** -1.599*** -0.379*** 0.882*** 0.787*** 0.759*** 0.717*** 0.673** 0.766*** 0.785*** 0.857*** 0.800*** 0.841*** 0.737*** 0.650** 0.763*** 1.831 203 0.212

0.228 0.228 0.489 0.418 0.550 0.611 0.152 0.297 0.288 0.290 0.284 0.286 0.284 0.292 0.293 0.293 0.295 0.282 0.279 0.297 0.644

0.005 0.022 0.004 0.054 0.007 0.009 0.013 0.003 0.006 0.009 0.012 0.018 0.007 0.007 0.003 0.006 0.004 0.009 0.020 0.010 0.004

Variable

Dy/Dx

Std. err.

P-value

Level of education Awareness of technology Primary responsibility for E. coli risk reduction – feedlots Primary responsibility for E. coli risk reduction – processor/packer Primary responsibility for E. coli risk reduction – consumers Primary responsibility of for E. coli risk reduction – regulators External locus of control Beef products can be traced to my farm Premiums available through branded programs Recommendation from my veterinarian Subsidy to compensate cost of adoption available Government recommending use of vaccine My reputation as producer is at risk My buyer requiring use of vaccine Feedlots providing assurance for a booster Duration of immunity greater than 6 months Through vaccination I can reduce supply disruptions My neighbors adopting vaccine Attraction of a new set of buyers I can include vaccine in an existing vaccination routine

0.203 0.161 -0.499 -0.232 -0.543 -0.576 -0.115 0.266 0.238 0.229 0.216 0.203 0.231 0.237 0.259 0.242 0.254 0.223 0.196 0.231

0.074 0.071 0.166 0.058 0.151 0.181 0.046 0.088 0.086 0.086 0.085 0.085 0.084 0.087 0.087 0.087 0.088 0.084 0.083 0.089

0.006 0.024 0.003 0.035 0.002 0.001 0.012 0.003 0.006 0.008 0.011 0.017 0.006 0.007 0.003 0.006 0.004 0.008 0.018 0.009

Dependent variable = willingness to adopt. = P<0.05; *** = P<0.01.

2 **

Table 4. Binary probit reduced model – marginal effects

International Food and Agribusiness Management Review

359

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

significant (P≤0.01) with beef packers significant at P≤0.05 (Table 3). The magnitude of the marginal effect for feedlots in Table 4 suggests that producers who believed feedlots, rather than cow-calf producers, held the primary responsibility were 50% less likely to be willing to adopt the vaccine, 23% less likely for those who believe packers bore the primary responsibility and 54% less likely for those who felt the primary responsibility lay with consumers. The clear inference here is that when cow-calf producers believe the primary responsibility for E. coli lies elsewhere in the beef supply chain, they are unlikely to adopt this type of vaccine, regardless of the effectiveness of the vaccine as a pre-harvest intervention strategy within food supply chains. The opportunity for consumers to mitigate the risks of exposure to E. coli contamination by properly cooking beef products might also explain the large marginal effect for consumers. As indicated in the discussion of the adoption environment, a slight majority of respondents felt that packers held primary responsibility, with relatively few placing that responsibility with consumers or regulators. Consistent with a priori expectations, the external locus of control variable is negative and highly significant (P≤0.01), suggesting that if a producer believes that E. coli risks lie beyond his/her control, he/she will less open to adopting this type of vaccine even given the right incentives. Cow-calf producers in this survey sample were 11.5% less likely to adopt the vaccine if they believed that controlling E. coli risks lay largely beyond their control. The remaining 13 variables in Tables 3 and 4 represent the effects of the adoption incentives as captured in the BWS scores. Consistent with a priori expectations, each of these variables is positive and significant at P≤0.01 with the exception of ‘Government recommending use of the vaccine’ and ‘attraction of a new set of buyers’ which are significant at P≤0.05 (Table 3).

6. Policy and managerial implications A number of policy and managerial implications flow from the analysis with respect to pre-harvest intervention strategies with the potential to reduce societal food safety risks, as well as informing future product development and commercialization strategies for similar products. Education and prior awareness of the technology are positive determinants of willingness to adopt, yet almost half of the sampled producers (42%) had not previously heard of the E. coli vaccine technology. Furthermore, if a producer believes that primary responsibility for E. coli risk reduction lies with other members of the supply chain he/she is significantly less likely to consider using the vaccine. Similarly, if producers believe that there is little they can do within their farm operation to mitigate E. coli risks, such that these risks lie largely outside of their control, they are significantly less likely to be willing to vaccinate their cattle. Taken together, these results suggest that steps to inform producers about the vaccine, its potential broader benefits, and the role of producers in mitigating E. coli risks will be important components of any strategy to encourage broader adoption of these types of technologies from a commercial perspective for firms in the animal health business or from a policy perspective if this is deemed to be in society’s best interest. Producer associations and veterinarians were previously identified as important sources of information and therefore may be particularly influential in the design and implementation of awareness programs for similar vaccines in the future. Efforts to enhance producer education and awareness regarding management techniques (including but not limited to use of an E. coli vaccine) to reduce risk of E. coli contamination or cross-contamination within the cow-calf operation, may also be warranted from a societal perspective. This may be particularly pertinent with respect to management practices that reduce contamination of irrigation or drinking water supplies, including removal of farm animals from the proximity of private water supplies (such as wells, boreholes), prevention of ground water contamination or contamination of ready-to-eat crops from farm run-off, or keeping livestock out of ready-to-eat crops using fencing (LeJeune and Wetzel, 2007; Soon et al., 2011). Further work could examine the potential role of information provision in encouraging adoption, including what is communicated, how and by whom. International Food and Agribusiness Management Review

360

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

The low adoption rates for the Econiche vaccine, and its eventual removal from the market, suggest that tangible economic incentives within the beef supply chain are necessary for the adoption of these types of technologies. In this study, willingness to adopt was framed in the context of thirteen potential adoption incentives (Table 2). Some of the adoption incentives take can be viewed as ‘carrots’, while others are ‘sticks’. Reputational effects include adoption incentives that take the form of a threat (the stick), such as the possibility that beef products could be traced to the producer’s farm or the producer’s reputation being at risk. Incentives related to the practicality of using the vaccine also influence willingness to adopt, including the duration of immunity, the ability to include the vaccination within existing vaccination routines, and a recommendation from the producer’s veterinarian. While market-driven incentives appear to dominate, the practicalities of using a new vaccine remain a relevant consideration; indeed Figure 2 shows that ‘lots of changes to my vaccination protocol’ would deter adoption for a non-trivial portion of respondents. The positioning of new vaccines therefore needs to take into consideration how the vaccine fits within producers’ existing vaccination routines in terms of timing, access and availability. The market-driven incentives (carrots) are of particular interest, reflecting requirements by new or existing buyers for vaccination, price premiums for vaccinated cattle, or an assurance that vaccination protocols are continued through the feedlot stage of production, and their importance is confirmed in other work (Ochieng’ and Hobbs, 2016)10. These incentives reflect the need for a whole supply chain approach to reducing E. coli risks within the food system, with interventions possible at the cow-calf, feedlot and packing plant levels of the chain. Indeed, if cow-calf producers vaccinate their calves but feedlots do not, then the potential societal gain from vaccination may be reduced. If only some cow-calf producers vaccinate their calves and those calves are subsequently mixed with non-vaccinated animals, the potential commercial benefits of reduced exposure to beef recalls and supply disruptions for those who vaccinate is likely to be considerably weakened. If vaccination by cow-calf producers induces beef packers to take less care and attention in mitigating E. coli risks during processing – a moral hazard effect – benefits from vaccination may be watered down. Exploration of these broader incentives to adopt risk mitigation interventions at other stages of the beef supply chain, and the extent to which they complement or substitute for risk mitigation actions at the cow-calf producer level lies beyond the scope of the present analysis, but represents a fruitful area for further research. Finally, this research is relevant for the development of similar livestock vaccinations or other animal health interventions with wider societal benefits for human or environmental health. Despite compelling scientific evidence of the efficacy of a vaccine in reducing E. coli colonization and shedding in cattle, adoption rates by cattle producers throughout the period in which the E. coli vaccine was commercially available were low, and the vaccine is no longer on the market. Since cattle are asymptomatic carriers of the E. coli pathogen, cattle producers receive little direct commercial benefit from the vaccine without stronger market or liability pressures to adopt. Early consideration of the economic incentives to adopt a vaccine could allow more informed research and development decisions into innovations that address food safety risks through preharvest interventions, as well as inform commercialization strategies for these products. In the absence of commercial pressure from downstream buyers for cow-calf producers and feedlots to adopt pre-harvest risk mitigation measures such as vaccination, there is little incentive for adoption. Future product development and commercialization strategies for these types of products should incorporate an understanding of the supply chain dynamics within the cattle/beef sectors in which the product is to be commercialized, including the extent to which downstream buyers are likely to require or encourage vaccination by their suppliers.

10 Ochieng’ and Hobbs (2016) present a detailed analysis of the underlying best-worst scaling experiment, finding the three most influential incentives

to be: ‘my buyer requiring use of the vaccine’, ‘subsidy to compensate costs of adoption’ and ‘premiums available through branded beef programs’. Heterogeneity in producer responses to adoption incentives is explored using a latent class cluster analysis which identifies a market/supply chain oriented cluster, a production protocol cluster and a risk averse producer cluster.

International Food and Agribusiness Management Review

361

Ochieng’ and Hobbs

Volume 20, Issue 3, 2017

Acknowledgements The authors gratefully acknowledge funding for this research provided by the Farm Credit Canada (FCC) MSc. Scholarship Program and the Alliance for Food and Bioproducts Innovation (AFBI) Scholars Program (Saskatchewan Ministry of Agriculture).

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2016.0177. Methods S1. Standard probit model specification. Table S1. Binary probit full model – parameter estimates.

References Adesina, A.A. and J. Baidu-Forson. 1995. Farmer’s perceptions and adoption of new agricultural technology: evidence from analysis in Burkina Faso and Guinea, West Africa. Agricultural Economics 13: 1-9. CanFax. 2014 Canadian beef cow population. CanFax Research Services. Available at: http://tinyurl.com/ gnwvq25. Cobanoglu, F. 2012. Is export orientation a major motivator for the adoption of food safety systems in the Turkish dried fig firms? Journal of Agriculture and Rural Development in the Tropics and Subtropics 113: 31-42. Ellis-Iversen, J., A.J.C. Cook, E. Watson, M. Nielen, L. Larkin, M. Wooldridge and H. Hogeveen. 2010. Perceptions, circumstances and motivators that influence implementation of zoonotic control programs on cattle farms. Preventative Veterinary Medicine 93: 276-285. Grier, K. and C. Schmidt. 2009. E. coli O157 risk reduction: economic benefit to Canada: report to Bioniche food safety. George Morris Centre, Guelph, Ontario, Canada. Heiman, K.E., R.K. Mody, S.D. Johnson, P.M. Griffin and H.L. Gould. 2015. Escherichia coli O157 Outbreaks in the United States, 2003-2012. Emerging Infectious Diseases 21: 1293-1301. Jayasinghe-Mudalige, K.U. and S. Henson. 2006. Economic incentives for firms to implement enhanced food safety controls: case of the Canadian red meat and poultry processing sector. Review of Agricultural Economics 28: 494-514. Jordan, D., S.A. McEwen, A.M. Lammerding, W.B. McNab and J.B. Wilson. 1999. Pre-slaughter control of Escherichia coli O157 in beef cattle: a simulation study. Preventive Veterinary Medicine 41: 55-74. LeJeune, J.T. and A.N. Wetzel. 2007. Pre-harvest control of Escherichia coli O157 in cattle. Journal of Animal Science 85: 73-80. Lewis, R.J., A. Corriveau and W.R. Usborne. 2013. Independent review of XL Foods Inc. beef recall 2012, Government of Canada. Available at: http://tinyurl.com/ly3k3b5. Loader, R. and J.E. Hobbs. 1999. Strategic responses to food Safety legislation. Food Policy 24: 685-706. Marette, S., B.E. Roe and M.F. Teisl. 2012. The welfare impact of food pathogen vaccines. Food Policy 37: 86-93. National Farm Attractors Network (NFAN). 2015. E-coli vaccine no longer available. Available at: http:// tinyurl.com/ho2z7t6. Ochieng’, B.J. 2015. Incentives for the adoption of socially beneficial technologies: the case of an E. Coli vaccine. MSc Thesis, University of Saskatchewan, Saskatoon, Canada. Available at: http://tinyurl. com/z6njbz8. Ochieng’, B.J. and J.E. Hobbs. 2016. Incentives for cattle producers to adopt an E. coli vaccine: an application of best-worst scaling. Food Policy 59: 78-87. Scharff, R.L. 2012. Economic burden from health losses due to foodborne illness in the United States. Journal of Food Protection 75:121-131. Signorini, M.L. and H.D. Tarabla. 2010. Interventions to reduce verocytotoxigenic Escherichia coli in ground beef in Argentina: a simulation study. Preventative Veterinary Medicine 94: 36-42. International Food and Agribusiness Management Review

362

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

Ochiengâ&#x20AC;&#x2122; and Hobbs

Volume 20, Issue 3, 2017

Smith, G.C., S.E. Goebel, C.R. Culbert and L.A. Guibault. 2013. Reducing the public health risk Escherichia coli 0157 exposure by immunization of cattle. Canadian Journal of Public Health 104:e9-e11. Smith, D.R., R.A. Moxley, R.E. Peterson, T.J. Klopfenstein, G.E. Erickson, G. Bretschneider, E.M. Berberov and S. Clowser. 2009. A two-dose regimen of a vaccine against type iii secreted proteins reduced Escherichia coli O157:H7 colonization of the terminal rectum in beef cattle in commercial feedlots. Foodborne Pathogens and Disease 6: 155-161. Sockett, P., S. Goebel, N.P. Valera, A. Guthrie, J. Wilson, L.A. Guilbault and W.F. Clark. 2014. Verotoxigenic Escherichia coli: cost of illness in Canada including long-term health outcomes. Journal of Food Protection 77: 216-226. Soon, J.M., S.A. Chadd and R.N. Baines. 2011. Escherichia coli O157:H7 in beef cattle: on farm contamination and pre-slaughter control methods. Animal Health Research Reviews 12: 197-211. Teisl, M.F. and B.E. Roe. 2010. Consumer willingness-to-pay to reduce the probability of retail foodborne pathogen contamination. Food Policy 35: 521-530. Tonsor, G.T. and T.C. Schroeder. 2015. Market impacts of E. coli vaccination in U.S. Feedlot cattle. Agricultural and Food Economics. 3:7. Verbeek, M. 2004. A Guide to Modern Econometrics (2nd Edition). John Wiley and Sons Ltd, West Sussex, UK. Withee, J., M. Williams, T. Disney, W. Schlosser, N. Bauer and E. Ebel. 2009. Streamlined analysis for evaluating the use of preharvest interventions intended to prevent Escherichia coli O157:H7 illness in humans. Foodborne Pathogens and Disease 6: 817-825.

International Food and Agribusiness Management Review

363

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0177 - Tuesday, May 09, 2017 8:30:27 PM - IP Address:24.21.169.207

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2015.0064

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Received: 12 May 2015 / Accepted: 21 December 2016

Possibility of exporting halal-certificated food in Hokkaido, Japan: acceptance by Malaysian consumers RESEARCH ARTICLE Satoko Kubota a, Hiroichi Konob, and Takuhiro Chibac aAssistant

Professor, and bAssociate Professor, Department of Animal and Food Hygiene, Obihiro University of Agriculture and Veterinary Medicine, Nishi 2-11, Inada-cho, Obihiro, Hokkaido 080-8555, Japan cStaff

member, Grain and Feedstuff Department, Kanematsu Corporation, Seavans-N, Shibaura 1-2-1, Minato-ku, Tokyo 105-8005, Japan

Abstract Expectations of exporting food and agricultural product to Malaysia have been substantially growing in Japan. Although Japan has two halal certification bodies recognized by Jabatan Kemajuan Islam Malaysia (JAKIM), it is unclear whether Malaysian consumers trust its certified processed foods. Thus, this study aims to clarify Malaysian consumers’ preference for Hokkaido ice cream by conducting a conjoint analysis. The results show no significant difference in the level of consumer trust in halal certifications, indicating a distribution based on both Malaysian and Japanese certification bodies. Therefore, it is possible that Malaysian consumers would accept Japanese products with a halal certification approved by JAKIM. In addition, the willingness to pay for Hokkaido ice cream reported a high satisfaction value. However, there is no significant effect on food functionality possibly because Malaysian consumers’ awareness of the health benefits remains low. Keywords: export possibility, halal certification, food functionality, conjoint analysis, Malaysia JEL code: D10, D11, D18 Corresponding author: skubota@obihiro.ac.jp

365

Kubota et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

1. Introduction Globalization has significantly advanced in recent years, leading to not only technological and economic progress but also a constantly changing food, agriculture, and food safety and security environment. While efforts to establish trade rules led by the World Trade Organization and further progress in global free trade (e.g. Free Trade Agreement, Economic Partnership Agreement, and Trans-Pacific Strategic Economic Partnership Agreement) are advantageous to create new business opportunities, such developments have highlighted the several risks associated with food and agriculture, which were originally characterized by an unclear food chain (Huynen et al., 2005). In Japan, the food self-sufficiency rate based on calorie supply remains low at less than 40% owing to its shrinking domestic markets for production lands and the shortage of successors in the agricultural sector (Yoshii and Oyama, 2011). Agricultural and livestock product exports have attracted much attention in Japan as a strategy to increase the competitiveness of its agricultural sector, which serves as a foundation for the local economy and promotes the concept of ‘strong agriculture’ aimed at improving the food selfsufficiency rate. The strategy gained further momentum in August 2013, when the Ministry of Agriculture, Forestry and Fisheries (MAFF) set a goal to increase the exports of agricultural and marine products to 9.7 billion dollars (1 trillion yen) by 2020 (MAFF, 2016). In addition, the global Japanese food boom and the expansion of markets in developing countries as a result of population growth and economic progress have become factors underlying increased exports. This situation has been particularly observed in Hokkaido, Japan, where many public and private industries are engaging in foreign export businesses. Among several identified foreign core markets, the halal market mainly operating in Asia is an emerging and prospective market viewed with high expectations. According to Thomson Reuters (2014), the global halal market was estimated at 2,001 billion dollars and the halal food market at 1,292 billion dollars in 2013, of which the halal food market in Asia accounted for 49.7% (642 billion dollars). The halal market is, thus, considered a key target market for Japan, a country with close geographical and economic ties with other Asian regions. As part of the Third National Agricultural Policy enacted in 1998, Malaysia developed systems and infrastructure to become a halal hub in the Muslim world (Othman et al., 2009), thus increasing Japan’s expectations to become a halal exporting partner. However, manufacturing halal products and guaranteeing that they meet the halal standards is difficult in Japan given that Islamic customs are rarely practiced by its population. Although obtaining a halal certification can serve as an entry into the halal market, the trustworthiness of halal certified-products in Japan and consumers’ willingness to pay (WTP) for such products are unclear. Thus, this study aims to quantitatively estimate consumer preferences for Japanese agricultural and livestock products from Hokkaido to discuss the possibility of exporting them to Malaysia and obtaining a mutual halal certification from the Malaysian government’s only halal certification agency, Jabatan Kemajuan Islam Malaysia (JAKIM), Department of Islamic Development Malaysia. To do so, a conjoint analysis is conducted of responses by Malaysia’s general consumers to surveys comparing the various characteristics of agricultural and livestock products. The article ‘The strategies for the export of agricultural products and foods by country and item,’ issued in 2013 by MAFF, Japan, identifies six types of core items exported to Malaysia: beef1, fruits and vegetables, confectioneries, condiments, marine products, and milk and dairy products (MAFF, 2013). This study focuses on milk and dairy products, particularly ice cream, given their potential for product expansion due to the relatively high processability.

Although beef exports have been currently banned owing to bovine spongiform encephalopathy, interviews with the Department of Veterinary Services revealed that discussions are underway to resume exports.

International Food and Agribusiness Management Review

366

Kubota et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

2. Literature review Research on halal history has been conducted within the area of Islamic studies, with particular emphasis on humanistic elements, such as Islamic jurisprudence and folklore. However, the scope of research rapidly expanded to the social and natural sciences as a result of improvements made to the halal certification system, mainly in Malaysia during the 2000s. The underlying reason for this development was the need to understand social conformity and validity as the halal designation became institutionalized and documented as halal certification. Abdul et al. (2009) identified differences in the levels of awareness about halal logos and food ingredients on the basis of individual religious beliefs. Rezai et al. (2012) conducted studies on the degree of consumer trust in processed foods with a halal logo and social factors comprising trust and showed that products manufactured in non-Muslim countries or without JAKIM’s halal logo received lower ratings of trust. Mohayidin and Kamarulzaman (2014) investigated processed food traits preferred by consumers and clarified that JAKIM’s halal logo had a more significant impact on consumer purchasing behavior than those by other countries. Akbiyik and Eroglu (2016) conducted a factor analysis in Turkey and found that six factors promoted halal-certified product awareness: promotion, location, attractiveness and quality, halal certification, price, and market demand. Ismoyowati (2015) found that while a halal certification was important, taste and nutrition still received highest priority in chicken-based food consumption in Indonesia. In their study on an organic product market in Switzerland, Janssen and Hamm (2011a) showed that WTP was significantly higher for almost all products with certification logos than those without one, suggesting that certification logos enhance credibility. Further, consumer perceptions significantly change when a logo exists, although in many cases, this is not based on the objective knowledge of organic labeling schemes (Janssen and Hamm, 2011b). Kamaruddin et al. (2012) examined consumers’ demand and WTP for the additional costs of halal logistics (certification requirements). These studies suggest that consumers in Malaysia consider products from non-Muslim countries and a halal certification (logos) from bodies other than JAKIM unfavorable from a credibility perspective. However, it is noteworthy that Japan has two organizations that have been certified by Malaysia’s JAKIM but consumers’ attitudes toward these bodies are yet to be clarified. Discussing non-Muslims’ views of halal designations, Rezai et al. (2010) explored the halal concept and consumer perceptions and attitudes toward halal food products and demonstrated that the perception of halal is influenced by factors such as the environment, sustainability, animal welfare, and food safety. Mathew et al. (2014) adopted a consumer psychological perspective and showed that non-Muslims also view the halal designation favorably by linking it to cleanliness and quality. It appears that the halal designation has gained a certain level of acceptance from non-Muslims, even in Malaysia, and therefore, obtaining a halal certification could increase a product’s market share. The evaluations in this study target the Muslim population as well as all Malaysian consumers. As Lever and Miele (2012) mentioned, the halal market has transcended Muslim countries to various European ones in the past 15 years, resulting in a significant growth in research findings (Ahmed, 2008; Bonne and Verbeke, 2008; Verbeke et al., 2013). Yusof and Shutto (2014) also examined the current status of halal markets in Japan. However, these studies focus on domestic consumption and do not investigate consumer preferences or markets from the perspective of an export partner country.

3. Halal concept and Malaysian halal system Shariah law, the Islamic customary law, views every ‘good thing’ in daily Muslim life as ‘halal’. It is appropriate to define ‘halal’ as ‘positive, good, or yes’ in the English language. Shariah law defines dietary customs (criteria) for Muslims and specifies foods that can be labeled as halal. On the other hand, ‘haram’ refers to forbidden acts, for example, the consumption of pork or alcohol. ‘Syubhah’ is a condition that does not clearly fall into either category. Halal foods cannot be contaminated with haram ingredients; more specifically, they must be free from all haram ingredients and not produced using haram-contaminated equipment, and thus, haram should be spatially separated from halal at every stage, that is, from production and processing to sales distribution. In addition, the term ‘halalan toyyiban’ refers to products that are not International Food and Agribusiness Management Review

367

Kubota et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

harmful and safe to be consumed as underlined by the Syariah law (Omar et al., 2013). As Allah says in the Quran, ‘O mankind! Eat of that which is lawful and good on the earth’ (Surah Al Baqarah 2: 172). They ask you (O Muhammad SAW) what is lawful for them (as food) ... Lawful unto you are at Tayyibaat (all kind of halal foods) (Surah Al Maidah 5: 4). As explained, Islam requires that Muslims find rizk (sustenance) and consume food that is halal and toyyiban because it ensures a healthy living that reflects good attitudes and behaviors (Yousef, 2010). It constitutes the concept of wholesomeness, which includes quality, cleanliness, and safety (Omar et al., 2013). The Malaysia Standard (MS) for ‘Halal Food -Production, Preparation, Handling and Storage- General Guidelines (MS1500: 2009)’ is used as a basis for halal certification in Malaysia. Halal certifications are currently given to not only food products but also various goods and services such as cosmetics and pharmaceutical drugs and those in the financial and tourist industries. Halal standards have also been recently established in the logistics area (MS2400: 2010). These developments suggest the continuous expansion of areas subject to halal certification requirements. JAKIM is the only certification agency under the direct authority of the Malaysian Prime Minister’s Office. The Halal Industry Development Corporation (HDC) was established in September 2006 by the Malaysian government to facilitate Malaysia’s halal hub project and assist in tasks related to halal businesses. The government’s intention is to not only view halal from a religious perspective but also promote Malaysia as a major player in the global halal industry (Zakaria, 2008). HDC coordinates the overall development of the halal industry in Malaysia, focusing on the development of halal standards, audit and certification, and capacity building for halal products and services. It promotes the participation and facilitates the growth of Malaysian companies in the global halal market (HDC, 2013). The Malaysian government’s intentional establishment of the certification agency separate from the business promotion agency is a unique step. JAKIM also assesses and audits halal certification bodies in other countries. The agency then approves of them as mutual certification bodies if they meet the requirements. In principle, obtaining a certification from a mutual certification body is necessary to export products with halal-certified labels. Japan has two mutual certification bodies: Japan Halal Association (JHA) and Japan Muslim Association (JMA) (JAKIM, 2015).

4. Methodology Questionnaire surveys This study proposes the following hypotheses to explore the possibility of exporting Hokkaido ice cream from Japan to Malaysia. First, consumer preferences do not differ by the halal certification body that grants the certification, provided the products are mutually certified by JAKIM. Past studies have shown that JAKIM’s halal certification or logo is highly effective in gaining consumer acceptance. JAKIM acknowledges that Malaysian consumers accept products as long as they are certified by mutual certification bodies. Therefore, this study hypothesizes that a halal logo issued by a mutual certification body will gain the same level of acceptance as one by JAKIM. Second, Hokkaido ice cream will be well-received by Malaysians. ‘Hokkaido’ as a brand is favorably viewed by Malaysian consumers who harbor desires for ‘snow.’ The Hokkaido Fair held frequently in Malaysia is also a factor contributing to Malaysian consumers’ familiarity with Hokkaido (Japan External Trade Organization, 2012). Thus, it is likely that Hokkaido ice cream will be positively perceived by consumers in Malaysia. Third, Malaysian consumers are concerned with the functional attributes of ice cream. Diabetes has become widespread in Malaysia (Mafauzy, 2005). Moreover, with Japan now facing a similar situation and given the rising demand for healthy food in recent years, many related technologies and research findings have become available. Therefore, it is possible to earn a good reputation among Malaysian consumers by using Japanese technology and adding diabetes-related functional properties.

International Food and Agribusiness Management Review

368

Kubota et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Muslims and non-Muslims in multiethnic Malaysia are likely to express differing views on halal certifications. However, contrasting views on halal certifications can be expected even among Muslims given the varying degrees of religious devotion. This study analyzes the proposed hypotheses by examining consumer characteristics, which are broadly divided into three categories: (1) a certification requesting group that always requests for halal certifications; (2) a voluntary certification requesting group that only requests for the exclusion of pork and alcohol in food products; and (3) an indifferent group that is indifferent to the halal designation. To test the abovementioned hypotheses, questionnaire surveys were conducted targeting Malaysian consumers. Since the Kuala Lumpur Metropolitan Area (Klang Valley) is the most significant and a large economic area and can serve as a basis for exporting Hokkaido products, this study targeted only those consumers who frequented supermarkets in Kuala Lumpur and adjacent areas. We conducted interview surveys in front of AEON Malaysia stores (Bandar Utama Store and Mid Valley Store) during November 3-8, 2014. Consumers were given the questionnaires at the store and randomly selected (haphazard sampling). A total of 205 consumers responded to the surveys. This way, the collected samples reflected consumers’ characteristics, data for which were gathered from the store officer. Design of conjoint analysis and model Since almost no agricultural and livestock product has been imported by Malaysia, it was not possible to conduct an analysis using actual market data. From this viewpoint, Hokkaido agricultural and livestock products can be viewed as goods that are not sold through monetary transactions in general society or with an assigned price (non-market goods). According to Bateman et al. (2002), the following two types of analysis can be performed for such non-market goods: a revealed preference analysis based on individuals’ actual behavior and a stated preference analysis of individual’s responses to expected outcomes if they were to take a certain action in the future. This study utilized one of the stated preference methods, a conjoint analysis. In a conjoint analysis, respondents are shown various profiles comprising combinations of attributes and levels and asked their preferences for each one. A conventional conjoint analysis uses an assessment method based on ranking or rating. Choice-based conjoint analysis has been adopted more frequently in recent years since it allows respondents to select a product they would consider purchasing and the option ‘will not purchase any product’ (e.g. Mohayidin and Kamarulzaman, 2014). Using the method, Hu et al. (2005) evaluated foods with genetically modified ingredient information and Iwamoto et al. (2012) examined the influence of hazard analysis and critical control point (HACCP) and eco-labeling on consumer preferences. Grunert et al. (2014) studied consumer preferences for foods with sustainability labeling that offered a sense of environmental awareness and morality. Studies have also evaluated consumer preferences on the basis of the characteristics of foods such as their appearance (Jervis et al., 2014; Oltman et al., 2014). Attributes and level of ice cream (Table 1) include country of origin, halal logo, functionality, and price. For a comparative analysis, this study includes Thailand as a level for the ‘country of origin’ attribute since Thai ice cream imports have a successful track record in Malaysia. In addition, unlike meat products, meeting halal requirements is not mandatory for ice cream products under customs regulations. Therefore, ‘without logo’ is added as a level for the ‘halal logo’ attribute. ‘Addition of oligosaccharides’ is used as a level for ‘functionality’ since oligosaccharides are believed to help inhibit diabetes. The three countries of origin are fixed. A total of eight pages comprising four choices, including ‘will not purchase any product,’ were prepared. However, two questionnaires each consisting of four pages were finally utilized since presenting eight pages to each respondent was too overwhelming.

International Food and Agribusiness Management Review

369

Kubota et al.

