Economics and Business Review, vol. 1(15), no. 2, 2015

Economics and Business Review

ISSN 2392-1641

Volume 1 (15) 2 2015 Volume 1 (15)Number Number 2 2015 CONTENTS ARTICLES A perspective on leading and managing organizational change Stanley J. Smits, Dawn E. Bowden Alternative configurations of firm-level employment systems: evidence from American companies Bruce E. Kaufman, Benjamin I. Miller How team leaders can improve virtual team collaboration through trust and ICT: A conceptual model proposition David Kauffmann International trade in differentiated goods, financial crisis and the gravity equation Udo Broll, Julia Jauer Tax revenues and aging in ex-communist EU countries Mihai Mutascu, Maciej Cieślukowski The analytics of the New Keynesian 3-equation Model Jean-Christophe Poutineau, Karolina Sobczak, Gauthier Vermandel Investments and long-term real interest rate in Poland. Study of investment structure, current account and their correlation with long-term real interest rates Jakub Krawczyk, Szymon Filipczak BOOK REVIEWS Paweł Marszałek, Systemy pieniężne wolnej bankowości. Koncepcje cechy, zastosowanie [Free Banking Monetary Systems. Concepts, Characteristics, Application], Wydawnictwo Uniwersytetu Ekonomicznego w Poznaniu, Poznań 2014 (Bogusław Pietrzak) Ewa Mińska-Struzik, Od eksportu do innowacji. Uczenie się przez eksport polskich przedsiębiorców [From Export to Innovation – Learning by Exporting in Polish Enterprises], Difin, Warszawa 2014 (Jan Rymarczyk)

Poznań University of Economics Press

Editorial Board Ryszard Barczyk Witold Jurek Cezary Kochalski Tadeusz Kowalski (Editor-in-Chief) Henryk Mruk Ida Musiałkowska Jerzy Schroeder Jacek Wallusch Maciej Żukowski International Editorial Advisory Board Udo Broll – School of International Studies (ZIS), Technische Universität, Dresden Wojciech Florkowski – University of Georgia, Griffin Binam Ghimire – Northumbria University, Newcastle upon Tyne Christopher J. Green – Loughborough University John Hogan – Georgia State University, Atlanta Bruce E. Kaufman – Georgia State University, Atlanta Steve Letza – Corporate Governance Business School Bournemouth University Victor Murinde – University of Birmingham Hugh Scullion – National University of Ireland, Galway Yochanan Shachmurove – The City College, City University of New York Richard Sweeney – The McDonough School of Business, Georgetown University, Washington D.C. Thomas Taylor – School of Business and Accountancy, Wake Forest University, Winston-Salem Clas Wihlborg – Argyros School of Business and Economics, Chapman University, Orange Jan Winiecki – University of Information Technology and Management in Rzeszów Habte G. Woldu – School of Management, The University of Texas at Dallas Thematic Editors Economics: Ryszard Barczyk, Tadeusz Kowalski, Ida Musiałkowska, Jacek Wallusch, Maciej Żukowski • Econometrics: Witold Jurek, Jacek Wallusch • Finance: Witold Jurek, Cezary Kochalski • Management and Marketing: Henryk Mruk, Cezary Kochalski, Ida Musiałkowska, Jerzy Schroeder • Statistics: Elżbieta Gołata, Krzysztof Szwarc Language Editor: Owen Easteal • IT Editor: Piotr Stolarski

© Copyright by Poznań University of Economics, Poznań 2015 Paper based publication

ISSN 2392-1641

POZNAŃ UNIVERSITY OF ECONOMICS PRESS ul. Powstańców Wielkopolskich 16, 61-895 Poznań, Poland phone +48 61 854 31 54, +48 61 854 31 55, fax +48 61 854 31 59 www.wydawnictwo-ue.pl, e-mail: wydawnictwo@ue.poznan.pl postal address: al. Niepodległości 10, 61-875 Poznań, Poland Printed and bound in Poland by: Poznań University of Economics Print Shop Circulation: 300 copies

Volume 1 (15) Number 2 2015 CONTENTS ARTICLES A perspective on leading and managing organizational change Stanley J. Smits, Dawn E. Bowden...................................................................................................

3

Alternative configurations of firm-level employment systems: evidence from American companies Bruce E. Kaufman, Benjamin I. Miller ........................................................................................... 22 How team leaders can improve virtual team collaboration through trust and ICT: A conceptual model proposition David Kauffmann ............................................................................................................................ 52 International trade in differentiated goods, financial crisis and the gravity equation Udo Broll, Julia Jauer ........................................................................................................................ 76 Tax revenues and aging in ex-communist EU countries Mihai Mutascu, Maciej Cieślukowski.............................................................................................. 95 The analytics of the New Keynesian 3-equation Model Jean-Christophe Poutineau, Karolina Sobczak, Gauthier Vermandel ......................................... 110 Investments and long-term real interest rate in Poland. Study of investment structure, current account and their correlation with long-term real interest rates Jakub Krawczyk, Szymon Filipczak ................................................................................................. 130 BOOK REVIEWS Paweł Marszałek, Systemy pieniężne wolnej bankowości. Koncepcje cechy, zastosowanie [Free Banking Monetary Systems. Concepts, Characteristics, Application], Wydawnictwo Uniwersytetu Ekonomicznego w Poznaniu, Poznań 2014 (Bogusław Pietrzak) ...................... 149 Ewa Mińska-Struzik, Od eksportu do innowacji. Uczenie się przez eksport polskich przedsiębiorców [From Export to Innovation – Learning by Exporting in Polish Enterprises], Difin, Warszawa 2014 (Jan Rymarczyk) ......................................................................................... 151

Economics and Business Review, Vol. 1 (15), No. 2, 2015: 3–21

A perspective on leading and managing organizational change1 Stanley J. Smits2, Dawn E. Bowden3

Abstract: Organizational change poses significant challenges. Change itself is changing; evolving in ways that present new rules, new strategies for winning, and more and more dynamic complexity. This paper presents the principal drivers of change as stand-alone entities and later discusses their interaction effects. Organizational Life Cycle Change, types of change, capacity for learning, and the common causes of change failures are explored to establish an understanding of the proclaimed enormity of the change-failure issue and our difficulty in quantifying it. The paper concludes with suggestions that will help organizational change agents improve their success rates.

Keywords: change management, organizational change, organizational culture, learning organization, strategic entrepreneurship.

JEL codes: M00, M1, M14, M19.

Introduction Change is ubiquitous: Our bodies change throughout life, as do the physical, economic, social, and technological environments in which we live. So it should come as no surprise that the organizations to which we belong also change. In fact, they are designed to change and that is why we employ leaders and managers and assign them responsibilities to be agents of change. So why does organizational change pose such a challenge for us? Why is successful organizational change the exception, not the rule? The short answer is because change 1

Article received 16 November 2014, accepted 19 February 2015.

From 2002–2014, dozens of participants in the doctoral-level Change Management Seminars offered by the International School of Management (ISM) , Paris, helped shape the perspective presented in this paper via their shared experiences, lively discussions, probing questions, and post-seminar project applications. Appreciation is extended to all.

2 Georgia State University, Department of Managerial Sciences, Atlanta, USA. 3 Johnson & Johnson, Department: Health Economics & Market Access, 9919 Sand Verbena Trl NE, Albuquerque, 87122 New Mexico, USA, corresponding author, e-mail: dawn.bowden@ gmail.com.

4

Economics and Business Review, Vol. 1(15), No. 2, 2015

itself is changing; evolving in ways that present new rules, new strategies for winning, and more and more dynamic complexity.

In The Fifth Discipline, Peter Senge [1990a] introduced us to the challenges posed by dynamic complexity: Systems thinking teaches us that there are two types of complexity – the ‘detail complexity’ of many variables and the ‘dynamic complexity’ when ‘cause and effect’ are not close in time and space and obvious interventions do not produce expected outcomes [p. 364].

The examples he used to introduce the concept help explain why organizational change is a continuous challenge for leaders and managers: When the same action has dramatically different effects in the short and the long run, there is dynamic complexity. When an action has one set of consequences locally and a very different set of consequences in another part of the system, there is dynamic complexity. When obvious interventions produce nonobvious consequences, there is dynamic complexity [Senge 1990a: 71].

In the next section of this paper, we examine four principal drivers of change and the interaction effects among them resulting in dynamic complexity.

In this paper, we do not accept the less than optimal success rates of organizational change as inevitable and argue that many of the failures are preventable. Specifically, this paper has three purposes. To: – Sensitize the reader to the dynamic complexity underlying organizational change; – Highlight the documented common causes of less-than-optimal change; and to – Provide practical, step-by-step, theory-based suggestions for leading and managing organizational change more effectively.

1. The dynamic complexity of organizational change The principal drivers of change are presented here as stand-alone entities and later discussed in terms of their interaction effects. The interaction effects are potential sources of dynamic complexity, described as “situations where cause and effect are subtle, and where the effects over time of interventions is not obvious” [Senge 1990a: 71].

1.1. Strategic change All organizations, even not-for-profit ones, exist in environments where gaining and sustaining competitive advantage is a prerequisite for success. Organizations survive and thrive when they have substantive advantages over their competitors. As Porter told us: “Competitive strategy is about being different. It means deliberately choosing a different set of activities to deliver a unique mix of

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

5

value” [1996: 6]. But in our high-tech, communication-rich, global environment, unique competitive strategies are challenged by a cadre of imitators. So organizations find themselves frequently changing or fine-tuning their strategies to ensure differentiation or cost advantage. Strategic change, however, has a cascading effect on the organization requiring changes to, and realignment of structures, systems, and processes to maintain required “fit” between strategy and operational effectiveness [Burke 2002; Nadler and Tushman 1989]. This is where timing can become an issue: Strategies can be changed faster than structures; and during the period of “catch-up” between strategy and structure, organizations lose both efficiency and effectiveness.

The “fit/timing” issue between strategy and structure is not new. Based upon his research in the 1950s, Alfred Chandler stated the issue in the form of a behavioral law in 1962: Structure follows strategy. Simply put: First we decide what we want to do, next we plan how we will do it, and then we organize the resources needed to do it. But today’s competitive world is quite different from the environment of Chandler’s research in the decade of the 1950s. Market competition is changing too fast for organizations to initiate a strategic change and then patiently work for a year or two to bring its structure, systems, operations, and human resources in line with it. To cope with today’s dynamic realities, organizations must adopt flatter, more flexible structures, and position their processes, systems, and people to have greater capacity for creativity and innovation, all at a much more rapid rate. In brief in Grant’s [2002] terms: Many operating organizations have morphed into innovating organizations to cope with the demands of rapidly changing external environments.

The demands for sustained competitive advantage and innovation are forging a new scholarly coalition, strategic entrepreneurship (SE), focused on the question: “[…] how do firms create and sustain competitive advantage while simultaneously identifying and exploiting new opportunities?” [Hitt et al. 2011: 57].

Building on research from multiple disciplines, SE operates at the nexus of strategic management and entrepreneurship designing actions “to exploit current advantages while concurrently exploring opportunities that sustain an entity’s ability to create value across time” [p. 57]. Hitt and his associates suggest that SE may be of particular value to large established firms that need to become more entrepreneurial to sustain competitive advantage.

Strategic entrepreneurship provides a convenient transition to our next driver of change: Life cycle change, the natural process whereby small, entrepreneurial, startup organizations grow and mature into large operating organizations less capable of maintaining competitive advantage through innovation.

1.2. Life cycle change Over time, startup organizations face similar crises and challenges, and grow and mature in predictable ways as they survive and prosper. These changes

6

Economics and Business Review, Vol. 1(15), No. 2, 2015

Organizational stages of development Source: [Daft 1992]; adapted from [Quinn and Cameron 1983; Greiner 1972]

have become commonly known as “Life Cycle Change” [Cawsey, Deszca, and Ingols 2012]. Multiple authors have described the life-cycle process of growth and development with three of the better known models/theories being a fivephase model proposed by Greiner [1972], a four-phase model developed by Quinn and Cameron [1983] based upon their synthesis of nine published lifecycle models, and the compilation and adaptation of both of these by Daft [1992: 164] (Figure).

Life cycle changes manifest themselves in various ways and are dealt with differently across industries and cultures causing critics to label Greiner’s model as “overly prescriptive” while at the same time acknowledging its applicability: “While Greiner’s model is prescriptive, it captures many of the issues faced by organizations both in growth and in dealing with the human side of organizational change” [Cawsey, Deszca, and Ingols 2012: 83].

For purposes of discussion here, we will use Daft’s [1992] four-stage model which incorporates features and inputs from Quinn and Cameron’s [1983] nine published life cycle models. The four stages (Entrepreneurial, Collective, Formalization and Control, and Elaboration of Structure) represent the passage of time from startup to full maturity and typically substantive increases in size.

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

7

– Stage 1: Entrepreneurial. The enterprise often starts with a novel idea that is launched as a small business under the direction of the entrepreneur/ owner whose creativity is now being made operational by a small group of trusted employees (sometimes friends and relatives). It is characterized by: • A marshalling of resources (often personal resources combined with whatever venture capital could be raised), • Lots of ideas (helping to shape and implement the original creative idea), • Entrepreneurial activities (to establish a customer base, provide reliable goods and/or services, produce cash flow; i.e., to do what is needed to survive), • Little planning and coordination (operating basically “free-form” to get things started), • “Prime Mover” has the power (Stage 1 organizations are run by the owner).

– Stage 2: Collectivity. This is also known as the success stage because of the rapid growth. Here the owner can remove her/himself from the active management role turning it over to professional managers or reinvest the profits into more growth while maintaining control. This stage is characterized by: • Informal communication and structure (we are too busy growing the business to worry about efficiency), • Sense of collectivity (we feel like a “family” and are often cross-trained to do each other’s jobs as needed), • Long hours spent at work (work comes first and there’s always more to be done than the time to do it), • Sense of mission (we are part of this, it is part of us), • Innovation continues (the original creativity improves through refinements), • High commitment (success drives success).

– Stage 3: Formalization. In this stage the organization transitions from entrepreneurial to professional. Structure and systems develop to bring the organization under control, to standardize and stabilize how it does business.

The formalization makes the organization more efficient thereby enabling it to move from a differentiation competitive advantage to a cost advantage should it choose to do so. It is characterized by: • Formalization of rules (rules substitute for the face-to-face, hands-on guidance and direction from leaders and managers), • Stable structure (structure and systems are refined to make operations more reliable and efficient), • Emphasis on efficiency and maintenance (change is replaced by order), • Conservatism and institutional procedures (tradition and past success are honored as existing order is maintained).

8

Economics and Business Review, Vol. 1(15), No. 2, 2015

– Stage 4: Elaboration. Here the mature organization attempts to cope with changing conditions in their business environment and find new ways to continue growing. With more diversification, they often decentralize decision-making and try to adapt to change via more teamwork. When adaptation fails, decline ensues.

• Elaboration of structure (to cope with continued growth through diversification), • Decentralization (to cope with elaboration of structure and the need to bring decisions closer to the point of implementation), • Adaptation (reactive change to continue and consolidate growth), • Renewal (transformational change to reposition the organization within its competitive environment).

The Life Cycle model of change presents a diagnostic framework for two common problems: – Leadership/structure mismatch: The entrepreneur who started the organization continues to lead it in a Stage 1 manner even though its growth, complexity, and management systems are now operating in a Stage 3 or 4 manner. As Patrick Canavan noted in his analysis of a Harvard Business Review case by Beer [2006]: “Charismatic CEOs […] can be good news for small companies, but they can be bad news for large ones” [p. 52].

– Entrepreneurial organizations morph into operating organizations as they grow and age: Rules and formulization of structure slowly produce an organization that finds it difficult to respond to ongoing or new opportunities for competitive advantage through innovation. In brief, they have become so good at what they do that it is difficult for them to do something different, that is to change.

1.3. Learning-based change People learn via experience. Organizations, as collections of people working together to achieve common objectives, also learn via experience. To exist is to experience, therefore organizations learn continuously throughout their life cycles: “Learning is a relatively permanent change in behavior or behavior potential resulting from direct or indirect experience” [Griffin and Moorhead 2007: 102]. We tend to think about learning as positive, but we are also capable of learning inappropriate, counter-productive ways of behaving, multiple forms of prejudice, and even resistance to change.

Given that learning can produce ways to increase performance and innovations essential to competitive advantage, some organizations strive to become learning organizations. While discussing strategic change, Rowden [2001] described the learning organization as one which is actively engaged in “[…] identifying and solving problems, enabling the organization to continuously ex-

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

9

periment, change, and improve, thus increasing its capacity to grow, learn, and achieve its purpose” [p. 15]. He then specified four defining characteristics of the learning organization presented here in summary form: – Constant Readiness • Exists in a constant state of readiness, anticipates change, • Is prepared for change in general, stays attuned to its environment, • Is willing to reevaluate past assumptions and future directions.

– Continuous Learning • Flexible plans are shared and embraced by the entire organization, • Maintains flexible, open strategic directions, revisions are expected.

– Improvised Implementation • Encourages experimentation throughout the organization, • Learns from successes; small wins are rewarded and institutionalized.

– Action Learning • Builds on previous learning of an individual and their peers and focuses on problems within the organization that are relevant and require immediate application [Vince and Martin 1993], • “[…] the learning organization takes action, reflects, and adjusts its course as it goes, seeking to enhance the speed and effectiveness by which it learns how to change” [Rowden 2001: 16].

Rowden’s [2001] portrayal of the learning organization is similar to the concepts underlying Argyris, Putnam and Smith’s [1985] Action Science, Anderson and Anderson’s [2001] Unfolding Change, Vera and Crossan’s [2004] 4I Framework of Organizational Learning: Intuiting, Interpreting, Integrating, and Institutionalizing, and Senge’s [1990b] challenge to leaders: The Leader’s New Work: Building Learning Organizations. The bottom line is that well-developed systems of learning allow the organization to capture knowledge important to effective change. And as knowledge acquisition and application gained recognition as the underpinnings of successful strategic change and organizational effectiveness, knowledge management became a formalized function with a “chief knowledge officer” as its designated leader/manager [Grant 2002].

In brief, the relationship between organizational knowledge systems and effective change that gained recognition in the decade of the 1990s and became more formalized in the decade of the 2000s seems to continue to escalate in importance: Together, economic, technological, and social events have changed the knowledge needed to manage organizations effectively. Complexity has increased as the economy has become more global, and information technology has changed the way people and organizations operate. […] Today, more than ever, organizations need research-based knowledge about organizational change, management, and effectiveness [Mohrman and Lawler 2012: 41].

10

Economics and Business Review, Vol. 1(15), No. 2, 2015

1.4. Leader-initiated change According to Warren Bennis: “Leaders are people who do the right thing; managers are people who do things right” [1989: 36]. In this brief statement, he has succinctly described and differentiated two important types of leader-initiated change: Transformational and transactional. In so doing, he joined a cadre of leadership experts who differentiate between “leadership” and “management” [for example: Bass 1985; Kotter 1990; Rowe 2001; Zaleznik 1977]; and took issue with experts who make no such differentiation [for example: Herold and Fedor 2008; Mintzberg 2004; Yukl 2010]. For our purposes, we will start by describing two types of change and their importance to organizational survival and prosperity.

Transformational change. This is change that is often radical, transforming entire organizations, creating a need for new structures and management systems, and that is often associated with breakthrough technologies, new products, and new markets. Transformational change has a visionary component taking the organization out of its present (known) comfort zone and moving it into a future (unknown) state therefore making it “unpredictable, uncontrollable change that must be shaped, and adapted as it unfolds” [Anderson and Anderson 2001: 4]. Transformational change realigns the organization with its external environment enabling it to pursue opportunities and avoid threats and in so doing disrupts the congruence among the internal components that evolved overtime to facilitate efficient functioning. When one part of the organization is changed, the other parts also need to adapt to restore congruence [Nadler and Tushman 1989]. In brief, transformational change necessitates immediate attention to needed transactional change [Burke 2002].