Volume 20, Issue 3, 2017

Table 1. Attributes and levels of ice cream. Attributes

Levels

Country of origin Halal logo Functionality Price (RM/90 ml)2,3

Malaysia, Japan (Hokkaido), and Thailand JAKIM (Malaysia), JMA (Japan), CICOT1 (Thailand), and None Addition of oligosaccharides, no addition of oligosaccharides 3, 6, 9, and 12

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

The Central Islamic Committee of Thailand (CICOT) is a halal certification body in Thailand and has been certified as a mutual certification body by JAKIM. 2 During the study period, 1 RM=0.2994 USD (Mitsubishi FUJI Research and Consulting Co. Ltd, 2016). 3 90 ml is the volume of an ice cream cup per person.

A choice-based conjoint analysis is a discrete choice model based on random utility theory. It assumes that respondent i gains utility Uij; accordingly, the following equation is formulated for four choices j including ‘will not purchase any product’. Uij = Vij + εij

(j = 1, 2, 3, 4)

(1)

where Vij is the observable and deterministic part of indirect utility determined by product and respondent attributes. εij is the error term that cannot be measured. Here, we assume that the respondent selects a choice with the highest utility. Probability pi( j) is defined such that respondent i’s utility from choice j is higher than that from other choices, q: Pi ( j) = Prob {(Vij + εij > Viq + εiq)

for all q ∈ C (j ≠ q)} (2)

when a set of choices is designated as C, which is assumed to be identical for all respondents. Here, we assumed that error term εij follows the Gumbel distribution with a location parameter of 0 and scale parameter of 1 and the independence of irrelevant alternatives. The probability that respondent i selects choice j is: Pi ( j) =

exp(Vij) ∑q∈C exp(Viq)

(3)

Equation 3 is a conditional logit model and the function parameter V can be estimated using maximumlikelihood estimation. In this study, measurable V of the utility function is set to be linear and the model is expressed as follows: Vij = β ASCm ASCijASCm + ∑n β n χijn + β price Priceij

= β ASCM ASCijASCM + β ASCH ASCijASCH + β ASCT ASCijASCT + β Halal_M Halal_Mij

(4)

+ β Halal_J Halal_Jij + βHalal_T Halal_Tij + β Func Funcij + β price Priceij

where ASCijASCm is the alternative-specific constant (m=M,H,T), χijn are attributes other than price, and Priceij is the price of choice j selected by respondent i. An explanation of the estimation variables is provided in Table 2. Marginal willingness to pay (MWTP) can be obtained from the following equation. The total differentiation of Equation 4 is:

∑ ∂χ∂V dχ n

∂V dPrice = dV (5) ∂Price

where the utility level is assumed to be fixed as an initial level (dV=0) and attributes other than focused attribute χl are fixed as initial levels (dχn=0, n≠l). MWTP, MWTPχl , for attribute χl is calculated using: International Food and Agribusiness Management Review

370

Kubota et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Table 2. Estimation variables. Variable

Explanation

ASCijASCM ASCijASCH ASCijASCT Halal_Mij Halal_Jij Halal_Tij Funcij Priceij

= 1 if the option is Malaysian ice cream = 1 if the option is Japan (Hokkaido) ice cream = 1 if the option is Thailand ice cream = 1 if the option is JAKIM (Malaysia) halal logo = 1 if the option is JMA (Japan) halal logo = 1 if the option is CICOT (Thailand) halal logo = 1 if the option is functionality (oligosaccharides) = price of choice

dPrice MWTPχl = =– dχl

∂V ∂χl βl = – Price (6) ∂V β ∂Price

where β l corresponds to attribute χl. The value for which the coefficient of ASCASCm is divided by – β Price is the WTP when the attribute variables are valued at 0.

5. Results Respondents’ attributes Since children could not be included in the surveys, a majority of the respondents were in their 20s as opposed to the population statistics for Kuala Lumpur (Table 3) compiled by the Department of Statistics, Malaysia (2011). Respondents were then classified as per responses based on individual religious beliefs: of the 203 respondents, 43 (21.2%) were categorized in the halal certification requesting group, 58 respondents (28.6%) in the voluntary certification requesting group, and 102 (50.2%) in the indifferent group. The mean monthly income for the sample data was 3,858 RM2, whereas that as per the Department of Statistics, Malaysia (2016) was 4,585 RM in 2014. In addition, while the mean monthly income for urban areas was 5,156 RM as per statistical data, our sample data showed a lower value. This can be attributed to responses primarily by students in their teenagers and 20s, who tend to have lower income earnings than other Mid Valley Store visitors. The average volume of ice cream purchased by respondents was 369 ml per month and the average price paid was 15.6 RM/90 ml. These figures are similar to those obtained through interviews held with retail store staff and buyers. Assessment of halal certification As per the halal certification system based on JAKIM’s MS1500:2009, a certification requires compliance with the International Organization for Standardization 9001, HACCP, good manufacturing practices, and good hygiene practice management systems along with related MS standards. These food safety management systems are believed to increase the safety of halal-certified products and strengthen consumer trust. Thus, the following two questions were presented regarding halal labeling and food cleanliness: ‘What is your impression of the halal certification?’ and ‘How much do you understand about the foreign halal certifications and do you check whether a product is halal certified before purchasing it?’ A cross-tabulation was performed to assess consumers’ view of halal certifications and the importance of a product being halal certified. 2

Participants were asked to choose from four household income categories: 1 for 0-15,000 RM, 2 for 15,001-45,000 RM, 3 for 45,001-105,000 RM, and 4 for greater than 105,001 RM. Then, the mean income was calculated by multiplying the average value of each category by each category’s sample size and dividing it by the entire sample size. Next, we estimated the mean monthly income. In the case of categories 1 and 4, we used values 15,000 RM and 105,001 RM.

International Food and Agribusiness Management Review

371

Kubota et al.

Volume 20, Issue 3, 2017

Table 3. Comparisons of respondent attributes and statistical data (adapted from the Department of Statistics, Malaysia (2011) and supplemented with survey data). Sex (n=205)

Religion (n=201)

Sample (%)

Statistics (%)1

64 141 14 78 49 28 22 14 61 37 89 4 10

31.2 68.8 6.8 38.1 23.9 13.7 10.7 6.8 30.3 18.4 44.3 2.0 5.0

50.9 49.1 30.7 22.5 16.7 13.4 9.0 7.6 46.4 5.8 35.7 8.5 3.6

2010 statistical data for Kuala Lumpur.

As per the results, the number of respondents who believed that a halal certification included the ‘cleanliness’ aspect was the highest, followed by those referencing ‘Shariah law’, in the certification requesting and voluntary certification requesting groups. In the indifferent group, on the other hand, the most common response was ‘nothing in particular’. Nevertheless, about 20% respondents from the indifferent group believed that the halal certification included the ‘cleanliness’ aspect. However, it appears that consumers from all groups had an insufficient understanding of the halal certification system, including traceability (Figure 1). As for foreign halal certifications, the results show that most respondents do not check whether a product is certified, despite them being aware of the certification. As mentioned, halal-certified products entering Malaysia must be certified by a JAKIM-recognized entity. However, since Malaysian consumers possibly place a high level of trust in JAKIM, they do not always check if the product has a foreign halal certification. Thus, it is reasonable to assume that halal-certified products distributed within Malaysia will not be treated differently by consumers on the basis of a domestic or foreign halal certification (Figure 2).

80 Response rate (%)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Age (n=205)

Male Female 10-19 20-29 30-39 40-49 50-59 Over 60 Muslim Christian Buddhist Hindu Others

Sample

60 Certification requesting group Voluntary certification requesting group Indifferent group

40 20 0

Cleanliness

Traceability

Sharia Law

Health

Nothing in particular

Figure 1. Respondent impressions of halal certification. The respondents were allowed to provide multiple answers. Each response rate indicates the percentage in relation to the total number of responses from each group. Certification requesting group (n=43), voluntary certification requesting group (n=58), and indifferent group (n=102).

International Food and Agribusiness Management Review

372

Kubota et al.

Volume 20, Issue 3, 2017

Response rate (%)

50 40

20 10 0

Checks products for certification before purchasing

Understands certification but does not check products for certification

Has no understanding of certifications

Figure 2. Degree of understanding of foreign halal certification and whether consumers check products for certification. Assessments of Hokkaido agricultural and livestock products Respondents were asked multiple choice questions on the type of Hokkaido agricultural and livestock products they would be interested in purchasing. The results show that, in general, most respondents were interested in purchasing marine products, particularly those from Japan, given the popularity of sushi. Next were confectioneries and dairy products (Table 4), indicating the potential for dairy products, such as ice cream, as a core export product from Hokkaido. In addition, 61.8% respondents from the indifferent group were interested in purchasing Hokkaido dairy products. This is higher than the 53.4 and 51.2% respondents from the voluntary certification and certification requesting groups. The results for respondentsâ&#x20AC;&#x2122; impression of Hokkaido agricultural and livestock products show that those in the certification requesting group perceive the products less favorably (Table 5). A possible reason is that Malaysian consumers do not have much access to Hokkaido agricultural and livestock products. In the indifferent group, on the other hand, a larger percentage considers the products to be palatable and safe and about 28% considered them to be expensive, which is the highest compared to the other two groups. About 46% respondents from the certification requesting group viewed the products as healthy, which is higher than those considering the products as palatable. Importantly, the results show that unfavorable opinions regarding safety are higher, which can be attributed to concerns about damages due to the Fukushima Daiichi nuclear plant disaster in 2011. Many countries have imposed import restrictions on foods from Japan and to this effect, the Malaysian government has increased the stringency of quarantine tests and reports on foods imported from Japan (Johnson, 2011). Given this situation, even if foods are tested for radioactive material in Japan (Bachev and Ito, 2014), the spread of damaging rumors across overseas markets are beyond oneâ&#x20AC;&#x2122;s control.

Certification requesting group Voluntary certification requesting group Indifferent group

32.6 29.3 30.4

23.3 24.1 17.6

9.3 8.6 11.8

32.6 31.0 36.3

51.2 53.4 61.8

39.5 41.4 43.1

Confectioneries

Marine products

Vegetables and fruits

Dairy products

Milk

Chicken

Other meat products

Table 4. Percentage of purchase interest in Hokkaido agricultural and livestock products (%).1,2

Beef

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Certification requesting group Voluntary certification requesting group Indifferent group

41.9 63.8 58.8

65.1 48.3 55.9

1 Respondents were allowed to provide multiple answers. Each response rate indicates the percentage in relation to the total number

of responses from each group. 2 Certification requesting group (n=43), voluntary certification requesting group (n=58), and indifferent group (n=102).

International Food and Agribusiness Management Review

373

Kubota et al.

Volume 20, Issue 3, 2017

Table 5. Impressions of Hokkaido agricultural and livestock products (%).1,2 Certification requesting group Voluntary certification requesting group Indifferent group

Palatable

Expensive

Healthy

Safe

34.9 34.5 48.0

18.6 24.1 28.4

37.2 46.6 35.3

18.6 17.2 21.6

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

1 Respondents were allowed to provide multiple answers. Each response rate indicates the percentage in relation to the total number

of responses from each group. 2 Certification requesting group (n=43), voluntary certification requesting group (n=58), and indifferent group (n=102).

Willingness to pay for ice cream The results for the first hypothesis – consumer preferences do not differ by the halal certification body that grants the certification – show no significant difference in the utility value of the three logos between the certification requesting and voluntary certification requesting groups (Table 6). However, a comparison of the two groups indicates that the utility value of each logo is higher for the certification requesting group, which is an expected outcome since the group has a higher demand for certifications. On the other hand, statistical significance was found only for the JAKIM logo in the indifferent group, suggesting low awareness of foreign halal certification (Figure 2) and that trends appear to lead to the statistical outcome. The second hypothesis – consumer preferences for Hokkaido ice cream are high – was supported for the voluntary certification requesting and indifferent groups. However, no statistically significant result was found for the certification requesting group. For the third hypothesis – consumers are concerned with the functional attributes of ice cream – no statistically significant difference was found for functionality in any group. This is possibly because consumers do not have a sufficient understanding of oligosaccharides and ‘functional’ food products. Table 6. Conjoint analysis results for ice cream.1 Variable names

Certification requesting group

Voluntary certification requesting group

Indifferent group

ASCASCM ASCASCH ASCASCT Halal_M Halal_J Halal_T Func Price Number of samples Log likelihood AIC2 AIC/N ρ2 Adj-ρ2

-0.23 (0.704) -0.804 (0.760) -2.239 (0.956)** 2.586 (0.492)*** 2.714 (0.638)*** 2.713 (0.846)*** 0.386 (0.375) -0.048 (0.044) 114 -105.827 227.658 1.997 0.229 0.211

2.183 (0.680)*** 3.137 (0.713)*** 1.127 (0.887) 1.258 (0.356)*** 1.360 (0.442)*** 1.377 (0.699)** 0.112 (0.359) -0.141 (0.047)*** 137 -117.115 250.162 1.826 0.139 0.122

1.148 (0.403)*** 2.400 (0.420)*** -0.555 (0.646) 0.530 (0.241)** 0.384 (0.289) -1.129 (1.139) -0.439 (0.277) -0.099 (0.032)*** 240 -221.856 459.600 1.915 0.051 0.040

1 ***

and ** denote statistical significance at the 1% and 5% level; the figures in parentheses indicate standard errors. = Akaike information criterion.

2 AIC

International Food and Agribusiness Management Review

374

Kubota et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Next, MWTP for ice cream was calculated for each consumer category. The findings indicate that MWTP for the halal logo is considerably high in the certification requesting group; however, that for the halal certification of Thai ice cream is negative. Although under the system, obtaining a halal certification is not necessary for ice cream products, the results appear to indicate that the certification requesting group requires that products meet halal requirements. It is noteworthy that the reliability of price variables after measuring the certification requesting groupâ&#x20AC;&#x2122;s MWTP is low since no statistically significant results were obtained. MWTP for Hokkaido ice cream is 22.2 RM/90 ml and 24.2 RM/90 ml for the voluntary certification requesting and indifferent groups, which is higher than that for Malaysian ice cream. Since the average price a respondent pays for ice cream is 15.6 RM/90 ml, it is reasonable to assume a higher WTP for Hokkaido ice cream and that it can be sold to these groups at a higher price even without a halal logo, although the logo may further increase their WTP for the product. Nevertheless, it is necessary to emphasize the Japanese halal certification and halal logo for the indifferent group. Possibility of exporting Hokkaido ice cream Next, we conduct a simple estimation of the expected retail price of ice cream exported to Malaysia and examine the possibility of exporting Hokkaido ice cream. Shojikiya, a retailer in Malaysia, sells ice milk manufactured by a major Japanese confectionery company at 7 RM/100 ml (price in Japan: 3 RM/100 ml). Accordingly, Hokkaido ice cream is expected to be sold at a retail price that is about twice that of the price in Japan when exported to Malaysia. Thus, since Hokkaido ice cream is sold at around 10 RM/90 ml in Japan, the retail price in Malaysia is expected to be 20 RM/90 ml. Consequently, it can be sufficiently competitive as MWTP for the voluntary certification requesting and indifferent groups is expected to be 22-24 RM/90 ml. MWTP further increases if the halal logo is placed on the product. Although it is possible to obtain a JAKIM certification by exporting Hokkaido raw milk ice cream liquid to Malaysia and processing it locally, a company must be able to first source and partner with a local company, which is rather labor and time intensive. In other words, placing a JAKIM logo on Hokkaido raw milk-based ice cream products may not be a realistic option. More than 50% Malaysian consumers stated that they would be interested in purchasing Hokkaido dairy products and it is essential to provide palatable ice cream products with emphasis on health factors (Table 4 and 5). Hokkaido produces a large amount of high quality beet sugar and raw milk used as ice cream ingredients. As for concerns over safety issues, companies have been actively implementing HACCP owing to encouragement from the Hokkaido Government Office and each municipality. In addition, the Japanese government has been conducting regular tests for radioactive material. Therefore, it is possible to convince Malaysian consumers that Hokkaido dairy products are safe.

6. Conclusions This study explored the possibility of exporting Hokkaido products and obtaining a halal certification with focus on ice cream products. The results showed no significant difference in the level of consumer trust in halal certification, which supports distribution based on both domestic and foreign halal certification bodies. In other words, it is possible that Malaysian consumers will accept Japanese products with a halal certification approved by JAKIM. In addition, Hokkaido ice cream products would be positively perceived on a satisfaction level. Since Malaysia imports various agricultural, livestock, and food products from various countries, it is likely that Malaysian consumers will view such products favorably. Further, since the level of health consciousness, including the functional attributes of ice cream products, is yet to develop in Malaysia, low demand for such products can be expected. Drawing on these findings, export strategies for Hokkaidoâ&#x20AC;&#x2122;s agricultural and livestock products were considered and MWTP for Hokkaido ice cream products was estimated to be higher than the expected retail price of imported ice cream products. However, the characteristics of our sample do not necessarily match International Food and Agribusiness Management Review

375

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Kubota et al.

Volume 20, Issue 3, 2017

those of the Kuala Lumpur statistics. As explained in the Methodology section, our questionnaire survey was conducted in AEON Malaysia to ensure a realistic trade situation. The female sample is larger than that of the statistical data, although most of women live with their family (93.6%) and often, the opinions at home are reflected in the results. Therefore, it is possible to say that the results do not significantly differ from reality. In addition, the income level is lower than that in the statistical data, possibly because AEON Malaysia does not target consumers with high income levels. Despite these limitations, the results indicate that even those with low income levels are willing to pay a higher price for Hokkaido ice cream. As for the number of participants by religion, Muslims are fewer in number since the store is believed to be frequented by Chinese consumers. Therefore, an overestimation of the results is possible. However, individualâ&#x20AC;&#x2122;s religious devotion tends to differ: Muslims were categorized under the voluntary certification requesting group (32.8%) and indifferent group (13.1%) and Christians and Buddhists requested for the halal certification (2.7 and 6.7%). In sum, high positive opinions regarding Hokkaido ice cream suggest the possibility of exporting other processed food products as well. In recent years, halal certification has been largely sensationalized because of the rapidly increasing interest in halal in Japan. The quantitative assessments in this study can serve as a reference when making objective decisions regarding the exports of processed food products with the premise of obtaining a halal certification. At the same time, the most important task for future Japanese export promotion is helping local residents gain a sufficient understanding of halal logos by JHA and JMA, both of which are certified by JAKIM as mutual certification bodies, and emphasizing the safety of Hokkaido agricultural and livestock products and processed products.

References Abdul, M., H. Ismail, H. Hashim and J. Johari. 2009. Consumer decision making process in shopping for Halal food in Malaysia. China-USA Business Review 8: 40-47. Ahmed, A. 2008. Marketing of halal meat in the United Kingdom: supermarkets versus local shops. British Food Journal 110: 655-670. Akbiyik, F. and A.H. Eroglu. 2016. A study to determine consumersâ&#x20AC;&#x2122; attitude, expectation, perceptions for Halal certified products: Isparta province as an example. International Journal of Social Science Studies 4: 56-68. Bachev, H. and F. Ito. 2014. Implications of Fukushima nuclear disaster for Japanese agri-food chains. International Journal of Food and Agricultural Economics 2: 95-120. Bateman, I.J., R.T. Carson, B. Day, M. Hanemann, N. Hanleys, T. Hett, M. Jones-Lee, G. Loomes, S. Mourato, E. Ozdemiroglu, D.W. Pearce, R. Sugden, and J. Swanson. 2002. The foundations of economic valuation. In: Economic valuation with stated preference techniques? A manual. Edward Elgar Publishing Limited, Cheltenham, UK, pp. 13-60. Bonne, K. and W. Verbeke. 2008. Muslim consumer trust in halal meat status and control in Belgium. Meat Science 79: 113-123. Department of Islamic Development Malaysia (JAKIM). 2015. The recognized foreign halal certification bodies and authorities. Available at: http://tinyurl.com/zt2u7f2. Department of Statistics, Malaysia. 2011. Population distribution and basic demographic characteristics report 2010. Available at: http://tinyurl.com/jzygocd. Department of Statistics, Malaysia. 2016. Principal statistics of household income. Available at: http:// tinyurl.com/h66ns3k. Grunert, K.G., S. Hieke and J. Wills. 2014. Sustainability labels on food products: consumer motivation, understanding and use. Food Policy 44: 177-189. Halal Industry Development Corporation Malaysia (HDC). 2013. Official website of global halal support centre, Malaysia. Available at: http://tinyurl.com/blwvs6l. Hu, W., M.M. Veeman and W.L. Adamowicz. 2005. Labelling genetically modified food: heterogeneous consumer preferences and the value of information. Canadian Journal of Agricultural Economics 53: 83-102.

International Food and Agribusiness Management Review

376

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Kubota et al.

Volume 20, Issue 3, 2017

Huynen, M.M.T.E., P. Martens and H.B.M. Hildetink. 2005. The health impacts of globalisation: a conceptual framework. Globalization and Health 1: 1-12. Ismoyawati, D. 2015. Halal food marketing: a case study on consumer behavior of chicken-based processed food consumption in central part of Java, Indonesia. Agriculture and Agricultural Sciences Procedia 3: 169-172. Iwamoto, H., Y. Yamamoto, K. Sato, and M. Sawada. 2012. Effects of safety and freshness of food on Japanese consumers’ choice of milk. In: Food consumption: empirical studies of Japanese dietary, edited by K. Sasaki. Tsukuba Shobo, Tokyo, Japan, pp. 146-155. Janssen, M. and U. Hamm. 2011a. Certification logos in the market for organic food: what are the consumers’ willingness to pay for different logos? In: EAAE 2011 Congress, Zurich, Switzerland, 30 August to 2 September 2011, pp. 1-12. Available at: http://tinyurl.com/h9lva4b. Janssen, M. and U. Hamm. 2011b. Product labelling in the market of organic food: consumer preferences and willingness-to-pay for different organic certification logos. Food Quality and Preference 25: 9-22. Japan External Trade Organization. 2012. Survey on Japanese food consumption trend in Malaysia (in Japanese). Japan External Trade Organization, Tokyo, Japan. Jervis, S.M., M.G. Jervis, B. Guthrie, and M.A. Drake. 2014. The efficacy of using photographs to represent attributes of sliced sandwich bread in an adaptive choice-based conjoint. Journal of Sensory Studies 29: 64-73. Johnson, R. 2011. Japan’s 2011 earthquake and tsunami: food and agriculture implications. Congressional Research Service. Available at: http://tinyurl.com/h25zntp. Kamaruddin, R., H. Iberahim and A. Shabudin. 2012. Willingness to pay for halal logistics: the lifestyle choice. Social and Behavioral Sciences 50: 722-729. Lever, J. and M. Miele. 2012. The growth of halal meat markets in Europe: an exploration of the supply side theory of religion. Journal of Rural Studies 28: 528-537. Mafauzy, M. 2005. Diabetes control and complications in private primary healthcare in Malaysia. Medical Journal of Malaysia 60: 212-217. Mathew, V.N., A.M.R.A. Abdullah and S.N.M. Ismail. 2014. Acceptance on halal food among non-Muslim consumers. Social and Behavioral Sciences 121: 262-271. Ministry of Agriculture Forestry and Fisheries (MAFF). 2013. The strategies for the export of agricultural products and foods by country and item (in Japanese). Available at: http://tinyurl.com/z6ymfer. Ministry of Agriculture Forestry and Fisheries (MAFF). 2016. FY 2015 annual report on food, agriculture and rural areas in Japan. Available at: http://tinyurl.com/gqqo425. Mitsubishi UFJ Research and Consulting Co., Ltd. 2016. Foreign exchange rate (in Japanese). Available at: http://tinyurl.com/z3zxa6g. Mohayidin, M.G. and N.H. Kamarulzaman. 2014. Consumers’ preferences toward attributes of manufactured halal food products. Journal of International Food and Agribusiness Marketing 26: 125-139. Oltman, A.E., S.M. Jervis and M.A. Drake. 2014. Consumer attitudes and preferences for fresh market tomatoes. Journal of Food Science 79: S2091-S2097. Omar, E.N., H.S. Jaafar and M.R. Osman. 2013. Halalan toyyban supply chain of the food supply industry. Journal of Emerging Economics and Islamic Research 1: 1-12. Othman, P., I. Sungkar and W.S.W. Hussin. 2009. Malaysia as an international halal food hub: competitiveness and potential of meat-based industries. ASEAN Economic Bulletin 26: 306-320. Rezai, G., Z. Mohamed and M.N. Shamsudin. 2012. Assessment of consumers’ confidence on halal labelled manufactured food in Malaysia. Pertanika Journal of Social Science and Humanities 20: 33-42. Rezai, G., Z. Mohamed, M.N. Shamsudin and E.F.C. Chiew. 2010. Non-Muslims’ awareness of halal principles and related food products in Malaysia. International Food Research Journal 17: 667-674. Thomson Reuters. 2014. Demand profile and market sectors. State of the global Islamic economy 2014-2015. Report 48-50. Thomson Reuters, New York, NY, USA. Verbeke, W., P. Rutsaert, K. Bonne, and I. Vermeir. 2013. Credence quality coordination and consumers’ willingness-to-pay for certified Halal labelled meat. Meat Science 95: 790-797.

International Food and Agribusiness Management Review

377

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2015.0064 - Tuesday, May 09, 2017 8:31:55 PM - IP Address:24.21.169.207

Kubota et al.

Volume 20, Issue 3, 2017

Yoshii, K. and T. Oyama. 2011. Applying mathematical programming food supply model for improving Japanâ&#x20AC;&#x2122;s food self-sufficiency ratio. In: 10th International Symposium on Operations Research and its Applications (ISORA 2011), Dunhuang, China, 28-31 August 2011, pp. 28-50. Available at: http:// tinyurl.com/gv8lnjh. Yousef, D. K. 2010. UAE: halal food numbers look tasty. Size of global Muslim population creates significant customer base. Available at: http://tinyurl.com/jn9f8m7 Yusof, S.M. and N. Shutto. 2014. The development of halal food market in Japan: an exploratory study. Social and Behavioral Sciences 121: 253-261. Zakaria, Z. 2008. Tapping into the world halal market: some discussions on Malaysian laws and standards. Shariah Journal 16: 603-616.

International Food and Agribusiness Management Review

378

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2016.0115

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Received: 5 July 2016 / Accepted: 15 November 2016

What do consumers think about farm animal welfare in modern agriculture? Attitudes and shopping behaviour RESEARCH ARTICLE Heinke Heise and Ludwig Theuvsen Chair-Management in Agribusiness, Georg-August-University of Goettingen, Department of Agricultural Economics and Rural Development, Platz der Göttinger Sieben 5, 37073 Göttingen, Germany

Abstract Several food crises damaged the image of the agricultural sector and consumers have lost trust, especially in animal production practices. Large parts of society believe that animal welfare standards in livestock production need to be improved. As a result, numerous animal welfare products have emerged on the market. This consumer paper identifies five clusters and, thus, strategic groups for the purchase of animal welfare products within the large group of consumers that differ significantly in their attitudes towards modern agriculture, their perception of animal welfare, their social acceptance of meat consumption and their shopping behaviour. Even personal differences are found between the clusters. Based on the results, we derived specific marketing implications for each cluster. These implications can help to develop a more differentiated market segment for animal welfare products in terms of animal welfare level and required price premium, enabling consumers to make product choices according to their preferences. Keywords: attitudes, consumers, cluster analysis, farm animal welfare, shopping behaviour JEL code: Q13, M31 Corresponding author: hheise@gwdg.de

379

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

1. Introduction Due to several food crises and food scares in the recent past, the image of the agricultural sector has heavily been damaged and consumers have lost trust, especially in animal production practices (e.g. Grunert, 2005; Kanis et al., 2003; Vanhonacker and Verbeke, 2014). The broader public is increasingly concerned about farm animal welfare (FAW), and large parts of society believe that animal welfare standards in livestock production need to be improved (De Jonge and van Trijp, 2013; European Commission, 2007a). Several studies have shown that German citizens are particularly attentive to issues concerning FAW (European Commission, 2005; Schulze et al., 2008). Thus, moral and ethical considerations increasingly determine consumers’ meat consumption (Mäkiniemi et al., 2011; Rozin et al., 1997; Schröder and McEachern, 2004). Moral and ethical value conflicts surrounding meat purchase decisions have deepened in Western societies in recent decades; in response, consumers have developed varying strategies to reduce their feelings of guilt (Te Velde et al., 2002). Some tend to ‘de-animalize’ meat so as not to be reminded of the fact that animals have to be slaughtered before becoming a food product (Buller and Cesar, 2007; Schröder and McEachern, 2004). Others respond to their growing concerns by decreasing their consumption of meat or even by becoming vegetarians or vegans (Harper and Henson, 2001; Vanhonacker et al., 2010). Additionally, the number of consumers who prefer meat from more animal-friendly production systems has been constantly increasing for the past several years (Blandford and Fulponi, 1999; Burda Community Network, 2009; Deimel et al., 2010; European Commission, 2007b; Lusk and Norwood, 2012; Schulze et al., 2008). As a result, a number of animal welfare products (AWPs) have emerged on the market. But, despite promising signals from market research studies, these products are still very rare and, with few exceptions (e.g. Switzerland, the United Kingdom and the Netherlands), AWPs have not yet attained any great importance in the European meat market. Even though the majority of EU consumers advocate higher animal welfare standards, they face a number of key barriers during the purchase of animal-friendly meat and meat products: lack of appropriate information on animal welfare standards, information overload and asymmetries, lack of availability of animal-friendly products in retail shops, perceived lack of individual influence on overall welfare standards in livestock production, disassociation from food production and high additional costs (Harper and Henson, 2001). These barriers often lead to attitude-behaviour discrepancies, which are also known as the ‘consumer-citizen gap’ (Coff et al., 2008; Harvey and Hubbard, 2013; Vanhonacker et al., 2010). Numerous studies have already investigated determinants of consumer behaviour in food purchase situations and concluded that this behaviour is crucially dependent on consumer attitudes and beliefs as well as on personal characteristics like sociodemographics and lifestyle (e.g. Nocella et al., 2010; Schulze et al., 2008; Vanhonacker and Verbeke, 2014). However, most studies have tended to focus only on specific aspects of this topic (e.g. the role of consumer trust in shopping behaviour or consumers’ conception of animal welfare) (Frewer et al., 2005; Nocella et al., 2010; Vanhonacker et al., 2010). These studies indicate clearly that, depending on various parameters, consumers differ considerably in their attitudes towards FAW and, thus, in their buying behaviour. Because of these differences, separate target groups can be identified regarding the purchase of AWPs. To meet the requirements of these heterogeneous consumer demands and to reduce the discrepancies between consumer attitudes and shopping behaviour, the market segment for AWPs prospectively needs to become more differentiated (e.g. Verbeke, 2009). The present study uses the broad literature base as a starting point for a comprehensive and representative quantitative survey of German consumers. The specific aspects analysed in the existing studies are brought together to acquire a detailed overview of consumers’ attitudes and personal characteristics as well as their shopping behaviour. Thus, this paper complements previous research with a comprehensive empirical study intended to identify different clusters and, thus, strategic groups for the purchase of AWPs within the larger group of consumers and helps to develop appropriate market segmentation.