Transactional change. This type of change is incremental in nature, designed to restore congruence among the components of the organization so it can function efficiently to meet the demands resulting from its transformed status in the external environment. This is planned change focused on improving specific systems, operations, and organizational components to restore equilibrium and stability thereby allowing the organization to function efficiently. Transactional change is reactive in nature focusing on goals that arise out of necessities that result from the need to merge people, systems, and technologies to restore or improve performance [Bass 1985; Cawsey, Deszca, and Ingols 2012; Rowe 2001].

These two essential types of change require two types of change leadership, whether performed by different individuals which is often the case in large organizations or by the same individual which is typical in small and mid-size organizations: – Transformational leadership: The process of keeping the organization in sync with the demands of its external environment, – Transactional leadership: The process of creating systems, policies, and practices to keep the organization functioning efficiently.

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

11

Interestingly, the interplay of these two types of change leadership causes the inherent paradox complexity theorists contend is found in dynamic organizations [Cawsey, Deszca, and Ingols 2012]. The propositions underpinning complexity theory state that organizations are a web of nonlinear feedback loops connected to other individuals and organizations through nonlinear feedback loops. These feedback systems can operate in both stable and unstable states [Stacey 1996]. Thietart and Forgues described the dynamic interplay succinctly: […] Organizations also have counteracting forces at play. Some forces push the system toward stability and order; these include the forces of planning, structuring, and controlling. Some other forces push the system toward instability and disorder: the forces of innovation, initiative, and experimentation.

The coupling of these forces can lead to a highly complex situation: a chaotic organization [1995: 23].

To put the leader-driven change dimension in perspective we turn to Complexity Theory which posits two competing forces at work in organizations: Change vs. Stability. It says transformational leaders create disequilibrium in the organization and transactional leaders (aka: managers) restore equilibrium by revamping structures, systems, technology, and people [Cawsey, Desca, and Ingols 2012; Thietart, and Forgues 1995]. Gersick [1991] synthesized theory from six fields of inquiry and referred to this phenomenon as the “Punctuated Equilibrium Paradigm” which she summarized as follows: “Systems evolve through the alternation of periods of equilibrium, in which persistent underlying structures permit only incremental change, and periods of revolution, in which these underlying structures are fundamentally altered” [p. 13]. These alternations are the context for leader-driven change.

1.5. Dynamic complexity The four principal drivers of change function in an interactive manner resulting in what Peter Senge [1990a] and others label dynamic complexity. Each principal driver of change triggers related changes via the other drivers. For example: A visionary leader sees an opportunity for competitive advantage, engages in futuristic, non-linear, “out-of-the-box” thinking, articulates his vision and obtains buy-in for a long-range strategic change that unfolds and must be reshaped over a multi-year period during which there are changes in the organization’s political, economic, social, and technical environments. While this strategic change is in play and is driven by transformational leadership, the organization matures and experiences substantial growth. The transformation demands new ways of performing which, in turn, requires a new knowledge base, new technology, and the acquisition of new skills. And while these obvious sources of disequilibrium are pushing the organization toward chaos, skilled managers engage in transactional leadership to restore order to keep the organization’s components functioning in a congruent manner.

12

Economics and Business Review, Vol. 1(15), No. 2, 2015

The dynamic complexity inherent in the above example must be acknowledged, leveraged, and managed to ensure success: If organizations give into the forces for stability, they become ossified and change impaired. If they succumb to the forces for instability, they will disintegrate. Success is when organizations exist between frozen stability and chaos [Cawsey, Deszca, and Ingols 2012: 84].

2. Avoiding the common causes of change failures Before we explore the common causes of change failures we need to understand the proclaimed enormity of the change-failure issue and our difficulty in quantifying it. Trautlein [2013] cited evidence that the failure rate for major changes in organizations has been about 70 percent for the last 20 years.

To define major changes, she provided the following breakdown from a survey of human resource professionals in 2010 who were asked to describe major changes their organizations faced in the next six months: – Organizational restructure: 51 percent, – New leadership: 20 percent, – Acquisition/merger: 13 percent, – New product launch: 10 percent, – New technology: 6 percent [Trautlein 2013: 9].

Adding to the perspective of a 70 percent persistent failure rate for changes such as these, she commented: “These are all large-scale changes that affect nearly every corner of an organization. Done right, they can enhance a company’s performance dramatically; mishandled, they can turn into costly disasters” [Trautlein 2013: 9]. Similarly, Herold and Fedor [2008] concluded that “between 67 percent and 80 percent of change efforts, large and small, fail” [p. 2].

One further comment before we discuss what we know about the common causes of change failures: Quantifying failure rates can be challenging. Change efforts seldom fail completely, nor are they completely successful. How do we rate an ambitious change effort that achieves 5 of 7 stated objectives, a success or failure? How do we rate a change that meets its stated objectives but has overruns of time and costs? How do we judge a change effort that does everything its planners set out to do but failed to save the organization, i.e., “winning the battle but losing the war”? Regardless of the measurement issues, we know that too many change efforts fail and we have considerable insight into why failures happen. So next we explore five common causes of failure described by Herold and Fedor [2008] that change agents should make conscious efforts to avoid.

Common Cause of Failure #1: A disconnect between the solution and the problem. The change was successful but did not address the problem it was intended to resolve.

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

13

The causation here may range from inadequate pre-change research to define the problem to a powerful organizational leader getting emotionally caught up in a popular fad, personal agenda, and everything in between. That said, Herold and Fedor [2008] suggested two major reasons: 1. What was changed failed to address the problem (for example, cosmetic changes, staffing changes, reorganizations, or new technologies that were not able to deal with the underlying issues or support the overall business strategy).

2. The change addressed the wrong or even a nonexistent problem (for example, still another restructuring when the basic strategy is not working) [Herold and Fedor 2008: 18].

How could reasonable change agents possibly initiate changes that even when successful do not help the organization solve a fundamental problem?

Two general managerial tendencies have been known for years to mitigate problem solution: First, looking for quick fixes instead of taking a long-term view; and second, implementing solutions piecemeal rather than taking a systems perspective [Neal and Tromley 1995].

Common Cause of Failure #2: Inadequate resources to complete/sustain the change. The change is valid in terms of what the organization needs but the capacity to see it through to a successful completion does not exist.

Change is often expensive. If an organization experiencing revenue losses for an extensive period decides to implement an appropriate but expensive change, the first question it should ask is “Can we afford it?” If the answer is clearly “No”, the search for a feasible alternative may be the only prudent course of action. Similarly, pre-change questions need to be answered honestly about the organization’s human resource capacity: “Can we do it?” Do we have the talent base needed to implement the proposed change? If not, can we acquire it in time to see the change through to completion? If the fiscal resources and/ or needed knowledge, skills, and abilities are absent, going ahead with a welldesigned, well-intended change will have little chance of success.

Common Cause of Failure #3: Change turbulence. Starting new changes before earlier changes are completed; mistakenly seeing organizational changes as independent events.

Complexity theory warns us that when organizations experience excessive instability, they disintegrate [Cawsey, Deszca, and Ingols 2012] and the congruence model of organizations [Nadler and Tushman 1989] argues persuasively that when one part of the organization changes other parts need to adapt to restore congruent functioning; yet it is commonplace to see new changes implemented while the organization is still attempting to cope with the disequilibrium of previous changes. Perhaps the most convincing presentation for avoiding change turbulence comes from Herold and Fedor’s [2008] depictions

14

Economics and Business Review, Vol. 1(15), No. 2, 2015

of the decline in performance due to the introduction of a major change and the duration of the recovery in production merely returning to the pre-change baseline [Figure 7.2; p. 89] and the cumulative effect of overlapping changes on performance recovery, that is “Change Turbulence”, on performance recovery [Figure 8.1; p. 109]. Based on their analyses, they state unequivocally: “Performance inevitably declines in the face of change” [p. 92] with the recovery dependent upon the slope of the learning curve, motivation, time and effort, and the assistance and resources committed to the recovery effort. Here we present their concerns about change turbulence: – Each change requires an expenditure of […] resources, and as these resources get diverted to new changes they are unavailable for application to previous changes, prolonging the duration of recovery to baseline and the realization of performance improvements. Furthermore, each change following on the heels of previous yet only partially digested or mastered change starts from a lower performance baseline [Herold and Fedor 2008: 109].

– People have a finite capacity for change [Herold and Fedor 2008: 110].

– Organizational changes cannot be contemplated as independent, isolated events […]. All changes cannot be ‘priority one’ [Herold and Fedor 2008: 112].

In summary, when multiple ongoing change initiatives compete for money, time, effort, and leadership, each is less likely to be successful.

Common Cause of Failure #4: Counter-cultural change. Change that is inconsistent with an established culture which has a powerful impact on perception, cognition, affect, and behavior.

While it is possible to develop cultures that facilitate change [Smits and Bleicken 1997], the general function of organizational culture is to maintain stability [Kotter 2012]. Using an analogy, Herold and Fedor [2008] described the stabilizing effect of culture succinctly: “When change is introduced, if it is seen as an attack on basic and valued aspects of the organism, the immune system will go into rejection mode” [p. 108]. Organizational culture, the result of extensive group learning, deeply ingrained via experiences of success and failure, and largely unconscious by nature, as a stabilizing force is especially disruptive to change in two situations: – Transformational change. Here the change is focused on what Gersick [1991] called the underlying, or deep, structures of the organization that can only be changed during “periods of revolution” because they function to reinforce “the basic activity patterns that will maintain its existence” [p. 14]. In brief, to transform the organization is to change its underlying culture. And when the transformation is completed, if it is not anchored in a new culture, the transformed organization will soon revert back to its former way of perceiving, thinking, and behaving [Kotter 2012].

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

15

– Mergers and acquisitions. Here the change often relies on a due diligence process that describes the organizations involved in much detail but seldom takes more than a cursory look at culture. As Schein [1985] warned us: “[…] a culture mismatch in an acquisition or merger is as great a risk as a financial, product, or market mismatch” [p. 34].

In summary, change agents need to be conscious of the organization’s culture and to extend their thinking from organizations as systems to organizations as complex, paradoxical entities (even ambidextrous) that will often not be controllable; where possible, leveraging it to promote the desired change; and where counter-cultural, taking the necessary steps to change the culture as part of the overall change strategy.

Common Cause of Failure #5: Inadequately led change initiatives. A failure to provide the different types of leadership at different levels of the organization needed to complete the change.

As Anderson and Anderson [2001] demonstrated with their model, meaningful change often cascades through three levels of the organization, strategic, managerial, and operational, requiring a coordinated change initiative among the levels. Herold and Fedor [2008] extend that argument insisting that three levels of informed, active, focused leadership are required for successful change efforts: – Strategic leaders are needed to set a clear direction, get the organization into change mode, and communicate via multiple channels what is to be done and why.

– Implementation leaders get their marching orders from the strategic leaders but have much discretion regarding what is to be changed within their units and how to go about making the change (time-tables, resource allocation, assignments).

– Process leaders are tasked with how to make it happen. They seldom have the latitude enjoyed by implementation leaders; the strategy is set, the parameters of the implementation are set, their leadership is focused on making it operational by getting the membership in general to do what needs to be done.

If the leadership at any level is not adequate, the change fails: “[…] to assume that senior leadership is the only leadership level that really counts […] to ensure a change’s success as it cascades down through the organization tends to be foolhardy and helps explain why so many changes go away” [Herold and Fedor 2008: 42].

In this section we briefly reviewed five common causes of change failures.

Change failures cannot be eliminated but we contend the frequency of their occurrence can be reduced and the severity of their impact ameliorated. In the next section, we offer suggestions to help change agents be more successful.

16

Economics and Business Review, Vol. 1(15), No. 2, 2015

3. A brief “to-do” list for successful change agents From our perspective, the following suggestions will help organizational change agents improve their success rates. As we know, change agents may take different routes to solving similar problems [de Caluuwe and Vermaak 2003; Trautlein 2013]. Therefore, while our “to-do” list appears prescriptive, it is intended more as a way to focus attention on issues of importance than as a recipe for a fixed process.

Understand your organization and the environment in which it functions.

Organizations are unique so change initiatives need to be tailored to fit their realities. Change agents, whether internal to the organization or outsiders brought in specifically to initiate change, need to have an in-depth understanding of its history, culture, and current life cycle stage/status. They must start with the organization as it is, not with a general model of similar organizations in its industry. This can be especially challenging for outsiders brought in as change agents. However, outsiders often have an objectivity advantage assuming they begin the pre-change organizational analysis with an open mind and gather the needed information from a variety of reliable sources.

In addition to knowing the organization’s strategy, structure, operating systems, and people, it is important to know its capacity to learn and to change.

Does it have demonstrated knowledge management capabilities? What are the knowledge, skills, and abilities levels of its human resources? Have they demonstrated a capacity to learn and implement new performance-oriented behaviors? What is their experience with change? How much change turbulence are they experiencing at this time? What is their change readiness?

Several models reviewed in earlier sections of this paper can help change agents understand the organization more fully. Life cycle status can be assessed using Daft’s Organizational Stages of Development [1992] which was adapted from Greiner’s Five Phases of Organizational Growth [1972] and Quinn and Cameron’s synthesis model of multiple life cycle theories [1983]. Organizational culture can be assessed using Cameron and Quinn’s user-friendly Organizational Culture Assessment Instrument based on the Competing Values Framework [2011]. Additionally, the 34-item, easy to score, Readiness to Change Scale [Cawsey, Deszca, and Ingols 2012] has subscales depicting six key dimensions of interest to change agents: – Previous Change Experience, – Executive Support, – Credible Leadership and Change Champions, – Openness to Change, – Rewards for Change, – Measures for Change and Accountability.

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

17

An in-depth understanding of the organization to be changed minimizes the occurrence of several common causes of change-failure described in the earlier section.

Guideline #1: Start the change process with the organization where reality-based indicators say it is.

Know what to change; and since not everything can be changed at once, set priorities for a change agenda.

Nadler and Tushman’s Organizational Congruence Model [1980: 35–51; 1989] provides a clear blueprint to help change agents focus on what to change. In summary, their model says: Change whatever is needed in order to (a) keep the organization’s strategy in sync with the political, economic, social, and technical (PEST) factors in its changing environment; (b) keep the organization’s internal components aligned with its strategy; and (c) achieve congruence among all components of the organization in order to meet desired outcomes efficiently and effectively.

Herold and Fedor’s [2008] contribution to our second suggestion is to remind us that we cannot do everything at once, often cannot afford to do expensive idealistic change, and must guard against the demoralizing impacts of change turbulence. This reminder suggests the need for a well-planned, realistic change agenda. Where can we get the most benefit for our expenditures of time, talent, energy, motivation and limited fiscal resources to achieve change?

Which changes are most crucial to our short-term and long-term survival and prosperity?

Change agendas have the additional benefit of showing progress, “generating short-term wins” to put it in the context of Kotter’s eight-step model [2012].

Guideline #2: Change what is possible starting with the changes likely to produce the most benefit with the least risk and cost.

Understand the nature of the change you are undertaking: Planned vs. Unfolding.

The type of change initiated determines how it is to be planned and executed for success. Planned change (aka: transactional change), typically designed to restore stability and promote efficiency after some disruptive event (like transformational change) can and should be carefully designed and managed within defined time and budget parameters with everyone accountable for achieving the planned outcomes. Traditional management tools applied by experienced, competent managers should take much of the ambiguity out of the change process and minimize the unintended consequences associated with dynamic complexity.

Unfolding change, to use Anderson and Anderson’s [2001] term is essentially different from planned change: Unfolding change (aka: transformational

18

Economics and Business Review, Vol. 1(15), No. 2, 2015

change), disrupts stability and takes the organization into a future state where key parameters in the external environment are themselves changing while the unfolding change plays out. Unfolding change is a long-term process during which unpredictable and uncontrollable events happen in the PEST Factors of the external environment [Nader and Tushman 1980; 1989] requiring the change agent to revamp the change as it unfolds. Planned vs. Unfolding change have different time perspectives and require a different mindset to see them through [Gersick 1991; Thoms and Greenberger 1995].

Applying planned change techniques to unfolding change and vice versa, is a certain recipe for disaster. Managing short-term change takes a different mindset and skill set than leading visionary unfolding change that may take several years to accomplish.

Guideline #3: Do not confuse stability-focused change intended to restore equilibrium with change designed to put the organization on a fundamentally different course.

Know your change leadership team: Change must be led at all levels of the organization (Strategic, Managerial, and Operational) in a coherent, integrated, consistent manner.

Change typically cascades down through the organization with a strategic initiative launched at the executive level, translated into relevant structural and systems changes at the managerial level, and made operational by supervisory personnel and employees at the operational level. Like a relay race, the hand-offs are important, and are likely to involve fewer setbacks when the leaders at each level know and trust each other and have practice working together.

Multi-level change implementation leadership teams should start by selecting members with “influence power”; that is leaders and managers who have the respect of those being led through the change and have expertise/credibility consistent with the nature of the change being made. Such individuals are often the most talented members the organization has to offer and therefore also its busiest members. For the change to be led and managed successfully, the members of the change leadership team need to have designated time away from their normal responsibilities and a clear mandate to make the change effort a top priority.

Guideline #4: Don’t skimp when staffing the change leadership team; pick the best available members at each level and give them the time and resources needed to insure a successful change.

Carefully manage the transition from the current modus operandi to the new mode of organizational performance.

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

19

Cawsey, Deszca, and Ingols [2012] emphasized the important of managing the transition from the old to the new and put the challenges of doing so in perspective by observing: “Change management is about keeping the plane flying while you rebuild it.” [p. 326]. Their primary recommendation is to select an experienced transition manager to lead a knowledgeable and respected transition team and to make sure it is fully coordinated with the change leadership team. Our suggestion is to make the transition manager a key member of the change leadership team.

Lengthy transitions from the old to the new are especially challenging and often require carefully planned midpoint goals and milestones to prevent lagging motivation and comprehensive communication systems to keep people informed, thereby reducing their natural anxiety about such personal concerns as: “Will my pay be affected?”, “Who is my new boss?”, or “What is my new job description?” [Cawsey, Deszca, and Ingols 2012: 327].

Guideline #5: Manage the transition carefully with respect and compassion for the insecurities that major changes are capable of causing among members of the transforming organization. Don’t crash!

Concluding statement In this paper, we have attempted to share a practical perspective on how to improve the success rates for organizational change efforts. Our perspective relies on selected theories that seem to us to capture the essence of the change challenges that leaders and managers face as a normal part of their responsibilities.

We conclude here with statements from Herold and Fedor [2008] from their concluding chapter entitled: SMART CHANGE LEADERS – THEY GET IT! In it they contend that searching for one “right” approach to change or one “right” type of change leader is futile arguing that the only thing successful change leaders have in common is that their behaviors were appropriate to the realities of the situations they were facing. That is the core of our message as well. We hope we have provided a perspective about change that is consistent with the following description: Smart change leadership is about recognizing, diagnosing, tailoring, balancing, and otherwise adapting one’s hoped-for outcomes and implementation strategies to the realities of the situation. […] Ultimately, the savvy change leader juggles all elements of the situation, engages in parallel rather than serial processing, and arrives at decisions about what to change and what implementation process to use […] [Herold and Fedor 2008: 130–131].

20

Economics and Business Review, Vol. 1(15), No. 2, 2015

References Anderson, L.A., Anderson, D., 2001, The Change Leader’s Roadmap: How to Navigate Your Organization’s Transformation, Jossey-Bass, San Francisco.

Argyris, C., Putnam, R., Smith, D.M., 1985, Action Science: Concepts, Methods, and Skills for Research and Intervention, Jossey-Bass, San Francisco.

Bass, B.M., 1985, Leadership and Performance Beyond Expectations, Harvard University Press, Cambridge, MA. Beer, M., 2006, Big Shoes to Fill, Harvard Business Review, May: 43–54.

Bennis, W., 1989, Why Leaders Can’t Lead, Training & Development Journal, April: 35–37.

Burke, W.W., 2002, Organizational Change: Theory and Practice, Sage, Thousand Oaks, CA. Cawsey, T.F., Deszca, G., Ingols, C., 2012, Organizational Change: An Action-oriented Toolkit, 2nd ed., Sage, Los Angeles, CA. Chandler, A, 1962, Strategy and Structure: Chapters in the History of the American Industrial Enterprise, Press, Boston, MI. Daft, R., 1992, Organizational Theory and Design, West Publishing, St. Paul, MN. de Caluwe, L., Vermaak, H., 2003, Learning to Change: A Guide for Organization Change Agents, Sage, Thousand Oaks, CA. Gersick, C.J.G., 1991, Revolutionary Change Theories: A Multilevel Exploration of the Punctuated Equilibrium Paradigm, Academy of Management Review, vol. 16, no. 1: 10–36.

Grant, R.M., 2002, Contempory Strategy Analysis: Concepts, Techniques, Applications, 4th ed., Malden, MA. Greiner, L., 1972, Evolution and Revolution as Organizations Grow, Harvard Business Review, July–August: 37–46.

Griffin, R.W., Moorhead, G., 2007, Organizational Behavior: Managing People and Organizations, 8th ed., Houghton Mifflin, New York.

Herold, D.M., Fedor, D.B., 2008, Change the Way You Lead Change: Leadership Strategies that Really Work, Stanford University Press, Stanford, CA. Hitt, M.A., Ireland, R.D., Sirmon, D.G., Trahms, C.A., 2011, Strategic Entrepreneurship: Creating Value for Individuals, Organizations, and Society, Academy of Management Perspectives, vol. 25, no. 2: 57–75.

Kotter, J.P., 1990, What Leaders Really Do, Harvard Business Review, vol. 90, no. 3: 103–111.

Kotter, J.P., 2012, Leading Change, Harvard Business Review Press, Boston, MA. Mintzberg, H., 2004, Managers not MBAs: A Hard Look at the Soft Practice of Managing and Management Development, Berrett-Koeler, San Francisco.

Mohrman, S.A., Lawler, E.E., III, 2012, Generating Knowledge that Drives Change, Academy of Management Perspectives, vol. 26, no. 1: 41–51.

Nadler, D.A., Tushman, M.L., 1980, A Model for Organizational Diagnosis, Organizational Dynamics, Autumn.

Nadler, D.A., Tushman, M.L., 1989, Organizational Frame Bending: Principles for Managing Reorientation, Academy of Management Executive, vol. 3, no. 3: 194–204.

Neal, J.A., Tromley, C.L., 1995, From Incremental Change to Retrofit: Creating Highperformance Work Systems, Academy of Management Executive, vol. 9, no. 1: 42–54.

S.J. Smits, D.E. Bowden, A perspective on leading and managing organizational change

21

Porter, M.E., 1996, What Is Strategy? Harvard Business Review, November–December: 1–20.

Quinn, R.E., Cameron, K., 1983, Organizational Life Cycles and Shifting Criteria of Effectiveness: Some Preliminary Evidence, Management Science, vol. 29, no. 1: 33–51.

Rowden, R.W., 2001, The Learning Organization and Strategic Change, Advanced Management Journal, vol. 66, no. 3: 11–16, 24.

Rowe, W.G., 2001, Creating Wealth in Organizations: The Role of Strategic Leadership, Academy of Management Executive, vol. 15, no. 1: 81–94.

Schein, E.A., 1985, Organizational Culture and Leadership, Jossey-Bass, San Francisco.

Senge, P.M., 1990a, The Fifth Discipline: The Art and Practice of the Learning Organization, Double/Currency, New York.

Senge, P.M., 1990b, The Leader’s New Work: Building Learning Organizations, Sloan Management Review, Fall: 7–23.

Smits, S.J., Bleicken, L.M., 1997, Human Resource Management in a Culture of Change, in: Phillips, J.J., Holton, E.F., III (eds.), In Action: Leading Organizational Change, American Society for Training and Development, Alexandria, VA: 157–171.

Stacey, R.D., 1996, Strategic Management and Organizational Dynamics, 2nd ed., Pittman, London.

Thietart, R.A., Forgues, B., 1995, Chaos Theory and Organization, Organization Science, vol. 6, no. 1: 19–31.

Thoms, P., Greenberger, D.B., 1995, The Relationship between Leadership and Time Orientation, Journal of Management Inquiry, vol. 4, no. 3: 272–292.

Trautlein, B., 2013, Change Intelligence: Use the Power of CQ to Lead Change That Sticks, Greenleaf, Austin, TX. Vera, D., Crossan, M., 2004, Strategic Leadership and Organizational Learning, Academy of Management Review, vol. 29, no. 2: 222–240.

Vince, R., Martin, L., 1993, Inside Action Learning: An Exploration of the Psychology and Politics of the Action Learning Model, Management Education and Development.

Yukl, G., 2010, Leadership in Organizations, 7th ed., Prentice Hall, Upper Saddle River, NJ. Zaleznik, A., 1977, Managers and Leaders: Are They Different? Harvard Business Review, May–June: 74–81.

Economics and Business Review, Vol. 1 (15), No. 2, 2015: 22–51

Alternative configurations of firm-level employment systems: evidence from American companies1 Bruce E. Kaufman2, Benjamin I. Miller3

Abstract: This paper examines the concept of employment systems, describes alternative models of employment systems, selects one for empirical examination and uses data on HRM practices for several hundred American firms to test the predictions of this model. We find considerable support for the existence of distinct ESs but weaker support for this particular ES model.

Keywords: internal labour markets, employment systems, HRM configurations.

JEL codes: L23, M51, M54.

Introduction Labour resources are coordinated, allocated and priced in both external labour markets (ELMs) and internal labour markets (ILMs). The idea that ELMs take on different structural forms, such as competitive, monopsony and dual is conventional in labour economics and goes back to at least the 1930’s [Robinson 1933]. Not so well known or theoretically developed in economics, however, is the parallel idea that ILMs also exhibit distinct structural forms. These forms have a commonly-used term – employment systems (ESs) – and a modest literature on ESs has sprung up since the late 1980’s. Examples include Osterman [1987], Marsden [1999], Baron, Burton, and Hannan [1999], Appelbaum et al.

[2000], Toh, Morgeson, and Campion [2008], Ross and Bamber [2009], Keefe [2009], and Kaufman [2013]. Little of this literature, however, has so far spilled 1

Article received 4 December 2014, accepted 19 February 2015.

Georgia State University, Department of Economics, Atlanta, GA 30303, USA, Department of Employment Relations and Human Resources and Centre for Work, Organization and Wellbeing, Griffith University, Brisbane, AUS. Corresponding author, e-mail: bkaufman@gsu. edu.

3 Bennett Thrasher LLP, Atlanta, GA30339, USA. 2

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

23

over into organizational economics and personnel economics [see Gibbons and Roberts 2013; Grandori 2013; Lazear and Oyer 2013].

The contributions of this paper to the employment systems research program are several-fold. First, we provide in Section II a multidisciplinary survey of this literature, thus bringing to attention of economists several complementary but frequently overlooked research streams. Second, in Section III we take a detailed look at the ES model that, out of all the studies reviewed, provides the most detailed and explicit set of predictions concerning the shape and features of alternative ES configurations. The third contribution in Section IV is empirical analysis. We take the ES model and parameterize the key structural characteristics of its alternative configurations and then, using detailed data from a survey of HRM practices amongst American firms, determine how well the data match the model’s predicted set of employment systems. The data show strong evidence of alternative ES configurations but weaker support for the predictions of this specific model. Section V provides a summary and conclusion.

1. Employment systems: a literature review Models of employment systems assume that different firms choose different bundles of human resource management (HRM) practices; these studies then endeavor to identify the reasons for this and the configuration of the specific bundles. We searched the literature to find the ES theory that would be most amenable to empirical testing with the idea of using our detailed data set on firm-level HRM practices, along with cluster techniques, to see if the predictions well-matched the data. The model, therefore, preferably needed to (1) delineate a number of separately distinguishable employment systems (rather than just two) and (2) provide specific predictions about the kinds and amounts of HRM practices that are found in each system.

We found a model (described later), but the process was not easy. What we encountered was a diverse and fragmented literature spread across three separate labour fields with little attempt at review and synthesis. We decided, therefore, to modestly expand the literature review section of this paper to provide economists and other interested readers with the first-start of a synthetic overview [also see Kaufman 2013].

A useful place to start this review is with another review article. Short, Payne, and Ketchen [2008] provide a comprehensive review of the literature on organizational configurations. The subject of organizational configurations is a general management topic with a long research tradition. This line of research endeavors to identify if firms sort into distinct organizational forms based on strategy, structure, and goals and the factors behind this organizational differentiation.

A classic study is Burns and Stalker [1961] who distinguish between “organic” firms and “mechanistic firms.”

24

Economics and Business Review, Vol. 1(15), No. 2, 2015

These authors examine 110 studies published since 1993. They conclude that these studies divide along two principle theoretical axes. The first is models of organizational configuration that make the primary explanatory variable strategy or, respectively, organizational structure; the second is whether the configurations are universal or contingent with respect to external and internal environmental variables. The strategy group accounts for about one half of the studies in their sample and the organizational structure group includes the other half. They also note another divide: that is, some studies focus solely on identification of alternative organizational configurations while others go further and endeavor to relate different configurations to measures of organizational performance. In terms of method, the most frequently used empirical technique in configurational research is cluster analysis.

Extant research on employment systems – in labour economics called ILMs [Waldman 2013] and in management HRM architectures [Lepak and Snell 1999] – crosses disciplinary lines and academic fields of study. The three principle locations are industrial/employment relations (IER); labour process – including industrial sociology and critical management studies; and strategic human resource management. Personnel economics is not included in this review for although ILMs and individual HRM practices are much studied [e.g. Lazear and Oyer 2013], employment systems per se are not.

We summarize research in each area, going in chronological order of development and with a modest effort to put the entire stream of research in each field in an historical perspective, with the consequent necessity of some large leap-frogging at places. An outline of alternative ES models is given with respect to their main structural features and theoretical orientation; we cannot for reasons of space, however, go into an in-depth discussion of underlying theoretical determinants or empirical issues or applications. The review is focused on firm-level (micro) studies of ESs. A separate national-level (macro) literature, such as in varieties of capitalism, is not included [e.g., Hall and Soskice 2001; Katz and Darbishire 2002; Hendry 2003] is not included.

Industrial/employment relations The first labour field is industrial/employment relations (IER), a term which also includes the institutional approach to labour economics. The academic founder of the American IER field, institutional labour economist John Commons, is probably the originator of ES typologies. In his book Industrial Goodwill [1919] Commons identifies five alternative “theories of labour” – commodity (demand/supply), machinery (scientific management), goodwill (commitment), public utility (a publicly protected resource), and citizenship (industrial democracy) – and discusses how each theory leads to a different model of people management and associated practices. Commons’ explanation of these systems is discursive but one finds sprinkled here and there factors that

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

25

he points to as determinative. The commodity model, for example, is a good fit for low-skill jobs where the work is standardized and easily monitored; the goodwill model, on the other hand, is good fit where jobs require higher level and more intangible-type skills and where cooperation and good citizenship behavior have particular value. Notably, Commons uses the term “competitive advantage” [p. 74] to describe why it pays firms to invest the extra funds necessary to create and operate a goodwill HRM system.

For the next major ES contributions in industrial/employment relations we fast forward to the 1950’s. One comes from economist Clark Kerr and the other from John Dunlop. Kerr [1950, 1954] first divides labour markets into “structured” and “unstructured.” The latter is equivalent to Commons’ “commodity model” where undifferentiated labour is traded back and forth, turnover is high, demand and supply determine wages and other terms and conditions of employment, and employer HRM practices are very bare-bones. The former correspond to what Kerr calls “institutional” labour markets that are structured by various rules, norms, and organizational features created by employers, unions and/or government. He divides structured labour markets into two generic ES types: “communal ownership” and “private property.” The former is exemplified by an occupational labour market or industry organized by a craft union where jobs require well-recognized skills but the skills are general and therefore portable (e.g., in nursing, construction); the latter is exemplified by a factory or company organized by an industrial union where job access is regulated by seniority and skills tend to be company-specific (e.g., in an auto plant). Of the two, the private property model has the more extensively developed internal labour market (ILM) and HRM system since the longevity of the employment relationship makes all aspects of HRM more important to organizational performance.

Dunlop’s contribution to the ES literature is the concept of an “industrial relations system.” In his well-known book Industrial Relations Systems [1958], Dunlop defines an IR system as “the complex of interrelations among managers, workers, and the agencies of government” where “the parts and elements are interdependent and each may affect other elements and outcomes of the system as a whole” [p. 13]. The primary factor Dunlop focuses on as determinative of the shape of ESs is the rules and regulations negotiated by unions and companies through collective bargaining.

IER theorizing on employment systems next moves to the 1970s and 1980s and two more contributions. The first is Doeringer and Piore [1971]. Their contribution is to delineate four factors that create ILMs: skill specificity, onthe-job training, unions, and customary law. Higher levels of these four factors promote more internalization of labour coordination and, hence, a more HRM intensive ES. The second study is by Osterman [1987]. He identifies four dominant ESs: industrial, salaried, craft, and secondary. The nature of these different ESs is evident from their names; what distinguishes Osterman’s paper is that he

26

Economics and Business Review, Vol. 1(15), No. 2, 2015

then links different ES models to different configurations of four HRM practice areas – job classification, deployment (staffing), security, and wage rules – and identifies five factors as the chief determinants of ES selection – goals of the firm (cost minimization, flexibility, predictability), production technology, social technology, labour force characteristics, and government policies.

Our review of the IER stream ends with recent contributions from the 1990’s to date. Four contributions merit brief mention [also see Applebaum et al. 2000; Orlitzky and Frenkel 2005]. First is Arthur [1992]. His article is noteworthy because it integrates the IR system idea; the role that strategy plays in shaping ESs; and the popularly-called high performance model of work organization.

He looks at steel minimills, differentiates between “cost minimization” and “product differentiation” business strategies [Porter 1980], and then examines if they map into a “command and control” or “high commitment” HRM/IR system. Cluster analysis reveals the ESs in the minimills sort into distinct types in line with the two strategies.

The second contribution is Marsden’s A Theory of Employment Systems [1999]. The book is almost entirely a work of theory and is based on ideas from institutional economics and, in particular, works by Coase [1937], Simon [1951] and Williamson [1975]. Marsden identifies the constraints on behavior that are necessary for the employment relationship to be a viable economic proposition for both employers and workers; from these constraints he identifies four permutations of job design and skill development in firms’ production systems, and in the last section of the book examines specific HRM practices that go with each production system. His theory indicates ILMs and “transformed” (high performance) HRM systems develop when transaction costs are high and knowledge requirements are high and firm-specific – conditions created in turn by factors such as interdependent job tasks, tacit knowledge, learning on the job, and difficulty of monitoring job performance.

A third contribution is Kaufman’s [2004, 2010] ES model. It is similar to Marsden in that it uses institutional economic theory and, in particular, the transaction cost (TC) concept. He demonstrates that in a world of very low TC organizations dis-agglomerate into small units and HRM systems are “externalized” and “simple;” in a world of very high TC organizations agglomerate into very large units and employment systems become “internalized” and “HRM intensive.” Variation of five variables, in turn, causes different ES permutations between these polar opposites: the degree of bounded decisionmaking, interdependence in utility functions, interdependence in production functions, indivisibilities and gaps in property rights and legal restrictions on trade by the sovereign.

Finally, note must be taken of a mini-symposium on employment systems in the telecommunication industry featured in the October 2009 issue of Industrial and Labour Relations Review. The three papers [Doelgast 2009; Ross and Bamber 2009; Keefe 2009] are primarily empirical; a common theme that

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

27

comes out of all three, however, is that theory must pay attention to the issue of complementarity and fit between the external market environment and the internal structure of ILMs. In particular, ESs in telecommunication companies are becoming considerably more decentralized, flexible and open in response to more turbulent markets.

Labour process The second stream of research on employment systems comes from the labour process (LP) field. It draws from diverse disciplines but is centered in neo-Marxist economics/sociology, industrial and organizational sociology, and critical management studies. The roots of ES research in LP are found in Marx, Weber and (to a lesser extent) Mayo; nonetheless, explicit work on ES models did not begin in earnest until the 1970’s. By all accounts [e.g., Wardell, Steiger, and Meiskins 1999], the pioneering LP study was by Harry Braverman in his book Labour and Monopoly Capital [1974]. Braverman’s perspective was avowedly radical, Marxist and class-based. In the intervening three decades many writers in LP have softened or abandoned all of these three premises; the overall perspective remains, however, critical of neoclassical economics and orthodox management [Thompson and Harley 2007].

The starting point in this line of theorizing is Marx’s [1847/1935] distinction between labour and labour power [Burawoy 1979]. Labour is a commodity input firms rent for a certain amount of money per hour; the labour, however, is in the form of a person and the person only contributes to production and profit by providing labour power – that is, physical, mental and emotional effort.

Thus, the goal of the employer is to pay as little as possible for the labour input and then get the most effort possible out of it; the goal of the worker, however, is just the opposite – to get the best compensation and benefits possible at the least cost in terms fatigue, boredom, and other causes of disutility. Employers and workers, in this perspective, are in a constant struggle (or “game”) where workers individually and collectively seek to evade, resist or minimize the work demands of employers and employers are continually searching for ways to more effectively extract labour power through more effective motivation, discipline and supervision.

The next key concept in LP is control and, in particular, regimes of control [Friedman 1977; Edwards 1979], Since workers evade, resist and minimize all the aspects of work that are a source of disutility (not only the tiring/boring aspects of the job itself but also taking orders, reporting on time, providing good customer service, etc.), the task of employers is to build into the production and HR systems a regime of control that maximizes the quantity and quality of labour power provided in the most cost effective way. A “regime of control,” however, is just a critical/Marxist term for an “employment system” and, indeed, objectively viewed the two are largely equivalent. Since there is evidently

28

Economics and Business Review, Vol. 1(15), No. 2, 2015

no single best control regime for all employment situations, researchers in the LP tradition are therefore led to theorize about different control regimes, their constituent parts, and the contextual factors that determine when one regime is better than others.

In the IER literature the most basic distinction in thinking about ESs is between external and internal labour market modes of coordination; in LP the most basic distinction – in keeping with Marx’s emphasis on the materialist basis of social relationships – is between Fordist and post-Fordist production systems. The contention is that the nature of the production system is the most important determinant of the complementary HRM system since the latter exists to serve the needs of the former. Thus, differences in production systems map into differences in ESs. Until the early 1980s, the core production system in advanced industrial countries was a Henry Ford-inspired model of mass production; surrounding it were other production systems outside manufacturing, in smaller firms and peripheral sectors. Since the early 1980’s, the mass production model has given way to a post-Fordist system of flexible specialization [Piore and Sabel 1980].

In the Fordist era of mass production, LP researchers identify several distinct control regimes (ESs). They are “simple,” “technical” “craft,” and “bureaucratic” [Edwards 1979]. Many smaller or less technologically dynamic firms continue to use variants of these today.

In a classic mass production operation, such as an auto plant, employers opt for a technical control system. The technical system gets its name because control and extraction of labour power is predominantly organized and enforced through the technology of production. That is, the assembly line, machine tools, conveyor belts and other parts of a mass production operation effectively control the pace of work while the quality of work is controlled by finely subdividing parts and tasks and putting in place supervisors to act as monitoring agents of management. The HRM system, in turn, complements the technical control system; that is, it tends to be a non-strategic function (except in the union avoidance area) that does large-scale but routinized hiring and placement of blue-collar workers, handles a variety of administrative and transactional personnel activities (e.g., job classification, payroll, benefits), and helps line managers in discipline and discharge.

Other firms use a simple, craft or bureaucratic form of ES. The simple control system is best for small firms where direct coordination and supervision by the owner or manager is possible. The HRM system is mostly informal and covers the basics of payroll, hiring and legal compliance. The craft system arises in areas where occupational or craft skills are strategic to the production system; the bureaucratic system arises where the production system is large-scale but jobs have more autonomy, the tasks require more discretion, and there are well-developed lines of upward progression (e.g., in an insurance company or university). The craft system is primarily controlled by peer monitoring, skill

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

29

qualifications and professional norms and HRM practices are formal but largely administrative and transactional; the bureaucratic system is typically associated with long-term employment relationships and a well-develop ILM and, hence, HRM is not only formal but takes on greater strategic importance and invests more on careful selection, training, and employee relations.