International Food and Agribusiness Management Review

380

Heise and Theuvsen

Volume 20, Issue 3, 2017

The remainder of the paper is organized as follows: In Section 2 we present the literature that forms the basis of the study. Section 3 introduces materials and methods, and Section 4 presents the empirical results. The paper closes with a discussion of the results (Section 5), some implications for food marketing (Section 6) and concluding remarks (Section 7).

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

2. Literature review: state of research Several studies have investigated consumers’ attitudes towards modern agriculture and FAW, their information behaviour and the effects of these factors on consumers’ shopping behaviour. The results of these studies are presented in the following section, as they constitute the basis for our empirical research. Overall, consumers are increasingly aware of grievances in agricultural production and have, therefore, become more and more critical towards modern agriculture. Kayser (2012) showed that 70% of press articles on agriculture and 85% of the comments on agricultural issues in online social media are negatively connoted. The deterioration of its image is especially true for intensive livestock production. Press and other media coverage of the criticism of animal husbandry, transportation and slaughtering have resulted in a highly emotional controversy about current practices in conventional livestock production (Deimel et al., 2011; Franz et al., 2012). Schulze et al. (2008) showed that only 18% of their respondents rated the husbandry conditions in modern livestock production as ‘rather good’ or ‘good’. Furthermore, several studies have found that consumers who are particularly critical of modern animal production systems show a higher willingness to pay a price premium for particularly animal-friendly products compared to unconcerned consumers (Schulze et al., 2008; Vanhonacker et al., 2007). In a very recent study, 89% of the German consumers surveyed expressed a willingness to pay more for more animal-friendly products, with young consumers showing the highest willingness to pay more for higher animal welfare standards (BMEL, 2016). These numbers clearly illustrate that large parts of society have lost trust in animal production practices (Grunert, 2005; Kanis et al., 2003; Vanhonacker and Verbeke, 2014). This is underlined by Busch et al. (2015), who reported that only 26.8% of German consumers trust livestock producers and only 38.2% believe that farmers take good care of their animals. Consumers use the information given on food products to make informed choices (Frewer et al., 2004; Mayfield et al., 2007). Since FAW is a credence attribute whose true level cannot be evaluated by consumers either before or after consumption, labels have become important communication tools about animal welfare standards. However, consumers still consider the level of animal welfare-related information to be insufficient (Vanhonacker et al., 2010). A European study determined that 41% of consumers wish to be better informed about the living conditions of farm animals (European Commission, 2007a). Furthermore, people often feel misled by labels and fear they will be cheated by retailers (Schröder and McEachern, 2004). This loss of trust constitutes a severe problem particularly for the marketing of AWPs, as not only animal welfare but also the reliability of labels is a credence attribute that cannot be verified by the consumers. Thus, credence goods such as food products labelled animal-friendly are only purchased if the consumers trust the information provided (Vanhonacker and Verbeke, 2014). The public controversy about livestock production focusses especially on large-scale farming, a term that has clearly negative associations for the majority of consumers (Boehm et al., 2010; Busch et al., 2013). Thus, the public is increasingly concerned about the conditions of farm animals in large-scale intensive production systems, where a high number of animals are kept per unit (Heyder and Theuvsen, 2009; Kanis et al., 2003; Kayser et al., 2012a; Vanhonacker et al., 2010). Busch et al. (2013) showed that consumers who are particularly concerned about large-scale farming are more willing to change their consumption habits than consumers who are less concerned about intensive livestock production methods. These concerns have led to an increasing number of people reducing their meat consumption in recent years (Harper and Henson, 2001, Vanhonacker et al., 2010). Even though the vast majority of consumers still International Food and Agribusiness Management Review

381

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Heise and Theuvsen

Volume 20, Issue 3, 2017

want to eat meat on a regular basis (BMEL, 2016), a growing segment of the population is convinced that animals should not be reared for slaughter (Povey et al., 2001) and attach high importance to ethical issues when buying animal-based food products (Vanhonacker and Verbeke, 2009). This increasingly threatens the social acceptance of meat production and consumption, as more and more consumers question the legitimacy of conventional livestock production systems (Kauppinen et al., 2010). Several studies have shown that the frequency of meat consumption crucially influences peopleâ&#x20AC;&#x2122;s attitudes towards FAW and their willingness to pay a price premium for AWPs. Furthermore, people with strong ethical attitudes towards animal production tend to attach minor importance to price and the availability of animal food products (Schulze et al., 2008; Vanhonacker et al., 2007; Vanhonacker and Verbeke, 2009). Even involvement in agricultural topics influences consumersâ&#x20AC;&#x2122; perceptions of modern livestock production and, therefore, affects their shopping behaviour. Verbeke and Vackier (2004) showed that consumers who are more involved reveal a smaller discrepancy between attitude and actual behaviour and attach greater importance to FAW. However, alienation from agricultural production has led to low involvement and thus low awareness of the actual design of production processes involved in animal-based products for large parts of society in Western industrialized countries (Frewer et al., 2005; Kanis et al., 2003; Te Velde et al., 2002; Vanhonacker et al., 2010). According to Busch et al. (2013), only 26% of consumers feel rather well- or well-informed about agricultural topics. Against this background, many consumers use mass media, like television, newspapers or the internet, as sources of information about agricultural topics (TNS Emnid, 2012). These media are often characterized by very critical, negatively connoted news coverage on agriculture (Kayser, 2012). Despite (or because of) the overall alienation from the agricultural sector, there is growing interest in FAW in livestock production, at least from certain parts of society (De Jonge and van Trijp, 2012; Harper and Henson, 2001; Vanhonacker and Verbeke, 2014). This interest is positively associated with pro-welfare behaviour and lower perceived barriers for the purchase of products subject to enhanced animal welfare standards (Vanhonacker and Verbeke, 2009). Additionally, greater awareness of agricultural production processes often negatively affects attitudes towards modern agriculture (Kayser et al., 2012b). Highly interested and well-informed consumers are, therefore, often more critical towards livestock production than less informed consumers (Busch et al., 2013). Consumer concerns about livestock production are often related to the perceived level of FAW during production processes (Deimel et al., 2011; Franz et al., 2012; European Commission, 2005). A cross-national survey showed that 77% of European consumers believe that the FAW standards need to be improved (European Commission, 2005). However, these sentiments run even higher in Germany. In a recent study by the German Ministry of Food and Agriculture, 88% of the respondents agreed that FAW needs to be improved (BMEL, 2016). Consumers often associate higher animal welfare standards with higher product quality. In this way, animal welfare standards are linked to intrinsic quality attributes like taste, healthiness or product safety (Meuwissen and van der Lans, 2004; Phan-Huy and Fawaz, 2003). Thus, consumers transform the credence attribute FAW into an experience attribute that can be confirmed after purchase (Nocella et al., 2010). Attitudes towards FAW and shopping behaviour are also influenced by many sociodemographic and other personal characteristics. Several studies indicate that women are more likely to buy AWPs than men, as they accept higher price premiums for these products (Kehlbacher et al., 2012; Makdisi and Marggraf, 2011; Nocella et al., 2010; Schulze et al., 2008; Vanhonacker and Verbeke, 2009). Even age influences consumersâ&#x20AC;&#x2122; attitudes and, thus, affects their shopping behaviour. However, results vary with respect to this characteristic. While most studies indicate that older people are more animal welfare-sensitive (Makdisi and Marggraf, 2011; Schulze et al., 2008; Nocella et al., 2010), other studies conclude that middle-aged persons show the greatest willingness to pay price premiums for products with enhanced animal welfare standards (Vanhonacker and Verbeke, 2009). Negative relationships between age and the accepted price premium have even been observed International Food and Agribusiness Management Review

382

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

(Kehlbacher et al., 2012). Furthermore, the income and education of consumers are also decisive. Bennet and Blaney (2002) found that concern about FAW and willingness to pay more for AWPs increase with level of education and income (Kehlbacher et al., 2012; Makdisi and Marggraf, 2011; Vanhonacker and Verbeke, 2009). Occupation, household size, place of residence, preferred shopping location for meat and frequency of meat consumption are further determinants of consumers’ shopping behaviour (Makdisi and Marggraf, 2011; Nocella et al., 2010; Schröder and Mc Eachern, 2004; Schulze et al., 2008; Vanhonacker et al., 2010). To sum up, there is a broad range of information on various aspects of consumers’ attitudes towards FAW and their consumption of meat and other products of animal origin, as well as their information behaviour and shopping habits. But the overall picture still tends to be scattered, and a comprehensive picture is missing. The information presented in this chapter constitutes the basis for our comprehensive empirical study. We also included consumers’ perceptions of the potential market effects of higher animal welfare standards, which may be helpful in better anticipating the prospective market development for AWPs and the consequences for intensive livestock production.

3. Material and methods Study design German consumers throughout the entire country were surveyed in October 2015 by means of a standardized online questionnaire. In order to acquire a sample distribution representative of the German population with regard to gender, age, place of residence, education and income, respondents were recruited via a panel company1. After purging incomplete questionnaires and outliers, 516 data sets were left for further analysis. The statements regarding consumers’ perceptions of FAW were mostly measured using five-point Likert scales from -2 = ‘I totally disagree’ to +2 = ‘I totally agree’. Furthermore, nominally scaled questions were used to inquire about respondents’ sociodemographics and other personal characteristics concerning lifestyle. The questionnaire was divided into several parts. The first part concentrated on sociodemographic and lifestyle characteristics. Next, participants were asked to indicate what they consider particularly important for the level of FAW. Then, they were asked to evaluate various statements concerning FAW and the need to enhance animal welfare standards in modern livestock production. In addition, consumers were asked to give their personal opinions about the potential market for AWPs and their perception of the social acceptance of meat consumption. Even the effects of higher animal welfare standards on other quality attributes like taste or healthiness were questioned. The final section of the questionnaire dealt with consumers’ behaviour when purchasing food products and their willingness to pay a price premium for AWPs. Statistical analysis The data were analysed using SPSS Statistics 22 (IBM, Armonk, NY, USA) (Backhaus et al., 2011). First, frequency distributions of sociodemographic and lifestyle characteristics were conducted to acquire a brief overview of the participants. Next, an explorative factor analysis was conducted to reduce the number of items and to capture the central dimensions based on consumers’ perceptions of FAW (Franz et al., 2012). The analysis contained all Likert-scaled statements mentioned in the previous paragraph, excluding the variables concerning the perceived quality of AWPs and consumers’ shopping behaviour. We conducted a factor analysis to identify latent constructs underlying the respondents’ attitudes towards modern agriculture and FAW. Variables that correlated highly were grouped together in one factor to separate them from less highly correlated groups. Then, principal component analysis was used to summarize the 1 To

see details of the sample composition compared to the basic population in Germany see Supplementary Table S1.

International Food and Agribusiness Management Review

383

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

variables that load highly on one factor. We removed from the analysis all variables that showed loadings ≥0.4 on more than one factor because a clear assignment to only one factor was not possible (Backhaus et al., 2011). In order to facilitate the interpretation of the factors, an orthogonal Varimax rotation was used to maximize the variance of the squared factor loadings by column (Backhaus et al., 2011). The quality of the factor analysis was verified using the Kaiser-Meyer-Olkin criterion and the Bartlett test for sphericity with subsequent reliability analysis (Brosius, 2011). Based on the extracted factors, we carried out a cluster analysis to classify the consumers into groups according to their attitudes towards agricultural topics and FAW. We applied a hierarchical clustering method using the single-linkage method. The optimal number of clusters was then identified using Ward’s method. Depicting the merger process graphically through a dendrogram and applying the elbow criterion helped us to find the best cluster solution (Backhaus et al., 2011). We refined the resulting solution, using a K-means cluster analysis (Bacher et al., 2010; Janssen and Laatz, 2007) and used discriminant analysis to check its results (Backhaus et al., 2011). To characterize the different groups in greater depth, analysis of variance with post hoc tests were used showing no variance equity (Brosius, 2011; Everitt, 1998). Means of cluster-building factors and further cluster-describing variables usually only deviate slightly within the respective cluster while means of a variable in different clusters usually differ greatly (Brosius, 2011). Results of the analysis of variance must not be interpreted as a full significance test as data for the analysis of variance statistic is the same as for clustering. Thus, the post hoc tests do not represent independent tests of significance. For this reason, results should be interpreted as an indication for the mean differences between the clusters. The post hoc tests are of great importance to show differences between the clusters and to characterize the clusters in detail (Brosius, 2011).

4. Results Factor analysis To reduce complexity and to identify the most important factors influencing consumers’ attitudes towards modern agriculture and FAW, we conducted an explorative factor analysis. The final factor solution included six factors with 24 variables (Table 1). The first factor, ‘Involvement in agriculture and livestock production’, describes the knowledge and interest in agricultural topics from the consumers’ point of view, combining statements concerning their perceived knowledge, their interest and their information sources. The second factor, ‘Perception of animal welfare in livestock production’, summarises seven statements about how consumers view FAW and their perception of how farmers treat livestock. The third factor, ‘Conception of animal welfare and animal welfare standards’, combines four statements about the importance of health, natural innate behaviour and husbandry system for the level of animal welfare as well as the enhancement of animal welfare standards for livestock production. The fourth factor, ‘Animal welfare, the market and stakeholders’, comprises variables that reflect the consumers’ positions towards the market effects of higher animal welfare standards and the products of more animal-friendly husbandry systems. Additionally, the statement ‘Politicians, journalists and consumers cannot evaluate whether or not farm animals are kept under good conditions’ is also included in this factor. The fifth factor ‘Social acceptance of meat consumption’ summarizes two statements concerning the acceptance of eating meat among family and friends and by society in general. The last factor, ‘Animal welfare and farm size’, consists of two statements inquiring whether consumers think that the level of animal welfare depends on farm size. The tests conducted to examine the quality of the factor analysis indicated that all factors meet the common requirements. The factor analysis explains 65.41% of the total variance among the 24 variables. These variables are well suited for the factor analysis as the Kaiser-Meyer-Olkin measure of sampling adequacy is relatively high at 0.820. Additionally, Bartlett’s test of sphericity was statistically significant, yielding correlation coefficients for the population with values different from zero. The reliability analysis showed that the internal consistency of the factors is adequate (Backhaus et al., 2011; Bühl, 2010). In the next step, we used the determined factors as cluster-building variables. International Food and Agribusiness Management Review

384

Heise and Theuvsen

Volume 20, Issue 3, 2017

Table 1. Results of the factor analysis.1

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Factors and underlying statements2

Agree (%)3 FL4

Involvement in agriculture and livestock production (Cronbach’s alpha=0.922) I know a lot about agricultural topics. I regularly use media to keep informed about agricultural topics. I am interested in agricultural topics. I know a lot about livestock production. I regularly use the agricultural trade press for information on agricultural topics. Perception of animal welfare in livestock production (Cronbach’s alpha=0.816) Farm animals in livestock production systems feel comfortable. Farmers take good care of their farm animals. The health status of farm animals is good. Farm animals can show their natural innate behaviour in agricultural husbandry systems. The agricultural sector is honestly interested in improving FAW. A farm animal with good performance also feels comfortable. A farm animal that is not used to an outdoor paddock does not miss it. Conception of animal welfare and animal welfare standards (Cronbach’s alpha=0.830) Good health is particularly important for the level of animal welfare. The structural-technical systems used in barns are particularly important for the level of animal welfare. Animals must be able to engage in their natural behaviour; only then can they feel comfortable. The animal welfare standards for farm animals should be enhanced. Animal welfare, the market and stakeholders (Cronbach’s alpha=0.662) Higher national animal welfare requirements will lead to more imported meat. Higher national animal welfare requirements will lead to competitive disadvantages for German farmers on international markets. Products from more animal-friendly production systems will always occupy only market niches. Politicians, journalists and consumers cannot evaluate whether or not farm animals are kept under good conditions. Social acceptance of meat consumption (Cronbach’s alpha=0.794) Eating meat is socially accepted. My friends and family accept the consumption of meat. Animal welfare and farm size (Cronbach’s alpha=0.670) Small farms are more easily able to implement higher animal welfare requirements than large farms. Animals feel more comfortable on small farms than on large farms.

19.9 36.2 46.2 26.9 15.9

0.900 0.890 0.867 0.850 0.833

12.0 19.6 22.5 19.9

0.828 0.789 0.734 0.721

28.2 30.8 6.0

0.651 0.551 0.489

95.8 90.9

0.869 0.857

92.1

0.818

88.4

0.610

26.7 37.8

0.804 0.792

29.1

0.574

26.5

0.486

77.1 81.0

0.887 0.875

44.6

0.837

64.5

0.805

KMO (Kaiser-Meyer-Olkin measure) = 0.820; explained variance = 65.41%. 2 Scale from +2 = ‘totally agree’ to -2 =‘totally disagree’. 3 Numbers refer to the sum of the percent data relating to categories +2 and +1; n=657. 4 FL= Factor loading; n=524.

Cluster analysis Based on the factors identified, a cluster analysis was conducted. The aim was to group the consumers into clusters according to their involvement in agricultural topics and attitudes towards FAW and meat consumption.

International Food and Agribusiness Management Review

385

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

First, the single linkage method was applied in order to eliminate eight outliers from the sample. Since the scree test and the dendrogram did not clearly show the optimal number of clusters, additional plausibility considerations were undertaken to determine the optimal number of clusters, yielding a five-cluster solution (Backhaus et al., 2011). The approximate solution of Ward’s method was optimised by 18 iterations using K-means clustering (Brosius, 2011). Several criteria indicate that the solution reached is of high quality. The clusters are quite homogeneous as all the F-values are less than one. Additionally, eta is 0.69 on average, showing that there are significant differences among the cluster-building factors and the variance within the clusters is low. Furthermore, eta 2 is 0.43, indicating that, on average, the cluster-building factors can explain 43% of the variance between the clusters. Moreover, discriminant analysis confirmed that the accuracy of classification is 96.5% and, therefore, meets the requirement stipulated in the literature (Backhaus et al., 2011). To describe the clusters, mean comparisons were conducted using one-way analysis of variance based on the factors. To describe the clusters in greater detail, the following were also used for variance analysis: the individual variables in the factors, sociodemographic and lifestyle characteristics and variables describing the perceived effects of higher animal welfare standards on other quality attributes and the shopping behaviour of the respondents. In order to examine significant differences among the means of the clusters, Tamhane’s T2 post hoc comparisons complemented the cluster description (Backhaus et al., 2011; Everitt, 1998). Table 2 gives the results of the cluster analysis, showing the means of the cluster-building factors2. The first cluster (cluster A) is characterized by a comparatively strong interest in agricultural topics and agreement with higher animal welfare standards. Consumers in this cluster are, therefore, called ‘interested animal welfare advocates’. With 118 participants, this cluster is the largest. In comparison to the other clusters, these persons are best informed about agriculture and livestock production in particular and know more about agricultural topics than the members of the other clusters (µ=0.33; SD=0.89). Their perceptions of FAW are diverse (µ=-0.02; SD=0.96), but they are quite sure that welfare standards for farm animals should be enhanced (µ=1.66; SD=0.59). Health, husbandry system and natural innate behaviour are very important to them. They are unsure about the market effects of higher national animal welfare requirements and rank all of the variables belonging to this factor in the range of zero (µ=-0.12; SD=0.79). Compared to the other clusters, they view the social acceptance of meat consumption most critically (µ=-1.07; SD=0.74). Furthermore, they slightly agree that the level of animal welfare is better on small farms (µ=0.07; SD=0.92). 2 To

see the means of the underlying statements of the cluster-building factors, see Supplementary Table S2.

Table 2. Results of the cluster analysis.1 Variables

Cluster A

Cluster B

Cluster C Cluster D Cluster E

Number of cluster members (n) Involvement in agriculture and livestock production2 Perception of animal welfare in livestock production2 Conception of animal welfare and animal welfare standards2 Animal welfare and the market2 Social acceptance of meat consumption2 Animal welfare and farm size2

118 0.33bde

109 -0.06a

115 0.18de

-0.02de

-0.18cde

0.33bd

-0.63abce

0.59abd

0.39ce

0.16abe

0.32ae

-1.80abcd

-0.12bcde -1.07bcde 0.07cd

-1.13acde 0.54ae 0.06cd

0.56ab 0.55ae -0.74abde

0.54ab 0.51ae 0.79abce

0.28ab -0.60abcd -0.08cd

97 -0.22ac

77 -0.42ac

1 Scale from +2 = ‘totally agree’ to -2 = ‘totally disagree’; superscript letters indicate a significant difference with the corresponding

cluster (Tamhane’s T2 post hoc test). 2 Variables are significant at P≤0.001. International Food and Agribusiness Management Review

386

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Heise and Theuvsen

Volume 20, Issue 3, 2017

Cluster B can be described as ‘uninformed animal welfare advocates’ and comprises 109 consumers. They are rather uninterested in agricultural topics and do not think that they know much about these issues (µ=-0.06; SD=0.93). Their perception of FAW is quite diverse (µ=-0.18; SD=0.90), ranking most of the statements around zero. Nevertheless, they are firmly convinced that the animal welfare standards for farm animals should be enhanced (µ=1.73; SD=0.50) as they attribute high importance to health, husbandry system and natural innate behaviour. They view the market for products from animal friendly production systems rather uncritically (µ=-1.13; SD=0.73); furthermore, participants in this cluster believe that meat consumption is socially accepted (µ=0.54; SD=0.66). They are unsure if small farms are more easily able to implement higher animal welfare standards (µ=0.24; SD=1.04), but they think that animals feel more comfortable on small farms (µ=1.03; SD=0.71). A total of 115 consumers belong to cluster C, ‘the market-conscious animal welfare proponents’. These consumers have diverse perceptions of their knowledge about and interest in agriculture and livestock production (µ=0.18; SD=1.02). On average, they have neither a negative nor a positive perception of FAW (µ=0.33; SD=0.94). In comparison to the other clusters, these consumers are most critical towards the market effects of higher animal welfare requirements as they agree that higher national animal welfare standards will cause problems for German farmers and that more animal-friendly products will continue to occupy only niche markets in future (µ=0.56; SD=0.81). They are convinced that meat consumption is socially accepted (µ=0.54; SD=0.62). Furthermore, members of cluster C are rather undecided as to whether farm size actually influences the level of FAW. They rate the influence of farm size on FAW lower than the other clusters (µ=-0.74; SD=0.92). Cluster D consists of ‘critical proponents of animal welfare on small farms’. Participants in this cluster are rather uninterested in agriculture and livestock production and rate their own knowledge comparatively low (µ=-0.22; SD=1.01). Members of cluster D are most critical concerning the current level of FAW (µ=-0.63; SD= 0.92) and consider the importance of good health, appropriate behaviour and the structural-technical equipment of barns as comparatively important for FAW. The enhancement of animal welfare standards is most important for them (µ=1.74; SD=0.50). They are pessimistic about the market for AWPs (µ=0.54; SD=0.80) and tend to see disadvantages for German agribusiness in the implementation of additional national animal welfare requirements. They strongly believe that meat consumption is socially accepted (µ=0.51; SD=0.70). Furthermore, they are convinced that small farms can more easily implement higher animal welfare standards and that animals feel more comfortable on small farms (µ=0.79; SD=0.72). The last cluster (cluster E) is the smallest group, with 77 participants. They are called ‘the uninvolved’ as consumers in this cluster have neither a negative nor a positive attitude towards most of the statements. They are the most uninterested in agricultural topics and perceive their own knowledge as rather low (µ=0.42; SD=1.01). They rate factor two statements mostly in the zero range (µ=0.59; SD=0.85); thus, they are the least critical cluster in their rating regarding the level of animal welfare. They are rather unsure about animal welfare standards (µ=-1.80; SD=0.97) and do not really have an opinion about the market effects of enhanced national animal welfare requirements (µ=0.28; SD=0.70). ‘The uninvolved’ tend to believe that meat consumption is socially accepted and slightly agree that FAW is dependent on farm size (µ=-0.08; SD=0.81). The clusters differ with regard to some interesting sociodemographic and lifestyle characteristics, which are shown in Supplementary Table S3. The ‘interested animal welfare advocates’ are the cluster with the highest proportion of women. Participants in this cluster are often responsible for the household and disproportionately often have a net household income between € 3,000 and € 3,999 per month. Compared to the other clusters, participants in cluster A had the most religious education as a child and are still more active in religion than participants in the other clusters, even though they slightly disagree with these statements. Compared to the rest of the sample, participants in cluster A most often grew up in a small city. Furthermore, this is the cluster with the lowest meat consumption. More than 10% of the cluster members do not eat meat. International Food and Agribusiness Management Review

387

Heise and Theuvsen

Volume 20, Issue 3, 2017

The ‘uninformed animal welfare advocates’ are characterized by the highest educational level, with more than 42% holding a high school certificate. A large proportion are employed and, therefore, not responsible for the household. Furthermore, this is the cluster with the highest percentage of people earning ≥€ 4,000 per month. Members of this cluster disproportionally often live in northern Germany.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Nearly all the sociodemographic and lifestyle characteristics of cluster C (‘the market-conscious animal welfare proponents’) lie between the other clusters, showing no extreme characteristics. The only exception is that the proportion of persons consuming meat is higher than in the other groups. Members of cluster D, ‘the proponents of animal welfare on small farms’, are slightly younger than members of the other clusters, but this difference is not at a significant level. Compared to the other clusters, cluster D is the group with the highest proportion of pupils, apprentices and persons who have not yet completed an apprenticeship. Their education and lifestyle is less religious compared to the other clusters, and they more often live in eastern Germany. In comparison to the other respondents, members of cluster D disproportionally often have no connection to agriculture, with no friends or family working in this business. ‘The uninvolved’ (cluster E) are characterized by the lowest share of women and the smallest proportion of persons holding a high school certificate. They less likely to have grown up in small cities compared to the other clusters; members of cluster E comparatively seldom have a connection to agriculture through friends or family. To analyse whether consumers perceive FAW as having an influence on other quality attributes like healthfulness or taste, Table 3 shows the means for the quality-related statements in each cluster. It is clear that consumers see close connections between the conditions under which farm animals are reared and meat quality since the overall means for all statements are greater than 1, indicating relatively strong agreement. Comparing the five clusters clearly shows that the perceived meat quality of AWPs differs significantly among the clusters. Clusters A, B and D rate all statements quite similarly and show comparatively strong commitment. Even the participants in cluster C positively evaluate the effects of higher animal welfare standards on meat quality, but they are not as convinced as participants in clusters A, B and D. Again, participants in cluster E tend to be undecided when it comes to the effects of FAW on meat quality. Table 3. Animal welfare and meat quality.1 Variables

Cluster A

Cluster B

Cluster C Cluster D Cluster E

Number of cluster members (n) Meat from animals kept under more animalfriendly conditions is of better quality than meat from conventionally kept animals2 Meat from animals kept under more animalfriendly conditions tastes better than meat from conventionally kept animals2 Meat from animals kept under more animalfriendly conditions is healthier than meat from conventionally kept animals2 Farm animal welfare crucially affects meat quality2 A longer fattening period positively affects meat quality2 Stress negatively affects meat quality2

118 1.31e

109 1.32e

115 1.12e

97 1.33e

77 0.40abcd

1.19e

1.18e

0.96e

1.14e

0.36abcd

1.17e

1.15e

0.93e

1.25e

0.52abcd

1.46e 1.34e

1.51e 1.28e

1.26e 1.00e

1.45e 1.46e

0.70abcd 0.73abcd

1.45e

1.59e

1.28e

1.58e

0.78abcd

Scale from +2= ‘totally agree’ to -2= ‘totally disagree’; superscript letters indicate a significant difference with the corresponding cluster (Tamhane’s T2 post hoc test). 2 Variables are significant at P≤0.001. International Food and Agribusiness Management Review

388

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

In order to investigate the possible effects of different attitudes towards FAW and modern agriculture and to derive potential target groups for the purchase of AWPs, aspects of consumers’ shopping behaviour as well as their willingness to pay a price premium for AWPs are analysed (Supplementary Table S4). Overall, 42% of the consumers in our study state that they buy AWPs on a regular basis (µ=0.38; SD=1.04). Nevertheless, many consumers feel that finding more animal-friendly products in retail outlets is not easy (µ=-0.42; SD=0.989). For 44% of the respondents, FAW is the basis of their decision-making when buying products of animal origin (µ=0.39; SD=1.021) and 38.5% always think about how the animals were kept when buying products of animal origin (µ=0.25; SD=1.101). About 29% pay attention to the animal-friendly labelling of the products they buy (µ=0.03; SD=1.024). For nearly 30%, the price is always the basis of their shopping decisions (µ=-0.03; SD=1.089), and 48.4% agree that they would love to buy products from AWPs more often but find these products too expensive (µ=0.46; SD=1.061). Only 15.1% feel that the labelling provided informs them sufficiently about the production process, and nearly 38% are undecided about this statement (µ=-0.42; SD=0.989). Of the respondents, 37.8% agree or somewhat agree that animal welfare labels try to cheat the consumer (µ=0.31; SD=0.884), and only 22.9% trust animal welfare labels (µ=-0.13; SD=0.970). Overall, 76.6% are willing to pay a price premium for AWPs. On average, 23.1% of the consumers prefer shopping at the discounter, 51.2% mostly purchase food in the supermarket. Nearly 16% buy their meat from a butcher, 2.9% buy directly from a farmer and nearly 5% prefer buying food at the weekly farmers’ market. However, the high standard deviations clearly show that respondents’ shopping behaviour is diverse and differs significantly among the five clusters. Clusters A and B exhibit quite similar shopping behaviour, as members of both clusters regularly purchase AWPs and the level of FAW is the basis for their shopping decision. Members of cluster D also buy organic products on a regular basis. Furthermore, participants in cluster A more often think about how the animals were kept and, compared to the other clusters, animal welfare labels are most important to them. This cluster has the highest willingness to pay more for AWPs: 90% are willing to pay a price premium. Additionally, persons in cluster A buy their food at the supermarket disproportionally seldom but comparatively often buy meat from a butcher. Even cluster B has a high willingness to pay more for AWPs (82.6%). However, they do not consider it easy to find more animal-friendly products. Furthermore, they feel the least well-informed by labels about the way the animals were kept. Compared to the other clusters, price is least important for these consumers, and they most often buy their meat from a butcher. Clusters C, D and E all have diverse attitudes towards the purchase of AWPs. However, clusters C and D would love to buy more AWPs, but consider these products too expensive. Members of cluster D regularly buy products from more broadly defined quality programmes. Compared to the other clusters, cluster D respondents are particularly critical of labels and do not really trust them. Cluster E especially shows a comparatively low willingness to pay more for animal friendly products. Compared to the other clusters, members of this group, place the greatest importance on price.