The Fordist mass production model dominated the core of industrial economies into the 1970’s; after that it was gradually displaced – particularly in manufacturing – by a post-Fordist production model of flexible specialization.

Traditional technical and bureaucratic control systems were particularly challenged in industries facing global competition and new production technologies [Rubery and Grimshaw 2003: Thompson and Harley 2007] and in their place was developed an alternative model called a high performance work system (HPWS). The HPWS is an alternative control regime that melds parts of the craft system (e.g., employee teams) and technical system (e.g., electronic monitoring of production and quality) with new components derived from a commitment form of employment. In effect, control is partly reconstituted so employees internalize management’s performance objectives through mutualgain pay systems, egalitarian organizational cultures, and participation forums.

The ES part of the HPWS is distinctive because it is very HRM intensive in the sense of considerable investment in sophisticated selection and placement procedures, extensive training, widespread and formalized employee involvement programs, and considerable attention to maintaining positive employee relations.

Strategic human resource management A third literature stream on employment systems comes from strategic human resource management (SHRM). It is the latest to be developed, emerging in the 1990’s.

SHRM grew out of two important developments. The first was the transition in the late 1970’s-early 1980’s from the traditional personnel management (PM) model and to a new human resource management (HRM) model.

Some writers perceive that HRM is largely a repackaged version of PM [Strauss 2001]; many others, however, regard HRM as a new and different philosophy and method of people management. At the forefront of the “new model” interpretation were a group of Harvard professors, including Michael Beer, Paul Lawrence, D. Quinn Mills, Burt Spector, and Richard Walton. In an interpretation largely parallel with that the of the LP group, Walton [1985] claims that traditional PM and IR (together PIR) rely on a model of command and control while the new HRM relies on a model of employee commitment and involvement. Beer and Spector [1984] worked out the logic of this bifurcation in much greater detail. In their typology, PIR assumes a conflict of interest; is based on a Taylorist production system of narrow and repetitive jobs, top-down man-

30

Economics and Business Review, Vol. 1(15), No. 2, 2015

agement, and tight supervision; and takes an administrative/reactive approach to people management. HRM, on the other hand, assumes a unity of interest; approaches people management from a proactive perspective; and is based on a more flexible, team-oriented, egalitarian and mutual-gains kind of production/HRM system. Hence, in this popular view HRM covers the generic field of people management but at the same time connotes an orientation toward a human capital, commitment and high involvement approach.

The second stream of influence occurred in the late 1980’s-early 1990’s and was the emergence of strategic human resource management as a subfield of HRM. Pioneered by people such as Fombrun, Tichy, and Devanna [1984] and with some carry-over from the IER literature [e.g., Kochan, Katz, and McKersie 1986], this research brought the strategy concept from general management and applied it to choice of individual HRM practices and entire HRM systems.

In particular, the idea is that HRM practices and systems need to align with the organization’s business strategy and with its internal structure and capabilities.

Further, the HRM practices need to align with each other so they fit together and generate maximum synergies [Bamberger and Meshoulam 2000]. These ideas became known as vertical and horizontal fit.

We have here the beginning of distinct SHRM typologies for employment systems. The most basic distinction in SHRM is between command/control and commitment/involvement people management systems. A command/ control strategy relies on a mix of technical and bureaucratic HRM practices (e.g., narrow job classifications, straight-time pay, transactional personnel activities) and a commitment/involvement strategy relies on HPWS-type employment practices (broad and autonomous jobs, pay-for-performance, employee participation).

As SHRM evolved, new ideas and typologies have led to more variegated models of employment systems. For example, Delery and Doty [1996] argue that SHRM models separate into three basic groups: universalistic, contingency and configurational. The universalistic perspective argues that one particular ES is everywhere best practice and this ES is associated with high commitment/ high involvement HRM. Strategic choice here is simple because an HPWS-type ES is always the best; usually argued based on propositions from the resourcebased view (RBV) of the firm. The contingent perspective argues that the best performing ES is contingent on various internal and external contextual factors, such as firm size, industry, production technology, and state of the economy.

Kaufman [2010] distinguishes between “weak” and “strong” contingency cases: weak contingency moderates but does not reverse the positive performance effect of high performance HRM practices; in strong contingency the contextual factors sometimes make an HPWS system the “low performing” option (as in small firms or an economic depression). The configurational perspective looks to see that the bundle of HRM practices in an employment system fit together for maximum complementarity and synergy. Thus, if a firm adopts an

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

31

HPWS the performance pay-off is hypothesized to be greater when employee involvement, pay-for-performance and job security are used together; while using the first two but hire-fire methods for the third would greatly reduce the system’s performance.

The universalistic perspective in SHRM thus argues that HRM practices should be converging to an HPWS-type employment system; the weak contingency perspective argues that an HPWS employment system is dominant but will take different second-order shapes and configurations depending on contextual factors; and the strong contingency case suggests employment systems will sort into diverse and perhaps polar opposite configurations.

SHRM theorists taking a strong contingency perspective have endeavored to fill-out alternative ESs that include an HPWS but also different forms. Two influential examples are Delery and Doty [1996] and Lepak and Snell [1999]. Delery and Doty contrast two polar opposite employment systems that correspond closely to those found in industrial/employment relations; that is, a “markettype” and “internal” ES. The former uses primarily market-based pay, features hire and fire staffing methods and provides little formal employee voice, while the latter provides forms of organizational gain-sharing, employment security, and formal voice mechanisms. Lepak and Snell [1999] develop a model of employment systems called HRM architectures. They identify two central attributes that distinguish human capital across firms, “value” and “uniqueness” and based on transaction cost and RBV considerations derive a four-fold typology of ESs: Commitment, Market-based, Compliance, and Collabourative.

The first three ESs correspond to versions already encountered (e.g., commitment = HPWS; compliance = technical/bureaucratic; market-based = simple/ external labour market); the “Collabourative” ES is a hybrid ES where firms hire the human capital in the form of products or services produced by employees in network or alliance firms.

A final ES example from the SHRM literature is the recent typology developed by Toh, Morgeson, and Campion [2008; also see Sheppeck and Militello 2000; Youndt and Snell 2004; Bae and Yu 2005; Kinnie, Swart, and Purcell 2005; Tsai 2006]. They identify five ES’s: cost minimizers, contingent motivators, competitive motivators, resource makers, and commitment maximizers. They then seek to match each of these ES types to a particular configuration of HRM practices based on four key HR functions: staffing, development, reward and evaluation. Consistent with much of the SHRM literature, these authors identify the goals/strategy of the firm as a key determinant of ES choice. For example, firms that seek competitive advantage through a low cost strategy are led to adopt a minimalist HRM system while those pursuing a high commitment strategy adopt an HRM intensive system. Cost minimizers and commitment maximizers, therefore, anchor the opposite ends of the HRM spectrum. The other three ES models fall between these end points: Contingent motivators, for example, use considerable incentive types of pay; competitive motivators

32

Economics and Business Review, Vol. 1(15), No. 2, 2015

purchase human capital from the labour market and use pay to elicit work effort; and resource makers invest more in training, development and empowerment. This study, however, is relatively more advanced than many others in the SHRM literature because it then incorporates an additional degree of fit by matching ES systems to other contingent factors, such as organizational structure and organizational values.

Synthesis This literature review reveals both commonalities and differences in research on employment systems. First are four commonalities.

Looking across fields, one shared characteristic is that researchers in IER, LP, and SHRM all start with a two-way model of employment systems. In each case, these represent polar opposites: external versus internal in IR, mass production versus flexible specialization in LP, and control versus commitment in SHRM. A second similarity is that researchers then introduce additional contingencies into the theory and derive a more nuanced and complex typology of employment systems, typically broadening the mix to somewhere between three and five distinct ESs. In IER, an important contingency is the goals and structure of labour unions (e.g., craft versus industrial unions); in LP an important contingency is the nature of technology (e.g., tight versus loose control of work pace); and in SHRM a key contingency is organizational structure (e.g., mechanistic versus organic).

Although researchers often use different labels to identify individual employment systems, a third commonality is that behind these labels are often fairly generic models. For example, in IER, LP and SHRM one finds some type of “mechanistic” ES (e.g., “machine” in IR, “technical control” in LP, and “cost minimizer” in SHRM); also prevalent is some kind of “occupational” ES (e.g., “communal ownership” in IR, “craft” in LP, and “collabourative” in SHRM).

A fourth commonality is that at a broad level these different ES models lead to fairly commonly predicted HRM configurations. For example, a market-based system has the fewest and least developed formal HRM practices; a mechanistic/control system has an intermediate range of HRM practices with emphasis on policing, monitoring and administrating (e.g., clearly specified tight job classifications, narrow task-specific training, straight-time pay or incentive piece rates); and a commitment/high performance system uses the largest breadth/depth of HRM practices and a set of practices aimed at building organizational motivation and capabilities (e.g., mutual-gain pay, employee participation, job security).

Now for the differences in ES research. We mention two that seem most important. One is that researchers in each field use a different theoretical construct as the principal tool to differentiate amongst employment systems. That is, in

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

33

IER the central construct is the labour market, in LP the central construct is the production regime, and in SHRM it is strategy. Second, researchers differ in the emphasis given to external versus internal determinants of ES structure.

IER researchers locate on the external end since they emphasize markets, laws, unions, culture, and other factors in the external environment of firms; SHRM researchers are on the internal end since they emphasize management goals, employee motivations and skills, organizational structure and other factors internal to firms; and LP researchers fall somewhere in the middle.

2. Begin’s model of employment systems This literature review provides a broad-based context for our empirical examination of one particular ES theory. We chose this theory because to the best of our knowledge it is the most fully developed in two important respects. The first is that it delineates more than two alternative employment systems and, second, gives the most detailed and specific predictions about the bundle of individual HRM practices that attaches to each. This ES model is contained in the work of James Begin, particularly his book Strategic Employment Policy [1991; also Begin 1993, 1997]. It has been cited and utilized in other more recent studies, such as Verberg, Hartog, and Koopman [2007].

Before getting into the specifics of Begin’s model, it is useful to locate it in terms of the models described in the previous literature review. Begins’ model has elements of IER, HRM and (most loosely) LP. He conceives of alternative ES structures as alternative ILM configurations, as in IER and labour economics; as in HRM he emphasizes that bundles must have good strategic fit with internal organizational characteristics and the external environment; and as in LP one of the chief functions of an ES is to coordinate and control the internal division of labour in firms. However, these are best viewed as complementary features to what is the main intellectual wellspring of Begin’s ES theory – alternative organizational structures. Recall from the literature review section that Short, Payne, and Ketchen [2008] surveyed 110 studies of organizational configurations (of which the ES is one component) and found they divided roughly in half in terms of one set emphasized the defining role of strategy and the other emphasized the role of organizational structure. It is the latter wing in which Begin’s work is based. In particular, Begin’s model is built on the typology of organizational structured developed by well-known management theorist Henry Mintzberg and, specifically, his influential book Structure in Fives: Designing Effective Organizations [1983].

Now let’s get into the specifics. The starting point for Begin’s ES theory are three connected propositions: (1) an ES is a necessary component of every organizational design; (2) firms choose the ES that best fits their organizational architecture so both are well aligned with each other and with effective accom-

34

Economics and Business Review, Vol. 1(15), No. 2, 2015

plishment of the organization’s objectives; and (3) diverse internal and external contingencies across firms lead them to design distinctly different organizational architectures which lead, in turn, to the design of distinctly different employment systems. The remainder of his book is devoted to elabourating and developing these three propositions.

Although Mintzberg [1983] identified six organizational configurations, two are mixtures or hybrids (e.g., a “divisionalized” organization) and Begin therefore drops them, yielding a core of four generic configurations. He argues that these four configurations are the product of two fundamental intersecting forces, one external to the organization and one internal. The first is the degree of volatility in the external market environment facing the organization, the second is the complexity of the technical production system internal to the organization. Mintzberg divides the market volatility dimension into “stable” and “dynamic” and the technical production system into “simple” and “complex,” thus yielding a two x two matrix with four cells. Each of these four cells yields a distinct organizational configuration [Begin 1991, Table 2-2; also see Verberg, Hartog, and Koopman 2007, Table 2]. They are: – Simple: simple/low-cost production technology, dynamic/competitive environment, smaller size, direct and often personal control/coordination from the top.

– Machine: a larger-scale but relatively routinized and sub-divided production technology, a moderate-to significant stable/planned environment, emphasis on narrow skills and task proficiency, tight top-down coordination through formal rules and supervision.

– Professional: a more loosely structured/regulated but larger-scale production technology utilizing complex/intangible skills and knowledge; typically a more stable/predictable environment; greater decentralized and discretionary coordination/control; formal rules complemented by professional/ social norms.

Adhocracy: a complex, knowledge intensive, and human-centered production technology, a rapidly changing environment fueled by innovation and learning, competitive advantage based on quality, speed and service, and flatter, looser and decentralized coordination and control.

Begin’s next step is to match an appropriate set of HRM practices to each organizational type. Thus, the key question is: for HRM activities, such as job classification, selection, training, compensation, performance appraisal and voice, what particular type of HRM practice best aligns with the organization’s structure? He considers a host of external and internal contingencies and contextual factors, including not only the environmental volatility and technical features of production focused on by Mintzberg but also characteristics of the command and control system (e.g., formal versus informal, horizontal versus vertical), the nature of jobs (e.g., narrow versus broad; simple versus complex), and numerous other factors. The end-product is four distinct configurations

35

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

of HRM practices that form, respectively, a simple ES, machine ES, professional ES, and adhocracy ES. Table 1 shows the four different ES types across the top and a group of nine HRM practices and organizational characteristics on the left-hand side, as taken from Begin [1991].4 It is the bundle of these nine HRM practices and characteristics (P&C) that collectively define each ES. Note that the nine P&C’s include many of the core sub-areas of the personnel/HRM function; also note that Begin’s model includes several other practices, such as job design, that are more engineering and less HRM oriented, leading us to therefore omit them from further consideration. Lastly, observe that the predicted HRM practices are for core employees of the organization.

Table 1. Begin’s four employment systems and predicted HRM characteristics HRM Characteristic/ES

Simple

Machine

Professional

Adhocracy

Recruiting/Staffing

LI

LF

LI

EF

Training/Development

LI

LI

LI

EF

Benefits/Rewards

LI

EF

EI

EF

Performance Management

N/LI

LF

LF

EF

Participation/Voice Mechanisms

Few

Few

Moderate

Many

Work Force Size

Small

Large

Varies

Varies

Unionization

No

Yes

No

Yes/No

Strategic Involvement

N

N/LI

N

EF

Little

Much

Degree of Formalization

Intermediate Intermediate

N – none, LI – limited (informal), LF – limited (formal), EI – extensive (informal), EF – extensive (formal).

Source: Table 2-2 and Chapter 5 of [Begin 1991].

As indicated earlier, one reason we chose Begin’s ES model is because more than other writers he makes predictions about how these specific, measureable types of HRM P&C vary across organization types.5 Unfortunately, these predictions are not made using a common metric but are a mix of quantitative, qualitative and descriptive measures. These different metrics are displayed in 4 We coded Degree of Formalization for the Professional and Adhocracy ESs as “Intermediate” since in Table 2-2. Begin lists them as “Little” but then in the Chapter 5 discussion of each ES (e.g., p. 108; p. 117) he suggests a much higher degree of formalization, albeit one that in the Adhocracy case exhibits considerable flexibility.

5 Verburg, den Hartog, and Koopman [2007] considerably expand Begin’s list of HRM practices for each ES; we follow Begin here partly to remain true to his original work and partly because our data set does not contain many of these other practices.

36

Economics and Business Review, Vol. 1(15), No. 2, 2015

the individual cells of Table 1. For example, the level/type of staffing for the four ES’s is, respectively LI (little/informal), LF (little/formal), and EF (extensive/formal), while for participation and voice mechanisms the metric is Few, Few, Moderate, Many. These metrics are measuring a mix of rules and activity levels; that is, “little/informal” means that the activity of staffing and the rules governing it are modest sized, not greatly detailed and offer significant room for discretion. Further, when an activity such as training is marked as “little/ informal” this does not mean that the employees are not necessarily highly trained; it does mean, however, that the employees have obtained the training elsewhere (e.g., a medical school) and the organization itself provides only modest and loosely structured training.

Before moving further toward empirical analysis, it is useful to examine the nature of the four ESs depicted in Table 1 and offer a brief summary.

The simple ES is exemplified by a restaurant, budget hotel or dry cleaners.

Begin’s model predicts such organizations should fall in the “no/little” and “informal” category for all nine HRM practices. Thus, recruiting/staffing and performance appraisal (development) are little and informal, HRM has no integration with organizational strategy, size of labour force is small and unionization is largely non-existent. These HRM characteristics seem to match with what we know about these kinds of firms.

A machine ES, on the other hand, is exemplified by a railroad, traditional auto assembly plant, newspaper, or government administrative agency. The machine ES configuration is similar to the simple ES in two cells (training, participation/voice), close in a third (strategy integration) but different in the other five. A machine ES uses limited and informal staffing methods, extensive and formal benefits, limited and formal training/development, few participation/voice mechanisms, employs a large labour force, and is more likely to be unionized. Again, these HRM practices seem in line with expectation.

The professional ES is found in organizations such as hospitals, large law firms, commercial building architects, and universities. Like a simple ES, the staffing/recruitment function for core employees is modest but also more formalized (e.g., selection of new nurses or professors is formalized but not technically complex), formal in-house training also ranges from little-to-modest (an average of values in Begin’s Table 2–2 and 5–6) and most professional ES’s are not good candidates for unions – at least amongst core employees. The professional ES is different, however, by having extensive benefits and greater employee participation/voice.

The adhocracy ES is exemplified by consulting firms, high-tech entrepreneurial firms, software design companies and firms with “high involvement” production systems. This type of ES is the most HRM intensive with respect to dollars spent on human capital although not necessarily in terms of formal HRM programs and practices. Because skills, motivation and creativity are important and the technology of production and tight supervision cannot tightly

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

37

regulate the work process, adhocracies find it necessary to invest substantial resources in carefully selecting, developing, rewarding and retaining talent.

Often, however, these HRM activities are decentralized, flexible and implemented outside a formal HR department. Adhocracies also provide the most participation/voice mechanisms for core employees, which in turn makes them poor candidates for unions.6 Because human capital is a major source of competitive advantage, HRM policies/practices are also more tightly and explicitly integrated with organizational strategy.

3. Empirical analysis A central object of this paper is to offer evidence pro and con on the existence of distinct groupings of firms based on differences in HRM practices. To structure this investigation and provide an opportunity to take the analysis even deeper, we have utilized the detailed theory of ES formation and structure developed

by Begin. We now proceed to investigate both issues in a four step process.

These four steps involve some complex and perhaps at places tedious manipulations and measurements; we do our best, however, to guide readers through them in an understandable and transparent fashion. Also, we here state and acknowledge an important but inevitable shortcoming of our empirical analysis; that is, in translating Begin’s theoretical concepts and predictions into a format capable of testing with data there are necessarily some “slips between the cup and the lip” due to imprecise measurement and dividing-up continuous variables into discrete categories.

Metric conversion The first step is to transform all the cell entries in Table 1 into a common metric. We start with the first three HRM practices (recruiting, training, benefits).

These three HRM practices are measured on a scale that shows whether the practice is predicted to be non-existent (N), limited/informal (LI), limited/formal (LF), extensive/informal (EI), or extensive/formal (EF). The measures N, LI and EF correspond in a relatively straightforward way to low, moderately low and high usage, respectively. The two remaining categories LF and EI, although less clear-cut, seem reasonably approximated by a ranking of moderate and moderate high.

We next turn these rankings into numerical scores. Each ranking is coded with a score covering a two point range (e.g., 1–2) in recognition that a score such as “low” has some internal variance. With a two point range, the five categories distinguished by Begin sort into a numerical range from 1–6 (6 = high6

Begin, however, gives a Yes score to the union variable for administrative support employees in an adhocracy [p. 117], leading us to put a Yes/No entry in this cell.

38

Economics and Business Review, Vol. 1(15), No. 2, 2015

est). Accordingly, N is given a score of 1–2, LI gets 2–3, LF gets 3–4, EI gets 4–5 and EF gets 5–6.

Next are the other six HRM P&C in Table 1. Each is again converted into a numerical score using a 1–6 ranking. First, participation and voice mechanisms are ranked in Table 1 as Few, Moderate, or Many. This three-way ranking converts neatly into Few = 1–2, Moderate = 3–4 and Many = 5–6. Next, the size of the work force, the level of unionization and the degree of formalization variables are ranked using binary measures, such as Small or Large, No or Yes, and Little or Much. These variables are, accordingly, assigned a score of 1–3 for Small, No and Little; 4–6 for Large, Yes and Much. A combination of Little/Much (“intermediate”) gets 3–4. The two cell entries “Varies” (work force size) in Table 1 are coded 1, 2, 3, 4, 5, 6. The strategic involvement variable uses the N, LI, LF, EI and EF ranking, as with the three HRM practices discussed above, and is similarly converted to the 1–6 scale. Finally, degree of formalization is relatively straightforward: Little is coded 1, 2, 3; Intermediate is 3, 4; and Much is 4, 5, 6. The results of this metric conversion are depicted in Table 2.

Table 2. Quantifying the ES characteristics (1–6 scale) HRM Characteristic/ES

Simple

Machine

Professional

Adhocracy

Recruiting/staffing

2, 3

3, 4

2, 3

5, 6

Training/development

2, 3

2, 3

2, 3

5, 6

Benefits/rewards

2, 3

5, 6

4, 5

5, 6

1, 2, 3

3, 4, 5

3, 4, 5

5, 6

1, 2

1, 2

3, 4

5, 6

Work force size

1, 2, 3

4, 5, 6

Unionization

1, 2.3

4, 5, 6

1, 2, 3

2, 3, 4

1, 2

1, 2, 3

1, 2

5, 6

1, 2, 3

4, 5, 6

3, 4

3, 4

Performance management Participation/voice

Strategic involvement Degree of formalization

1, 2, 3, 4, 5, 6 1, 2, 3, 4, 5, 6

Data set We use data from the USA provided by the Bureau of National Affairs (BNA) from its 2005 and 2006 reports, HR Department Benchmarks and Analysis.

These data come from annual surveys of hundreds of American companies in which they are asked a wide range of questions about the structure, organization and strategic involvement of their HR function; the use of and expenditure on a variety of individual HRM practices, and other information such as industry, sector, work force size, and unionization. The BNA surveys provide a particularly detailed and in-depth information source of firm-level HRM

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

39

practices; they are also the most recent data source available. These data for the most part come from company administrative records and are not self-reports, thus boosting confidence in their reliability.

The 2005 and 2006 surveys include a total of 641 observations. Some observations are for entire firms, others represent an autonomous business unit (e.g., a division or subsidiary). Only firms (or sub-units) with a minimum employment size of twenty-five are included. We removed all duplicates between the two years and all observations with missing information on the HRM P&C variables in Table 2. The remaining observations are 264. The sample of firms is distributed by industry sector as follows: manufacturing (23%), non-manufacturing (43%), and public and non-profit (34%). Twenty-seven percent had some collective bargaining representation.

A key advantage of the BNA data is that it asks not only about the use of specific HRM practices but also on the dollar expenditure on each practice. In particular, the survey instrument asks each respondent to provide “your HR department’s 2004 expenditures” and then in the next question asks them to “approximate the percentage” of this expenditure that goes into ten discrete areas of HR practice (training, compensation, etc.).

A problem in other studies is that they only have data on the presence of the HRM practice, but not the expenditure level. Two firms may both report they utilize an HRM practice (e.g., employee participation) yet one is simple and bare-bones (a suggestion box) and another is extensive and state-of-theart (a plant productivity/quality council). A yes/no presence measure gives an equal value to both the bare-bones and advanced systems which, evidently, may be quite misleading. An expenditure measure, on the other hand, is likely to much better capture differences in the depth and formalization of the practice.

Clustering technique The statistical method used to perform the cluster analysis builds on but improves the technique used by Appelbaum et al. [2000]. They use K-clustering which requires that the number of clusters and their centroids be specified a priori. We instead let the data inform the choice of centroids and number of clusters by first using a hierarchical agglomerative clustering method and then apply to these results the K-clustering as a second step.

We start with a hierarchical agglomerative clustering technique. Initially, each firm is its own cluster (e.g., the number of clusters g = the n observations).

The n clusters are then reduced to n-1 by agglomerating the two least dissimilar clusters and this process continues in an iterative fashion until only one cluster emerges (g = 1). The next step is to use these results to choose the number of clusters (g) that best fit the data. We restrict attention to g ≤ 6 in order to identify a manageable number of discrete employment systems. Beginning with g = 6 and moving to g = 1, we search through each set of clusterings for what is

40

Economics and Business Review, Vol. 1(15), No. 2, 2015

known in the literature as the “sharp step” – that is, the value of g where combining one more cluster leads to a substantial change in fit but where moving beyond this leads to a small change.

The problem with agglomerative clustering is that in early stages of the clustering (a high g) certain observations may be placed in a particular cluster but when the number of clusters shrinks to a lower g these observations may achieve a larger reduction in dissimilarity if they are moved to a new cluster.

But the technique prevents this, leading to potential mis-classification. At this point, therefore, we introduce K-clustering. As earlier noted, K-clustering requires that the number of clusters g and their centroids be specified a priori.

We use the agglomerative results to specify the initial value of g (at the sharp step) and centroid values. Then the K-clustering partitions the observations across the g groups in a manner that minimizes the squared distance of each observation from its centroid. Once the firms have been so grouped, the centroids for the clusters are recalculated and the firms are re-clustered. If any of the firms move from one cluster to another, the centroids are recalculated and the firms are accordingly re-clustered. The K-means clustering method is complete when no firms move between clusters.

Results To put the clustering analysis into better context, we first plot in Figure the distribution of the 264 organizations by their level of HRM expenditure per

Frequency

A

0

2000

4000 6000 8000 Per Capita Expenditures on HRM Practices

Frequency distribution, HRM expenditures per employee Source: [Bureau of National Affairs, 2005/2006]

10000

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

41

employee. The data generate a bell-shaped frequency distribution with a distinctly skewed right-hand tail.7 Per capita HRM expenditure ranges from a low of $152 to a high of $8,709; however, roughly half of employers – those in the middle range between the 25th percentile and the 75th percentile – spend between $615 per employee and $2,069 per employee for HR activities.

Figure 1 provides two pieces of information supporting the ES and configurational concepts. The first is the very large variance in total HRM expenditures per employee. The universalistic or best practice SHRM model [Delery and Doty 1996; Pfeffer 1998] renders the concept of ES’s mostly moot since it predicts that only one set of HRM practices – typically associated with some version of a high involvement or high performance work system (extensive training, mutual-gains pay, employee involvement, etc.) – maximizes performance, suggesting longrun competitive selection pressures should concentrate organizations around this model. Since a best practice ES entails considerable investment in HRM activities and programs [Huselid 1995], the universalistic hypothesis predicts that the HRM frequency distribution should have a relatively concentrated variance centered on a relatively high (but not necessarily highest) level of HRM expenditure. The data clearly do not support this hypothesis, at least in a relatively close interpretation of the best practice model; indeed, a transformed HRM intensive employment system appears to be a distinct minority phenomenon located in the region of right-hand tail. The lack of empirical support for the universalistic hypothesis may be because the hypothesis is false or, alternatively, because competitive selection pressures are weak, obstructed, or erratic. Whatever the case, the existence of a large dispersion in HRM expenditure levels among the organizations in Figure 1 provides prima facie evidence that firms sort not into one kind of ES but potentially numerous ones.

A second piece of evidence emerges from a more detailed look at one particular point in the HRM frequency distribution. We found in the BNA data a group of three firms with nearly identical levels of HRM expenditure ($1,010, $1,013 and $1,014). They are marked as Point A in Figure 1. A weaker version of the universalistic model is that ES’s may vary over the entire HRM frequency distribution (due to some contingency factor), but that at a point (or range) firms adopt a relatively homogeneous ES. The data also do not support this hypothesis. We show in Table 3 for each of the three firms the level of their HRM expenditures (ranked Low, Medium, High) allotted to each of nine standard HRM practices (not all in the Begin ES typology). A firm is classified as “Medium” if the percentage of its expenditure on the respective HRM practice is within half a standard deviation of the mean for the dataset; those that are greater (less) are classified as High (Low).

7 A similar but less skewed distribution is obtained when a count of HRM practices is used.

A similar looking HRM practice count distribution is shown in Exhibit 5–2 (p. 96) of Freeman and Rogers [1999].

42

Economics and Business Review, Vol. 1(15), No. 2, 2015

Table 3. HRM practice usage for three firms with similar total expenditure Type of HRM Practice

Firm 1 ($1,010)

Firm 2 ($1,013)

Firm 3 ($1,014)

High

Medium

Medium

Medium

High

Medium

Compensation

Low

Low

Medium

Benefits

High

Medium

High

Employee Relations

Medium

High

Medium

External Relations

Low

Low

Medium

Performance Management

Low

High

High

Medium

Low

Medium

Low

Low

Medium

Recruitment Training

OSHA Strategic Planning

Source: Authors’ calculations using data collected from Bureau of National Affairs (2005/2006).

It is clear that all three firms have selected substantially different HRM bundles. For example, Firms 1 and 2 share in common only three practice levels out of nine (Compensation, External Relations, Strategic Planning); Firms 1 and 3 share four (Training, Benefits, Employee Relations, and OSHA); and Firms 2 and 3 share only two (Recruitment and Performance Management). Again, the data provide prima facie evidence against the universalistic SHRM model and in favor of distinct ES’s.

Figure 1 suggests the organizations in the BNA data set do not sort into only one ES, even as a rough approximation or central tendency; likewise, the data do not show any spike or discernible grouping of firms in the right-hand tail of the distribution that would potentially represent an HPWS node or attraction point. Even if there is no apparent sizable cluster of high performance HRM firms, it is also true on the other hand that the range of possible ES’s extends from a low of 2 to a high of 264. Figure 1 also provides no evidence on the merits of Begin’s ES typology. It is to these twin matters we now turn.

We proceed through another three steps. The first is to apply the clustering techniques described above to the BNA data in order to identify distinct ES’s.

The key question from the BNA survey we use is the one earlier cited: “Please approximate the percentage of your HR budget allocated to each area below […]”. The answers for nine of the categories (the catch-all tenth category “All other areas” is omitted) are reported in Table 3 for all respondents who gave reasonably complete data.

We used the budget share answers from this question to perform the clustering analysis. The best fit (location of the sharp step) is obtained with four clusters (g = 4). Thus, the 264 organizations – using the budget share numbers for the nine HRM practice/activity areas – sort into four distinct groups of

43

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

HRM practices. Technically speaking, all that the clustering has demonstrated is the existence of four distinct patterns of association in the data; nonetheless, it does not seem to too-large a leap to go further and put forward as an operating hypothesis that the data provide discernible evidence that these firms sort into four distinct employment systems – just as models such as Begin’s predict.

Table 4 shows the nine major HRM practices listed in the BNA report along the vertical left-hand side, the four cluster groups (Group 1, Group 2, etc.) are arrayed along the top (with number of observations for each), and the individual cells show the average budget share numbers (in percents). The cell entries for each group do not sum to 100.0 because of the omitted “All other areas” category.

Table 4. Clustering of organizations by hrm budget share (nine BNA practices) HRM Practice/ES

Group 1 (78 obs)

Group 2 (67 obs)

Group 3 (63 obs)

Group 4 (56 obs)

Population average (264 obs)

Employment/Recruiting

15.08

16.68

15.02

15.44

15.55

Training/Development

7.32

10.76

11.07

9.02

9.54

Compensation

23.08

15.62

15.74

13.55

17.00

Benefits

22.71

22.08

21.77

25.51

23.02

Employee Relations

4.74

5.00

5.38

6.40

5.38

External Relations

1.34

1.69

2.46

1.80

1.82

Health and Safety

3.61

3.91

3.60

4.30

3.78

Personnel/HR Records

3.18

3.64

4.24

3.55

3.65

Strategic Planning

2.92

4.47

4.10

3.41

3.97

We now want to transition from the results in Table 4 to a “test” of the Begin model. To do so, we present in Table 5 the nine HRM P&C that Begin explicitly identifies as part of his ES typology. Note there is some divergence between the nine HRM practices in Table 4 and the nine in Table 5. The reason is that Begin does not include in his ES typology certain of the HRM practices/activities listed in the BNA survey’s HR budget share question used for the clustering reported in Table 4, while on the other hand he does include other HRM/ organizational characteristics (unionization, size, etc.) not in the budget share data but for which data are provided in other parts of the BNA survey and which can therefore be used for Table 5. The first four HRM characteristics in Table 5 are measured in terms of expenditures per employee (e.g., recruiting/ staffing expenditure per capita). Employment and percent unionized are numerical values taken directly from the BNA survey. Strategic involvement is

44

Economics and Business Review, Vol. 1(15), No. 2, 2015

measured in the survey on a 1–5 scale (5 = highest) in response to the question: “How would you describe the HR function’s strategic involvement in key business decisions made by our organization?;” it is also directly used in Table 5. The characteristic in Begin’s typology that has the least explicit counterpart in the BNA data set is “degree of formalization” of the HRM program. Based on the supposition that more formalization requires more HR headcount and expenditures on programs and services, we used from the survey “total HRM expenditures per employee” as a proxy measure.

To proceed, we must next scale the numerical values in Table 4 so they match the 1–6 scale used earlier (Table 2) to delineate the four ES’s in the Begin model.

This scaling process is relatively straightforward since the results of the cluster analysis are already characterized using an ordinal measure. The scale is determined in the following manner. First, the mean and standard deviation are calculated for each of the nine characteristics for the entire population of firms.

These terms are represented by μj and σj, respectively, with j = {1, 2, …, 9}. Then, the mean values of these nine variables are determined for each cluster. These values are given by the term cij where i represents the cluster (i.e., i = {1, 2, 3, 4}), and j represents the nine characteristics (i.e., j = {1, 2, …, 9}). Finally, each cij is assigned a score of 1–6, referred to as vij so they are comparable to the predictions shown above. This is done by using μj and σj to determine the value of cij relative to the rest of the population of firms in the dataset. The criteria used to determine each vij is as follows: if… cij < μj – 0.25*σj, then vij = 1, μj – 0.25*σj < cij < μj – 0.1*σj, then vij = 2, μj – 0.1*σj < cij < μj, then vij = 3, μj < cij < μj + 0.1*σj, then vij = 4, μj + 0.1*σj < cij < μj + 0.25*σj, then vij = 5, cij > μj + 0.25*σj, then vij = 6.

Therefore a value of vij = 1 implies that the per employee expenditures on recruitment practices for a particular group of firms are less than 0.25 standard deviations from the population average. Similarly, a value of vij = 2 indicates that the group’s average level of per employee expenditures on recruitment practices is between 0.10 and 0.25 standard deviations from the population mean. The reason that a difference of 0.1 and 0.25 standard deviations from the population mean are used as the threshold values is because they create the best differentiation in outcomes. Although a data-driven choice, the results change little using other proximate values (e.g., 1.0 and 0.5).

Table 5 shows the results of the scaling process.

To help make these rankings more concrete and user-friendly, we converted them into numerical values, as shown in Table 6. Each cell entry shows the average numerical value for the indicated HRM P&C for the organizations in that ES group, as calculated from the BNA data. These data show in real life

45

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

Table. 5. Scaled values of HRM practices in BNA data, by cluster group HRM characteristic/ES

Group 1

Group 2

Group 3

Group 4

Recruiting/staffing

2

6

3

5

Training/development

2

3

2

6

Benefits/rewards

3

3

3

5

Performance management

2

5

3

5

Participation/voice

3

4

2

5

Work force size

3

2

6

2

Unionization

4

2

4

4

Strategic involvement

1

2

3

5

Degree of formalization

2

3

4

4

Table 6. Numerical values of HRM practices, by cluster group Simple Group 1

Machine Group 2

Professional Group 3

Adhocracy Group 4

Population Average

Population St. Dev.

Recruiting

223.55

660.53

341.03

513.88

395.43

992.54

Training

119.69

288.55

138.18

530.35

293.37

942.59

Benefits administration

515.96

481.72

521.39

998.71

667.29

2,506.66

Performance management

46.22

126.62

39.06

119.65

84.17

246.58

Employee relations

99.21

154.09

50.09

220.50

138.31

414.20

Employment

1209.57

881.54

3124.59

1566.84

1666.55

5,379.92

Unionization

0.29

0.20

0.27

0.30

0.27

0.44

Strategic involvement

3.24

3.52

3.61

3.92

3.71

1.08

1676.05

3349.71

3851.98

6543.09

3572.21

15,193.44

HRM characteristic/ES

Degree of formalization

terms the patterns of differentiation among the ES groups and individual HRM P&C, as well as the large diversity that exists across ES groups.

We have identified discrete groups of firms based on different configurations of HRM practices; the remaining question is whether the specific HRM components in each configuration match the predicted components in Beginâ&#x20AC;&#x2122;s model. Toward this end, Table 7 shows in parentheses the scaled predictions

46

Economics and Business Review, Vol. 1(15), No. 2, 2015

from Begin’s ES typology (Table 2) and above them in bold type the actual level (scaled) reported for the 264 organizations in the BNA data set (Table 5). The table has 36 cells (9 × 4).

The predicted values of the HRM P&C match the actual values in 24 of the 36 cells – a relatively modest “success rate” of 66.6 percent. The predictive success of the model increases noticeably when applied to the two polar opposite ESs (simple and adhocracy); here it scores 14 out of 18 (78 percent). Further, seven of the “misses” across all four groups are off by only one scaled value. We would label this degree of conformance between theoretical predictions and empirical evidence as “discernible but loose.” Several reasons may explain the loose fit. Probably most important is the large number of steps we had to go through to parameterize the model, no doubt introducing some-to-considerable noise and mis-measurement. Another possible reason is that Begin’s theoretical model needs recalibration, particularly for the intermediate ES cases (machine and professional), and also because his model was developed for organizations circa the 1980’s and possibly organizations’ ILMs and ESs have since Table 7. Begin’s predictions (parentheses) and HRM characteristics (bold), by ES HRM characteristic/ES Recruiting/ staffing Training/ development Benefits/rewards Performance management Employee relations Work force size

Unionization Strategic involvement Degree of formalization

Simple Group 1 (78)

Machine Group 2 (67)

Professional Group 3 (63)

Adhocracy Group 4 (56)

2

6

3

5

(2, 3)

(3, 4)

(2, 3)

(5, 6)

2

3

2

6

(2, 3)

(2, 3)

(2, 3)

(5, 6)

3

3

3

5

(2, 3)

(5, 6)

(4, 5)

(5, 6)

2

5

3

5

(1, 2, 3)

(3, 4, 5)

(3, 4, 5)

(5, 6)

3

4

2

5

(1, 2)

(1, 2)

(3, 4)

(5, 6)

3

2

6

2

(1, 2, 3)

(4, 5, 6)

(1, 2, 3, 4, 5, 6)

(1, 2, 3, 4, 5, 6)

4

2

4

4

(1, 2, 3)

(4, 5, 6)

(1, 2, 3)

(2, 3, 4)

1

2

3

5

(1, 2)

(1, 2, 3)

(1, 2)

(5, 6)

2

3

4

4

(1, 2, 3)

(4, 5, 6)

(3–4)

(3–4)

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

47

then changed considerably. Finally, the data set is also not ideal in all respects, particularly because it includes observations that span different organizational levels (e.g., plant, division, firm).