5. Discussion A variety of factors affects consumers’ attitudes towards FAW and their actual shopping behaviour. This statement reflects the results of several prior studies that investigated various aspects of consumers’ attitudes towards FAW or their purchase habits and willingness to pay a price premium for particularly animal friendly products (cf. Schulze et al., 2008; Vanhonacker et al., 2007). The present paper compiles the issues already taken into account in earlier studies and analyses them in a more comprehensive large-scale empirical study. The results confirm previous studies but also reveal new aspects of this highly relevant and frequently debated topic. International Food and Agribusiness Management Review

389

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Heise and Theuvsen

Volume 20, Issue 3, 2017

Overall, our study shows that consumers rate their own knowledge about agriculture rather low and are poorly informed about agricultural topics. The results, therefore, closely match the findings of other studies, which indicated overall low consumer involvement in agricultural topics (Busch et al., 2013; Frewer et al., 2005; Kanis et al., 2003; Te Velde et al., 2002; Vanhonacker et al., 2010). However, nearly half the consumers we surveyed are generally interested in agricultural topics, confirming that large parts of society care about these issues (De Jonge and van Trijp, 2012; Harper and Henson, 2001; Vanhonacker and Verbeke, 2014). More than 36% regularly access information in the mass media about agriculture. Thus, our results in this regard are also in line with previous studies, suggesting that mass media constitutes the most important platform for consumers’ information (TNS Emnid, 2012). Large parts of consumers perceive current livestock production systems very critically. Our results indicate that there is even a negative trend for this issue. A European study from 2005 found that 77% of European consumers think that animal welfare standards should be enhanced (European Commission, 2005). In our sample, more than 88% of the respondents believe there is a need for higher animal welfare standards. Thus, within ten years, there has been an increase of more than 10 percentage points, putting the responsible stakeholders along the food supply chain under huge pressure to respond to these ever-growing consumer demands for improved husbandry conditions in livestock farming. Consumers’ criticism particularly focusses on large-scale farming, where a high number of animals is kept (Busch et al., 2013; Heyder and Theuvsen, 2009; Kanis et al., 2003; Kayser et al., 2012a; Vanhonacker et al., 2010). Our results confirm these earlier findings but also show that only 46.6% of consumers agree or somewhat agree that higher animal welfare standards can more easily be implemented on small farms. Many consumers are unsure about this statement. However, most of the respondents believe that animals feel more comfortable on small farms. The concept of small-scale farming with a modest number of animals kept per farm still seems to comply with consumers’ idea of good animal welfare (Kayser et al., 2012a). It can be assumed that this notion very much reflects the current public debate or idealized pictures provided by marketing campaigns, media coverage or other sources. However, contrary to the results of Busch et al. (2013), a clear relationship between involvement and criticism of large-scale farming was not observed in our sample. What could turn out to become a problem in the future is that investments in higher animal welfare standards are subject to considerable economies of scale, which clearly favours larger farms, whereas consumers prefer small-scale farming. Furthermore, our results corroborate those of Busch et al. (2015), who showed that major parts of society do not trust farmers with regard to FAW. In this respect, too, consumers’ evaluation has become increasingly more negative in the recent past. While Busch et al. (2015) found that 38% of German consumers think that farmers take good care of their animals, only 19% of the respondents in our sample agree or somewhat agree with this statement. This steep decline in the level of trust in farmers increasingly threatens producers’ license to operate, that is, the social acceptance of their production processes and standards (Hiss, 2006). Our findings also confirm those of Schröder and McEachern (2004), who showed that consumers often feel misled by labels. Our results suggest several reservations concerning labels and the information provided by them. Firstly, many people think that animal welfare labels try to cheat them, and consumers’ trust in these labels tends to be low. Secondly, many people do not feel that the labels provide useful information. This loss of trust and the perceived information asymmetry have led to considerable purchase barriers for AWPs and prevent consumers from translating their concerns regarding FAW into appropriate shopping behaviour. In this way, the so-called citizen-consumer gap, which is often deplored by industry representatives from the agrifood sector, is further strengthened. Moreover, this severe loss of consumer trust increasingly restricts the options of the actors along the food supply chain when responding to consumer preferences, as efforts to improve animal welfare are not recognized as such by consumers. Currently, only 28% of consumers believe that the agricultural sector is honestly interested in the improvement of animal welfare standards.

International Food and Agribusiness Management Review

390

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

The growing concerns and mistrust of consumers is also reflected in a relatively low social acceptance of meat consumption – an issue which has yet not been directly investigated in empirical studies. However, our results confirm those of other studies that showed many consumers question the legitimacy of conventional livestock production systems (Kauppinen et al., 2010). Loss of trust and low social acceptance can both become starting points for reduced meat consumption, which has been observed in recent consumer studies (e.g. BMEL, 2016), and go hand-in-hand with a decrease in potential customers for AWPs (Spiller and Schulze, 2008). This paper also evaluates consumers’ perspective on the potential market effects of higher national animal welfare standards. Up to now, this aspect has not been investigated in other consumer studies. Our results show that considerable consumer segments see competitive disadvantages for farmers through enhanced national animal welfare standards. Thus, a certain understanding for farmers’ economic constraints can be identified. However, the results of Busch et al. (2013) showed that this understanding does not always lead to changes regarding consumers’ purchase habits. Nearly 30% of the respondents in our sample believe that products from more animal-friendly production systems will still remain a niche market in future. The current study confirms previous results, which showed that consumers cannot be understood as one homogeneous group (e.g. Schulze et al., 2008; Vanhonacker and Verbeke, 2014). Instead, it is possible to differentiate between five groups of consumers, which differ significantly according to their involvement, their attitudes towards FAW, the perceived social acceptance of meat consumption and the perceived market effects of higher animal welfare standards. These clusters also differ in the perceived relevance of higher animal welfare standards for other quality attributes and in their sociodemographics, lifestyle characteristics and shopping behaviour. This study supports Vanhonacker and Verbeke (2009), who found that interested persons who attach high importance to ethical issues when buying animal-based food show a higher willingness to pay a price premium for AWPs. However, in our study, this relationship is not as clear as has sometimes been mentioned in previous studies (Schulze et al., 2008; Vanhonacker et al., 2007). While cluster B is only ordinarily interested in agricultural topics, members of this cluster have an above-average willingness to pay a price premium for AWPs and somewhat agree that they buy these products on a regular basis. Furthermore, these participants most often purchase organic products, which tend to be associated with higher animal welfare standards and characterized by high price premiums (Harper and Makatouni, 2002; Makatouni, 2002; Oekobarometer, 2010). The findings by Kayser et al. (2012b) that persons with good knowledge are most critical of agricultural topics are not completely confirmed by our results. Clusters A and C are comparatively well-informed about agricultural topics. However, members of cluster C are somewhat unconcerned about FAW. The members of cluster D perceive their own knowledge about agriculture and animal production as comparatively low. Nevertheless, these persons are the most critical about the current living conditions in livestock production systems. This could indicate a trend that it is no longer only the well-informed who are critical about intensive livestock production but that this has become a social mega-trend in which many people participate regardless of their personal knowledge or educational status. Furthermore, the connection between critical perception of conditions in current livestock production systems and consumers’ actual willingness to pay a price premium for AWPs is not as clear-cut as mentioned in other studies (Schulze et al., 2008; Vanhonacker et al., 2007). Members of cluster D are most critical of the level of animal welfare in current livestock production systems; furthermore, members of this cluster evidence the strongest support for higher animal welfare standards. Nevertheless, cluster D shows only an ordinary willingness to pay more for particularly animal friendly products and does not clearly confirm buying these products regularly. Thus, these people may be the consumer group showing the greatest discrepancy between attitude and actual behaviour. This could be explained by the fact that these people attach comparatively high importance to product prices, maybe due to budget constraints, or experience other obstacles, such as lack of trust in labels, which prevents them from changing their buying behaviour. International Food and Agribusiness Management Review

391

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Our study bears out the results of Schulze et al. (2008), who showed that persons who attach high importance to farm animals’ opportunity to engage in natural innate behaviour are also willing to pay high prices for more animal-friendly products. Beyond this, our results suggest that these people also rate the health and the structural-technical equipment of barns as particularly important for animal welfare. The perceived importance of a good human-animal relationship did not affect the willingness to pay more in our study. All the consumers we surveyed associate higher animal welfare standards with other quality attributes. These results strengthen previous studies showing that consumers often associate higher animal welfare standards with higher product quality (Meuwissen and van der Lans, 2004; Phan-Huy and Fawaz, 2003). Similar results were observed with regard to other quality attributes, such as organic (Wier and Calverley, 2002). However, the clusters differ significantly with regard to the strength they attribute to the relationship between animal welfare and other quality attributes. Our results confirm those of other studies, which found that consumers’ attitudes towards FAW are also influenced by sociodemographic and other personal characteristics. Our findings suggest that women are more attentive to animal welfare issues and are more willing to pay higher prices for AWPs. Thus, our results closely match the findings of earlier studies (Kehlbacher et al., 2012; Makdisi and Marggraf, 2011; Nocella et al., 2010; Schulze et al., 2008; Vanhonacker and Verbeke, 2009). However, the relationship between other sociodemographic characteristics, including age, education, income or place of residence, and attitudes towards animal welfare or shopping habits is not as close as stated in earlier studies (Bennet and Blaney, 2002; Kehlbacher et al., 2012; Makdisi and Marggraf, 2011; Nocella et al., 2010; Schröder and Mc Eachern, 2004; Schulze et al., 2008; Vanhonacker et al., 2010; Vanhonacker and Verbeke, 2009). Even the literature indicates that the relationship between sociodemographic characteristics, on the one hand, and consumers’ attitudes and animal welfare-orientated shopping behaviour, on the other, considerably differs among studies. Several researchers have already expressed doubts concerning the usefulness of these variables to explain consumers’ shopping behaviour (Dagevos, 2005; Diamantopoulos et al., 2003, Pouta et al., 2010). Furthermore, Vanhonacker et al. (2009) and Toma et al. (2012) found that attitudes towards animal welfare are much more strongly influenced by individual experiences, involvement in agricultural practices and other lifestyle characteristics. Our results indicate that these findings also apply for our sample. Like most non-experimental studies, ours has some limitations that need to be taken into account when interpreting the results. Firstly, the aided nature of our variables could affect the response behaviour of the survey participants. Heise and Theuvsen (2015) and Heise et al. (2015) clearly showed that different methodological approaches (qualitative vs quantitative) lead to considerable differences in farmers’ and veterinarians’ definition of farm animal welfare. A similar pattern might also exist among consumers. Due to effects of social desirability, participants’ answers might not always honestly reflect their personal opinions but also include social expectations from an ethical and moral perspective. This might lead to biased answers. Secondly, from previous studies we already know that attitude-behaviour discrepancies are existent among consumers (Coff et al., 2008; Harvey and Hubbard, 2013; Vanhonacker et al., 2010). These discrepancies may result in an overestimation of the market potential for more animal-friendly products as positive attitudes and the indicated willingness to buy these products do not always lead to a corresponding behaviour when buying products of animal origin.

6. Marketing implications The different attitudes towards agriculture and animal welfare as well as the diverse sociodemographic characteristics and shopping habits indicate that consumers have distinctive profiles. According to Verbeke (2009), appropriate market segmentation could help to address this heterogeneity in consumer demand and to transfer consumers’ concerns and their expectations concerning FAW into corresponding shopping behaviour. Programmes that are differentiated in terms of animal welfare level and the price premiums it entails could, therefore, lead to a broader segment for AWPs and enable consumers to make product choices according to their individual preferences (De Jonge and van Trijp, 2012). International Food and Agribusiness Management Review

392

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Heise and Theuvsen

Volume 20, Issue 3, 2017

Four of the clusters identified (the ‘interested animal welfare friends’, the ‘uninformed animal welfare friends’, the ‘market-conscious animal welfare proponents’ and the ‘critical proponents of animal welfare on small farms’) generally favour the enhancement of FAW standards. Clusters A and B already regularly buy products from more animal-friendly production systems and, thus, currently constitute the most important target groups for AWPs. Clusters C and D would prefer to buy AWPs more frequently, but high prices currently prevent them from purchasing these products. Nevertheless, these groups constitute potential initial target groups for the purchase of more animal friendly products. At the same time, they represent a need for innovations that will make it possible to provide higher animal welfare standards at lower costs. Even cluster E, the ‘uninvolved’, could potentially become a target group, as they do not oppose the improving the animal welfare level. However, preferences and attitudes are said to be relatively stable and durable cognitive orientations (Weber et al., 2005). Therefore, it is very likely that there are mobility barriers between the individual clusters and that consumers will remain in the same group for the long term. For that reason, specific marketing implications should be derived for each cluster. Members of cluster A most often buy more animal-friendly products and show the highest willingness to pay for these products. Cluster A is the group that is most involved in agricultural topics. Retailers should use the interest of these persons to advertise precisely defined product segments to these consumers. Mass media should be included for advertising campaigns, as this is consumers’ main source of information. Furthermore, improvements concerning animal health, the structural-technical equipment of barns and animals’ opportunity to engage in natural innate behaviour should be particularly highlighted, as these aspects are quite important to cluster A. An above average number of the individuals in cluster A are women with a comparatively low frequency of meat consumption who are responsible for the household. Furthermore, the overall social acceptance of meat consumption is strongly questioned by this sub-group. To avoid even more members of this cluster further reducing their meat consumption, AWPs with standards significantly above legal requirements should be implemented and placed on the market. As these persons comparatively often buy their meat from a butcher, butcher shops and fresh meat counters in the supermarkets should provide products tailored to the needs of this sub-segment. Members of cluster B also buy AWPs on a regular basis and show an above average willingness to pay for these products. They also place the highest trust in labels. This is the only other cluster that buys organic animal-based products on a regular basis. Furthermore, the price is least important to cluster B members since they report the highest net income per month. Members of cluster B also have the highest education level, but are generally not responsible for the household. Thus, employed high earners characterize this cluster. As this cluster indicates a high social acceptance of meat consumption and an above-average share of members who eat meat, it constitutes a very attractive target group for high-priced AWPs. However, working people often have only limited time for grocery shopping, and this group of people most often agreed that animal-friendly products are not easy to find at the supermarket. Lack of time to shop and the perceived low availability of specific AWPs could lead to the purchase of organic products, which are much easier to find in retail and are often associated with higher animal welfare standards (Harper and Makatouni, 2002; Makatouni, 2002; Oekobarometer, 2010). To preserve this consumer group specifically for the market segment of AWPs, prime importance should be given to the practical applicability. Retail must ensure that labelling is clear and easy to recognize and that AWPs are easily available in the markets. As these individuals most often buy their meat from a butcher, the market segment for high-priced AWPs should be extended to this shopping location. Cluster C is characterized by average involvement in agricultural topics and moderate attitudes towards FAW and animal welfare standards. Members of cluster C do not associate animal welfare with small farms. Furthermore, members of cluster C have the highest social acceptance of meat consumption and most often eat meat. For this reason, they constitute an attractive target group for the meat market. Currently, these individuals seldom purchase AWPs but would like to buy them more often. So far, the high price constitutes a shopping barrier for this cluster. For this reason, this cluster constitutes a target group for a middle segment in the market for more animal-friendly products in which products are priced slightly to moderately above International Food and Agribusiness Management Review

393

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

standard goods. As mentioned before, this may require process innovations to make it possible to provide AWPs at moderate prices. Mass media could be used to communicate the improvements concerning animal welfare to this cluster, as these persons most often use the media to inform themselves about agricultural topics. Cluster D reports comparatively little involvement in agricultural topics but rates the current conditions of FAW most critically. Furthermore, members of cluster D most strongly agree that animal welfare standards should be enhanced. They currently do not purchase AWPs on a regular basis. Nevertheless, members of cluster D report that they would like to buy more animal-friendly products but find these products too expensive. Furthermore, cluster D sees comparatively strong relationships between animal welfare standards and other quality attributes like taste or healthfulness. Moreover, these persons report buying products from quality assurance programmes on a regular basis. Thus, cluster D is an appropriate target group for products from more broadly defined quality assurance programmes that include higher animal welfare standards as an additional benefit in addition to other quality attributes, such as better taste. Even a relatively lowpriced market segment of specific AWPs could provide a suitable range of products for these consumers. Furthermore, cluster D prefers animal production on small farms. Thus, programmes supporting small-scale agriculture could also be attractive for these people. As members of cluster D most strongly mistrust labels, transparent communication strategies should be implemented to avoid information asymmetries and gain the trust of this consumer group. As many young people who are still at school or in apprenticeship are in cluster D, this group of persons prospectively constitutes a very important future target group for the retail sector. Cluster E, the ‘uninvolved’, are the only sub-group that does not clearly favour higher animal welfare standards. Their attitudes towards FAW and meat consumption and towards the relationship between animal welfare and other quality attributes are diverse and often undecided, with most statements in the zero range. Thus, these people do not have a pronounced opinion about FAW. Moreover, this sub-group reports below average willingness to pay for AWPs. Furthermore, they currently do not regularly buy AWPs or show an interest in buying them more often in future. Thus, members of cluster E currently do not constitute an interesting target group for AWPs. As this cluster is the smallest, with only 77 participants, specifying a market segment or targeting marketing campaigns for this group does not seem useful at the moment. Despite the target group-specific implications, responsible persons from retail should in general try to reduce consumers’ purchase barriers for AWPs in order to decrease the discrepancy between consumers’ attitudes and their actual shopping behaviour. Animal welfare is a credence attribute to consumers. Thus, they are crucially dependent on appropriate information about the conditions under which animals were kept (Lagerkvist and Hess, 2010). Beyond this, consumers’ lack of trust could be decreased through the reduction of information asymmetries. Transparency and trustworthiness during the entire production process should be ensured in order to gain the trust of the wider public. Furthermore, retailers should guarantee easy availability of AWPs. If consumers cannot find the products they prefer, they are easily frustrated and feel powerless to influence the level of FAW (Schröder and McEachern, 2004). The overall alienation of consumers from agricultural production leads to low consumer involvement in agricultural topics. Targeted advertising in the media could counteract the mostly negative headlines about agriculture and food production in general in the recent past (Kayser, 2012). Furthermore, transparency campaigns from the agricultural sector, like guided visits to farms and livestock barns could help consumers acquire a more accurate impression of modern agriculture (Windhorst, 2016). High additional costs constitute massive hurdles for large numbers of consumers, especially if additional benefits from buying these products remain unclear (Enneking, 2004). In Germany, an industry solution has been established whereby retailers temporarily bear the additional costs of higher animal welfare standards. In this way, consumers do not bear the initial brunt of higher prices, and a broad market segment for meat with higher animal welfare standards is created. The long-term aim is to shift the additional costs little by little to consumers. Kahneman and Tversky (1979) showed that it is more important to individuals to avoid the ‘bad’ than to gain the ‘good’. Therefore, improvements in FAW have to be clearly communicated to International Food and Agribusiness Management Review

394

Heise and Theuvsen

Volume 20, Issue 3, 2017

consumers, allowing them to become used to the higher level of animal welfare and, after a while, to perceive the higher level as new standard. Thus, even consumers with low animal welfare preferences could develop a certain willingness to pay a price premium rather than fall back to the lower standards, which no longer meet their adjusted expectations.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

7. Concluding remarks Consumersâ&#x20AC;&#x2122; attitudes towards agriculture and FAW and the effects on their shopping behaviour have been widely studied in recent decades (e.g. Schulze et al., 2008; Vanhonacker and Verbeke, 2014). We used this information as a starting point for a comprehensive and representative empirical study and added new aspects that had not been investigated before. It was the aim of the present study to find different consumer groups that differ in their attitudes and their shopping behaviour and to derive target groups for AWPs. This goal was achieved by grouping the consumers into five clusters according to their involvement in agricultural topics, their attitudes towards FAW and their social acceptance of meat consumption. These clusters also differ with regard to their sociodemographics and other lifestyle characteristics, their perception of the relationship between FAW and other quality attributes and their shopping behaviour. Based on our results, we derived specific marketing implications for each cluster. These implications can help develop a more differentiated market segment for AWPs in terms of animal welfare level and required price premium, enabling consumers to make product choices according to their preferences (De Jonge and van Trijp, 2012). In this way, a broader market segment for AWPs can be established, the overall level of animal welfare in livestock production can be enhanced and the demands of large segments of society can be met. This can help to counteract the overall low reputation of the agricultural sector and regain consumer trust in agricultural production and the information provided by such means as labels. However, the long-term success of animal welfare concepts is not determined only by consumersâ&#x20AC;&#x2122; attitudes and behaviour but also by the acceptance of other stakeholders along the supply chain and their willingness to participate (Deimel et al., 2010; Franz et al., 2010; Golan et al., 2000; Gulbrandsen, 2006). Future studies should consider this aspect and investigate in greater depth the attitudes of various stakeholders along the meat supply chain towards FAW and specific animal welfare criteria. As an initial step, these studies could analyse the importance of different animal welfare criteria for different stakeholders. Next, the practical applicability of these criteria could also be analysed from different points of view. These results could help develop animal welfare concepts that are accepted by all stakeholders and effectively enhance the level of animal welfare.

Supplementary material Supplementary material can be found online at https://doi.org/10.22434/IFAMR2016.0115. Table S1. Sample composition compared to basic population in Germany. Table S2. Results of the cluster analysis. Table S3. Personal characteristics and lifestyle habits of the respondents. Table S4. Purchase behaviour of the consumers.

International Food and Agribusiness Management Review

395

Heise and Theuvsen

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

References Bacher, J., A. Pöge and K. Wenzig. 2010. Clusteranalyse: anwendungsorientierte Einführung in Klassifikationsverfahren. 3rd Edition. Oldenbourg Wissenschaftsverlag, Munich, Germany. Backhaus, K., E. Erichson, W. Plinke and R. Weiber. 2011. Multivariate analysemethoden: eine anwenderorientierte Einführung. Springer, Berlin, Germany. Bennett, R.M. and R.J.P. Blaney. 2002. Social consensus, moral intensity and willingness to pay to address a farm animal welfare issue. Journal of Economic Psychology 23: 501-520. Blandford, D. and L. Fulponi. 1999. Emerging public concerns in agriculture: domestic policies and international trade commitments. European Review of Agricultural Economics 26: 409-424. Boehm, J., M. Kayser and A. Spiller. 2010. Two sides of the same coin? Analysis of the web-based social media with regard to the image of the agri-food sector in Germany. International Journal of Food System Dynamics 1: 264-278. Brosius, F. 2011. SPSS 19. Mitp, Heidelberg, Germany. Bühl, A. 2010. SPSS 18: Einführung in die moderne Datenanalyse. Pearson, Munich, Germany. Buller, H. and C. Cesar. 2007. Eating well, eating fare: farm animal welfare in France. International Journal of Sociology of Agriculture and Food 15: 45-58. Bundesministerium für Ernährung und Landwirtschaft (BMEL). 2016. Deutschland, wie es isst: Der BMELErnährungsreport 2016. Berlin, Germany. Available at: http://tinyurl.com/joeu79x. Burda Community Network. 2009. Typologie der Wünsche 2009: Menschen-Medien-Märkte. Burda Community Network, Offenburg, Germany. Busch, G., M- Kaiser and A. Spiller. 2013. Factory farming from a consumer’s perspective: associations and attitudes. Jahrbuch der Österreichischen Gesellschaft für Agrarökonomie 22: 61-70. Busch, G., C. Schwetje and A. Spiller. 2015. Bewertung der Tiergerechtheit in der intensiven Hähnchenmast durch Bürger anhand von Bildern: Ein Survey-Experiment. German Journal of Agricultural Economics 64: 131-146. Coff, C., M. Korthals and D. Barling. 2008. Ethical traceability and informed food choice. ethical traceability and communicating food. In: The international library of environmental, agricultural and food ethics 15, edited by C. Coff, D. Barling, M. Korthals and N. Thorkild. Springer, Dordrecht, the Netherlands, pp. 1-18. Dagevos, H. 2005. Consumers as four-faced creatures: looking at food consumption from the perspective of contemporary consumers. Appetite 45: 32-39. De Jonge, J. and H.C.M. van Trijp. 2013. Meeting heterogeneity in consumer demand for animal welfare: a reflection of existing knowledge and implication for the meat sector. Journal of Agricultural and Environmental Ethics 26: 629-661. Deimel, I., A. Franz and A. Spiller. 2011. Das ‘Animal Welfare’-Verständnis Deutscher Schweinemäster: Eine Analyse landwirtschaftlicher Frames. Jahrbuch der Österreichischen Gesellschaft für Agrarökonomie 20: 191-200. Deimel, I., A. Franz, M. Frentrup, M. Von Meyer, A. Spiller and L. Theuvsen. 2010. Perspektiven für ein Europäisches Tierschutzlabel. Available at: http://download.ble.de/08HS010.pdf. Diamantopoulos, A., B.B. Schlegelmilch, R.R. Sinkovices and G.M. Bohlem. 2003. Can sociodemographics still play a role in profiling green consumers? A review of the evidence and an empirical investigation. Journal of Business Research 56: 465-480. Enneking, U. 2004. Willingness-to-pay for safety improvements in the german meat sector: the case of the Q&S label. European Review of Agricultural Economics 31: 205-223. European Commission. 2005. Attitudes of consumers towards the welfare of farmed animals. special eurobarometer 229, wave 63.2. TNS opinion and social. Available at: http://tinyurl.com/jdt58dg. European Commission. 2007a. Attitudes of EU Citizens towards animal welfare (special eurobarometer 270). Available at: http://tinyurl.com/ogkgtwd. European Commission. 2007b. Attitudes of consumers towards the welfare of farmed animals. special eurobarometer, Vol. 229 No. 2. Available at: http://tinyurl.com/hlhp64d. Everitt, B.S. 1998. The Cambridge dictionary of statistics. Cambridge University Press, Cambridge, UK. International Food and Agribusiness Management Review

396

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Heise and Theuvsen

Volume 20, Issue 3, 2017

Franz, A., I. Deimel and A. Spiller. 2012. Concerns about animal welfare: a cluster analysis of german pig farmers. British Food Journal 114: 1445-1462. Franz, A., M. Meyer and A. Spiller. 2010. Einführung eines Animal Welfare Labels in Deutschland: Ergebnisse einer Stakeholder Befragung. Jahrbuch der Österreichischen Gesellschaft für Agrarökonomie 19: 41-50. Frewer, L.J., A. Kole, S.M.A. Van De Kroon and C. De Lauwere. 2005. Consumer attitudes towards the development of animal-friendly husbandry systems. Journal of Agricultural and Environmental Ethics 18: 345-367. Golan, E., E. Kuchler and L. Mitchell. 2000. Agricultural economic report, No. 793: economics of food labeling. Economic Research Service, USDA. Available at: http://tinyurl.com/gopn744. Grunert, K.G. 2005. Food quality and safety: consumer perception and demand. European Review of Agricultural Economics 32: 369-391. Gulbrandsen, L.H. 2006. Creating markets for eco-labeling: are consumers insignificant? International Journal of Consumer Studies 30: 477-489. Harper, G. and S. Henson. 2001. Consumer concerns about animal welfare and the impact on food choice. EU-Project EU-FAIR-CT-98-3678. Final Report. Centre for Food Economics Research, Department of Agriculture and Food Economics. The University of Reading, Reading, UK. Harper, G. and A. Makatouni. 2002. Consumer perception of organic food production and farm animal welfare. British Food Journal 104: 287-299. Harvey, D. and C. Hubbard. 2013. Reconsidering the economy of FAW: an anatomy of market failure. Food Policy 38: 105-114. Heise, H., N. Kemper and L. Theuvsen. 2015. Was verstehen Tierärzte unter Tierwohl? Ergebnisse einer empirischen Erhebung. Tierärztliche Umschau 70: 299-304. Heise, H. and L. Theuvsen. 2015. Biological functioning, natural living oder Welfare Quality: Untersuchungen zum Tierwohlverständnis Deutscher Landwirte. Berichte über Landwirtschaft, 93: 1-19. Heyder, M. and L. Theuvsen. 2009. Corporate social responsibility in agribusiness. In: Die Ernährungswirtschaft im Scheinwerferlicht der Öffentlichkeit, edited by J. Böhm, F. Albersmeier and A. Spiller. EuL Publishing, Lohmar and Cologne, Germany, pp. 47-73. Hiss, S. 2006. Warum übernehmen Unternehmen gesellschaftliche Verantwortung: Ein soziologischer Erklärungsversuch. Campus, Frankfurt am Main, Germany. Janssen, J. and W. Laatz. 2007. Statistische Datenanalyse mit SPSS für Windows. Springer, Heidelberg, Germany. Kahneman, D. and A. Tversky. 1979. Prospect theory: an analysis of decision under risk. Econometrica 47: 263-291. Kanis, E., A.B.F. Groen and K.H. De Greef. 2003. Societal concerns about pork and pork production and their relationships to the production system. Journal of Agricultural and Environmental Ethics 16: 137-162. Kauppinen, T., A. Vainio, A. Valros, H. Rita and K.M. Vesala. 2010. Improving animal welfare: qualitative and quantitative methodology in the study of farmers’ attitudes. Animal Welfare 19: 523-536. Kayser, M. 2012. Die Agrar- und Ernährungswirtschaft in der Öffentlichkeit: Herausforderungen und Chancen für die Marketing-Kommunikation. Cuvillier Verlag, Goettingen, Germany. Kayser, M., J. Boehm and A. Spiller. 2012b. Zwischen Markt und Moral: Wie wird die deutsche Landwirtschaft in der Gesellschaft wahrgenommen? Schriften der Gesellschaft für Wirtschafts- und Sozialwissenschaften des Landbaues e.V. 47: 329-341. Kayser, M., K. Schlieker and A. Spiller. 2012a. Die Wahrnehmung des Begriffs ‘Massentierhaltung’ aus Sicht der Gesellschaft. Berichte über Landwirtschaft 90: 417-428. Kehlbacher, A., R. Bennett and K. Balcome. 2012. Measuring the consumer benefits of improving farm animal welfare to inform welfare labelling. Food Policy 37: 627-633. Lagerkvist, C.J. and S. Hess. 2011. A meta-analysis on consumer willingness to pay for farm animal welfare. European Review of Agricultural Economics 38: 55-78. Lusk, J.L. and F.B. Norwood. 2012. Speciesism, altruism and the economics of animal welfare. European Review of Agricultural Economics 39: 189-212. International Food and Agribusiness Management Review

397

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Heise and Theuvsen

Volume 20, Issue 3, 2017

Makatouni, A. 2002. What motivates consumers to buy organic food in the UK? British Food Journal 104: 345-352. Makdisi, F. and R. Marggraf. 2011. Consumer willingness-to-pay for farm animal welfare in Germany: the case of broiler. Available at: http://tinyurl.com/zbcmu28. Mäkiniemi, J.P., A.M. Pirttilä-Backman and M. Pieri. 2011. Ethical and unethical food: social representations among Finnish, Danish and Italian students. Appetite 56: 495-502. Mayfield, L.E., R.M. Bennett, R.B. Tranter and M.J. Wooldridge. 2007. Consumption of welfare-friendly food products in Great Britain, Italy and Sweden, and how it may be influenced by consumer attitudes to, and behavior towards, animal welfare attributes. International Journal of Sociology of Food and Agriculture 15: 59-73. Meuwissen, M.P.M. and L.A. van der Lans. 2004. Trade-offs between consumer concerns: an application for pork production. Proceedings of 84. EAAE-Seminar ‘Food Systems in a Dynamic World’. Zeist, the Netherlands. Available at: http://tinyurl.com/z78j9ph. Nocella, G., L. Hubbard and R. Scaroa. 2010. Farm animal welfare, consumer willingness to pay, and trust: results of a cross-national survey. Applied Economic Perspective and Policy 32: 275-297. Oekobarometer. 2010. Repräsentative Bevölkerungsbefragung im Auftrag des Bundesministeriums für Ernährung, Landwirtschaft und Verbraucherschutz. Available at: http://tinyurl.com/jhokfhn. Phan-Huy, S.A. and R.B. Fawaz. 2003. Swiss market for meat from animal friendly production: responses of public and private actors in Switzerland. Journal of Agricultural and Environmental Ethics 16: 119-136. Pouta, E., J. Heikkila, S. Forsman-Hugg, M. Isoniemi and J. Makela. 2010. Consumer choice of broiler meat: the effects of country of origin and production methods. Food Quality and Preference 21: 539-546. Povey, R., B. Wellens and M. Conner. 2001. Attitudes towards following meat, vegetarian and vegan diets: an examination of the role of ambivalence. Appetite 37: 15-26. Rozin, P., M. Markwith and C. Stoess. 1997. Moralization and becoming a vegetarian: the transformation of preferences into values and recruitment of disgust. Psychological Science 8: 67-73. Schröder, M.J.A. and M.G. McEachern. 2004. Consumer value conflicts surrounding ethical food purchase decisions: a focus on animal welfare. International Journal of Consumer Studies 28: 168-177. Schulze, B., D. Lemke and A. Spiller. 2008. Glücksschwein oder arme Sau? Die Einstellung der Verbraucher zur modernen Nutztierhaltung. In: Zukunftsperspektiven der Fleischwirtschaft: Verbraucher, Märkte, Geschäftsbeziehungen, edited by A. Spiller and B. Schulze. University Publishing, Goettingen, Germany, pp. 465-488. Spiller, A. and B. Schulze. 2008. Trends im Verbraucherverhalten: ein Forschungsüberblick zum Fleischkonsum. In: Zukunftsperspektiven der Fleischwirtschaft: Verbraucher, Märkte, Geschäftsbeziehungen, edited by A. Spiller and B. Schulze. University Publishing, Goettingen, Germany, pp. 230-271. Te Velde, H.T., N. Aarts and C. Van Woerkum. 2002. Dealing with ambivalence: farmers’ and consumers’ perception of animal welfare in livestock breeding. Journal of Agricultural and Environmental Ethics 15: 203-219. TNS Emnid. 2012. Das Image der deutschen Landwirtschaft. Available at: http://tinyurl.com/hv57che. Toma, L., A.W. Stott, C. Revoreda-Giha and B. Kupiec-Teahan. 2012. Consumers and animal welfare: a comparison between European countries. Appetite 58: 597-607. Vanhonacker, F., E. Van Poucke, F. Tuyttens and W. Verbeke. 2010. Citizens’ views on farm animal welfare and related information provision: exploratory insights from Flanders, Belgium. Journal of Agricultural and Environmental Ethics 23: 551-569. Vanhonacker, F. and W. Verbeke. 2009. Buying higher welfare poultry products? Profiling flemish consumers who do and do not. Poultry Science 88: 2702-2711. Vanhonacker, F. and W. Verbeke. 2014. Public and consumer policies for higher welfare food products: challenges and opportunities. Journal of Agricultural and Environmental Ethics 27: 253-171. Vanhonacker, F., W. Verbeke, E. Van Poucke and F.A.M. Tuyttens. 2007. Segmentation based on consumers’ perceived importance and attitude towards farm animal welfare. International Journal of Sociology of Agriculture and Food 15: 91-107.

International Food and Agribusiness Management Review

398

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

Heise and Theuvsen

Volume 20, Issue 3, 2017

Verbeke, W. 2009. Stakeholder, citizen and consumer interests in farm animal welfare. Animal Welfare 18: 325-333. Verbeke, W. and I. Vackier. 2004. Profile and effects of consumer involvement in fresh meat. Meat Science 67: 159-168. Weber, H., T. Rammsayer and J. Bengel. 2005. Handbuch der Persönlichkeitspsychologie. Hogrefe, Goettingen, Germany. Wier, M. and C. Calverley. 2002. Market potential for organic foods in Europe. British Food Journal 104: 45-62. Windhorst, H.-W. 2016. Transparenz zahlt sich aus. Geflügelhaltung im Spiegelbild von Medien und Gesellschaft. Deutsche Geflügelwirtschaft und Schweineproduktion 68: 34-36.

International Food and Agribusiness Management Review

399

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0115 - Tuesday, May 09, 2017 8:32:26 PM - IP Address:24.21.169.207

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2016.0143

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Received: 18 August 2016 / Accepted: 11 January 2017

Influence of product type and individuals’ perceptions on the geographic boundary for local products RESEARCH ARTICLE Madiha Zaffoua, Alicia L. Rihnb, Benjamin L. Campbell c, Hayk Khachatryand, and Omer Hokee aGraduate

Research Assistant, Department of Agricultural and Resource Economics, University of Connecticut, 1376 Storrs Road, Unit 4021, Storrs, CT 06269-4021, USA bPostdoctoral

Research Associate, and dAssistant Professor, Food and Resource Economics Department, University of Florida, Mid-Florida Research and Education Center, Apopka, UF/IFAS Extension, Gainesville, FL 32611, USA cAssistant

Professor, Department of Agricultural and Applied Economics, University of Georgia, 314A Conner Hall, Athens, GA 30602-7509, USA

eTransfer

Pricing Associate, Pricewaterhousecoopers, 488 Almaden Blvd 1800, San Jose, CA 95110, USA

Abstract Over the past couple of decades, consumers have begun to increase purchasing of locally labeled products. However, research has shown their definition of local production is not always accurate and varies by product category. This study investigates consumers’ perceptions of the geographic boundary for local fruits/ vegetables and ornamental plants. A multinomial logit model is used to assess how consumers’ perceptions and perceived characteristics of local product attributes/factors (e.g. freshness, support local community, etc.) influence their understanding of geographic boundaries of local. Results are applicable to producers and retailers in their efforts to obtain portions of the local market. They are also pertinent to policy makers as they determine relevant regulations and definitions of local products. Keywords: green industry, local labeling, produce JEL code: Q13 Corresponding author: bencamp@uga.edu

401

Zaffou et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

1. Introduction During the last decade, growing environmental and social awareness has increased consumers’ desire for products that are perceived as environmentally and/or socially responsible. Locally and organically grown products have seen an upsurge in demand given they generally are perceived to convey these characteristics (Campbell et al., 2014; Darby et al., 2008; Essoussi and Zahaf, 2008; Onozaka et al., 2010; Ritson and Oughton, 2007; Yue and Tong, 2009). Particularly, the term ‘locally grown’ has received a large amount of attention which is demonstrated by every state having some type of marketing program devoted to increasing local sales (Onken and Bernard, 2010). However, many of the state-based local marketing and regulatory efforts have been focused on food-related products. For instance, by 2020 the state of Connecticut aims to have 5% of their total food sales sourced locally (Governor’s Council for Agricultural Development, 2011). Similarly, the National Grocers Association (2015) indicates that ‘more locally grown foods’ is the second most desired improvement among grocery shoppers at 32.1%, behind ‘price/cost savings.’ In 2012, 7.8% of U.S. farms (163,675 farms) were producing/marketing local foods and U.S. demand for local food was valued at $6.1 billion (Low et al., 2015). Perceived benefits of local foods include less transportation miles, more environmentally friendly, lower carbon footprint, longer shelf life, less greenhouse gas emissions, fresher, more community support (i.e. revenue and jobs), and not being associated with corporate production/ marketing entities (Campbell et al., 2014; Darby et al., 2008). Consequently, consumers are often willing to pay premiums for local produce (Darby et al., 2008; Onozaka and McFadden, 2011). Despite reported preference, perceived benefits, and premiums associated with local foods, little research has focused on the meaning of local when applied to non-food products such as ornamental plants. State promotional programs often do not include ornamental plants within their guidelines even given the economic importance of the green industry (Hodges et al., 2015). Three notable exceptions are the ‘Jersey Grown,’ ‘Texas Superstar’ and ‘Fresh from Florida’ programs. Research has shown that using the state promotional program on ornamental plants increases consumers’ willingness-to-pay (Collart et al., 2010) and purchase likelihood (Rihn et al., 2015). Specifically, Collart et al. (2010) determined the Texas Superstar program has low consumer awareness but garners a 10% premium compared to unbranded plants. Rihn et al. (2015) found consumers were more likely to purchase indoor foliage plants that were promoted using the ‘Fresh from Florida’ brand. Other studies investigated local origins of plants without using a state specific marketing program or geographical boundaries. Yue et al. (2011) and Behe et al. (2013) found that consumers prefer local plants to those grown elsewhere. However, consumers’ interest in local plants (Behe et al., 2013; Collart et al., 2010; Rihn et al., 2015; Yue et al., 2011) does little to explain how they define local when considering plants and what benefits influence their definitions. As noted in Feldmann and Hamm’s (2015) literature review of local, the influence of product type on consumer perceptions of local needs to be investigated. Notably, the perceived benefits and geographic boundaries associated with local produce versus local plants should be compared. This study attempts to fill some of the gap in the literature. Our overarching objective was to examine whether local produce would have more perceived benefits and a more confined geographic boundary compared to local plants. Specifically, we were interested in the perception that purchasing local supports the local economy. We hypothesize that consumer’s view supporting the local economy as a larger factor in purchasing local produce compared to plants given the visibility of local labeling campaigns on produce. Furthermore, we hypothesize that demographics and retail outlet will impact the perception of the geographic boundary for local produce and plants. Following the results of Campbell et al. (2013), we expected race, education, gender, and retail shopping environment to contribute to a consumer’s boundary of local produce. We expect females to have a state or regional boundary of local since they are potentially more exposed to local products. Higher educated consumers were anticipated to perceive state and 50-100 mile boundaries given increased education may result in more exposure to media. We expected that consumers shopping at supermarkets, grocery stores, and wholesale clubs would have a broader boundary for local produce given these stores might have their own definitions. In comparison, we expect the factors driving the local boundary for plants to be similar in significance and magnitude as those driving the boundary for local produce. Answering the above questions will not only International Food and Agribusiness Management Review

402

Zaffou et al.

Volume 20, Issue 3, 2017

assist in clarifying consumers’ definitions of local, but can be used by researchers, policy makers and retail firms as they study, educate, legislate and promote local products to the end consumer. The next section discusses literature related to the definition of local, followed by a section on the study’s methodology, the empirical results, and a conclusion.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

2. Literature review – defining local The term local can have diverse connotations depending on place, culture and lifestyle (Carroll and Fahy, 2015). According to the 2008 Food, Conservation, and Energy Act (2008 Farm Act), local or regionally produced food must have travelled less than 400 miles or be sold within the state where it was produced (Martinez et al., 2010). Despite having a governmental issued description, many studies show there is a clear lack of definition of what local food is among consumers and supply chain members (Campbell et al., 2013, 2014; Carroll and Fahy, 2015; Dunne et al., 2011; Martinez et al., 2010; Pearson et al., 2011). However, previous definitions and perceptions of local have consistently been connected to transportation distance(s), other production method attributes, and retail outlets. Less transportation distance is one of the main attributes associated with local production (Campbell et al., 2013, 2014; Carroll and Fahy, 2015). Given local is defined as decreased transportation distance, many consumers perceive local to be associated with helping the local economy and increased product quality, freshness, and shelf life (Campbell et al., 2014; Darby et al., 2008). However, the definition of distance by consumers and businesses varies greatly due to a variety of factors including product type and availability, connections to local producer/community, population density, and relative geographic size of respondents’ location (Carroll and Fahy, 2015; Dunne et al., 2011). Many studies use state and regional boundaries to define origin, but distance tends to be subjective and dependent upon state size. For instance, Carroll et al. (2013) found consumers in geographically larger states (i.e. Virginia, Pennsylvania, Maryland) prefer tomatoes promoted as local over those promoted using the state promotional program. Conversely, consumers in smaller states (i.e. New Jersey, Delaware) prefer the state promotional program to local promotions. Darby et al. (2008) found consumers do not differentiate between in-state and nearby (‘within an undefined sub-state region’) origins. But when ‘in-state’ and ‘nearby’ are combined (termed local), consumers were willing to pay premiums for local products when compared to those grown in the U.S. or of an unidentified origin. This variance also extends to food retailers who frequently include several states in their definitions of local (Dunne et al., 2011). Consumers’ definitions and perceptions of local are often intermixed with other production methods/ characteristics that are perceived as environmentally friendly (Campbell et al., 2013; Onozaka and McFadden, 2011). In a recent study, Campbell et al. (2013) investigated consumers’ understanding and perceptions of local and organic foods in regard to production characteristics. Local was predominantly defined as decreased transportation miles, whereas organic was defined as food produced without synthetic pesticides. Furthermore, consumers with accurate definitions of local and organic shared a similar demographic profile, while consumers with misguided perceptions did not. Onozaka and McFadden (2011) determined locally grown tomatoes were valued the most by consumers which was heightened when local production was combined with fair trade certification. However, if local production negatively impacted the environment (i.e. carbon-intensive), the products had a much more severe discount than those from other origins suggesting that environmental concerns could be a core component to local products (Onozaka and McFadden, 2011). Adams and Salois (2010) argue that the demand for local products arose due to organic production becoming incorporated and therefore not benefiting the small farmer or local community. Supporting evidence from Lang et al. (2014) shows that (in addition to less transportation distance) consumers associate smaller independent growers/ manufacturers, family owned and operated, and unique to the region as elements of local. Retail outlet selection also influences consumers’ perceptions of the availability of local products. For instance, most consumers expect farmers market to sell local foods, while natural food stores and local independent supermarkets are ranked equally likely to sell local foods (Lang et al., 2014). Consumers are International Food and Agribusiness Management Review

403

Zaffou et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

often willing to pay premiums for local products sold at farmers markets (Onken et al., 2011), even though local products at farmers markets are not perceived as being more expensive than products from other origins (Brown, 2003). Shopping at farmers markets has perceived benefits of higher food quality/freshness, safer food, and support for local farmers (Brown, 2003; Conner et al., 2010). Nevertheless, there is very little research on the cumulative interactions of consumers’ perceptions of local and how they vary by product type, particularly plants. Notably, there is a lack of information on what local means when applied to plants and the geographic boundary associated with local plants.

3. Methodology Data was collected via a web-based survey. A total of 1,124 residents of Connecticut completed the survey which equated to a 90% response rate. Connecticut was chosen in order to comply with funding agency requirements. Additionally, Connecticut offers an interesting case given it has a strict definition of local that is heavily geared toward food products, but there is no plant-specific definition. Respondent were obtained from the database of Global Market Insight, Inc. (GMI) and were diverse but representative of Connecticut’s population as a whole. Respondents were emailed an invitation to participate in the survey by GMI. Upon agreeing to participate, participants were directed to the online survey. Survey questions included standard socio-demographic questions, as well as questions on perceived benefits of local produce and plants, purchase behavior, and perceptions of geographic boundaries of local produce and ornamental plants. In order to assess local geographic boundary perceptions, respondents were asked to choose which boundary best corresponded to their view of local. Of the 1,124 respondents that purchased produce, only 648 answered the plant questions given only those indicating they had recently purchased plants were asked to answer the plant questions. The geographic boundary options included my neighborhood/town/city or in a neighboring town/city, within 50-100 miles, Connecticut, Connecticut or a small distance into a neighboring state, Northeastern U.S., East Coast of the U.S., and anywhere in the U.S. These boundaries are similar to those used by Campbell et al. (2013) in their assessment of local food. Respondents were also asked to rate (on a 0-100 point scale where 0=not important; 100=very important) how potential benefits of local (i.e. freshness, price, supports the local economy, safe to eat, environmentally friendly, healthy, open space preservation, and preserve farming for future generations) impacted their decision to purchase local produce and plants. The survey instruments for this study were evaluated and approved by the institution’s Internal Review Board. Econometric model The estimation was conducted using a multinomial logit model (MNL). Participants were presented with j = 1, ...J different geographical definitions of local and they were asked to choose the boundary that they perceived as correct. The MNL model was chosen over the ordered logit model given that the geographical boundaries were not strictly ordered, meaning that some boundaries overlap and cannot be definitively ordered. Further, a deviance test indicated that MNL was an appropriate model compared to ordered logit. Considering a respondent i facing j=1, ..., j definitions of local, the indirect utility function for respondent i, from choosing definition j, is given by the following expression: Uij = Χijβ + εij

i=1,....,J

(1)

where Χij represents the participant’s demographic characteristics, different shopping locations, and respondent’s beliefs about local products. The respondent will choose the definition that gives him/her the highest utility. As noted by Greene (2003: 721), the probability of choosing choice j can be represented as: Prob(Yi = j) | xi ) =

e β’j xi (2) 1+∑Jk=1 e β’k xi

International Food and Agribusiness Management Review

404

Zaffou et al.

Volume 20, Issue 3, 2017

The marginal effect were then calculated by taking partial derivatives with respect to each explanatory variable. Marginal effects for continuous variables were calculated at the mean while categorical explanatory variable marginal effects were calculated based on a change from zero to one.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

4. Data Table 1 summarizes the sampleâ&#x20AC;&#x2122;s socio-demographic variables. Based on the sample characteristics, the sample was 88% Caucasian with an average income of $87,704. The sampleâ&#x20AC;&#x2122;s median age was 53 and 63% were female. Given variances are not provided for the Connecticut (CT) Census data estimates, testing whether our sample is statistically different from the CT population cannot be performed. However, our sample appears to over-represent Caucasians (82% CT) and higher income households ($75,000 CT). Our sample also has a higher median age of 53 years (compared to 43 years from CT census). However, the age variable is not directly comparable as the CT census counts persons under 18 years of age while our survey was limited to persons 18 years of age or older. Furthermore, oversampling females is in line with shopping patterns since they are the primary household shoppers (Private Label Manufacturers Association, 2013). In addition to the socio-demographic questions, respondents were given several geographical distances to define the term local and were asked to select the one that they consider the most appropriate, based on their personal views of local. Table 2 summarizes the different geographical definitions of local used in the survey and the corresponding percentage of people choosing each boundary. The highest percentage of respondents indicated that CT was local (32% for produce and plants, respectively), followed by CT and a small portion of the surrounding states, and then 50-100 miles between production and point-of-sale. Roughly a quarter of respondents selected a broader geographical boundary (i.e. surrounding states or greater) at 27% for produce and 24% for plants. Based on CT law, the correct geographical boundary is produced in CT or within 10 miles of point-of-purchase, which can include a small area in neighboring states (Connecticut General Assembly, 2012). Finally, given that environmental, moral and health concerns are often important for local food choice, participants indicated the importance of perceived benefits when purchasing locally grown produce and plants (Table 3). One of our main hypothesis was that supporting the local economy would be seen as a larger benefit for produce compared to plants. However, we find no statistical difference between the rating for produce and plants. With respect to the other perceived benefits, freshness was most important, regardless of product type. Consumption characteristics were more important for produce (i.e. safe to eat, healthiness). Conversely, for local plants, supporting the local economy and price were perceived as more important than consumption related traits. Table 1. Demographic characteristics of the sample. Number of respondents Gender Age Average income Ethnicity

Primary shopper

Male Female

Caucasian African American Hispanic Yes

1,124 37% 63% 53 $87,704 88% 4% 8% 70%

International Food and Agribusiness Management Review

405

Zaffou et al.

Volume 20, Issue 3, 2017

Table 2. Respondents’ boundary of ‘local’.1,2

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Definition Produced in: 1. My neighborhood, town/city, or in a neighboring town/city 2. Within 50-100 miles of point-of-sale 3. Connecticut 4. Connecticut or a small distance into a neighboring state 5. A neighboring state 6. Northeastern U.S. 7. East coast of the U.S. 8. Anywhere in the U.S. Total responses3

Fruit/vegetable characteristic

Plant characteristic

6% 17% 32% 18% 9% 7% 5% 6% 1,124

7% 17% 33% 18% 9% 9% 4% 3%*** 648

The survey question was ‘Which definition below do you feel is an appropriate boundary for fruit/vegetables to be considered locally produced?’ For the plant question, plant was substituted for fruit/vegetables. Measurement error could have been introduced if respondents answered in a way that was inconsistent with their perceptions; however, there is no way to accurately measure this error if it exists. 2 A t-test was used to test for significant differences; *** denotes significance at the 0.01 level. 3 Only individuals who had purchased plants recently (n=648; 58% of the sample) were asked to define the boundary for local plants.

Table 3. Respondents’ perceived benefits of ‘local’ products.1,2 Perceived benefits

Fruit/vegetable (mean %)

Plant (mean %)

Freshness Price Supports the local economy Safe to eat Environmentally friendly Healthy Open space preservation Preserve farming for future generations

88 74 81 84 69 81 61 74

83*** 79*** 80 64*** 71 77*** 64* 72

Beliefs were phrased as ‘How important are the following characteristics to your decision to purchase locally grown fruits and vegetables (plants)?’ with a 0-100 point scale (0=not important; 100=very important). 2 A t-test was used to test for significant differences as significance at the 0.01, 0.05, and 0.1 level are denoted by ***, **, *.

5. Empirical results In this section the marginal effects of the MNL regression analysis are discussed. The marginal effects demonstrate how the independent variables (e.g. expenditures, perceived benefits, socio-demographics, and retail outlets) influence respondents’ geographical boundary of local (as described in Table 2). The relationships and boundaries of local for produce and plants are described separately. Local fruit and vegetable boundary results Local produce expenditures influence respondents’ geographical boundaries of local. A $1 increase in a respondent’s expenditure on local produce decreases the likelihood of defining local as his/her own neighborhood by 0.01% (Table 4). Interestingly, respondents who spent more on local produce were more likely to select East Coast and ‘anywhere in the U.S.’ as local. International Food and Agribusiness Management Review

406

Zaffou et al.

Volume 20, Issue 3, 2017

Total expenditure2 Perceived benefits Freshness Price Supports local economy Healthiness Open space preserving Environ. friendliness Farm preserving Demographics Age Number of children Income Male Caucasian African American Single family house Suburban area Metropolitan area

Entire U.S. (8)

East coast U.S. (7)

Northeast U.S. (6)

Neighboring state (5)

Connecticut + portion of neighboring state (4)

Connecticut (3)

50-100 miles (2)

Variables

My neighborhood (1)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Table 4. Marginal effects of utility parameters for the boundary of local fruits and vegetables.1

-0.0001** -0.0001 (0.032) (0.461)

0.0000 (0.970)

0.0000 (0.618)

0.0000 (0.328)

0.0000 (0.680)

0.0000* (0.085)

0.0001*** (0.002)

-0.0008 (0.179) -0.0002 (0.652) -0.0006 (0.258) 0.0001 (0.845) -0.0003 (0.358) 0.0007 (0.122) -0.0004 (0.383)

-0.0002 (0.842) 0.0000 (0.965) -0.0013 (0.121) 0.0006 (0.500) 0.0006 (0.338) 0.0000 (0.973) -0.0009 (0.232)

0.0006 (0.647) 0.0004 (0.579) 0.0003 (0.757) -0.0001 (0.937) 0.0000 (0.996) -0.0015* (0.096) 0.0017* (0.084)

0.0014 (0.190) -0.0007 (0.262) 0.0009 (0.278) -0.0011 (0.173) -0.0005 (0.444) -0.0006 (0.398) 0.0003 (0.682)

0.0008 (0.157) -0.0002 (0.392) 0.0006 (0.150) -0.0003 (0.423) -0.0006** (0.037) 0.0007** (0.017) 0.0001 (0.773)

0.0005 (0.303) 0.0000 (0.842) 0.0002 (0.477) 0.0002 (0.637) 0.0001 (0.551) 0.0003 (0.350) -0.0004 (0.136)

-0.0017*** (0.000) 0.0006** (0.029) 0.0005 (0.134) 0.0005 (0.185) -0.0001 (0.654) 0.0003 (0.401) -0.0005 (0.174)

-0.0007 (0.111) 0.0000 (0.926) -0.0008* (0.068) 0.0000 (0.922) 0.0008*** (0.005) 0.0002 (0.620) 0.0001 (0.721)

-0.0006 (0.278) 0.0083 (0.239) 0.0002* (0.072) -0.0121 (0.365) -0.0215 (0.415) 0.0623 (0.330) -0.0054 (0.803) -0.0540*** (0.007) -0.0238 (0.148)

0.001 (0.320) -0.0163 (0.339) -0.0001 (0.804) 0.0619** (0.036) 0.0258 (0.577) 0.014 (0.875) -0.0286 (0.459) -0.0537* (0.093) -0.0103 (0.837)

0.0002 (0.880) 0.0032 (0.868) 0.0004 (0.257) -0.0552 (0.115) -0.0539 (0.368) -0.112 (0.214) 0.0159 (0.717) 0.0604 (0.105) 0.0031 (0.964)

0.0003 (0.760) 0.0151 (0.301) -0.0004 (0.136) -0.0048 (0.872) 0.0879** (0.044) 0.0477 (0.681) 0.0027 (0.942) 0.0511* (0.090) -0.0077 (0.892)

0.0004 (0.438) -0.005 (0.498) 0.0001 (0.543) -0.0232* (0.058) 0.0151 (0.434) -0.0847*** (0.000) 0.0093 (0.539) 0.0048 (0.710) 0.0013 (0.958)

-0.0008** (0.035) -0.0151* (0.063) 0.0000 (0.974) 0.0162 (0.202) 0.0013 (0.948) 0.0664 (0.299) 0.0103 (0.430) -0.0058 (0.644) -0.0158 (0.331)

-0.0012*** (0.006) -0.0042 (0.599) 0.0001 (0.355) 0.013 (0.373) 0.0096 (0.625) -0.0023 (0.949) -0.013 (0.460) 0.0139 (0.357) 0.0133 (0.694)

0.0007 (0.115) 0.0139*** (0.004) -0.0003** (0.019) 0.0043 (0.748) -0.0644* (0.063) 0.0086 (0.754) 0.0087 (0.519) -0.0167 (0.289) 0.0399 (0.189)

International Food and Agribusiness Management Review

407

Zaffou et al.

Volume 20, Issue 3, 2017

Education Some college Bachelor degree Graduate degree

-0.008 (0.674) -0.0225 (0.232) -0.03 (0.114)

Local retail outlets If primary shopper -0.0182 (0.272) Farms -0.0293 (0.124) Other locations 0.0907 (0.402) Specialty stores -0.0469*** (0.000) Supermarket -0.0255 (0.212) Grocery stores -0.0208 (0.317) Wholesale clubs -0.0342* (0.085) Number of obs. 1,124

Entire U.S. (8)

East coast U.S. (7)

Northeast U.S. (6)

Neighboring state (5)

Connecticut + portion of neighboring state (4)

Connecticut (3)

50-100 miles (2)

Variables

My neighborhood (1)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Table 4. Continued.

-0.0067 (0.878) -0.0047 (0.915) 0.0564 (0.251)

-0.0405 (0.424) -0.1136** (0.024) -0.0998* (0.053)

0.0588 (0.286) 0.1644*** (0.005) 0.1165* (0.056)

0.0189 (0.438) 0.0109 (0.636) -0.0011 (0.959)

0.0008 (0.966) 0.016 (0.448) 0.019 (0.390)

0.0062 (0.755) -0.0194 (0.255) -0.0273* (0.093)

-0.0295** (0.016) -0.0313** (0.020) -0.0336** (0.012)

0.0161 (0.567) 0.0312 (0.694) 0.017 (0.908) 0.0012 (0.988) 0.0304 (0.587) 0.0635 (0.261) 0.1466 (0.143) 1,124

-0.0300 (0.419) 0.0083 (0.925) -0.1247 (0.416) 0.1664 (0.100) 0.0424 (0.520) 0.0413 (0.528) -0.0966 (0.263) 1,124

0.0202 (0.509) 0.0664 (0.396) -0.1268 (0.115) -0.0113 (0.877) -0.0108 (0.836) -0.022 (0.672) 0.0133 (0.861) 1,124

-0.0067 (0.634) -0.0306** (0.050) -0.0252 (0.469) -0.0212 (0.312) -0.0372** (0.031) -0.0350** (0.041) -0.0173 (0.375) 1,124

0.0037 (0.748) -0.0247 (0.153) 0.0486 (0.510) -0.0766*** (0.000) 0.0009 (0.966) -0.0001 (0.997) -0.013 (0.591) 1,124

0.0073 (0.579) -0.0096 (0.747) 0.0975 (0.496) -0.0104 (0.703) 0.0105 (0.696) -0.0053 (0.833) 0.0032 (0.931) 1,124

0.0076 (0.547) -0.0117 (0.603) 0.0229 (0.755) -0.0012 (0.968) -0.0107 (0.583) -0.0217 (0.259) -0.0021 (0.942) 1,124

Number in brackets indicates the coefficient’s P-value. *, **, and *** denote statistical significance at 10, 5 and 1%, respectively; dependent variables were the various definitions of ‘local’ for fruits and vegetables as defined in Table 2. 2 We used the method discussed in Terza et al. (2008) to better understand whether expenditure explains boundary or vice versa. Results of the test indicate that expenditure explains boundary; thereby, expenditure was treated as an explanatory variable in the model.