Although Begin’s model perhaps gets no more than a modest passing grade from the data analysis, the evidence that firms cluster into distinct groups of HRM practices gains stronger support and the evidence that no quantitatively large group of HPWS-type firms exists is stronger still [also see Blasi and Kruse 2006]. These results bear, therefore, on the continuing debate about the relative merits of a universalistic, contingency, and configurational approach in SHRM [Becker and Huselid 2006; Boxall and Purcell 2008]. In particular, our study appears to provide additional reasons to be cautious about a relatively straight-forward version of the universalistic “one best ES” model for the data in Tables 6 and 7 show that HRM practices and characteristics across these 264 organizations exhibit considerable diversity both across and within groups.

Second, although little evidence supports the existence of a dominant HPWStype employment system, it is also the case that firms’ bundle of HRM expenditures and practices do not appear haphazardly chosen or randomly distributed. Rather, a cluster analysis reveals that firms appear to sort into a few relatively well-defined employment systems and that these systems form along lines predicted in broad outline by Begin and other ES theorists. Theory at this stage does not provide a foolproof guide to constructing a real life employment system; nonetheless, it is our judgment that extant theory has made a first step toward identifying useful first principles with regard to the forces that create and structure ESs across organizations. As a concrete illustration, if firms’ environmental volatility increases (e.g., from globalization) then a prediction is that the organizational structure needs to become more flexible, implying in a complex technology environment a move toward an Adhocracy ES and in a simple technology environment a move toward a Simple ES. Having stated these implications, we must also emphasize their contingent nature due to various limitations of this study. In addition to issues pointed out above, five concerns merit further comment. The first is that cluster analysis inherently contains a subjective data-driven element (e.g., choice of centroids). A second shortcoming is that our empirical methods provide evidence on correlation and patterns of association but do not reveal how tight the degree of association is; likewise, these methods do not permit causal inferences and direct statistical tests of hypotheses. Third, we are unable to empirically test for a relationship between ES configurations and organizational performance due to lack of performance data. Fourth, the Begin model gives only secondary attention to the role of alternative business strategies as a determinant of ES configurations; further, while our empirical analysis incorporates a measure of alternative HRM strategies a more detailed and construct validated measure would certainly be desirable. And, fifth, our results may well not generalize beyond the USA. Given all of these limitations, caution must be

48

Economics and Business Review, Vol. 1(15), No. 2, 2015

attached to specific results and interpretations. Nonetheless, on the other side are various pluses – the study presents an innovative and never-before-used empirical strategy for studying ESs, the empirical analysis clearly supports the idea that ILMs sort into distinct structural forms/architectures, and the Begin model can be said to have captured and explained at least some of the important features of real life employment systems.

Conclusions Recent research in organizational economics, industrial/employment relations, and human resource management takes a strategic perspective and looks at HRM practices in terms of synergistic bundles that align with and support organizational goals. These bundles and associated organizational characteristics have become known as employment systems. A variety of theoretical models of employment systems have been advanced and researchers have made exploratory progress in empirically identifying the existence and structure of alternative ESs. This paper advances this line of research along several fronts. First, we survey the ES literature and highlight common ideas and findings. Second, we draw attention to the ES theory developed by James Begin and its implications regarding the link between the structure of organizations and the structure of HRM systems. Third, we test the predictions of the Begin ES model using an innovative mapping technique, improved clustering methods and a unique and highly detailed data set on HRM practices. Fourth, our empirical analysis reveals that these organizations sort into four distinct sets of HRM practices and characteristics, thus giving relatively strong support for the existence of distinct employment systems and modest but arguably discernible support for Begin’s particular theory of ESs. Evidently further theory development and integration across alternative models are obvious next steps in the ES research program; so too is further investigation of the relationship between alternative employment systems and firm performance.

References Appelbaum, E., Bailey, T., Berg, P., Kalleberg, A., 2000, Manufacturing Advantage: Why High-Performance Work Systems Pay Off, Ithaca, NY, Cornell University Press.

Arthur, J., 1992, The Link between Business Strategy and Industrial Relations Systems in American Steel Minimills, Industrial and Labour Relations Review, 45: 488–506.

Bae, J., Yu, G., 2005, HRM Configurations in Korean Venture Firms: Resource Availability, Institutional Force and Strategic Choice Perspective, International Journal of Human Resource Management, 16: 1759–82.

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

49

Bamberger, P., Meshoulam, I., 2000, Human Resource Strategy: Formulation, Implementation, and Impact, Sage, Thousand Oaks CA. Baron, J., Burton, D., Hannan, M., 1999, The Road Not Taken: Origins of Employment Systems in Emerging Companies, in: Carroll, G., Teece, D. (eds.), Firms, Markets, and Hierarchies, Oxford University Press, New York: 428–64.

Becker, B., Huselid, M., 2006, Strategic Human Resource Management: Where Do We Go from Here?, Journal of Management, 32: 898–925.

Beer, M., Spector, B., 1984, Human Resources Management: The Integration of Industrial Relations and Organizational Behavior, Research in Personnel and Human Resource Management, vol. 2, JAI Press, Stamford, CT: 261–297.

Begin, J., 1991, Strategic Employment Policy: An Organizational Systems Perspective, Prentice-Hall, Englewood Cliffs, NJ. Begin, J., 1993, Identifying Patterns in HRM Systems: Lessons from Organizational Theory, Research in Personnel and Human Resource Management, supplement, JAI Press, Stamford, CT: 3–20.

Begin, J., 1997, Dynamic Human Resource Systems: Cross-National Comparisons, New York, De Gruyter.

Blasi, J., Kruse, D., 2006, U.S. High Performance Work Practices at Century’s End, Industrial Relations, 45: 457–478.

Boxall, P., Purcell, J., 2008, Strategy and Human Resource Management, 2nd ed., Palgrave Macmillan, New York.

Braverman, H., 1974, Labour and Monopoly Capital: The Degradation of Work in the Twentieth Century, Monthly Review Press, New York.

Burawoy, M., 1979, Manufacturing Consent: Changes in the Labour Process under Capitalism, University of Chicago Press, Chicago.

Bureau of National Affairs, 2005/2006, HR Department Benchmarks and Analysis Report, Bureau of National Affairs, Washington, D.C. Burns, T., Stalker, G., 1961. The Management of Innovation, Tavistock, London.

Coase, R., 1937, The Nature of the Firm, Economica, 4: 386–405.

Commons, J., 1919, Industrial Goodwill, McGraw Hill, New York.

Delery, J., Doty, D., 1996, Modes of Theorizing in Strategic Human Resource Management: Tests of Universalistic, Contingency, and Configurational Performance Predictions, Academy of Management Journal, 39: 802–35.

Doelgast, V., 2009, Still a Coordinated Model? Market Liberalization and the Transformation of Employment Relations in the German Telecommunications Industry, Industrial and Labour Relations Review, 63: 3–23.

Doeringer, P., Piore, M., 1971, Internal Labour Markets and Manpower Analysis, Lexington Books, Lexington, MA. Dunlop, J., 1958, Industrial Relations Systems, Holt, New York.

Edwards, R., 1979, Contested Terrain: The Transformation of the Workplace in the Twentieth Century, Basic Books, New York.

Fombrun, C., Tichy, N, Devanna, M., 1984, Strategic Human Resource Management, Wiley, New York.

Freeman, R., Rogers, J., 1999, What Workers Want, Cornell University Press, Ithaca, NY. Freidman, A., 1977, Industry and Labour: Class Struggle and Monopoly Capitalism, Macmillan, London.

50

Economics and Business Review, Vol. 1(15), No. 2, 2015

Gibbons, R., Roberts, J., 2013, Handbook of Organizational Economics, Princeton University Press, Princeton.

Grandori, A., 2013, Handbook of Economic Organization, Elgar, Northampton.

Hall, P., Soskice, D., 2001, Varieties of Capitalism: The Institutional Foundations of Comparative Advantage, Oxford University Press, Oxford, UK. Hendry, C., 2003, Applying Employment Systems Theory to the Analysis of National Models of HRM, International Journal of Human Resource Management, 14: 1430–1442.

Huselid, M., 1995, The Impact of Human Resource Management Practices on Turnover, Productivity, and Corporate Financial Performance, Academy of Management Journal, 38: 635–672.

Katz, H., Darbishire, O., 2000, Converging Divergences: Worldwide Changes in Employment Systems, Cornell University Press, Ithaca.

Kaufman, B., 2004, Employment Relations and the Employment Relations System: A Guide to Theorizing, in: Kaufman, B. (ed.), Theoretical Perspectives on Work and the Employment Relationship, Industrial Relations Research Association, Champaign, IL: 41–75.

Kaufman, B., 2008, Paradigms in Industrial Relations: Original, Modern and Versions in-Between, British Journal of Industrial Relations, 46: 314–39.

Kaufman, B., 2010, SHRM Theory in the Post-Huselid Era: Why It Is Fundamentally Mis-specified, Industrial Relations, 49: 286–313.

Kaufman, B., 2013, The Economic Organization of Employment: Systems in Human Resource Management and Industrial Relations, in: Grandori, A. (ed.), Handbook of Economic Organization, Elgar, Northampton: 289–311.

Keefe, J., 2009, Is Digital Technology Reshaping Employment Systems in U.S. Telecommunications Network Services?, Industrial and Labour Relations Review, 63: 42–59.

Kerr, C., 1950, Labour Markets: Their Character and Consequences, American Economic Review, 40, 278–91.

Kerr, C., 1954, The Balkanization of Labour Markets, in: Bakke, E. (ed.), Labour Mobility and Economic Opportunity, MIT Press, Cambridge: 92–110.

Kinnie, N., Swart, J., Purcell, J., 2005, Influences on the Choice of HR System: The Network Organization Perspective, International Journal of Human Resource Management, 16: 1004–1028.

Kochan, T., Katz, H., McKersie, R., 1986, The Transformation of American Industrial Relations, New York, Basic Books.

Lazear, E., Oyer, P., 2013, Personnel Economics, in: Gibbons, R., Roberts, J. (eds.), The Handbook of Organizational Economics, Princeton University Press, Princeton: 479–519.

Lepak, D., Snell, S., 1999, The Human Resource Architecture: Toward a Theory of Human Capital Allocation and Development, Academy of Management Review, 24: 31–48.

Marsden, D., 1999, A Theory of Employment Systems: Micro Foundations of Societal Diversity, Oxford University Press, Oxford.

Marx, K., 1847/1935, Wage Labour and Capital, International Publishers, New York.

Mintzberg, H., 1983, Structure in Fives: Designing Effective Organizations, PrenticeHall, Englewood Cliffs, NJ.

B.E. Kaufman, B.I. Miller, Alternative configurations of firm-level employment systems

51

Orlitzky, M., Frenkel, S., 2005, Alternative Pathways to High Performance Workplaces, International Journal of Human Resource Management, 16: 1325–1348.

Osterman, P., 1987, Choice of Employment Systems in Internal Labour Markets, Industrial Relations, 26: 46–67.

Pfeffer, J., 1998, The Human Equation, Harvard University Business School Press, Boston.

Piore, M., Sabel, C., 1980, The Second Industrial Divide: Possibilities for Prosperity, Basic Books, New York.

Porter, M., 1980, Competitive Strategy, Free Press, New York.

Robinson, J., 1933, The Economics of Imperfect Competition, Macmillan, London.

Ross, P., Bamber, G., 2009, Strategic Choices in Pluralist and Unitarist Employment Relations Regimes: A Study of Australian Telecommunications, Industrial and Labour Relations Review, 63: 24–41.

Rubery, J., Grimshaw, D., 2003, The Organization of Employment: An International Perspective, Palgrave Macmillan, London.

Sheppeck, M., Militello, J., 2000, Strategic HR Configurations and Organizational Performance, Human Resource Management, 39: 5–16.

Short, J., Payne, G., Ketchen, D., 2008, Research on Organizational Configurations: Past Accomplishments and Future Challenges, Journal of Management, 34: 1053–1079.

Simon, H., 1951, A Formal Theory of the Employment Relationship, Econometrica, 19: 293–305.

Strauss, G., 2001, HRM in the USA: Correcting Some British Impressions, International Journal of Human Resource Management, 12: 873–897.

Thompson, P., Harley, B., 2007, HRM and the Worker: Labour Process Perspectives, in: Boxall, P., Purcell, J., Wright, P. (eds.), Oxford International Handbook of Human Resource Management, Oxford University Press, Oxford: 147–65.

Toh, S., Morgeson, F., Campion, M., 2008, Human Resource Configurations: Investigating Fit with the Organizational Context, Journal of Applied Psychology, 93: 864–82.

Towers, B., 2003, Industrial Relations: Time to Move On?, in: Ackers, P., Wilkinson, A. (eds.), Understanding Work & Employment, Oxford University Press, Oxford: xii-xvi.

Tsai, C., 2006, High Performance Work Systems and Organizational Performance: An Empirical Study of Taiwan’s Semiconductor Design Firms, International Journal of Human Resource Management, 17: 1512–1530.

Verberg, R., Hartog, D., Koopman, P., 2007, Configurations of Human Resource Management Practices: A Model and Test of Internal Fit, International Journal of Human Resource Management, 18: 184–208.

Waldman, M., 2013, Theory and Evidence on Internal Labour Markets, in: Gibbons, R., Roberts, J. (eds.), The Handbook of Organizational Economics, Princeton University Press, Princeton: 520–71.

Walton, R., 1985, From Control to Commitment in the Workplace, Harvard Business Review, 63: 77–84.

Wardell, M., Steiger, T., Meiskins, P., 1999, Rethinking the Labour Process, State University of New York Press, Albany, NY. Williamson, O., 1975, Markets and Hierarchies: Analysis and Antitrust Implications, Free Press, New York.

Youndt, M., Snell, S., 2004, Human Resource Configurations, Intellectual Capital, and Organizational Performance, Journal of Managerial Issues, 16: 337–360.

Economics and Business Review, Vol. 1 (15), No. 2, 2015: 52–75

How team leaders can improve virtual team collaboration through trust and ICT: A conceptual model proposition1 David Kauffmann2

Abstract: The purpose of this paper is to present a conceptual model to facilitate the development of collaboration within virtual teams. The model claims that a high level of communication through ICT is an important antecedent for collaboration mediated by the level of trust among the team. Furthermore and according to the model, Team Leaders can have a major impact on the communication effectiveness and on the level of team trust.

Keywords: virtual team, distributed team, collaboration, trust, ICT, team leader.

JEL codes: D83, M12, M15, O32, O33.

Introduction Collaborative teams are most effective at achieving and enhancing an organization’s strategy. Much research has been conducted to identify the antecedents of collaboration in order to increase team effectiveness and the level of its outcomes. The effectiveness of the team and the level of its outcomes will allow assessment as to whether specifying if the team is a successful one or not.

With the Internet revolution of the 1990’s the world became a global village.

The distance separating people shrank and a new era of organization began.

One of the changes that this revolution brought to organizations is the creation of a new kind of team in addition to the conventional face-to-face team: the virtual or distributed team.

1

Article received 22 June 2014, accepted 19 February 2015.

I want to thank the anonymous reviewers for their comments. All remaining errors are those of the author.

2 Jerusalem College of Technology, Faculty of Business Administration, Havaad Haleumi 21 St., Givat Mordechai, Jerusalem, Israel, e-mail: davidk1970@gmail.com.

D. Kauffmann, How team leaders can improve virtual team collaboration

53

In the current highly competitive climate organizations must be dynamic, innovative and able to adapt quickly to new situations. Therefore, 21st century organizations need teams to solve problems and conflicts, to share information and knowledge, to make the right decisions, to be innovative and creative.

The quality and level of these attributes will define the nature of team collaboration and then this collaboration will lead to improved team performance [Peters and Manz 2007].

Five key factors have been identified by Bergiel, Bergiel, and Balsmeier [2008] as vital to the formation of a successful virtual team. These five factors are: trust, communication, leadership, goal setting and technology. Boughzala, de Vreede, and Limayem [2012], based on a model of a Collaborative Work Practice designed by de Vreede, Briggs, and Mossey [2009] also observe five critical elements for developing a successful collaboration: these elements are: leadership, people, technology, information and process.

The conceptual model that is proposed in this paper claims that team communication, based on Information and Communication Technology (ICT) channels and the level of trust have an impact on the quality of virtual team collaboration. The effective and proper use of ICT with the mediation of trust will act as an antecedent of the virtual teams’ collaboration. Furthermore, the model also claims that the team leader’s behavior will have a significant effect on the effective and proper use of ICT and on the team’s trust level.

The paper is divided into five sections. The first section is a literature review of papers that deal with the six elements at focus of this paper: Collaboration, Virtual Teams, Communication, Trust, ICT and finally, Team Leader. The second section verbalizes the theoretical background and articulates the hypotheses on which the model is based. The third gives a description of the model itself whose purpose is to help team leaders to improve virtual team collaboration through trust and ICT. The fourth section proposes a method of research to validate the model which includes two stages. Finally, the last section gives a summary of the model and raises its importance for modern organizations.

1. Literature review Collaboration is an essential ingredient in the success of the organization [Boughzala, de Vreede, and Limayem 2012]. Many of them organize training and seminars for their teams on a periodical basis in order to increase the level of collaboration and cooperation. Collaboration has been recognized as a process that can create outcomes that cannot be achieved by an individual alone [Peters and Manz 2007]. Virtual Teams, because of distance between the teammates, need to develop ways for creating a successful collaboration without face to face training or seminars, but with the help of E-collaboration tools [Hosley 2010]. “Communication and collaboration are the two most important factors

54

Economics and Business Review, Vol. 1(15), No. 2, 2015

in team success. A virtual environment fundamentally transforms the ways in which teams operate” [Duarte and Snyder 2011]. These E-collaboration tools are built on ICT which allow teammates to communicate with each other on social and task dimensions. At first Virtual Teams were created for limited time projects or task purposes [Jusrud 2008]. Therefore swift trust [Meyerson, Weick, and Kramer 1996], based on cognitive trust only, has been developed in this environment because of the temporary nature of the team. But in the last decades, other Virtual Teams called Distributed Work Groups [Jusrud 2008], have been created for on-going tasks which have a permanent character and therefore swift trust will not be enough to maintain a high level of trust. Affective-based trust, besides cognitive-based trust, will be a necessary ingredient to maintain a high level of trust in such team [De Jong and Elfring 2010]. Communication and trust have been raised several times as components for team collaboration building [Barczak, Lassk, and Mulki 2010]. In order to develop and maintain good communication [Sivunen 2008] and trust [Webber 2002], the team leaders have to play a positive role among their team. Through a review of literature, I will define the main concepts I have raised: Collaboration, Virtual Team (Temporary and On-going), Communication (Task- and relationship-oriented) especially via ICT, Trust (Cognitive- and Affective-based) and finally, Team leaders and their influence among their team to build effective communication, trust and collaboration.

1.1. Collaboration Collaboration is a complex process which as a result of communication and interaction between parties, creates relationships between them, allowing the sharing and synchronizing of information for the purpose of decision making and achieving common matters or goals. Thomson and Perry [2006] define it as a process in which autonomous actors interact through formal and informal negotiation, jointly creating rules and structures governing their relationships and ways to act or decide on the issues that brought them together; it is a process involving shared norms and mutually beneficial interactions.

Gray and Wood [1991] develop a theoretical framework for studying collaboration. This theoretical framework allows understanding of the process of collaboration that yields particular outcomes. They argue that scholars need to investigate three areas: antecedents to collaboration, the process of collaboration itself, and the outcomes of that process. However, during their research on collaboration, scholars often simultaneously associate antecedents with collaboration processes and outcomes. These lead to failure in differentiating the mediating variables from the outcome ones [Thomson, Perry, and Miller 2010].

For the purpose of this paper, I will focus only on one of these areas which is the collaboration antecedents. According to Mattessich, Murray-Close, and Monsey [2001], collaboration depends on twenty factors that influence the

D. Kauffmann, How team leaders can improve virtual team collaboration

55

success of collaboration. Trust and Communication have a major role among these factors: “Collaboration depends on the existence of trust, shared vision, communication, and other ingredients”. [Mattessich, Murray-Close, and Monsey 2001]. Collaboration requires a dynamic relationship across various members and groups [Hosley 2010], Trust and communication will facilitate this dynamic relationship.