When considering the correlation between respondents’ definitions of local and the perceived benefits of local production, several patterns emerge. Respondents who perceive local as fresher are 0.17% less likely to define local as the East Coast. If respondents perceived local as being more expensive they were more likely to define local as including the East Coast. Not surprisingly, perceptions of local produce helping the local economy reduce the likelihood of respondents’ defining local as the entire U.S. by 0.08%. However, if respondents perceive local produce as preserving open space, they are 0.06% less likely to define local as the neighboring state, but 0.08% more likely to define local as the entire U.S. Respondents who perceive local products as environmentally friendly are less likely to define local as just CT but more likely to include neighboring states in their definition of local produce. If farm preservation is a perceived benefit, respondents are 0.17% more likely to select CT as their geographical definition of local. Regarding the correlation between the boundary of local and respondents’ socio-demographic variables, respondents who had a higher income or were less likely to live in a suburb were more likely to define the International Food and Agribusiness Management Review

408

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Zaffou et al.

Volume 20, Issue 3, 2017

local boundary as one’s neighborhood (Table 4). Males and respondents living in non-suburban areas were more likely to view the boundary of local as 50-100 miles from the retail outlet. Respondents with bachelors and graduate degrees were less likely to define the boundary of local as grown within CT. Caucasian, living in a suburb, and highly educated respondents were more likely to perceive CT and a small portion of the neighboring state(s) as the boundary for local. Females and non-African Americans were more likely to view the local boundary as a neighboring state. Younger respondents and respondents with fewer children in the household were more likely to perceive the boundary for local to be the Northeast U.S. Younger respondents and those not having a graduate degree were more likely to view the East Coast of the U.S. as the local boundary. Lower income respondents, those having more children living in the household, being non-Caucasian (compared to the other ethnicity), and not having some college perceived the boundary for local to be the entire U.S. Respondents’ typical shopping locations influenced their probability of selecting the various local geographic boundaries. If respondents primarily shop at specialty stores or wholesale clubs they were less likely to define the boundary of local as ‘my neighborhood’. If they purchase their produce from farms, supermarkets or grocery stores, they were less likely to define local as including the neighboring states. Lastly, if they shop at specialty stores, they were 7.66% less likely to define local as ‘Northeast U.S.’. Taking all the results above in aggregate provides some interesting implications. Notably, our hypotheses were accurate for the most part. Caucasian respondents tended to have a smaller boundary for local which correlated with the CT state mandated definition. Females tended to have an increased boundary for local while higher educated respondents had a smaller boundary. However, contrary to our expectation, retail outlet provided little information that can be viewed to provide implications. Based on these results, firms need to examine their clientele to better understand the demographics they serve in order to provide information that educates their respondents with the definitions of local the firm is using. Local plant definition results The investigation of local plant definitions showed that expenditures on local plants influence respondents’ geographical boundary definitions of local plants. Specifically, a $1 increase reduced the likelihood of defining the boundary of local as their neighborhood by 0.02% but increased the selection of neighboring states by 0.01% (Table 5). Most of the coefficients of expenditure were not significant, except for the neighboring state boundary whereby increased expenditure had a positive effect. Respondents that define local as preserving farmland were less likely to consider 50-100 miles from the retail outlet as being local. Respondents who perceived price as being an important part of the purchase decision for local produce were more likely to consider grown in CT as the boundary of local plants. Respondents’ perceiving supporting the local economy as important to the local purchase decision were less likely to consider CT as the boundary for local. Individuals who perceive local as preserving open space were less likely to believe that local was defined as CT and a portion of the neighboring state(s); however, respondents who viewed local as environmentally friendly, were more likely to have an accurate, as defined by the state of CT, definition that local is CT grown and a portion of a neighboring state(s). If the respondent identified local as being less expensive or supporting the local economy, then s/he was less likely to perceive that neighboring state(s) are local. Regarding the relationship between respondents’ socio-demographics and their definitions of local plants, younger respondents were more likely to define local plants as being grown in ‘my neighborhood’, while respondents that males and non-metropolitan respondents were more likely to consider 50-100 miles from the retail outlet as the boundary for local. Furthermore, females were more likely to consider grown in CT as the boundary for local. Higher income respondents were more likely to perceive the CT plus a portion of a neighboring state as the boundary. African Americans were less likely to perceive regional (i.e. ‘Northeast

International Food and Agribusiness Management Review

409

Zaffou et al.

Volume 20, Issue 3, 2017

Total expenditure2 -0.0002*** (0.000) Perceived benefits Freshness -0.0005 (0.445) Price 0.0000 (0.937) Supports local -0.0009 economy (0.128) Healthiness 0.0005 (0.275) Open space 0.0005 preserving (0.481) Environ. -0.0009 friendliness (0.104) Farm preserving 0.0008 (0.216) Demographics Age -0.0014** (0.022) Number of -0.0038 children (0.702) Income -0.0001 (0.614) Male -0.0015 (0.928) Caucasian -0.0214 (0.503) African American -0.0084 (0.826) Single family 0.009 house (0.644) Suburban area -0.0272 (0.186) Metropolitan area -0.0192 (0.323)

Entire U.S. (8)

East coast U.S. (7)

Northeast U.S. (6)

Neighboring state (5)

Connecticut + portion of neighboring state (4)

Connecticut (3)

50-100 miles (2)

Variables

My neighborhood (1)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Table 5. Marginal effects of utility parameters on the boundary of local plants.1

0.0000 (0.873)

0.0001 (0.267)

0.0001 (0.541)

0.0001** (0.028)

0.0000 (0.828)

0.0000 (0.441)

0.0000 (0.123)

-0.0005 (0.602) -0.0008 (0.329) 0.0008 (0.415) 0.001 (0.276) 0.0011 (0.261) -0.0004 (0.639) -0.0025*** (0.009)

0.0008 (0.540) 0.0025** (0.035) -0.0027* (0.051) -0.0013 (0.189) 0.0006 (0.602) -0.0017 (0.208) 0.0017 (0.209)

-0.0004 (0.667) -0.0004 (0.576) 0.0006 (0.581) 0.0003 (0.699) -0.0017* (0.065) 0.0021* (0.057) 0.0011 (0.334)

0.0008 (0.313) -0.0010** (0.026) 0.0013* (0.072) -0.0001 (0.826) -0.0004 (0.403) -0.0003 (0.658) 0.0006 (0.320)

0.0003 (0.727) -0.0004 (0.558) 0.0004 (0.565) -0.0003 (0.679) -0.0003 (0.683) 0.0007 (0.277) -0.0012 (0.130)

-0.0003 (0.404) 0.0000 (0.950) 0.0007* (0.099) 0.0000 (0.866) 0.0004* (0.070) 0.0002 (0.560) -0.0006 (0.110)

-0.0001 (0.724) 0.0001 (0.703) -0.0003 (0.241) -0.0001 (0.665) -0.0001 (0.728) 0.0003 (0.247) 0.0001 (0.732)

0.0014 (0.323) -0.0022 (0.926) -0.0004 (0.305) 0.0828** (0.030) 0.0347 (0.602) -0.1067 (0.177) -0.0152 (0.812) -0.063 (0.106) -0.0890* (0.065)

0.0000 (0.997) -0.0017 (0.948) 0.0004 (0.310) -0.0813* (0.086) 0.0956 (0.241) -0.0607 (0.727) -0.0621 (0.428) 0.0523 (0.297) 0.0897 (0.380)

0.0015 (0.306) 0.0056 (0.802) 0.0005* (0.089) 0.0452 (0.245) 0.0337 (0.626) 0.0589 (0.734) 0.0454 (0.414) 0.0453 (0.224) -0.0121 (0.882)

-0.0003 (0.774) -0.0047 (0.752) -0.0003 (0.150) -0.0364* (0.070) 0.0424 (0.194) 0.2053 (0.377) 0.0604*** (0.006) -0.0044 (0.847) 0.0278 (0.637)

-0.001 (0.262) -0.0095 (0.575) -0.0001 (0.795) 0.0014 (0.960) -0.0974 (0.146) -0.0693** (0.048) 0.0225 (0.517) -0.0054 (0.858) -0.0015 (0.979)

0.0000 (0.994) 0.0125* (0.076) 0.0001 (0.571) -0.0104 (0.454) -0.0429 (0.277) -0.0122 (0.538) -0.0578 (0.138) 0.0029 (0.850) 0.0039 (0.887)

-0.0002 (0.598) 0.0039 (0.480) -0.0002* (0.069) 0.0003 (0.976) -0.0446 (0.236) -0.0068 (0.654) -0.0023 (0.851) -0.0004 (0.972) 0.0003 (0.984)

International Food and Agribusiness Management Review

410

Zaffou et al.

Volume 20, Issue 3, 2017

Education Some college Bachelor degree Graduate degree Local retail outlets If primary shopper

-0.0096 (0.682) -0.0017 (0.947) -0.0148 (0.567)

0.0448*** (0.002) Home -0.0052 improvement store (0.776) Nursery 0.0034 greenhouse (0.857) Supermarkets -0.0049 (0.913) Farms 0.015 (0.771) Wholesale clubs -0.0317 (0.151) Observations 648

Entire U.S. (8)

East coast U.S. (7)

Northeast U.S. (6)

Neighboring state (5)

Connecticut + portion of neighboring state (4)

Connecticut (3)

50-100 miles (2)

Variables

My neighborhood (1)

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Table 5. Continued.

-0.0454 (0.387) -0.0535 (0.310) -0.0137 (0.808)

-0.0797 (0.300) -0.0952 (0.222) -0.1203 (0.119)

-0.0488 (0.398) -0.0299 (0.632) -0.0348 (0.566)

0.0823 (0.248) 0.0724 (0.310) 0.0969 (0.203)

0.1100 (0.261) 0.1169 (0.236) 0.0920 (0.307)

0.0167 (0.617) 0.0059 (0.832) 0.0026 (0.927)

-0.0255*** (0.006) -0.0147 (0.166) -0.0079 (0.429)

-0.0086 (0.816) 0.0146 (0.762) -0.0184 (0.629) -0.1297** (0.011) 0.0164 (0.858) -0.0285 (0.666) 648

-0.0374 (0.454) -0.0949* (0.085) -0.0528 (0.290) 0.143 (0.169) -0.1798** (0.030) 0.0175 (0.851) 648

0.0492 (0.180) 0.0044 (0.929) 0.0464 (0.303) -0.0951 (0.144) 0.0397 (0.680) 0.0617 (0.484) 648

-0.0569* (0.052) 0.0159 (0.618) -0.0096 (0.697) 0.0225 (0.669) -0.0064 (0.898) -0.0519* (0.075) 648

-0.0265 (0.402) 0.0532 (0.232) 0.0118 (0.708) 0.0326 (0.680) 0.039 (0.625) -0.0236 (0.659) 648

0.0228 (0.103) -0.0014 (0.941) 0.0237 (0.242) 0.0033 (0.930) 0.0033 (0.915) 0.0014 (0.959) 648

0.0127 (0.167) 0.0134 (0.456) -0.0044 (0.676) 0.0283 (0.558) 0.0728 (0.362) 0.0551 (0.299) 648

Number in brackets indicates the coefficient’s P-value. *, **, and *** denote statistical significance at 10, 5 and 1%, respectively; dependent variables were the various definitions of ‘local’ for plants as defined in Table 2. 2 We used the method discussed in Terza et al. (2008) to better understand whether expenditure explains boundary or vice versa. Results of the test indicate that expenditure explains boundary; thereby, expenditure was treated as an explanatory variable in the model.

U.S.’) as a local boundary for plants. Finally, lower income respondents were more likely to perceive ‘entire U.S.’ as a geographic boundary for local plants. Shopping behavior also influenced respondents’ definitions of local plants. If the respondent was the household’s primary shopper, (s)he was 4.48% more likely to define local as his/her neighborhood but 5.69% less likely to define local as ‘including neighboring states’. Supermarket shoppers were 12.97% less likely to define local as 50-100 miles, while home improvement store and farm shoppers were less likely to perceive CT plants as local. Implications from the plant results are that the factors driving the boundary for local plants do not seem to be the same as factors for local produce. As with produce, female respondents were more likely to choose a larger boundary (i.e. state). However, we could find little pattern for race or education. As with produce, the retail outlet some significance, but provided little information that could guide industry stakeholders. Based on these findings it appears that there are differences in what drives a respondent’s perception of local

International Food and Agribusiness Management Review

411

Zaffou et al.

Volume 20, Issue 3, 2017

produce and plants. Firms (and policy makers) need to be aware of this fact and make decisions understanding that a one size fits all (for produce and plants) strategy that targets specific demographics may not work.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

6. Conclusions Our results provide insights into respondent perceptions of local in terms of geographical distances. Most respondents indicated that the boundary for local products are those grown in CT or CT and a portion of the surrounding states. This is not surprising when one considers the United States Department of Agriculture (USDA) definition of local (i.e. less than 400 miles or within the state of production) and the size of CT (roughly 70 miles wide and 110 miles long) which easily falls within the USDA definition of local. Future studies could address similar questions in states that exceed the USDA’s definition of local to test the robustness of the results. With respect to benefits, we find the perceived boundary for local produce and plants is similar. However, the degree by which respondents view certain benefits as important to their purchasing of local produce and plants vary. Given this difference between produce and plants, laws seemingly meant for food are being applied by consumers to non-food products. Further, there were several similarities in respondents’ perceptions of local products, regardless of product type (i.e. produce, plants). Notably, in the empirical results respondents who used broader geographical areas to define local perceived local production as a means of preserving open space. However, although respondents rated preserving open space positively, they did not rate this benefit nearly as highly as other perceived benefits (e.g. freshness, supporting the local economy, etc.) There is an opportunity in future studies to assess respondent tradeoffs of perceived benefits of local and how those benefits influence purchasing behavior and willingness-to-pay. Additionally, a closer investigation of the relationships between perceived benefits of local production and the consistency of those perceptions across product categories could help explain this finding. Results of this study are relevant for researchers, policy makers and retail firms. Results provide researchers with more information about the relationship between respondent perceptions and their definitions of local production. This information can be used in future studies on local production, labeling and marketing. From the policy maker’s perspective, the results can help decision makers implement programs that provide support for the purchase of locally grown products. Our findings result in a better understanding of respondents’ perceptions of locally grown products and can be used by policy makers to refine laws/regulations to insure local production aligns with respondent expectations. Results of this study will also provide businesses with critical information about why/how respondents are responding to local labeling. Notably, it allows retailers to understand the respondent trade-offs between local products so they can address respondent needs more appropriately.

Acknowledgements ‘Funding has been provided by the Specialty Crop Block Grant Program of the Agricultural Marketing Service, U.S. Department of Agriculture, awarded and administered by the Connecticut Department of Agriculture’. Dr. Hoke’s contribution to this paper is a result of his efforts and resources during his time as a Ph.D. student at the University of Connecticut.’

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: 331-341. Behe, B.K., B.L. Campbell, C.R. Hall, H. Khachatryan, J.H. Dennis and C. Yue. 2013. Consumer preferences for local and sustainable plant production characteristics. HortScience 48: 200-208. Brown, C. 2003. Consumers’ preferences for locally produced food: a study in southeast Missouri. American Journal of Alternative Agriculture 4: 213-224. International Food and Agribusiness Management Review

412

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Zaffou et al.

Volume 20, Issue 3, 2017

Campbell, B.L., H. Khachatryan, B.K. Behe, J. Dennis and C.R. Hall. 2014. U.S. and Canadian consumer perceptions of local and organic. International Food and Agribusiness Management Review 17: 21-40. Campbell, B.L., S. Mhlanga and I. Lesschaeve. 2013. Perception versus reality: Canadian consumer views of local and organic. Canadian Journal of Agricultural Economics 60: 531-558. Carroll, K.A., J.C. Bernard and J.D. Pesek Jr. 2013. Consumer preferences for tomatoes: the influence of local, organic, and state program promotions by purchasing venue. Journal of Agricultural and Resource Economics 38: 379-396. Carroll, B.E. and F. Fahy. 2015. Locating the locale of local food: the importance of context, space and social relations. Renewable Agriculture and Food Systems 30: 563-576. Collart, A.J., M.A. Palma and C.R. Hall. 2010. Branding awareness and willingness-to-pay associated with the Texas Superstar and Earth-Kind brands in Texas. HortScience 45: 1226-1231. Connecticut General Assembly. 2012. Connecticut general statutes title 22 – Agriculture. Chapter 423 – Grading and marketing of farm products, section 22-38 – Advertising of Connecticut-grown farm products. Advertising of locally-grown farm products. Available at http://tinyurl.com/z7fncby. Conner, D., K. Colasanti, R.B. Ross and S.B. Smalley. 2010. Locally grown foods and farmers markets: consumer attitudes and behaviors. Sustainability 2: 742-756. Darby, K., M.T. Batte, S. Ernst and B. Roe. 2008. Decomposing local: a conjoint analysis of locally produced foods. American Journal of Agricultural Economics 90: 476-486. Dunne, J.B., K.J. Chambers, K.J. Giombolini and S.A. Schlegel. 2011. What does ‘local’ mean in the grocery store? Multiplicity in food retailers’ perspectives on sourcing and marketing local foods. Renewable Agriculture and Food Systems 26: 46-59. Essoussi, L.H. and M. Zahaf. 2008. Decision making process of community organic food consumers: an exploratory study. Journal of Consumer Marketing 25: 95-104. Feldmann, C. and U. Hamm. 2015. Consumers’ perceptions and preferences for local food: a review. Food Quality and Preference 40: 152-164. Governor’s Council for Agricultural Development. 2011. An act concerning the governor’s council for agricultural development. Available at: http://tinyurl.com/7mbco2c. Greene, W.H. 2003. Econometric Analysis, 5th ed. Prentice Hall, Upper Saddle River, NJ, USA. Hodges, A., C. Hall, M. Palma and H. Khachatryan. 2015. Economic contributions of the green industry in the United States in 2013. HortTechnology 25: 805-814. Lang, M., J. Stanton and Y. Qu. 2014. Consumers’ evolving definition and expectations for local food. British Food Journal 116: 1808-1820. Low, S.A., A. Adalja, E. Beaulieu, N. Key, S. Martinez, A. Melton, A. Perez, K. Ralston, H. Stewart, S. Suttles, S. Vogel and B.B.R Jablonski. 2015. Trends in U.S. local and regional food systems. U.S. Department of Agriculture, Economic Research Service, Washington, WA, USA. Martinez, S., M. Hand, M. Da Pra, S. Pollack, K. Ralston, T. Smith, S. Vogel, S. Clark, L. Lohr, S. Low and C. Newman. 2010. Local food systems: concepts, impacts, and issues. USDA, Economic Research Report Number 97. U.S. Department of Agriculture, Economic Research Service, Washington, WA, USA. National Grocers Association. 2015. National Grocers Association Supermarketguru® consumer survey report. Available at: http://tinyurl.com/gnwgl56. Onken, K.A. and J.C. Bernard. 2010. Catching the local bug: a look at state agricultural marketing programs. Choices 25(1): article 112. Onken, K.A., J.C. Bernard, and J.D. Pesek. 2011. Comparing willingness to pay for organic, natural, locally grown, and state marketing program promoted foods in the Mid-Atlantic region. Agricultural and Resource Economics Review 40: 233-247. Onozaka, Y. and D.T. McFadden. 2011. Does local labeling complement or compete with other sustainable labels? A conjoint analysis of direct and joint values for fresh produce claims. American Journal of Agricultural Economics 93: 693-706. Onozaka, Y., G. Nurse, and D.T. McFadden. 2010. Local food consumers: how motivations and perceptions translate to buying behavior. Choices. 25(1): article 109.

International Food and Agribusiness Management Review

413

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0143 - Tuesday, May 09, 2017 8:33:03 PM - IP Address:24.21.169.207

Zaffou et al.

Volume 20, Issue 3, 2017

Pearson, D., J. Henryks, A. Trott, P. Jones, G. Parker, D. Dumaresq and R. Dyball. 2011. Local food: understanding consumer motivations in innovative retail formats. British Food Journal 113: 886-899. Private Label Manufacturers Association (PLMA). 2013. Todayâ&#x20AC;&#x2122;s primary shopper. Society may be radically changing but women still dominate the marketplace. PLMA Consumer Research Study. Available at: http://tinyurl.com/jbslb4p. Rihn, A., H. Khachatryan, B. Campbell, C. Hall and B. Behe. 2015. Consumer response to novel indoor foliage plant attributes: evidence from a conjoint experiment and gaze analysis. HortScience 50: 1524-1530. Ritson, C. and E. Oughton. 2007. Food consumers and organic agriculture. In: Understanding Consumers of Food Products, edited by L. Frewer and H. Van Trijp. Woodhead Publishing, Cambridge, UK. Terza, J.V., A. Basu and P.J. Rathouz. 2008. Two-stage residual inclusion estimation: addressing endogenity in health econometric modeling. Journal of Health Economics 27: 531-543. Yue, C. and C. Tong. 2009. Organic or local? Investigating consumer preferences for fresh produce using a choice experiment with real economic incentives. HortScience 2: 366-371. Yue, C., J.H. Dennis, B.K. Behe, C.R. Hall, B.L. Campbell and R.G. Lopez. 2011. Investigating consumer preference for organic, local, or sustainable ornamental plants. HortScience 46: 610-615.

International Food and Agribusiness Management Review

414

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2016.0101

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

Received: 26 May 2016 / Accepted: 23 November 2016

Managing the pork supply chain through a cooperative: the case of Jinzhong Food Co. Ltd. CASE STUDY – WITH TEACHING NOTE Chen Jia, Fu Jia

b,c,

and Jacques Trienekensd

aAssistant

Professor, China Academy for Rural Development (CARD), Zhejiang University, Hangzhou310058, Zhejiang, China P.R.

bProfessor,

cSenior

School of Economics and Management, Minjiang University, Wenxian Rd, Minhou Xian, Fuzhou Shi 350108, Fujian, China P.R.

Lecturer, Business School, University of Exeter, Rennes drive, Exeter EX4 4PU, United Kingdom dProfessor,

Management Studies Group, Wageningen University, P.O. Box 8130, 6700 EW Wageningen, the Netherlands

Abstract Food security and supply continuity have become main concerns of food companies and societies nowadays. To address these concerns, Jinzhong Food Co. Ltd. (hereafter abbreviated to Jinzhong) was one of the first Chinese meat companies to establish and integrate a pig production cooperative in 2005. Over the last decade or so, Jinzhong has successfully developed the cooperative to stabilize and improve the quality of pig supply, by building relationships with pig producers (farmers) and achieving an exceptional financial performance in the process. The company-led cooperative represents an innovative supply chain governance mechanism in a Chinese context. However, the pork industry has evolved significantly and the time has come for the senior management team at Jinzhong to decide whether or not to keep the cooperative. This case study is aimed at senior undergraduate students and postgraduate students specializing in agricultural economics/ agribusiness and can also be used for executive training for the management of food companies. Keywords: pork supply chain, farmer’s cooperative, case study JEL code: Q13 Corresponding author: fu.jia@exeter.ac.uk

This case was prepared for class discussion rather than to illustrate either effective or ineffective handling of an agribusiness management situation. The authors may have disguised names and other identifying information presented in the case in order to protect confidentiality. IFAMA prohibits any form of reproduction, storage or transmittal without its written permission. To order copies or to request permission to reproduce, contact the IFAMA Business Office. Interested instructors at educational institutions may request the teaching note by contacting: ifamr@ifama.org. © 2017 Ji et al.

415

Ji et al.

Volume 20, Issue 3, 2017

1. Introduction

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

Jinzhong Food Co. Ltd (hereafter abbreviated to Jinzhong), located in Qionglai, a prefecture belonging to Chengdu, the capital of Sichuan Province in China, is a leading pork production/processing company in south west China. On a hot summer morning in June 2005, Mr. Gou Mingjun, the Deputy Bureau Chief at the livestock bureau in Qionglai, rushed into the office of Mr. Liu Xiang, president of Jinzhong saying: ‘look Lao Liu, some epidemic diseases are prevailing in our surrounding areas, and I am really concerned because they may soon reach our county. We have got to work together to prevent these diseases. What do you suggest we do?’ Mr. Liu responded: ‘firstly, the government needs to encourage farmers to use epidemic vaccinations so that their farms will not be affected. Secondly, we could create an organization so that remote farmers are organized; then we can get to know their pig production processes better, and we can train them on a large scale to prevent disease on their farms in the long term’. Liu added: ‘establishing an organization would also be good for our company, because we could contract farmers to sell their pigs to us. As a result, we would obtain a more stable and high-quality pig supply’. ‘So, what kind of organization shall we build, and how shall we build it?’ Mr. Gou’s words got them both thinking. After this meeting, Mr. Gou and Mr. Liu had several rounds of discussions. Gou discovered that there were some cooperatives in the fruit and vegetables industries, and thought it would be a good idea to establish a cooperative led by Jinzhong. Liu was excited by this idea, and in August 2005 they decided to set up a cooperative called ‘Jinli’ with the support of Jinzhong company, the Qionglai county government and local small pig farmers. It was the very first cooperative in the pig industry in China. Between 2005 and 2015, Jinli cooperative brought benefits for both Jinzhong and local pig farmers. Small pig producers were able to improve their pig production thanks to the guidance and training offered by Jinli cooperative; and Jinzhong has a more stable source of healthy pigs by contracting with local farmers. However, there are problems facing Jinzhong during its coordination with Jinli cooperative. As there are huge price fluctuations on the market in China, small pig farmers in the cooperative regularly breach their contract with the cooperative depending on changes in the market. In addition, it is impossible for Jinzhong to control the quality of the pigs from smallholder farmers completely, because Jinzhong is not able to monitor the whole production process of its member farmers. This raises potential quality risks for the company. For Mr. Liu Xiang, food security and a stable supply of high quality pork are top priorities on his agenda. The case will illustrate how Jinzhong has managed its supply chain with Jinli cooperative, describing the benefits and problems Jinli cooperative has brought in the last decade. Thus, it aims to trigger the readers help Mr. Liu Xiang to make a decision: should Jinzhong continue, side-line or even phase out the cooperative in the future?

2. Pork supply chain in China The pork supply chain includes many different chain actors, such as feed producer, pig farmer/producer, slaughterhouse, pork processing companies, various kinds of pork distributors and pork consumers (Figure 1). Breeders in the pork chain breed pigs and sell piglets. They choose breeds with the best reproductive and biological characteristics to produce good piglet breeds. The three most famous breeds in China are Changbai, Taihu and Rongchang, which are found mainly in North-eastern China, Yangzi area and Chongqing respectively. In the last 30 years, China has imported many foreign breeds, such as Duroc and Yorkshire, which are crossed with Chinese pig breeds. Feed producers are companies that produce feed products for the livestock and/or the aquaculture industry. The feed industry is approaching market saturation due to over-production in China, so competition is concentrated on price. Pig producers are farmers who fatten pigs and sell these to slaughterhouses. Slaughterhouses are companies that slaughter live pigs and cut these into meat products. In China, companies wanting to operate in this industry must obtain a slaughterhouse permit, which is difficult to acquire. The processing industry International Food and Agribusiness Management Review

416

Ji et al.

Volume 20, Issue 3, 2017

Processing industry

Feed producer

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

Breeder

Slaughter house

Producer

Retailer

Consumer

Transporter

Figure 1. A typical pork supply chain (adapted from Trienekens et al., 2009). converts the meat into end products, such as sausage and bacon. Slaughterhouses and processing industries in China’s pork chain tend to be the focal companies (playing the central role in the chain) as they collaborate with pig farmers upstream in the chain to ensure pig supply and safety and work with retailers downstream in the chain for branding and marketing. China is the world’s leading consumer of pork. In the countryside live pig transportation is still common, and is considered one of the causes for the transmission of pig diseases, due to unhygienic transport conditions and the spread of diseases from sick pigs to healthy pigs during transportation. Fresh pork transportation is usually carried out by slaughtering and processing or logistics companies in vehicles equipped with professional cooling systems. In China, there are five main types of governance structure between focal companies (processors and slaughterhouses) and farmers, which are illustrated in Figure 2. From left to right, a company may buy from the spot market where no contract is required (e.g. transactions take place at livestock trading venues); purchase pigs from brokers without contracts, who in turn buy from individual pig farms (middleman relationship); use contract farming signing production or marketing contracts with pig farmers; buy from cooperatives, and vertically integrate pig farmers to produce pigs at their own base. As the relationship types move from left to right in the figure, they become more vertically integrated, with the spot market being the loosest and vertical integration being the most integrated type. Focal companies may adopt more than one type of relationship with pig farmers simultaneously. In general, focal companies tend not to buy from the spot market due to traceability and quality issues.

3. Challenges in China’s pig industry There are several challenges in China’s pig industry, such as food safety concerns, pig/pork price fluctuations, environmental stress and fragmented pig production. Food safety is the first concern. Epidemic diseases have always been a risk for pig farms. During the past 10 years, several epidemic diseases have affected the Sichuan area1, raising public concern and causing major economic losses among pig producers. Compared 1

Epidemic diseases occurring frequently in Sichuan area are blue-ear disease and Streptococcus suis disease.

Middleman (pig broker)

Contract farming

Spot market relationship

Vertical integration

Pig production cooperative

Figure 2. The relationship continuum between pig farmers and pig slaughter/slaughtering companies in China. International Food and Agribusiness Management Review

417

Ji et al.

Volume 20, Issue 3, 2017

Third and finally, pig production has created environmental stress, which raises concerns in the central as well as local governments. Traditional pig production farms discharge faeces directly into rivers causing serious water pollution. In Zhejiang province, the government started closing down farms situated near rivers; however, this approach doesnâ&#x20AC;&#x2122;t solve the environmental problems in the long run as it simultaneously decreases pig supply. The government in Sichuan provides policy levers to incentivize large-scale farms and stop subsidizing small and medium-sized ones, which creates another problem for the small-scale pig farmers, who are faced with the decision of whether to produce something else (rearing pigs will therefore cease to be their main business) or become employees on large farms. Neither the pig producers nor the government have yet found an effective way to solve the problem of water pollution.

4. Pig industry in Sichuan and Qionglai Sichuan Province is located in the southwest of China; its pig production volume ranks first in the country, accounting for more than 15% of total pig production (Figure 4). Qionglai prefecture is a pig production base in this province, also supplying pigs to other areas besides Chengdu, which is the capital of Sichuan province. 2

1 CNY = 0.15 $USD. 15.5 15.0 14.5 CNY/kg

14.0 13.5 13.0 12.5 12.0 11.5 11.0

Jan

Feb

Mar

April

May

June

July

Aug

Sep

Oct

Nov

Dec

2014

Figure 3. Average price of fattened pigs in Sichuan Province from January to December, 2014 (1 CNY = 0.15 $USD) (adapted from Livestock Industry Development Bureau in Sichuan, 2016).

Pigs sold in Sichuan province

50,000 # pigs Ă&#x2014; 10,000

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

with large-scale pig producers, small-scale farmers suffered more due to the limited availability of capital to hedge against the risk. Second, fluctuations in live pig prices pose a serious risk for pig producers in Sichuan, and were particularly significant between 2000 and 2015 (Supplementary Figure S1). Figure 3 shows the price volatility within the year 2014. To illustrate further, in 2015, the average loss of a medium-scale farm was CNY2 300 per pig in the first half of the year, and the average net profit of a medium-scale farm was CNY 200 per pig in the second half of the year.