Five of the concepts associated with collaboration which are most frequently mentioned by these scholars are: Knowledge & Information Sharing [Osman 2004; Bell and Kozlowski 2002; Evans 2012; Van Gelder 2011; Ghaznavi et al.

2013], Conflict Management [Osman 2004; Pazos, Ustun, and DelAguila 2011; Atteya 2013; De Dreu and Beersma 2005], Problem Solving [Casalini, Janowski, and Estevez 2007; Ghaznavi et al. 2013; Dillenbourg 1999], Decision Making [Bell and Kozlowski 2002; Michie, Dooley and Fryxell 2006; Turban, Liang, and Wu 2011], and Innovation and Creativity [Osman 2004; Evans 2012; Ghaznavi et al. 2013]. Therefore, these five concepts will be used to define the level of collaboration within the virtual team.

Information sharing is defined as “a process of making one’s own stored and updated information accessible for other members of a group. Sharing presupposes consensus of a group about the interaction and is a necessary condition to be effective” [Den Otter 2005]. Knowledge sharing is defined as “the willful application of one’s ideas, insights, solutions, experiences [i.e. knowledge] to another individual either via an intermediary, such as a computer-based system, or directly” [Turban et al. 2006].

Relationship conflict is defined as “interpersonal incompatibilities among group members, which typically includes tension, animosity, and annoyance among members within a group” [Jehn 1995]. Conflict management is defined as “behavior oriented toward the intensification, reduction, and resolution of the tension” [De Dreu, Harinck, and Van Vianen 1999] and it will hopefully lead to an opportunity to improve situations and strengthen relationships.

Problem solving is defined as a process used to obtain a best answer to an unknown, or a decision subject to some constraints [Mourtos, De Jong Okamoto, and Rhee 2004]. Collaborative problem solving is problem-solving done by peers, performing the same actions, having a common goal and working together [Dillenbourg 1999].

Decision making is defined as a group’s “ability to integrate information, use logical and sound judgment, identify possible alternatives, select the best solution, and evaluate the consequences” [O’Neil 1999]. Collaborative Decision Making typically evolves from either formal or informal deliberations in groups where the group members consider and debate various possible decision options. The decision issues are resolved through discussions, where argumentative logic and persuasive presentation are critical [Raghu et al. 2001].

Innovation is a dynamic process through which problems and challenges are defined, new and creative ideas are developed, and new solutions are selected

56

Economics and Business Review, Vol. 1(15), No. 2, 2015

and implemented [Sørensen and Torfing 2012]. Collaborative Innovation is defined as “The recursive interaction of co-creativity, knowledge, and mutual learning between two or more people working together toward a common goal of generating new sources of growth or wealth in an organization” [Lynch 2007].

All these concepts are closely related to Collaboration where a high level of cooperation between teammates is crucial for the success of the processes.

1.2. Virtual team Salas et al. [1992] provided a good working definition of a team as “a distinguishable set of two or more people who interact dynamically, interdependently, and adaptively towards a common and valued goal/objective/mission, who each have been assigned specific roles or functions to perform”. Salas et al.

[1992] also add in the definition that virtual teams “have a limited life-span of membership”, however, in the last decades, on-going virtual teams also emerge as I will define later.

Traditional teams are known as face-to-face teams, in which the whole team is mostly working in the same space-time. Virtual teams are different in several ways. Many researchers have tried to characterize the differences between virtual teams and face-to-face teams. According to Chudoba et al.

[2005] there are six discontinuities – geography, time, culture, work practices, organization, and technology – that capture distinctive aspects of the virtual team environment. It is widely agreed by scholars that the main elements which define a Virtual Team are groups of people who work together and are often dispersed across space, time, and/or organizational boundaries; furthermore these groups of people collaborate and communicate through electronic technologies commonly called ICT [Ebrahim, Ahmed, and Taha 2009; Hertel, Geister, and Konradt 2005]. Most organizations have teams which are working as Virtual Teams across distance, especially global ones. Martins, Gilson, and Maynard [2004] in a major review of the literature on virtual teams, conclude that “Virtual Teams are increasingly prevalent in organizations and, with rare exceptions, all organizational teams are virtual to some extent”.

In his research, Justrud [2008] refers to three kinds of teams working in a virtual environment. The first is known as a virtual task force. This group initially forms as a result of an acute or unexpected situation. The second kind of team defined by Justrud as a virtual team is a group formed for a limited period of time in order to solve certain pre-defined tasks. Both of these kinds of teams are temporary most of the time. Finally, Justrud dubs the third kind of team a distributed work group. This group contains people from different geographical units within the same organization. Such teams are usually of a more permanent nature than virtual teams, they work on an on-going basis.

In the last decade of the 20th century and the first few years of the 21st century, virtual teams were mostly based on temporary teams. Most of these teams

D. Kauffmann, How team leaders can improve virtual team collaboration

57

were project teams [Mankin, Cohen, and Bikson 1996; Pulnam 1992], task forces [Hackman 1990], or short-term project teams [Cohen 1993]. Usually temporary teams are working on non-routine, highly skilled technical or administrative projects, such as developing a new product or information system [Saunder and Ahuja 2006].

Over the past few years, the second kind of team – the ongoing or longterm team – has also become more prevalent in the virtual context. This kind of team is dubbed functional team [Hellriegel, Slocum, and Woodman 1998] or work team [Pulnam 1992; Mankin, Cohen, and Bikson 1996]. These teams are typically characterized by cyclically recurring activities, and their members expect to be working together on future tasks [De Jong and Elfring 2010].

Saunder and Ahuja [2006] well define these two kinds of team as follow: “Temporary teams engage in a single task, or, at most, a few tasks, to accomplish their goal. Their tasks are concrete and finite. On the other hand, ongoing teams are long term, often requiring multiple or repeated tasks to accomplish the many or recurring goals that are established at their inception or evolve over time”.

Most scholars have based their works on the temporary virtual teams and therefore have developed theories like swift trust [Meyerson, Weick, and Kramer 1996] – based on cognitive trust for quick team trust building. Ongoing teams tend to be more focused on interpersonal relationships, which increase the impact of trust dynamics on team member interactions [Karau and Kelly 2004].

Unlike swift trust, which is highly fragile and temporal, on-going teams must develop trust not only based on the cognitive dimension, but also on the affective dimension. These two dimensions of trust will be developed later in this chapter.

1.3. Communication Scholarly literature provides evidence that quality of communication has effects on team collaboration and performance [Hassall 2009]. These effects can be positive or negative depending on communication channels and styles.

Therefore communication was identified as an important process for any team.

However, it is especially important for virtual teams [Saunder and Ahuja 2006; Zofi 2012]. “At the core of any virtual team process is communication” [Powell, Piccoli, and Ives 2004]. Communication is not only an important process; it’s a real challenge in a virtual environment [Mumbi 2007] due to different cultures and time zones, and distance. Grabner-Kräuter and Kaluscha [2003] argue that the lack of physical contact makes it more difficult to establish strong relationships and bonds that lead to high levels of trust, making the communication process more challenging.

Literature often differentiates between two aspects of communication within the team, Task-oriented communication and Social/Relationship-oriented com-

58

Economics and Business Review, Vol. 1(15), No. 2, 2015

munication [Huang 2010; Lau, Sarker, and Sahay 2000; Jarvenpaa and Leidner 1999]. Task dimension focuses on how well project information, tasks and deliverables are being handled through the communication. In other words, task-oriented communication moves the team forward in the accomplishment of their task and includes such communication as “planning and scheduling work, coordinating subordinate activities, and providing necessary supplies, equipment, and technical assistance” [Yukl 2012]. Social dimension provides the basis and desire for team members to communicate with each other over time. Relationship-oriented communication’s aim is to maintain a positive psycho-social dynamic within the team such as “showing trust and confidence, acting friendly and considerate, trying to understand subordinates’ problems, helping to develop subordinates and further their careers, keeping subordinates informed, showing appreciation for subordinates’ ideas and providing recognition for subordinates’ accomplishments” [Yukl 2012].

1.4. Trust There are different definitions of trust in academic literature. Marguin [2010] refers to two of the most widely accepted definitions. The first is “one party’s willingness to be vulnerable to another party based on the belief that the latter party is competent, open, concerned and reliable” [Mishra 1996]. The second widely accepted definition is “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party” [Mayer, Davis, and Schoorman 1995]. These represent two definitions of trust in terms of the dyadic relationship.

Cummings and Bromily [1996] observed that trust also exists in collective relationships [groups, teams, and organizational units]. They defined collective trust as: “A common belief among a group of individuals that another individual or group: a] makes good-faith efforts to behave in accordance with any commitments [...] b] is honest in whatever negotiations preceded such commitments and c] does not take excessive advantage of another even when the opportunity is available”.

Jarvenpaa, Knoll, and Leidner [1998] developed a model of trust in virtual teams based on the two theories of dyadic and collective relationships, as quoted above. Their model extends the dyadic trust relationship between trustor and trustee based on perceived ability, benevolence and integrity of the trustee [Mayer, Davis, and Schoorman 1995] to all team members. The baseline hypothesis of their work was that, in a global virtual team, team trust is a function of the other team members’ perceived ability, integrity and benevolence, as well as of the members’ own propensity of trust.

In order to trust and therefore be willing to depend on another party [McKnight, Cummings, and Chervany 1998], to take risks [Jones and

D. Kauffmann, How team leaders can improve virtual team collaboration

59

George 1998] and to be vulnerable [Mayer, Davis, and Schoorman 1995], we must create social and interpersonal relationships with the other. One of the main challenges in virtual teams, as opposed to face-to-face teams, is “overcoming the isolation caused by the separation of the telecommuter from the social network in the traditional work space” [Kurland and Bailey 1999]. Similarly, Grabner-Kräuter, and Kaluscha [2003] argue that the lack of physical contact makes it more difficult to establish strong relationships and bonds that lead to high levels of trust.

Over the years, many trust models have been developed. Based on the concept that trust may have rational and emotional roots, a model of cognitive and affective dimensions in trust has been developed by McAllister [1995].

This theory was recently used by Schaubroeck, Lam, and Peng [2011] in their research on the relationship between team performance and cognition-based and affect-based trust.

When trust is based on cognition, individuals employ rational thought in order to trust others. Cognition-based trust refers to trust that is based on performance-relevant cognitions such as competence, responsibility, reliability, and dependability [Schaubroeck, Lam, and Peng 2011]. We hope that other people will fill their roles and that their actions are consistent with their speech [Erdem and Ozen 2003]. But when the interaction between the parties is intense, the emotional and mutual investment in the relationship becomes primordial; this is where the affective side of trust comes into play [Erdem and Ozen 2003].

The emotional attachment created by this intense interaction emphasizes empathy, affiliation and rapport, based on a shared regard for the other person [Schaubroeck, Lam, and Peng 2011].

In family relationships, such as spouse-partner, and even more so in parent-child relationships, the affective side is very strong and forms the basis for most of the trust in the relationship. In contrast, when we need the services of a specialist – such as a technical expert or consultant – the cognitive side is predominant. In a work environment, where colleagues work together toward a common goal, trust is initially cognition-based. However, to maintain this trust in the long run, we must develop the affective aspect of the relationship [McAllister 1995]. Cognitive and affective dimensions are often tightly intertwined in work relationships.

Trust is assumed to develop gradually over time based on direct personal interaction and communication [Mayer, Davis, and Schoorman 1995; Lewicki and Bunker 1995]. Individuals need time in order to trust another person. We need to develop both cognitive and affective trust. Other research has gone so far as to add other dimensions, such as the “early trust” suggested by Webber [2002] as an antecedent to both cognitive and affective trust, or the “intended behavior” defended by Cummings and Bromily [1996] as a third dimension.

However, high levels of trust at an early stage are possible and may be driven by cognitive cues from group membership and reputation. Affective trust

60

Economics and Business Review, Vol. 1(15), No. 2, 2015

has been thought to develop later in the life of an interpersonal relationship [Williams 2001].

1.5. Information and communication technologies As I outline above, ICTs are almost the only means to collaborate and communicate in a virtual environment since face to face meeting is nearly non-existent.

The impact of technology on collaboration has been a topic for several items of research including empirical findings [Dennis, Wixom, and Vandenberg 2001; Fjermestad and Hiltz 1998]. Thomas and Bostrom [2008] declare that they “found evidence that Virtual Team leaders do manage information and communication tools (ICTs) in order to affect changes in team cooperation, through trust and relationship improvements”. Technology is evolving at an exponential pace leading to new collaboration tools like Web2.0 tools and social media. Former research [Dennis, Wixom, and Vandenberg 2001] suggested that use of different collaboration technology could influence outcomes differently. The different technology characteristics may influence differently the level of collaboration of the team and therefore have various impacts on the performance of the team and his outcome [Ustun and Pazos 2012].

ICTs provide support for both synchronous and asynchronous communication [Warkentin, Sayeed, and Hightower 1997; Ashley 2003]. Synchronous systems enable interpersonal contact that simulates face-to-face contact. It has been argued that “asynchronous Computer-Mediated Communication (CMC) is closer to writing due to the fact that it allows for more syntactic complexity than synchronous CMC” and that “synchronous CMC is closer to speaking than asynchronous CMC because numerous communication strategies and a wide range of discourse patterns are found in the synchronous environment” [Hirotani 2009]. This difference will impact upon the optimal use of these channels. There are three different levels of channels as defined by Bos et al. [2002]; the first is based on text-like writing or online presentation, the second on vocal contact and the third includes vocal and visual contact. The advantage of asynchronous systems is that they allow people to think before answering and to establish the reason behind a particular decision. Asynchronous systems also have the three levels of contact (Text, Vocal and Visual).

ICTs can facilitate both task-based and relationship-based communication. Of the large range of ICT channels, some are more suitable for task-oriented communication and some for relationship-oriented communication [Kauffmann and Carmi 2014]. Kauffmann and Carmi [2014] also argue that depending on the type of trust the Team Leaders wants to develop (Cognitive or Affective) and depending on the kind of Virtual Team (Temporary or Ongoing), the appropriate ICT channel has to be used in order to develop trust in a more effective way.

D. Kauffmann, How team leaders can improve virtual team collaboration

61

1.6. Team leader The definition of team leader that I will use in the model is based on the functional leadership theory [McGrath 1962]. According to Morgeson, DeRue, and Karam [2010], this theory is the most prominent and well-known team leadership model. Bell and Kozlowsky [2002] and Zaccaro, Rittman, and Marks [2001] have also supported this observation. This theory suggests that the leadership role is “to do, or get done, whatever is not being adequately handled for group needs” [McGrath 1962]. Morgeson, DeRue, and Karam [2010] defined team leadership as “[…] oriented around team need satisfaction [with the ultimate aim of fostering team effectiveness]”. Several studies have focused on understanding the principal functions of the team leader. Zaccaro, Rittman, and Marks [2001] define this leadership as social problem solving, where leaders are responsible for (a) diagnosing any problems that could potentially impede group and organizational goal attainment, (b) generating and planning appropriate solutions, and (c) implementing solutions within typically complex social domains. Bell and Kozlowsky [2002] split the team leader function into two primary categories: (a) the development and shaping of team processes, and (b) the monitoring and management of ongoing team performance. On the one hand, team leaders must act as managers and be task-oriented [Gray 2004] and on the other they must act as leaders and be people-oriented [Abbas and Asghar 2010] in order to extract better performance and effectiveness from their teams.

2. Theoretical background and research propositions The key factors to a successful team in general and in a virtual team in particular, are High levels of trust, Clear communication, Strong leadership and Appropriate levels of technology [Bergiel, Bergiel, and Balsmeier 2008]. The model is based on these four key factors to develop the level of team collaboration to achieve success. I will use two elements of leadership which are the abilities of team leaders to encourage and develop communication skills of the team and to facilitate team trust building.

2.1. Team leaders as mentors and facilitators Quinn et al. [2010] argue that there are eight managerial roles for team leaders on their way to becoming a master manager (Figure 1). Two of these managerial roles are of them acting as a mentor and a facilitator based on their Human Relation Model. As a mentor, team leaders need to develop subordinates and to communicate effectively. Team leaders need to teach how and encourage teammates to communicate effectively. As a facilitator, they need to build the

62

Economics and Business Review, Vol. 1(15), No. 2, 2015

Figure 1. Quinn et al. [2010] Model

team, to encourage decision making and resolve team conflict. Trust is an element of the team building, and decision making and team conflict management are part of team collaboration.

Sivunen A. [2008] conducted research on the communication of leaders in virtual teams. The fourth finding of her research was that virtual team members expect their team leaders to give instructions for the use of communication technology and about computer-mediated communication practices in general. DeRosa and Lepsinger [2010: 44] and Duarte and Snyder [2011: 18] also defend this argument and claim that team leaders have impacts on the team’s communication skills. My first hypothesis of the model is: H1a: The greater the knowledge of the Team Leader of ICT, the greater Communication skills of the virtual team will be.

Webber [2002] in a paper examined the challenges faced by Cross-Functional Teams and why these challenges facilitate the need for development of a team climate of trust. On one hand Cross-Functional Teams differ from Virtual Teams, Virtual Teams have common goals while Cross-Functional Teams can have different goals but, on the other hand, they have much in common such as not working in the same space and time. Webber [2002] concludes that team leaders are major agents for building quick trust within the team. Hsu [2006] supported the hypothesis that the relationship between team transformational

D. Kauffmann, How team leaders can improve virtual team collaboration

63

leadership and team trust has a significant correlation in software development teams which also supports Webber’s argumentation of a positive correlation between team leaders’ behavior and team trust. The importance of team leaders as mentors and contributors to the virtual team trust level has also been outlined

by DeRosa and Lepsinger [2010: 92], Duarte and Snyder [2011: 83], and Zofi [2012: 102]. They argue that the virtual environment makes their roles in trust building more crucial than a regular team. The next hypothesis of the model is: H1b: Team Leader behavior has an impact on Trust among the virtual team.

2.2. Communication, Trust, Collaboration and the relation between them In 2010, Roth conducted research to analyze Virtual Teams Effectiveness as a Function of using Computer-Mediated Communication (Figure 2). His model was formed of three main parts: Inputs, Processes and Outputs.

Figure 2. Roth’s Model [2010]

The collaboration processes is characterized by trust and communication richness while the communication in virtual teams is mostly, if not entirely, based on Computer-Mediated Communication. Roth explores the links between Inputs, Processes and Output but does not explore the links between Communication, Trust and Collaboration (his hypothesis only proposed a link between communication and trust). My model intends to explore the connections between these three elements.

Research has found that communication and coordination are fundamental elements associated with the collaboration in virtual teams [Mattessich, Murray-Close, and Monsey 2001; Qureshi, Liu, and Vogel 2006; Hosley 2010].

ICTs have positive effects on collaboration where the type of media [e.g. synchronous and asynchronous technologies] and the purpose of the communi-

64

Economics and Business Review, Vol. 1(15), No. 2, 2015

cation have impacts on the effectiveness [Qureshi, Liu, and Vogel 2006; Hosley 2010; Lau, Sarker, and Sahay 2000]. Each type of technology has benefits and constraints due to the nature of the technology [Lau, Sarker, and Sahay 2000; Kauffmann and Carmi 2014]. In other words, various media meet differing needs for the purposes of collaboration. As I have noted in the literature review, there are two kinds of communication that act out different aspects of the communication: Task-oriented and Social/Relationship-oriented [Huang 2010; Lau, Sarker, and Sahay 2000; Jarvenpaa and Leidner 1999]. In a virtual team context, Lau, Sarker, and Sahay [2000] refer to the task aspect as the part of communication that is specifically directed toward getting the project work done on time and within budget, and the Social aspect as the communication that is directed toward building social relationships and solidarity among virtual team members. Zaccaro, Rittman, and Marks [2001] refer to two task-oriented processes (team cognitive and coordination processes) and two relationshiporiented processes (team motivational and affective processes) as necessary for team effectiveness (Figure 3).