Pigs sold in the whole country

40,000 30,000 20,000 10,000 0

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

Figure 4. Pigs sold by the end of year in China and in Sichuan Province from 2004-2013 (adapted from China statistical yearbook, 2014). International Food and Agribusiness Management Review

418

Ji et al.

Volume 20, Issue 3, 2017

The pig production volume of Qionglai has been on the rise for the past 10 years (Figure 5). Qionglai city is well-known for its agricultural development, which includes the pig, wine and dairy industries.

In Table 1, we can see that backyard pig production increased from 2014 to 2015 in terms of both numbers of pigs produced and percentage of pigs produced in Qionglai. The main reason is that there are low entry barriers for backyard pig producers. With the big price surge in the second half of 2014, backyard farmers were highly motivated to raise more pigs. Small-scale and medium-scale pig production scaled down in 2015 compared to 2014, in number of farms, number of pigs produced and percentage of pigs produced. Farms of 100-499 heads have been built historically on the basis of backyard farms, with no standardized and modernized farm facilities. The reason is that the local government does not give financial support to small and medium-size farms anymore, but tends to encourage the development of big farms by giving them financial benefits3. Therefore, although the number of pigs produced increased in 2015 compared to 2014, the number of big farms stays the same but shows a substantial increase in average production capacity. Table 2 shows the number of pig farms of different scales in Sichuan from 2012 to 2014, and backyard farms dominate pig production in Sichuan province. 3 There are many types of benefits that large farms can get from the government, such as subsidies for enlarging their infrastructure, introducing good breeds, disposing of faeces to prevent pollution, and loan interests, etc.

Pigs sold by the end of year in Qionglai

2,000,000 # pigs

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

Pig production in Qionglai nicely represents the pig production situation in China, which is still dominated by small-scale and backyard pig producers. In the pig industry farms with a production scale smaller than 50 heads of fattened pigs per annum are called backyard farms, between 50 and 100 heads small-scale farms, between 100 and 500 heads medium-scale farms, and more than 500 heads large-scale farms.

1,500,000 1,000,000 500,000 0

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Figure 5. Pigs sold by the end of year in Qionglai from 2004-2014 (adapted from Livestock Industry Development Bureau of Qionglai, 2015). Table 1. Pig farms and production volume in Qionglai in 2014 and 2015 (adapted from Livestock Bureau of Qionglai, 2015).1 Scale

Number of farms

Number of pigs produced

Percentage of pigs produced

2014

2015

2014

2015

2014

2015

n.a. 393 651 62 n.a

354,438 134,973 181,413 79,176 750,000

416,873 64,062 112,556 106,509 700,000

47.25 18.00 24.19 10.56 100.00

59.55 9.15 16.08 15.22 100.00

Backyard (<50 heads) n.a. Small scale (50-99 heads) 2,069 Medium scale (100-499 heads) 1,038 Large scale (>500 heads) 61 Total n.a. 1

n.a. = not available International Food and Agribusiness Management Review

419

Ji et al.

Volume 20, Issue 3, 2017

Table 2. Number of pigs farms of different scales in Sichuan Province from 2012 to 2014 (adapted from China livestock statistics yearbook, 2015).

2014 2013 2012

7,373,861 224,432 7,645,726 221,391 9,433,336 287,920

57,832 55,825 64,481

11,651 10,753 12,401

3,583 3,626 4,063

715 720 810

371 370 457

259 256 290

>50,000

10,000-49,999

5,000-9,999

3,000-4,999

1000-2,999

500-999

100-499

50-99

No. farms per year

<50

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

No. of animals per farm

4 4 6

5. Background of Jinzhong Jinzhong was founded in 1994 as a small pig slaughtering and processing factory (named Jinli originally) by Mr. Chen Naizhong, and has been a family-owned company and developing fast over the past 20 years. Mr. Chen reshaped the Jinli slaughtering factory into an incorporated company and changed the group name to Jinzhong in 1999. The incumbent president of Jinzhong is Mr. Liu Xiang, a son-in-law of Mr. Chen. Jinzhong is located in the industrial park of Qionglai prefecture, where many branded food processing companies have established factories, such as Yili Group (one of the Chinese dairy industry giants), Kinliufook or Jinliufu Co. Ltd. (China’s top 10 white wine brand), Wenjun (best tea brand in Sichuan Province) among others. Jinzhong now owns a slaughtering line imported from Meat Processing System (MPS)4, with a slaughter capacity of 2,000,000 heads annually. It has more than 20,000 tons of frozen pork reserved constantly at a national level for urgent and future use. There are several series/levels of products, which are sold in different channels (Supplementary Table S1 and Supplementary Figure S2). Table 3 shows the key development stages for Jinzhong. The MPS slaughtering and processing line is considered one of the key competences of Jinzhong as it is a modern and advanced processing technology system providing temperature control5 throughout the whole slaughtering process. Jinzhong is a leading company in the pork sector at a national level6 and creates a significant impact on the local economy and sustainable development in terms of employment and poverty alleviation. Its businesses include feed production, breeding, and pig production, pig slaughtering and processing, cold chain logistics (Jinzhong has its own logistics truck fleet) and retailing. Jinzhong’s pig supply comes from four channels in 2015 (Figure 6). More recently, in 2014, Jinzhong started collaborating with ItoYokato supermarket to sell its premium quality pork products. It also aims to make increasing use of Jinzhong’s own distribution channels (Jinzhong own branded stores) to distribute high-end pork products in the future. In the next five years starting from 2015, Jinzhong plans to further integrate its supply chain towards downstream.

MPS Group is a Dutch company that produces and sells meat slaughtering and processing lines (machinery) for ‘cold-fresh meat (0-4 °C fresh meat)’ internationally. It is a leading company in this industry which offers sophisticated meat processing technology that ensures meat safety. 5 Temperature control is extremely important for meat safety because when a pig is freshly slaughtered, its body temperature could reach 40 °C in the slaughter line. If it is not sent to places with a cooling system, microorganisms may reproduce fast, which negatively affects meat safety and quality. 6 Jinzhong is included in China’s agricultural industry leading companies list.

International Food and Agribusiness Management Review

420

Ji et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

Table 3. Stages of development of Jinzhong Food Co. Ltd.1,2 Period

Business development milestones

1994 1999-2004

Started as slaughtering and processing factory (named Jinli). Expanded business of Jinli Co. Ltd. into pig production, established a 1000-head pig (reproducible sows) farm and expanded the slaughter line capacity to 1,000,000 heads. Took the initiative to sell fresh/chilled pork and processed pork products to brand stores (either franchised or self-owned) in Chengdu area. Group name changed to Jinzhong. Jinli pig production cooperative was established. Self-owned production base is scaled up to 100,000 heads fattened per year. MPS slaughtering line was introduced from the Netherlands. Annual slaughtering capacity reaches 2,000,000 heads. Started to invest more in meat processing and diversification of processed meat products. Jinlai Feed Co. Ltd. was founded as a subsidiary company of Jinzhong group and has 100,000 tons of feed production capacity annually. Number of brand stores in Chengdu grew up to 200 in 2012. Penetrated the high-end pork market. Has taken advantage of Chengdu government’s project to establish ‘Tianfu (paradise) modern agricultural area’. By investing in a new processed food production line and working with ItoYokato, a Japanese high end retailer. Jinzhong vertically integrated the pork supply chain more deeply.

2005-2009

2010-2012 2012-2016

MPS = meat processing system. information can be found in Supplementary information S1.

2 Additional

10% Self-owned production base Middleman and spot market 40%

20%

Pig production company Jinli cooperative

30%

Figure 6. Percentage of pig supply from different channels in 2015 (Jinzhong Food Co. Ltd.: http://group. jzfoods.cn).

6. Jinli pig production cooperative Motives and resources to initiate a company-led cooperative Before 2005, farmers sold fattened pigs to Jinzhong based on live spot market prices and smallholder farmers were the dominant type of suppliers of Jinzhong. During that time, Jinzhong had to buy from all sources, i.e. backyard farmers, middlemen (who tend to buy from backyard farmers), small and medium as well as large-scale pig farms (Table 4). Contagious Streptococcus suis infection7 prevailed in Ziyang, Jianyang and Lezhi areas of Sichuan Province in 2005, and it raised major food safety concerns in government, industry and among the public at large. Even rural consumers switched to branded pork products sold in supermarkets rather than the wet market to ensure product safety and quality. Supermarkets have their own quality requirements

Streptococcus suis is a bacterial infection that could affect pigs’ meninges and respiratory passage and thus causes septicemia and quick death.

International Food and Agribusiness Management Review

421

Ji et al.

Volume 20, Issue 3, 2017

Table 4. Percentage of different scaled farmer members in Jinli in 2005 and 2015.1 Year

Large scale farmers members (>500 heads) (%)

Small and medium scale farmer Backyard scale farmer members members (50-499 heads) (%) (<50 heads) (%)

2005 2015

10 20

30 50

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

1 Data

60 30

supplied by Ms. Tian Bishan, Executive Manager of Jinli cooperative.

as well as cooling systems for pork products that are sold in their stores, while in wet markets, there is no quality standard for pork products and not all the sellers have cooling systems in their stores. To deal with the vulnerable market, Jinzhong had to think not only about how to stabilize its pig supply, but also how to ensure sourcing of healthy and safe pigs. However, since it started as a pig slaughterhouse and pork processor, Jinzhong had little experience in breeding and fattening pigs. Compared with professional pig production companies (e.g. Wenshi8), Jinzhong does not have the advantage of producing pigs on a large scale. In 2005, Mr. Gou Mingjun, Deputy Director of the livestock bureau in Qionglai discussed the idea of establishing a company-led cooperative with Jinzhong’s general manager at the time. It didn’t take long for the bureau, Jinzhong and farmers to decide to jointly invest CNY 1,000,000 to start the cooperative. Qionglai government invested CNY 400,000, Jinli Group CNY 400,000 and pig farm members CNY 200,000 based on their production scale (CNY 1 per member for small holders). One million CNY was raised to start an insurance foundation which was primarily used to compensate farmers for their losses due to epidemic diseases. All the operational costs incurred by the cooperative have been covered by Jinzhong. As there was no cooperative law and regulations set up at that time (laws and regulations on cooperatives in China were set up in 2007), Jinli pig production cooperative was registered as a civil society. Including the director of Jinli Cooperative, who is also the general manager of Jinzhong, Jinli cooperative has six staff members, who are responsible for managing cooperative members, collecting pig industry information, offering and organizing pig production training courses, financial services, and coordinating logistics and the transportation of live pigs. Jinli cooperative coordinates with different departments of Jinzhong company, including Jinlai feed, Jinzhong processing and Jinzhong logistics. Members of Jinli cooperative purchase feed from Jinlai feed, and sell pigs to Jinzhong processing; logistics/transportation is provided by Jinzhong logistics (Supplementary Figure S3). The costs incurred for running Jinli cooperative are the transaction costs in the form of staff time involved, including ex ante cost of farmer/supplier selection, contracting with farmers on behalf of Jinzhong and ex post cost of training provided to farmers, contract fulfilment by the farmers, resolution of conflicts between Jinzhong and farmers and coordination between farmer members and internal departments. The governance structure of Jinli Cooperative and relationship with members Jinli’s cooperative membership has decreased from 6,000 (in 2007) to 4,000 (in 2012) to 2,000 (in 2015). In the beginning, Jinli cooperative worked with farm members of different scales but predominantly sourced from backyard farmers. Later, it was found that many backyard farmers behaved opportunistically (e.g. didn’t raise pigs as required; didn’t sell their pigs to Jinli cooperative when they had been fattened – side selling), so Jinli cooperative moved to working with medium- and larger-scale pig farmers. Farmer members are 8

Wenshi Food Co. Ltd. is a listed company on the Shenzhen stock market (stock code 300498). Its businesses include poultry farming, pig farming, duck farming and cattle farming as well as feed and veterinary medicine production. It raises pigs in more than 23 provinces in China with a mode called ‘collaborating with local farmers’. It uses its technology and services to help local farmers construct pig farming facilities, train them in pig-raising skills and purchases the pigs of the farmers to sell to slaughter companies like Jinzhong or big pig brokers. Its advantage lies in the professional rearing of livestock animals.

International Food and Agribusiness Management Review

422

Ji et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

distributed in five towns in Qionglai and surrounding areas. Table 4 shows the percentage of different types of farms from which Jinli cooperative sourced live pigs in 2005 and 2015. It can be seen that compared to 2005, the number of large farms doubled (from 10 to 20%); the number of backyard farms halved (from 60 to 30%); and the number of small and medium-sized farms increased from 30 to 50% in 2015. For cooperative governance, Jinli cooperative set up a member committee, a board of directors, and a board of supervisors. Decisions are not made according to the ‘one member one vote’ principle, but rather concentrated in Jinzhong’s senior management. Farmer members do not have real democratic decisionmaking power. They have a legal right to exit Jinli cooperative at any time if they notify the cooperative in advance. However, Jinli cooperative also has the right to stop farmers’ membership if they have built up a bad credit record during their collaboration9. The relationship between Jinli cooperative and pig farmers is a double-edged sword. On the one hand, both the company and farmers achieve the expected benefits; on the other hand, there are problems such as breach of contract. Although limiting its cooperation with backyard farmers, Jinli cooperative doesn’t stop contracting with small farmers, but closely monitors their behavior and develops the small farmers to meet Jinzhong’s requirements, by communicating with them, checking their pigs produced by the end of the year and monitoring changes in the number of pigs. When small-scale pig production becomes medium- or even large-scale, Jinli will change its transactional approach aiming for long-term mutual goals. Therefore, compared with contract farming, buying through the Jinli cooperative means that both the company and the farmers are less concerned about hold-up problems, and transaction costs (monitoring cost) are lower for both sides. In the Jinli cooperative, members are allowed to purchase feed on credit. Farmers purchase Jinlai feed without paying cash at the moment of transaction, and when they sell pigs to Jinli cooperative, Jinli deducts the feed cost from the payment for the pigs they purchase. This means that if Jinzhong does not fulfil the contract (i.e. does not buy farmers’ pigs in poor market conditions), the farmers do not have to pay back the feed cost and the company will suffer a loss. On the other hand, farmer members need to pay CNY 4.0 per head to the cooperative as a guarantee. If they eventually sell pigs to Jinzhong as stipulated in the contract, they will get this money back; otherwise this is a sunk cost for them. Jinli cooperative’s roles Mr. Guo Hongxiang, Manager of Public Relations Department of Jinzhong, describes the role of Jinli cooperative as follows: ‘Jinzhong endows product quality and safety to its brand value. This is why customers become loyal to our brand. Jinli Cooperative is one of the means to ensure product safety and quality from the origin’. Ms. Tian Bishan, Executive Manager of Jinli Cooperative provided a similar view: ‘Jinli Cooperative serves as a bridge between the company and farmers. It helps farmers in Qionglai region to make more profits from pig production and also trains farmers to produce pigs in a safer and more scientific way’. Jinli cooperative does not aim to make profits from the transactions with farmers. Instead, it serves as a bridge that connects Jinzhong and pig farmers and is embedded in Jinzhong to coordinate between various departments of Jinzhong and farms (Supplementary Figure S3). When purchasing pigs from farmers, Jinli pays a higher price than the market price. 9

According to the definition of International Cooperative Alliance, Jinli cooperative is not a ‘real’ cooperative as it does not follow the ‘one member one vote’ principle.

International Food and Agribusiness Management Review

423

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

Ji et al.

Volume 20, Issue 3, 2017

In summary, Jinli cooperative assumes multiple roles benefiting both Jinzhong company and member farmers: ■■ Provides information and training to farmers including best practices in pig production like disease prevention and cure. It also trains farmers to raise pigs in a modern way using information technologies (e.g. ear tags to record pig production information) applicable to pig production. ■■ Acts as an information channel between Jinzhong and farmers and between farmers. For example, through Jinli, farmers’ problems and requests can be effectively passed on to Jinzhong. ■■ Plays an important role in building a long-term relationship between Jinzhong and farmers and makes both parties commit to the relationship to achieve mutual benefits. ■■ Offers vaccination and veterinarian services at a low price. Jinli cooperative has a vet station, which provides free professional veterinary consulting services to member farmers and only charges farmers for the drugs or vaccinations they use. ■■ Coordinates and provides free transportation to the farmers through Jinzhong logistics. ■■ Provides discounted Jinlai pig feed, which is well formulated and cheaper than feeds of similar quality on the market. Farmers do however have the choice to use other branded feed (such as Tongwei). In order to encourage pig farmers to use Jinlai feeds, Jinli cooperative recently allowed medium-scale pig producers (100-500 heads) with no membership (non-formal member farmers) to integrate with existing members to enjoy the benefits of Jinlai feed (Figure 7). ■■ Is responsible for negotiating contract prices with Jinzhong on behalf of its members to get a guaranteed price and higher profits for members. Based on the market price, a fixed protective price has been contracted between member farmers and Jinli. When the market price10 is lower than the contracted price, then the contracted price is applied; otherwise the market price is applied. In this scenario, Jinli purchases pigs from farmers at CNY 0.1-0.5 per kg higher than the market price. In this way, farmers can make more profits than selling to the spot market. Jinli cooperative gives member farmers returns in two ways. The first is to purchase farmers’ pigs with a protective price; the other is to give a bonus to farmers of CNY 5 per head sold to the cooperative. Farmers who raise more than 1000 heads annually will get an extra CNY 0.5 per head and farmers who raise more than 2,000 heads will get an extra CNY 1 per head. Jinzhong pays all the returns given to the farmers through Jinli cooperative.

Market price refers to the live pig average price of Qionglai at the transaction time point between farmers and the cooperative. The cooperative will let the farmers know the average price of the market published by livestock bureau every day.

Jinli

Member A (big pig producer)

Member B (big pig producer)

Smaller pig producers (non-member farmers)

Member N (big pig producer)

………

Smaller pig producers (non-member farmers)

Figure 7. Jinli’s federation cooperative model.

International Food and Agribusiness Management Review

424

Ji et al.

Volume 20, Issue 3, 2017

7. Benefits for Jinzhong brought by Jinli cooperative

From 2005 to 2015, the percentage of Jinzhong’s pig supply through Jinli cooperative increased from 10 to 40% (Figure 8). In the last ten years, Jinli has built stable relationships with its members, who are used to working with Jinli and selling pigs to Jinzhong, thereby ensuring Jinzhong’s pig supply. Compared with 2005, Jinzhong purchased fewer live pigs through the free market, which means Jinzhong is affected less by supply uncertainty and suppliers’ opportunistic behavior. At the same time, Jinzhong has increased its own pig production base scale (from 10,000 heads in 2004 to 100,000 heads in 2014) and has assured food safety through better quality control. This again contributes to the stabilization of pig supply. In addition, with its own production base experience and the cooperative production volume on the rise, Jinzhong is more knowledgeable about pig production processes and is better able to control food safety. There have never been any major food safety incidents in the Jinzhong pork supply chain. With consistent quality control and better brand management, Jinzhong has successfully linked premium quality and safe products with its brand image, which leads to better customer loyalty.

8. Jinzhong looks forward As commented by Ms. Tian Bishan, who is the leader of Jinli cooperative: ‘though Jinli cooperative has helped to stabilize pig supply, farmers are not obligated to sell pigs to Jinzhong. Pig purchasing prices are very unstable. When the market for live pigs improves (for example, in the second half of 2015), farmers tend to sell pigs to pig brokers because they may offer a higher price than Jinzhong. At the moment, it’s even easier for Jinzhong to purchase pigs from pig production companies such as Wenshi, since these kinds of transactions are simple and hassle-free. So, we are not sure whether we should work out a new mechanism to deal with the local farmers as members of Jinli cooperative or simply move to sourcing pigs more from other channels’. It is therefore a key decision for the senior management of Jinzhong to make on whether to keep the Jinli cooperative or move to other types of supplier governance.

45 40 35 30 %

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

Stabilizes pig supply and ensures pig safety

25 20 15 10 5 0

2005

2006

2007

2008

2009

2010

2011

2012

2013

Figure 8. Pig supply from Jinli cooperative to Jizhong (2005 to 2015).

International Food and Agribusiness Management Review

425

2014

2015

Ji et al.

Volume 20, Issue 3, 2017

Supplementary material

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0101 - Tuesday, May 09, 2017 8:37:50 PM - IP Address:24.21.169.207

Supplementary material can be found online at https://doi.org/10.22434/IFAMR2016.0101. Figure S1. Average live pig prices in Sichuan Province from 2000 to 2015. Figure S2. Jinzhong Product series and example pictures. Figure S3. Jinli cooperativeâ&#x20AC;&#x2122;s functions and its relationship with other departments at Jinzhong. Table S1. Jinzhong product ranges and examples, and their distribution channels. Supplementary information S1. Financial performance of Jinzhong. Teaching note

Acknowledgements The authors acknowledge the funding from Natural Science Foundation of China Projects (No. 71403243, 71333011, 71273136, 71403077), and Philosophy and Social Sciences Planning Project of Zhejiang Province (15NDJC100YB) to support the case study. We also appreciate the financial support of Newton Caldas Institutional Link with Colombia funded by British Council and Newton Mobility Grant with CASS, China. The comments from anonymous reviewers are also high appreciated. We also feel highly appreciated of the assistance provided by managerial board of Jinzhong and by staff of Qionglai Livestock Bureau to complete our case study.

References China livestock statistics yearbook. 2015. China Agriculture Press. Beijing, China. China statistical yearbook. 2014. Complied by National Bureau of Statistics of China. China Statistics Press. Beijing, China. Livestock Department of the Ministry of Agriculture. 2016. Livestock association in China. Available at: www.caaa.cn. Livestock Industry Development Bureau in Sichuan. 2016. Available at: http://www.scagri.gov.cn. Livestock Industry Development Bureau of Qionglai. 2015. Available at: http://tinyurl.com/j5hlf8z. Trienekens, J., B. Petersen, N. Wognum and D. Brinkmann. 2009. European pork chains: diversity and quality challenges in consumer oriented production and distribution. Wageningen Academic Publishers, Wageningen, the Netherlands.

International Food and Agribusiness Management Review

426

OPEN ACCESS International Food and Agribusiness Management Review Volume 20 Issue 3, 2017; DOI: 10.22434/IFAMR2016.0167

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

Received: 19 October 2016 / Accepted: 3 January 2017

Management lessons learned in supply chain development: the experience of PICS bags in West and Central Africa INDUSTRY SPEAKS Theodore Nouhoheflin a, Jeanne Y. Coulibalyb, Stephen D’Alessandroc, Codjo C. Aitchédjid, Maiyaki Damisae, Dieudonné Baributsaf, and James Lowenberg-DeBoerg aConsultant, bAgricultural

Purdue University, 615 West State Street, 47907 West Lafayette, IN, USA

Economist, World Agroforestry Center, P.O. Box 30677-00100, UN Avenue Gigiri, Nairobi, Kenya cIndependent dProgram

eAgricultural

Consultant, 4 rue Stanislas Torrents, 13006 Marseille, France

Manager, CDA-WEKE, BK Sarl, Zopah, BP 1971, Abomey-Calavi, Benin

Economist, Institute for Agricultural Research, Ahmadu Bello University, Zaria, Nigeria fAssociate

Professor, Entomology Department, Purdue University, 901 West State Street, 47907 West Lafayette, IN, USA

gProfessor

of International Entrepreneurship, Global Engineering Programs, Purdue University, 516 Northwestern Ave, 47907 West Lafayette, IN, USA

Abstract In order to facilitate the transformation of African agriculture, Africa’s smallholders will require more and better access to input markets that deliver context-specific and well adapted innovations. This article summarizes the management lessons learned in developing supply chains for Purdue Improved Crop Storage (PICS) bags in West and Central Africa. PICS bags are hermetic grain storage bags that are leading a revolution in the post-harvest handling of grain on smallholder farmers in Africa. The bags have been very popular with farmers, but the major challenge has been developing supply chains to make those bags available for purchase at the village level. This article: (1) describes the PICS supply chain in West and Central Africa; (2) identifies constraints (i.e. financial, structural, social) that impede the development of the PICS supply chain; and (3) analyzes strategies used by businesses to expand distribution channels and increase bag sales. Keywords: PICS bags, supply chain development, cowpea, West and Central Africa, input distribution, hermetic storage, storage innovations JEL code: O13, Q12 Corresponding author: tnouho@gmail.com

427

Nouhoheflin et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

1. Introduction As smallholder farms in Africa become more market oriented, they will inevitably use more technology, buy more inputs and drive the need to develop better supply chains to supply the technology and related inputs. This article summarizes the management lessons learned in developing the supply chains for Purdue Improved Crop Storage (PICS) bags in West and Central Africa. PICS bags are hermetic grain storage bags that are leading a revolution in the post-harvest handling of grain on smallholder farmers in Africa. In general 20 to 30% of the grain in Africa is lost between harvest and the consumer. With PICS bags farmers and the grain trade can reduce those losses to almost zero. The bags have been very popular with farmers, but the major challenge has been developing supply chains to make those bags available for purchase in rural areas. This article: (1) describes the status of the overall supply chain in countries covered by the PICS1 project including who is manufacturing the PICS bags and why, as well as the organization of the distribution network; (2) identifies constraints (i.e. financial, structural, social) that impede the development of the PICS supply chain; and (3) analyzes strategies used by local businesses (manufacturers, distributors, vendors, etc.) in expanding distribution channels and increasing bag sales. PICS bags are a simple, low-cost (about US$Â 2-3) triple-layer plastic bag that allows small-scale farmers to protect their harvested grain from insect damage without insecticides (Murdock and Baoua, 2014). The PICS bag consists of an outer layer of ordinary woven polypropylene and two inner liners of high density polyethylene 80 microns thick. Better storage improves household food security and it gives farmers more flexibility in marketing leading to higher incomes. Commercialization of the bags was launched with funding from the Bill and Melinda and Gates Foundation (BMGF) in the PICS1 project from 2007-2014. During this period the technology was disseminated to millions of cowpea farmers in 10 countries in West and Central Africa including Ghana, Nigeria, Burkina Faso, Niger, Senegal, Mali, Benin, Togo, Cameroon and Chad. Subsequently the PICS2 project (2011-2014) showed that the technology could be used to store any dry grain and the PICS3 project (https://www.picsnetwork.org) is extending use of hermetic storage in maize growing areas of Nigeria, Burkina Faso and Ghana, and into East Africa in the 2014 to 2019 period. To scale up PICS bag adoption, Purdue University worked with many partners, including local and international nongovernmental organizations (NGOs), national agricultural research systems, national extension organizations, farmer-based organizations, womenâ&#x20AC;&#x2122;s groups and for-profit businesses (i.e., manufacturers, distributors and agricultural input retailers). The extension effort involved direct contact with millions of farmers in almost 45,000 villages. In developing countries plastic manufacturers produced and local entrepreneurs sold almost five million PICS bags from 2007-2015. The origin of the PICS technology is in the Bean/Cowpea Collaborative Research Support Program (CRSP) which was funded by the United States Agency for International Development. The Bean/Cowpea CRSP identified cowpea storage as the key constraint to greater productivity in the 1980s. In village meetings farmers had indicated that they could produce more cowpea, but did not because harvest time prices were too low and storage losses were too high. Starting in 1987, Purdue entomology Professor Larry Murdock and colleagues developed a portfolio of non-chemical storage technologies. Adoption surveys showed that the most widely accepted improved storage technology was the triple layer storage bag. Grain coming from the field is commonly infested by low numbers of pest insects whose numbers increase exponentially during postharvest storage. When such grain is sealed in an airtight container, the insects quickly consume the oxygen in the bag, cease feeding and stop reproducing. Depending on the species insects either die or become dormant. In either case they cause no more grain damage. Subsequent research has shown that PICS bags can be used cost-effectively to store any dry grain, reducing storage damage to negligible levels. The PICS regional adoption studies in 2010 and 2012 showed that about 46% of respondents used some type of hermetic storage and about 44% of the quantity of cowpea stored on farms was in hermetic containers (Moussa et al., 2014). This is substantially more than the 30% of cowpea grain in hermetic storage reported in 2003-2004 survey (Moussa et al., 2011). Given the continued growth in PICS bag sales, the persistent

International Food and Agribusiness Management Review

428

Nouhoheflin et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

use of metal drums in some areas and the increased use of plastic jugs, it is likely that cowpea grain stored on farms in the region exceeded the 50% hermetic benchmark in 2013. In West and Central Africa PICS bags are used by more respondents and for a higher percentage of cowpea than any other hermetic container. Alternative hermetic containers include metal drums, plastic jugs and ‘double bags’ (i.e. two layer bags, with a single liner of low density polyethylene). PICS adoption surveys show that PICS bags are used by 18% of farmers in the region and for about 20% of cowpea quantity. Some insecticide use still occurs with hermetic containers of all kinds – as a sort of insurance – but insecticide use is down since 2003-2004. When PICS bags are available they are used by women as well as by men. A survey of adult women in randomly selected villages in Burkina Faso, Niger, and Nigeria, has shown that over half of the women use some kind of hermetic storage for their cowpea (Ibro et al., 2014), but PICS bags are the most common hermetic storage in the region. In these surveys about 28% of rural women in these three countries use PICS bags. PICS has resulted in greater cash income for farmers, new businesses, new jobs, less insecticide exposure, and increased cowpea availability for families, schools and orphanages. Hermetically stored cowpea commands a 10 to 15% premium price in some markets because consumers know it is less likely to be contaminated by insecticide residues. Surveys indicate that most PICS bags are used at least three years before being repurposed for non-hermetic uses. The survey respondents who store in PICS bags gain on average $ 26.58/bag at sale (Moussa et al., 2014). Given the sales of 1,277,470 PICS bags in the 2010-2012 period, a conservative estimate of the extra cash flow for African farmers and households attributable to PICS is $ 33,955,153 for the 2012-13 storage year. From a technology transfer perspective, the PICS project differed from many other agricultural development projects in the sense that from the beginning it treated African farmers as customers, not as charity cases (Sonka et al., 2016). Other than a few bags used in village demonstrations, almost all of the 5 million PICS bags that were manufactured were sold through commercial channels. When PICS started in 2007, the team thought that finding African manufacturers for the bags and implementing demonstrations in thousands of villages would be the major constraints. They quickly learned that African manufacturers were eager to find new products in high demand by their customer base and national extension services and NGOs were quite effective at implementing farmer training programs. The real bottleneck was development of supply chains down to the village level. An early PICS adoption study showed that farmers were reluctant to travel more than about 7 km to buy PICS bags (Moussa et al., 2010). This study identifies how the PICS team attempted to meet the challenge of developing a supply chain that extended to within 7 km of every cowpea farmer in West and Central Africa and the lessons to be learned from that experience.

2. Methods Given the nature of the PICS experience and the approach used to implement and develop the PICS supply chain, a research case study approach was adopted as the most appropriate analytical framework for the present assessment. This approach drew to some extent from Prahalad’s (2004) seminal work exploring strategies to deliver innovative goods and services to economically marginalized, yet vast consumer markets. The research case study approach helps document what is happening based on direct input from stakeholders. It is often used to explore, describe and explain real life experiences where traditional statistical methods may be less appropriate. The case study methodology used in this research draws on social science and business case study methods and is structured along the lines defined by Yin (2003). This research was carried out in three major phases: pre-field preparation, data collection, and data analysis and reporting. The first phase or pre-field preparation consisted of reviewing the existing literature and project materials (trip reports), interviewing in-country PICS team members, developing the interview guides and other data collection tools and identifying key informants. The second phase of the study consisted of collecting data in the ten PICS1 project countries through key informant interviews with in-country PICS International Food and Agribusiness Management Review

429

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

Nouhoheflin et al.

Volume 20, Issue 3, 2017

project staff and supply chain participants including bag manufacturers, vendors (i.e. wholesalers, semiwholesalers, and retailers), partner NGOs and research institutes, and PICS business consultants. Data collection focused on capturing history and current status of the PICS supply chain in each country; the key factors that influenced private sector actors’ willingness to invest in building up the supply chain; the institutional environment that impacted business and marketing strategies; and ‘market failures’ that needed to be addressed in launching the supply chain. Other information collected was related to factors that hinder the sustainability of the supply chain, and strategies used by the project team at Purdue and partners in expanding the distribution system and in increasing bag sales. The third phase of the study consisted of data analysis and reporting. A more detailed description of the methodology and country-by-country results are available in Coulibaly et al. (2012).

3. Results The PICS project began in 2007 with the launching of activities in Burkina Faso and Niger. The experiences and lessons learned from these countries were used to refine the intervention and management strategies for the next countries. However, the approach remained the same across countries and consisted of three major steps including: (1) the identification of in-country partners including agricultural research institutions and national and international NGOs; (2) the selection of the manufacturer and the national distributors to manufacture and to distribute the PICS bags in the country; and (3) the recruitment of a business consultant to help distributors build up the distribution network and to linkup bags buyers to distributors to allow an efficient distribution of the bags. In each country, the first year of the PICS1 project was the promotional year used to build up public awareness for the bags. The awareness building consisted of stimulating the demand for the bags through information campaigns using radio advertisements, posters, and village demonstrations. During the awareness building campaigns, the PICS in-country team selected five to six volunteer farmers in each village who were willing to store their cowpea in PICS bags. The volunteers were not allowed to remove the stored cowpea until an agreed date of opening the bags, typically within three to six months. On an agreed date (known as the ‘open the bags ceremony’) the villagers were assembled to open the stored cowpea. During the promotional year, other farmers who wanted to get bags were able to buy them at a price fixed by negotiation between the PICS project and manufacturers. Year two was characterized by the development of the supply chain and the scaling up of the project. The supply chain actors including bag manufacturers, national distributors, wholesalers, retailers, and volume buyers coordinated the mass production, procurement, and distribution of the bags to end users (e.g. farmers, grains and legume traders, farmers’ organizations). Four groups of constraints were identified across the supply chain: market constraints, logistics constraints, enabling environment, and social constraints. Constraints were prioritized according to their relative impact on one or more actors in the supply chain. Market constraints ■■

■■

Poor rainfall observed across countries in 2011 significantly reduced cowpea production and reduced demand for PICS bags. In 2011 the PICS technology was on a path toward adoption in almost all the project countries. A good rainfall in 2011 would have shown the demand for PICS bags since farmers were convinced of the effectiveness of the bags. PICS bags distributors had also projected high demand for that year. The excessive inventory of bags observed with distributors during this research is mainly explained by the poor rainfall. Better weather and crop forecasting by national and regional institutions could have helped improve the PICS bags supply chain management. Poor access to better and timely information as well as information sharing among bags distributors was another important constraint affecting the development of the supply chain. The lack of access to timely information for farmers such as where the bags were, and who were the closest distributors influenced the demand for PICS bags. On the vendors’ side, poor access to timely information International Food and Agribusiness Management Review

430

Nouhoheflin et al.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

■■

Volume 20, Issue 3, 2017

handicapped profit-maximization and efficient allocation of resources especially the number of bags to procure from the immediate upstream actor in the chain. The direct consequence for the bags distribution was the localized disruption of bag availability, especially with downstream actors during the harvest, as was the case with FASO KABA, a PICS distributor in Mali, and with the regional distributors in Ghana. The initial focus of PICS market development on cowpea limited demand because it made the bags a niche product. Given the limited volume of cowpea produced and stored by each household, the number of bags needed per household often is only one or two. Marketing the bags for other grains and legumes from the beginning could have increased sales in early years. PICS2 showed the bags can be used for storing all types of dry grain and subsequent initiatives helped increase demand for the bags. The variability and high price of raw materials increased manufacturing costs – A high proportion of the PICS bag cost depends upon petroleum prices. This constraint was mentioned by PICS bag manufacturers in Mali, Ghana and Burkina Faso. Even though the manufacturers strived to keep prices constant to make the bags affordable to resources constrained farmers, this variation in petrobased raw materials prices is a potential threat for the development and sustainability of the chain. Low margins at the retail level along with the slow turnover of bag inventory raised risks, increased capital costs and discouraged downstream investments (Figure 1 and 2). Vendors reported that average retail margins were modest, especially given retailers’ inability to take advantage of efficiencies of scale and their high marketing cost they incur per unit sold. Evidence suggests that limited margins, particularly downstream, were influenced by price fixing programs implemented by project partners during the early years of the PICS project. To facilitate widespread adoption of the bags among cash constrained farmers, the retail selling price of the bags was set and advertised on radios and TVs. This reference price was used as a benchmark in subsequent years during trainings in villages and markets. Retailers from isolated areas were harmed because of their high transport costs and disproportionate low share of the revenue. The slow rate of sales and inventory turnover relative to other products increased the risks and costs of capital across the supply chain. Considering the slow turnover of PICS bags, private actors and agro dealers were less inclined to invest their own money in PICS business. Figure 3 illustrates a rough analysis comparing different interest rates against three illustrative gross margins of PICS bags at the retailer level. A loan amount of 50,000 FCFA (or $ 103.74) and a 6-month linear repayment period for the procurement of approximately 45 bags at a retail price of $ 2.28/bag were used as key assumptions for the simulation. The analysis also assumed the financing of loan origination costs. The simulation indicates that at an average gross margin of $ 0.21/bag, a retailer can only make a profit after full loan repayment at rates roughly 9% and below. All assumptions

100%

80%

60%

40%

Retailer margin Semi-wholesaler margin Wholesaler margin Transport/handling Manufacturer margin Taxes Other cost Electricity Labor

20% Raw material 0%

Figure 1. Supply chain cost structure. International Food and Agribusiness Management Review

431

Nouhoheflin et al.

Volume 20, Issue 3, 2017

Costs $0.00

Margin

$ 1.00

$ 2.00

$ 3.00

Retailer

$ 2.45

Semi-wholesaler

$ 2.25

Transportation

$ 2.15

Nat’l Distributor

$ 2.05 $ 1.90

Manufacture (Emballage Miankala)

$ 1.80 raw materials 16%

plastic liners 51%

elec. 4%

VAT 16%

margin 10% labour 3%

Figure 2. Supply chain margin/cost structure (per 100 kg bag).

US$ 0.10

US$ 0.21

US$ 0.41

10 5 US$

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

Transportation/handling

0 -5

12% 24%

18%

-10 -15 -20 -25

6-month interest rate

Figure 3. Gross margin vs interest rate (measured in six-month intervals).

■■

being equal, if it were assumed that the retailer were able to turn over the capital once within the first 2 months, then at the average $ 0.21/bag gross margin the retailer’s ‘break even’ point would be roughly 16%. Given the current status of the supply chain and prevailing borrowing rates, formal credit is not an option for most PICS bag vendors. The upstream payment defaults soured relationships with manufacturers and with the PICS project. Failure by some PICS distributors to fully pay manufacturing costs and/or repay loans and advances severely disrupted the growth and development of the supply chain in Burkina Faso, Ghana, Benin, and Togo. In Burkina Faso, the agreement between PICS project and the national distributor ended in 2010 because of the non-reimbursement of financing that the project provided to help him with his first order of PICS bags. Because of this unpaid debt, he was not able to order the bags in 2011 and there was an impasse in Burkina Faso’s PICS market. With increasing demand for the PICS bags and weak supply of official PICS bags in the market, the PICS manufacturer in Burkina Faso, FASOPLAST, decided to produce generic triple bags branded Sosso Boro. These bags were labeled with the FASOPLAST logos and the logo of the Burkinabé national agriculture research institute. The texture of the Sosso Boro inner liner was slightly stiffer than that of the PICS bags, but the thickness was equal (80 microns) to that of the PICS bags. The FASOPLAST bags were sold at a price lower than the PICS bags by $ 0.10/unit. But FASOPLAST bags sold poorly because many customers noticed the lack of a PICS logo and considered them ‘fake’ bags, which could not be trusted to store cowpea safely. To overcome this situation, a new distributor was identified to sell PICS bags in Burkina Faso in 2011.

International Food and Agribusiness Management Review

432

Nouhoheflin et al.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

■■

Volume 20, Issue 3, 2017

Poorly adapted vendor networks (e.g. grain traders, farmers’ based organizations) also hindered the development of the supply chain. This was the case of the farmer’s association in Senegal and the Establishment Gansou (ETS Gansou) in Benin. Evidence suggests that the Senegalese farmers’ based organization was poorly adapted to an effective and efficient distribution of PICS bags. This was largely because it has a socially-driven organizational mandate that allowed insufficient incentives to vendors. Most of the margins from bag sales were captured by the farmer organization and not by the individual vendors engaged in PICS bag marketing. In Benin the ETS Gansou network was designed to buy grain from aggregators at regional or district levels and supply to urban markets. The flow of products and services was in the opposite direction to that required for PICS bags distribution. In addition, there was a conflict of interest in Benin. ETS Gansou’s business model was buying grain cheaply at harvest, storing it and selling it later at a substantially higher price. ETS Gansou was not motivated to sell a product which allowed farmers to store grain on-farm and capture much of that seasonal price change. Initial costs of investment in PICS technology vs those of more conventional storage methods discourages uptake among cowpea traders. Most volume cowpea buyers preferred relying on their traditional method of cowpea storage in woven sacks fumigated with pesticides like Phostoxin to control postharvest pests. This is due to what they perceive as comparatively high up-front costs of PICS technology. In addition to this initial investment, the extra labor cost associated with hermetic bagging of the cowpea discouraged trader use of PICS bags. On-farm PICS bags are tied with string or rope. Farmers usually have only a few bags and so time devoted to tying bags is not onerous. Traders often have thousands of bags, are accustomed to quickly sewing bags (usually with treadle sewing machines) and find tying bags to be very labor intensive. PICS bags cannot be sewn because sewing would puncture the liner and it would no longer be air tight.

Logistics constraints PICS bags vendors reported logistics constraints which caused significant delays in the delivery of PICS bags to different outlets. The delays were often the consequence of a poor organizational and inventory management capacity. PICS experience supports the argument by Kelly et al. (2006) that donors should invest in upgrading the legal and market regulatory institutions, transportation and communication infrastructure that support agro dealers and that enable input markets to function. For instance: ■■ In Niger in 2010, there was a disruption in the supply of the PICS bags because of underestimation of the demand by the vendors. The production of cowpea in 2010 was characterized by an increase in area planted and production. The total demand of PICS bags estimated by the vendors before harvest was evaluated at 100,000 bags with a first order of 50,000 bags delivered at the beginning of harvest, but the second order was delivered very late, almost in December, because of the late payment. The delay in the bags procurement led to a break in the supply of PICS bags from October to the end of November. The delay was explained by the fact that some wholesalers did not pay their cash advance on time. Figure 4 illustrates the complexity of the PICS supply chain for Niger. ■■ In Cameroon, the local PICS distributor did not have shops or outlets in villages and therefore was unable to efficiently reach farmers and meet the demand for PICS bags in these villages. Field technicians sometimes were forced to take money from famers and procure the bags for them. ■■ In Nigeria most of the PICS bags vendors in Nigeria are concentrated in the urban areas because of the poor transportation network to the rural areas. This raised the price of the bags to farmers and restricted availability of PICS bags in the rural areas. For example, the retail price of PICS bags depended on the distance and usually varies between $ 1.80 and $ 3.00 in the rural areas. Occasionally the price can go as high as $ 6 per bag as was the case in the Maigatari area of Jigawa State. ■■ In Burkina Faso, one challenge was related to the high cost of transport and tax paid to import the bags from Nigeria to Burkina Faso. Because of the high cost of manufacturing in Burkina Faso, the PICS distributor chose to import PICS bags from Nigeria. These bags passed through Niger to enter Burkina Faso. From the Niger border to the warehouse in Ouagadougou, the transportation cost and other fees payment is evaluated at about $ 20 per bale of 300 bags. International Food and Agribusiness Management Review

433

Volume 20, Issue 3, 2017

NIGER

NIGERIA

Nouhoheflin et al.

Omarou DAN MALKA (Niamey, Tillabery)

Husa’a Enterprise (Dosso) Ibrahim ZANGUINA (Dosso)

Assoumane SALIFOU (Tahoua)

BURKINA

Maman NA IYA (Maradi)

NIGER

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

Lela Agro (Kano, Nigeria)

Fasoplast (Ougadougou, BF)

Manufacturing

ETS Nassamou ABDOUL (Niamey)

Malan Manirou OUMAROU (Zinder)

NGOs OPVN

N=2

National Distribution

N=5

N=61

SemiWholesaling

Regional Wholesaling

Retailing

Farmers Coops/traders Processors Food banks

Figure 4. Organizational mapping of the supply chain. ■■

■■

In Chad in 2011, the order of PICS bags from Nigeria was delayed because of lack of money to pay for the transport. The shipment of 15,000 bags was held at Maiduguri, Nigeria, and did not reach the distributor at Moundou in time for the harvest. Moreover, the 300 bag bales were too big to be handled by Chadian public transportation. To bring the bales at Moundou, the bags were repackaged in bales of 150 bags. This repackaging alone cost $ 700 and the shipping to Moundou cost about $ 45 per bale. The national distributor of Chad complained about the high transaction costs which include transport, repackaging, travel negotiate shipping arrangements, and communication fees. In Ghana the lack of coordination between the technical, promotional and supply chain actors was a challenge. In Ghana there were cases in which PICS radios and TV messages continued to be broadcast even though there was no bags available in the country. This was especially in 2011 when the national distributor could not order bags locally because of unpaid manufacturing bills from the previous year.

Enabling environment constraints ■■

■■

In Niger in 2008 and 2009 a government food security program inflated prices at harvest and this discouraged on-farm storage. This program was implemented by the National Food Products Office of Niger (French acronym is OPVN). The stated purpose was to constitute a food reserve to be used during lean periods. OPVN purchased cowpea from producers in every region of Niger. It was purchased at a price of $ 0.56/kg when the market price was $ 0.27/kg. This cowpea was stored in the OPVN warehouses. Buyers of these stocks initially targeted were foreign government agencies (i.e. World Food Program) and private domestic buyers. This policy, even though it was popular among producers because it increased their revenues, discouraged farmers from storing cowpea for a higher price in the post-harvest period. The weak contract and regulatory enforcement mechanisms also hindered the development of the supply chain. Despite the formal agreement among the national distributors, the manufacturers and the PICS project, not all the bags pre-financed by the PICS project were reimbursed. As noted by Biggs and Shah (2006) input dealers are hampered by underdeveloped financial and insurance markets, weak public institutions for contract law and enforcement, and poor infrastructure. This was the case in Burkina Faso, Ghana, and Benin where the national distributors failed to comply with the terms of the agreements with the PICS project or the manufacturers. In Benin, the consequence was the

International Food and Agribusiness Management Review

434

Nouhoheflin et al.

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

■■

Volume 20, Issue 3, 2017

deadlock of the supply chain while in Niger, another national distributor was selected to take over the distribution of the bags. In both cases, the nonpayment resulted in a breakdown in bag supply. The relatively high import tariffs/taxes on raw materials along with the price fluctuation of raw materials constituted a challenge faced by the manufacturers of PICS bags. All the manufacturers import raw material from South Africa, Europe or the Middle East. In Mali, the import tariffs at Malian border were estimated at 30% of the free on board cost while in Burkina, they were estimated at 46% of the raw materials cost. In Ghana, on top of the price fluctuation of raw materials the volatility of the exchange rate of Ghana currency (Cedis) in the market was another concern for the Ghanaian manufacturer. These fluctuations and tariffs increase the retail price of the bags.

Social constraints There were some social constraints and belief systems evoked by the PICS supply chain actors which impeded the development of the supply chain to some extent: ■■ In Niger for instance, wholesalers mostly used their own financial resources to invest in the PICS business. Few took bank loans because their Muslim religious beliefs discourage paying interest rate. ■■ In Senegal, social considerations affected the development of the supply chain. Some farmers were reluctant to provide cowpea for the public storage demonstrations because they did not want neighbors to know their cowpea production, or because of perceived differences in social status. In some cases the field technicians were forced to do the storage demonstrations in their household compound with only their family members.

4. Management lessons learned With the assistance of the BMGF, the PICS1 project pioneered investments across ten countries in West and Central Africa in the development of commercially driven supply chains for its innovative on-farm storage technology. The project tested various strategies designed to encourage investments by manufacturers and local, independent vendors to expand the PICS supply chain. There were a number of successes, yet progress across the PICS landscape was uneven. This study is an effort by the project and its partners to document and learn from the experience. The following is a list of key lessons learned based on a cross-country analysis of the ten case studies. Critical bottlenecks and notable achievements are highlighted along with some recommendations on how best to support the future growth and sustainability of supply chains for PICS bags and other agricultural innovations. Lessons learned include: 1. PICS bags are well adapted to smallholder farmer storage needs – The PICS triple-ply, hermetically sealed storage bags were designed specifically to meet the needs of smallholder farmers for affordable on-farm storage of cowpea. Evidence of strong and growing demand for PICS bags across the region suggests that the technology responds well to farmers’ needs, offering them a compelling value proposition. Data also suggest that retail price does not significantly impinge upon the average farmer’s willingness to invest his or her limited resources in PICS technology. 2. PICS bags are not as well accepted by large scale traders – The high, up-front cost of PICS technology relative to other conventional, chemical-based storage methods is a significant obstacle, and adoption has been relatively limited. This is especially true among high-volume traders who want to maximize their profits and who view storage expenditures as a sunk cost. Based on a rough comparison of initial costs and not taking into account the storage period, PICS bags are initially 80-90% more expensive per 100 kg of stored cowpea than the fumigation usually used by large scale grain traders. In addition, anecdotal evidence from Northern Nigeria, Benin, and elsewhere suggests that many cowpea traders view PICS technology as a threat to their business model, which includes insecticide sales. Nonetheless, growing health concerns about the use of pesticides for the storage of cowpea could in time increase the demand for untreated, chemical-free cowpea, and thus, the use of PICS technology among cowpea traders. 3. Incremental rollout was a good strategy – The PICS1 project benefited from a progressive rollout over five years (2007-2012) and ten countries. This step-wise expansion became a significant International Food and Agribusiness Management Review

435

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

Nouhoheflin et al.

Volume 20, Issue 3, 2017

learning asset and contributed to an ever-expanding knowledge base of best practice over time. As the project’s geographical scope expanded into new cowpea producing countries, improvements in the project’s design and approach were made based on notable successes and setbacks encountered and associated lessons learned from past experience. This allowed for a gradual refinement and validation of strategies, an evolution that becomes evident when looking at the experience across the entire PICS timeline. Future expansion into new cowpea production areas within the region as elsewhere will no doubt benefit from this existing knowledge base. Yet, market dynamics vary considerably from one place to the next. Any future expansion should thus be preceded by a comprehensive market study to determine potential surplus storage capacity, to identify manufacturers and other supply chain partners, to target major production zones, and to better understand the competitive landscape and market preferences (e.g. consumer, norms/standards), in addition to other information key to planning and strategy development. Risk sharing needed for manufacturers and vendors – Investing in new technologies like PICS bags poses considerable risks for manufacturers, vendors, and other supply chain actors. Strategies are needed to help them address these risks via risk sharing, risk transfer or risk mitigation mechanisms. How to structure and finance the procurement also poses a significant challenge. The project’s initial engagement in tripartite arrangements with manufacturers and vendors involved direct financial support for the procurement of PICS bags. This approach carried significant risks for the project. Such risks were beyond the scope of the project to manage effectively and led in some cases to considerable but temporary supply chain disruptions (e.g. Burkina Faso, Niger). Zero-risk or no-risk approaches such as ‘guaranteed markets’ (e.g. via the direct purchase of demo bags from vendors or use of vouchers) and cost-sharing of promotional activities offer alternative approaches that carry limited risk to the project, but may be effective in incentivizing private sector participation and investments in the supply chain. What type of vendors sells the most PICS bags? – Full-line agro-dealer networks selling a range of complementary inputs such as seeds, fertilizers, and crop protection products can be effective distribution channels to get PICS bags into the hands of rural farmers; however, they are not the only game in town. Evidence suggests that alternatives such as mobile cell phone card vendors and school supply shops can be equally effective in marketing PICS bags. The most crucial qualities are that vendors’ business model is appropriately oriented toward the provision of good and/or services rather than toward output markets. For example, the PICS experience in Benin and Nigeria suggests that grain traders’ output market orientation and revenue model is in many respects inimical to the activity of marketing PICS bags. Many grain traders see the PICS technology as a direct threat to their business. In Niger and Togo, by contrast, successful vendor networks have piggy-backed on the channel distribution of mobile phone cards and school supplies. Is there a public sector role in the PICS supply chain? – The PICS project worked closely with national research institutes and public extension services in all ten countries. These partnerships were vital to securing necessary local commitment and reaching farmers. Yet, while such institutions can be an effective way to expose farmers to the technology and broaden awareness, they do not constitute a viable long term distribution channel. Extension agents, in particular, are close to farmers and wield considerable influence over farmers’ technology adoption and investment decision-making. Thus, the project worked closely in partnership with national extension services in most countries to facilitate local, ‘last mile’ distribution to farmers in rural areas on a consignment stock basis. However, widespread incidences of non-payment of cash collected from sale of consigned bags presented significant challenges to the supply chain’s viability. In some countries, however, agents acting independently from the national extension service and sourcing directly from PICS bag vendors on an ad-hoc basis became effective mobile vendors of PICS bags at the village level. Scope exists to leverage extension agents more broadly as retail sales channels by facilitating direct relationships between individual agents and local PICS bag vendors. Credit is not a key constraint, but low profitability for vendors and risk are – Lack of access to institutional credit does not in itself constitute a binding constraint to supply chain growth. Many vendors active in the PICS supply chain, particularly those operating upstream at the wholesale and International Food and Agribusiness Management Review

436

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

Nouhoheflin et al.

Volume 20, Issue 3, 2017

semi-wholesale levels, have banking relationships and are able to meet basic collateral requirements. However, high interests rates, the slow rate of turnover, and risk discourages vendors from using borrowed capital to finance their stocks. An analysis of borrowing costs indicates that at prevailing rates vendors would need to more than double their current sales volume in order to simply meet their loan repayment obligations. General distrust and lack of understanding about how banks work further discourages institutional borrowing, most notably among retailers. In such an environment, identifying ways to more fully leverage informal social capital already established within vendor networks is fundamental to supply chain development. Identifying ways to encourage wholesalers to sell PICS bags more broadly on credit or partial credit and on consignment stock could be explored. Vendor training to increase book-keeping, inventory management, and communications skills is one example. 8. First year reference prices have surprisingly long term effects – Reference price strategies designed to protect buyers and discourage market speculation can significantly dampen investment in the supply chain in subsequent years, especially downstream. In all countries, the project in collaboration with vendor partners identified a price ceiling at which PICS bags would be marketed to farmers the first year. This price was initially communicated to the market via promotional events and advertising. In many cases, these price targets carried over into subsequent years, orienting consumers around a target ‘benchmark’ price. An analysis of how margins are distributed across the supply chain in all ten countries suggests that upstream vendors (e.g. wholesalers, semi-wholesalers) capture (on a proportional basis) the vast majority of available margins, and this pattern is exacerbated under any reference price plan. Retailers have difficulty in securing sufficient margin from the sale of PICS bags to cover their marketing costs and make a reasonable profit from their activities. Facilitating open market price discovery after the first year is preferred. In particular, reference prices should not be part of the project supported advertising and promotional materials in the second and subsequent years. 9. Crop forecasts are key to timely availability of bags – Demand for PICS bags in any given year depends on the harvest, but manufacturers and distributors lack the information and tools to forecast inventory needs. In many cases, distributors waited until after harvest started to order bags. This contributed to delayed delivery of bags, debilitating ruptures in supply, and unsatisfied demand. Miscalculations of demand also led to costly ‘dead stock’ when poor and uneven rainfall in many areas contributed to significant declines in bag demand. The project addressed this challenge in part by facilitating yearly vendor meetings when supply chain actors meet up to discuss challenges and set up a procurement plan for the coming season. In order to strengthen the ability of vendors to procure and stock the right number of bags at the right time, further and deeper investments are needed in procurement and inventory management. Facilitating access among vendor groups to information from weather and crop forecasting agencies and public extension agencies that can help vendors better gauge planting acreages, expected output, and thus, demand for PICS bags is illustrative of such investments. 10. Trademarking the PICS logo was a useful step, but it is no substitute for patent protection – At the suggestion of PICS manufacturers and distributors, the project initiated the process of trademarking the PICS logo in 2009. The trademark provides one tool for the PICS supply chain to discourage substandard bags. Patent protection would have been an even stronger tool, but it was originally not pursued by the university and is no longer an option since the unpatented technology is already out in the public domain. Anecdotal evidence from Burkina Faso suggests that farmers seeing value in the PICS brand as a certification of quality are willing to pay a slightly higher price for PICS-branded bags over competing, lower priced products. Potential scope may now exist to leverage the PICS trademark via innovative licensing arrangements to safeguard PICS quality standards and encourage new investments in PICS bag production and distribution in the region as elsewhere. 11. Project business consultants played a key role which should be gradually shifted to the private sector – Beginning in 2007, the project recruited one or more local consultants in nine of the ten PICS countries to assist with supply chain development efforts. Input from stakeholders indicates that PICS business consultants have made a significant contribution to the project’s success in most countries. International Food and Agribusiness Management Review

437

Nouhoheflin et al.

Volume 20, Issue 3, 2017

http://www.wageningenacademic.com/doi/pdf/10.22434/IFAMR2016.0167 - Tuesday, May 09, 2017 8:38:51 PM - IP Address:24.21.169.207

Their role has focused on working with manufacturers and wholesalers to identify new downstream vendors and expand distribution networks, facilitate bag procurement and delivery, and increase bag sales. Given their important role, incremental cost-sharing arrangements with manufacturers and/ or wholesalers should be explored to ensure an appropriate and non-disruptive exit strategy for the project and its funding. To be competitive in todayâ&#x20AC;&#x2122;s globalized food markets, smallholder farmers need new and better technologies. Once developed, agriculture innovations need to be mainstreamed into the marketplace to make them available to rural farmers when and where they need them. In addition to reliable and timely delivery, affordability and accessibility are key. Availability in farm communities and access to information on their use are also critical to uptake. In order for innovations to have a positive impact on farmer productivity and income, supply chains must effectively get these tools to the right place, at the right time, in the right quantities, and at the right price. The PICS project provides important agricultural input supply chain development examples which should be the foundation for future improvements in seed, fertilizer, pesticide, and other farm input marketing in Africa.

Acknowledgements The authors would like to thank the BMGF for the funding that made this research possible. All opinions expressed are those of the authors, not of the BMGF.

References Biggs, T. and M.K. Shah. 2006. African SMES, networks, and manufacturing performance. Journal of Banking and Finance 30: 3043-3066. Coulibaly, J.Y., T. Nouhoheflin, C. Aitchedji, D. Maiyaki, S. Dâ&#x20AC;&#x2122;Alessandro, D. Baributsa and J. LowenbergDeBoer. 2012. Purdue improved cowpea storage (PICS) supply chain study, Staff Paper 12-4, Department of Agricultural Economics/International Programs in Agriculture, Purdue University, West Lafayette, IN, USA. Available at: http://tinyurl.com/jqluqvu. Ibro, G., M.C. Sorgho, A.A. Idris, B. Moussa, D. Baributsa and J. Lowenberg-DeBoer. 2014. Adoption of cowpea hermetic storage by women in Nigeria, Niger and Burkina Faso. Journal of Stored Product Research 58: 87-96. Kelly, V.A., E.W. Crawford and T.S. Jayne. 2003. Agricultural input use and market development in Africa: recent perspectives and insights. Policy Synthesis No. 70, Department of Agricultural Economics, Michigan State University, East Lansing, MI, USA. Moussa, B., T. Abdoulaye, O. Coulibaly, D. Baributsa and J. Lowenberg-DeBoer. 2014. Adoption of on-farm hermetic storage for cowpea in West and Central Africa. Journal of Stored Product Research 58: 77-86. Moussa, B., D. Baributsa and J. Lowenberg-DeBoer. 2010. Adoption of hermetic storage for cowpea in Niger and Burkina Faso in 2010. Poster presented at the 5th World Cowpea Conference, 27 September - 1 October 2010, Saly, Senegal. Available at: http://tinyurl.com/jhvxcy7. Moussa, B., J. Lowenberg-DeBoer, J. Fulton and K. Boys. 2011. The economic impact of cowpea research in West and Central Africa: a regional impact assessment of improved cowpea storage technologies. Journal of Stored Products Research 47: 147-156. Murdock, L.L. and I.B. Baoua. 2014. On Purdue improved cowpea storage (PICS) technology: background, mode of action, future prospects. Journal of Stored Product Research 58: 3-11. Sonka, S., Y.-T. Cheng and G. Kenney. 2016. The role of markets and enterprises in postharvest loss reduction. Paper presented at the 20th International Food and Agribusiness Association (IFAMR) Annual Meeting, 19-23 June 2016, Aarhus, Denmark. Prahalad, C.K. 2004. The fortune at the bottom of the pyramid: eradicating poverty through profit. Wharton School Publishing, Upper Saddle River, NJ, USA. Yin, R.K. 2003. Case study research: design and methods, 3rd ed. Sage, Thousand Oaks, CA, USA.

International Food and Agribusiness Management Review

438