Figure 3. A model of leader performance functions contributing to team effectiveness Source: [Zaccaro, Rittman, and Marks 2001]

Based on the research on the correlation between Communication/ICT and collaboration, and the two aspects of communication. My third and fourth Hypothesis in the conceptual model are: H2a: The greater Relationship-oriented Communication via ICT, the greater Collaboration among the team members.

H2b: The greater Task-oriented Communication via ICT, the greater Collaboration among the team members.

Within a virtual environment, trust is mainly created via a communication behavior established in the first few keystrokes. To maintain this trust, it seems

D. Kauffmann, How team leaders can improve virtual team collaboration

65

to be necessary for the communication to gather team members around the project and tasks. Social communication that complements rather than substitutes for task communication may strengthen trust [Jarvenpaa and Leidner 1999]. Kasper-Fuehrera and Ashkanasy [2001] argue that without appropriate ICT to communicate trustworthiness, trust building in a virtual organization is compromised. Roth [2010] finds a high correlation between the richness of communication and the level of trust when the working hours and days of the team members overlap. Thomas [2010] supported the hypothesis that a significant relationship exists between virtual team trust and the use of communication technologies which also supports Roth’s finding of a positive correlation between trust and richness of ICT. My fifth Hypothesis is: H2c: The greater Richness of ICT, the greater level of trust among the Team.

2.3. Trust as a mediating factor for collaboration Trust has been identified by several scholars as an important ingredient for collaboration. In collaboration between two companies, trust has been identified as the primary basis for collaboration to be successful [Johnston et al.

2004]. In their analysis, they found that there is a relation between the degree of trust and the level of cooperation behavior. This finding was confirmed by research conducted by Osman [2004] where he also argued that without trust companies will not engage in business relationships at all. Through research on working teams, the impact of trust was tested on several performance variables like “level of collaboration”, “quality” and “timeliness”. The variable most affected by trust was “level of collaboration” [Martínez-Miranda and Pavón 2012]. This relationship between trust and collaboration has been also examined in virtual environments and findings confirm that trust has a positive impact on collaboration in such environments too [Leitch Peters 2003; Peters and Manz 2007]. Some scholars argue that trust has a direct, well-defined impact on collaboration and performance. In Trainer’s [2012] definition, “Trust, or more precisely perceived trustworthiness, is a crucial ingredient of effective and productive collaborations”. Others believe that the relationship is still illdefined. “All these studies show evidence that, in some way, the trust relationship between the members of a work team affects the performance of the team in its tasks or activities” [Martínez-Miranda, and Pavón 2012]. In her research, Marguin [2010] brings two different points of view expressed in academic studies about the relationship between trust and performance in virtual teams. The first point of view sees trust as an antecedent to success [DeRosa et al. 2004; Sarker and Valacich 2003]. The second argues that trust is a moderator-mediator factor and therefore has an indirect effect on success [Dirks 1999; Dirks and Ferrin 2001; Brahm and Kunze 2012; Qureshi, Liu, and Vogel 2006]. Based on the argument that trust is a moderator-mediator factor, and that communication, trust and collaboration are all linked [Roth 2010]: communication to

66

Economics and Business Review, Vol. 1(15), No. 2, 2015

Figure 4. The McAllister [1995] Model, outlining the role of trust in interpersonal relationships within an organization

collaboration [Qureshi, Liu, and Vogel 2006; Mattessich, Murray-Close, and Monsey 2001; Hosley 2010], trust to collaboration [Trainer 2012; MartínezMiranda and Pavón 2012] and communication to trust [Roth 2010; Thomas and Bostrom 2008], I claim that trust acts as a mediating factor between communication and collaboration.

Moreover, from the theory of cognitive and affective trust [McAllister 1995] (Figure 4) and from the distinction of two kinds of virtual team – Temporary and On-going – [Jusrud 2008], trust building and its development will be of a different nature if we are managing a temporary team as opposed to an ongoing team. In a virtual temporary team, focus must be on the cognitive dimension, whereas, in a virtual on-going team, we will need to develop both the cognitive and the affective dimensions. In the virtual on-going team, the affective dimension must play a primordial role if we wish to foster good interpersonal relationships throughout the team’s lifetime. As a result of these observations – trust acts as a mediator between communication and collaboration, and the differentiation between cognitive and affective trust in temporary and on-going virtual teams, my sixth and seventh Hypotheses are: H3a: Cognitive Trust will act as a mediating factor between communication and collaboration in Temporary Virtual Teams but Affective Trust will have an insignificant impact on it.

H3b: Both Cognitive and Affective Trust will act as a mediating factor between communication and collaboration in On-going Virtual Teams.

These seven hypotheses based on the research papers I have cited are the theoretical background of my conceptual model.

D. Kauffmann, How team leaders can improve virtual team collaboration

67

3. Conceptual model The models, theories and research I have raised in the preceding paragraphs allow the presentation of a conceptual model to define antecedents of collaboration and to determine “How Team Leaders can improve Virtual Team Collaboration through trust and ICT” (Figure 5).

Figure 5. A conceptual model for antecedents of collaboration within virtual environment

The model claims that the team leaders, by facilitating trust within their team and by mentoring by means of the use of the right ICT channel to communicate, can consequently improve the level of team collaboration. In order to accomplish it in the most effective way, team leaders have to take into consideration multiple factors. First the virtual team must be defined as a temporary or an on-going team. According to this team definition, team leaders decide what kind of trust is needed. If the team is a temporary one, team leaders need to focus their efforts on cognitive trust team building while in an on-going team both cognitive and affective trust are critical. Then depending on the required message, relationship- or task-oriented, the team members will be able to choose the right ICT channel to communicate thanks to team leaders’ mentoring. The more the team has a high level of trust, the greater the level of collaboration which will improve at every team communication event. Therefore the frequency of team communication is critical to develop collaboration aspects such as knowledge and information sharing, conflict management, problem solving, decision making, and innovation and creativity.

68

Economics and Business Review, Vol. 1(15), No. 2, 2015

4. Proposed method of research In order to validate the model and test its hypotheses, I propose to conduct research divided into stages.

The first stage will be to corroborate previous research that has been conducted on the correlation between team leaders and their impact on team communication (H1a) [Sivunen 2008], between team leaders and team trust building (H1b) [Webber 2002; Hsu 2006], between communication and collaboration (H2a, H2b) [Mattessich, Murray-Close, and Monsey 2001; Qureshi, Lim, and Vogel 2006; Hosley 2010] and finally between ICT and trust (H2c) [Jarvenpaa and Leidner 1999; Kasper-Fuehrera and Ashkanasyb 2001; Roth 2010; Thomas 2010]. This research has been conducted in a similar environment or context and their findings have already been validated. Therefore I will corroborate them in the context of the model through a qualitative research approach by an individual in-depth interview of team leaders and members of virtual teams. The variables that will be used to validate the qualitative correlations will be based on the variables used in the previous research. For example: virtual team leading practices, the communication routines and habits of the team, communication technology use and choice in the team used by Sivunen [2008] to investigate the interaction between team leaders’ behavior and team communication. Another example is: relationship building, responsiveness, team cohesion and accountability, and frequency and type of ICT used [Thomas 2010] to investigate the interaction between trust and ICT richness.

The second stage will be to test hypotheses that have never been validated to my knowledge. Therefore I propose to use the mix method of both quantitative and qualitative approaches. This method improves understanding arises when quantitative [numbers, trends, generalizability] and qualitative [words, context, meaning] approaches offset the different weakness of the two approaches [Brewer and Hunter 1989]. If we are examining the same phenomenon using multiple perspectives that represent different but complementary views, then we are more likely to gain a better, more complete understanding [Hesse-Biber and Leavy 2008]. These hypotheses (H3a, H3b) claim that cognitive and affective trust have a mediator impact between communication (social- and task-oriented) and collaboration depending also on the type of virtual team (temporary or on-going). For the qualitative approach, I propose an individual in-depth interview of team leaders and members of virtual teams and for the quantitative approach, a web-based questionnaire based on the Likert scale for an online survey. The correlation between Relationship and Task communication variables (as independent variables) and collaboration variables (as dependent variables) that are based on the five concepts associated with collaboration that I raised before: Knowledge & Information Sharing, Conflict Management, Problem Solving, Decision Making and, Innovation and Creativity will be used for measuring this empirical research. While cognitive and affective trust vari-

D. Kauffmann, How team leaders can improve virtual team collaboration

69

ables will be defined as mediator variables that alter the strength of the causal relationship between communication and collaboration. The linear multiple regression analysis method will be used for the measurement.

These two stages will allow corroboration of previous findings that are similar to my first hypotheses of the model and to support or reject the last hypotheses of the mediator role of the trust in the interaction between team communication through ICT and the team level of collaboration within a virtual environment.

Conclusions Due to the fast evolution of technology, Virtual Teams are more common every day. Organizations develop such teams because of many benefits, some of which have been raised by several scholars. Such Teams “facilitate around-theclock work and allow the most qualified individuals to be assigned to a team” [Wakefield, Leidner, and Garrison 2008] or “the availability of a flexible and configurable base infrastructure” [Ebrahim, Ahmed, and Taha 2009] are some of these advantages. However, companies meet several difficulties in order to make these teams as effective as they first thought. Indeed, these teams have not only positive sides but also due to the lack of communication, a high level of collaboration becomes a real challenge.

Based on research to date, the model proposes the identification of antecedents for collaboration in a virtual environment. According to the model, a high level of communication via ICT using both aspects of communication which are task-oriented communication and social/relationship-oriented communication will lead to a higher level of collaboration and an increase in its effectiveness. The impact strength of communication on collaboration is mediated by the level of trust existing between the team members. Depending on the nature of the team (temporary or on-going), the level of cognitive or affective trust will mediate differently. Cognitive trust will be crucial for both temporary and on-going teams but affective trust will have a minor influence on the temporary team whereas in on-going teams it will be also crucial to the maintenance of trust over time. Finally, Team leaders are a major agent for building trust within the team and for mentoring the team to increase its ICT utilization skills in a more efficient way.

This model can help organizations and team leaders to overcome the collaboration challenges by getting a better understanding of the virtual environment. Team leaders will be able to increase the level of collaboration within their teams by using the right communication channel (ICT media) and the right type of communication (Social or Tasks oriented). They will be aware of the importance of the trust team building process (including its cognitive and affective aspect) and its role as a mediating factor depending on the kind of virtual team with which they are working.

70

Economics and Business Review, Vol. 1(15), No. 2, 2015

References Abbas, W., Asghar, I., 2010, The Role of Leadership in Organizational Change: Relating the Successful Organizational Change with Visionary and Innovative Leadership, University of Gavle, June.

Ashley, J., 2003, Synchronous and Asynchronous Communication Tools, Executive Update Online, December.

Atteya, N.M., 2013, Examining the Effect of the Conflict Management Strategies on Job Performance, North American Business Press, West Palm Beach.

Barczak, G., Lassk, F., Mulki, J., 2010, Antecedents of Team Creativity: An Examination of Team Emotional Intelligence, Team Trust and Collaborative Culture, Creativity and Innovation Management, vol. 19, no. 4: 332–345.

Bell, B.S., Kozlowski, S.W., 2002, A Typology of Virtual Teams: Implications for Effective Leadership, Cornell University ILR School.

Bergiel, B.J., Bergiel, E.B., Balsmeier, P.W., 2008, Nature of Virtual Teams: A Summary of Their Advantages and Disadvantages, Management Research News, vol. 31, no. 2: 99–110.

Bos, N. et al., 2002, Effects of Four Computer-Mediated Communications Channels on Trust Development, Collaboratory for Research on Electronic Work [CREW].

Boughzala, I., de Vreede, G.-J., Limayem, M., 2012, Team Collaboration in Virtual Worlds: Editorial to the Special Issue, Journal of the Association for Information Systems, vol. 13, no. 10: 714–734.

Brahm, T., Kunze, F., 2012, The Role of Trust Climate in Virtual Teams, Journal of Managerial Psychology, vol. 27, no. 6: 595–614.

Brewer, J., Hunter, A., 1989, Multimethod Research: A Synthesis of Styles, Newbury Park, Sage.

Casalini, M.C., Janowski, T., Estevez, E., 2007, A Process Model for Collaborative Problem Solving in Virtual Communities of Practice, UNU-IIST, United Nations University International Institute for Software Technology, April.

Chudoba, K.M., Lu, M., Watson-Manheim, M.B., Wynn, E., 2005, How Virtual Are We? Measuring Virtuality and Understanding Its Impact in a Global Organization, Information Systems Journal, December, vol. 15, no. 4: 279–306.

Cohen, S.G., 1993, New Approaches to Teams and Teamwork, in: Organizing for the Future: The New Logic for, Jossey-Bass Publishers, San Francisco: 194–226.

Cummings, L.L., Bromily, P., 1996, The Organizational Trust Inventory [OTI], Trust in Organizations: 302–330.

De Dreu, C.K.W., Beersma, B., 2005, Conflict in Organizations: Beyond Effectiveness and Performance, European Journal of Work and Organizational Psychology, vol. 14, no. 2: 105–117.

De Dreu, C.K.W., Harinck, F., Van Vianen, A.E., 1999, Conflict and Performance in Groups and Organizations, International Review of Industrial and Organizational, vol. 14: 369–414.

De Jong, B.A., Elfring, T., 2010, How Does Trust Affect the Performance of Ongoing Teams? The Mediating Role of Reflexivity, Monitoring, and Effort, Academy of Management Journal, vol. 53, no. 3: 535–549.

D. Kauffmann, How team leaders can improve virtual team collaboration

71

Dennis, A., Wixom, B., Vandenberg, R., 2001, Understanding Fit and Appropriation Effects in Group Support Systems Via Meta-analysis, MIS Quarterly, vol. 25, no. 2: 167–197.

Den Otter, A.F., 2005, Design Team Communication and Performance Using a Project Website, Bouwstenen series of the Faculty of Architecture, Building and Planning of the Eindhoven University of Technology, Issue 98.

DeRosa, D.M., Hantula, D.A., Kock, N., D’Arct, J., 2004, Trust and Leadership in Virtual Teamwork: A Media Naturalness Perspective, Human Resource Management, vol. 43, no. 2–3: 219–232.

DeRosa, D.M., Lepsinger, R., 2010, Virtual Team Success: A Practical Guide for Working and Leading from a Distance, First ed. San Francisco, Jossey-Bass.

De Vreede, G.J., Briggs R.O., Massey A.P., 2009, Collaboration Engineering: Foundations and Opportunities, Journal of the Association of Information Systems, vol. 10, no. 3: 121–137.

Dillenbourg, P., 1999, What do You Mean by ‘Collaborative Learning’?, in: Dillenbourg, P. (ed.), Collaborative-learning: Cognitive and Computational Approaches, Elsevier, Oxford: 1–19.

Dirks, K.T., 1999, The Effects of Interpersonal Trust on Work Group Performance, Journal of Applied Psychology, vol. 84: 445–455.

Dirks, K.T., Ferrin, D.L., 2001, The Role of Trust in Organizational Settings, Organization Science, vol. 12: 450–467.

Duarte, D.L., Snyder, N.T., 2011, Mastering Virtual Teams, 3rd ed., Jossey-Bass, San Francisco.

Ebrahim, N.A., Ahmed, S., Taha, Z., 2009, Virtual Teams: a Literature Review, Australian Journal of Basic and Applied Sciences, vol. 3, no. 3: 2653–2669.

Erdem, F., Ozen, J., 2003, Cognitive and Affective Dimensions of Trust in Developing Team Performance, Team Performance Management: An International Journal, vol. 9, no. 5–6: 131–135.

Evans, N., 2012, Destroying Collaboration and Knowledge Sharing in the Workplace: A Reverse Brainstorming Approach, Knowledge Management Research & Practice, vol. 10: 175–187.

Fjermestad, J., Hiltz, S., 1998, An Assessment of Group Support Systems Experimental Research: Methodology and Results, Journal of Management Information Systems, vol. 15, no. 3: 7–149.

Ghaznavi, M., Toulson, P., Perry, M., Logan, K., 2013, Organisational Learning and Problem Solving through Cross-firm Networking of Professionals, Academic Conferences International Limited, Kidmore End. Grabner-Kräuter, S., Kaluscha, E., 2003, Empirical Research in on-line Trust: A Review and Critical Assessment, International Journal of Human-Computer Studies, vol. 58, no. 6: 783–812.

Gray, B., Wood, D.J., 1991, Collaborative Alliances: Moving from Practice to Theory, Journal of Applied Behavioral Science, vol. 27, no. 2: 3–22.

Gray, R., 2004, How People Work and How You Can Help Them to Give Their Best Harlow, FT, Prentice Hall.

Hackman, J.R., 1990, Groups that Work [And Those that Don’t] – Creating Conditions for Effective Teamwork, Jossey-Bass Publishers, San Francisco.

72

Economics and Business Review, Vol. 1(15), No. 2, 2015

Hassall, S.L., 2009, The Relationship Between Communication and Team Performance: Testing Moderators and Identifying Communication Profiles in Established Work Teams, Faculty of Business, School of Management, Qeensland University of Technology, Brisbane, Australia.

Hellriegel, D., Slocum, J., Woodman, R., 1998, Organizational Behavior, s. l, SouthWestern, Cincinnati, OH. Hertel, G.T., Geister, S., Konradt, U., 2005, Managing Virtual Teams: A Review of Current Empirical Research, Human Resource Management Review, vol. 15: 69–95.

Hesse-Biber, S.N., Leavy, P., 2008, Handbook of Emergent Methods, Guilford Press, New York.

Hirotani, M., 2009, Synchronous versus Asynchronous CMC and Transfer to Japanese Oral Performance, CALICO Journal, vol. 26, no. 3: 413–438.

Hosley, C.F., 2010, The Perceived Effects of Technology on Product Management Team Collaboration, UMI Dissertation Pulishing, Ann Arbor.

Hsu, S.-Y., 2006, Team Transformational Leadership, Trust, Empowerment, Satisfaction, and Commitment: Testing a Structural Equation Model in Software Development Teams, s. l, UMI Dissertations Publishing.

Huang, M., 2010, Behavior, Trust and Leader Emergence in Virtual Teams, The Stevens Institute of Technology, Hoboken, New Jersey, United States.

Jarvenpaa, S.L., Knoll, K., Leidner, D.E., 1998, Is Anybody Out There? Antecedents of Trust in Global Virtual Teams, Journal of Management Information Systems, vol. 14, no. 4: 29–64.

Jarvenpaa, S.L., Leidner, D.E., 1999, Communication and Trust in Global Virtual Team, Organization Science, vol. 10, no. 6: 791–815.

Jehn, K.A., 1995, A Multimethod Examination of the Benefits and Detriments of Intragroup Conflict, Administrative Science Quarterly, vol. 40: 256–282.

Johnston, D.A., McCutcheon, D.M., Stuart, F., Kerwood, H., 2004, Effects of Supplier Trust on Performance of Cooperative Supplier Relationships, Journal of Operations Management, vol. 22: 23–38.

Jones, G.R., George, J.M., 1998, The Experience and Evolution of Trust: Implications for Cooperation and Teamwork, Academy of Management Review, vol. 23: 532–546.

Jusrud, T.E., 2008, Trust Across Distance – A Network Approach to the Development, Distribution and Maintenance of Trust in Distributed Work Groups, Norwegian University of Science and Technology, Faculty of Social Sciences and Technology Management, Trondheim.

Karau, S.J., Kelly, J.R., 2004, Time Presssure and Team Performance: An Attentional Focus Integration, Research on Managing Groups and Teams, vol. 6: 185–212.

Kasper-Fuehrera, E.C., Ashkanasyb, N.M., 2001, Communicating Trustworthiness and Building Trust in, Journal of Management, vol. 27: 235–254.

Kauffmann, D., Carmi, G., 2014, How Team Leaders Can Use ICT to Improve Trust among Virtual Teams To Increase Collaboration?, International Journal of Engineering and Innovative Technology [IJEIT], vol. 3, no. 9: 204–221.

Kurland, N.B., Bailey, D.E., 1999, Telework: The Advantages and Challenges of Working Here, There, Anywhere, and Anytime, Organizational Dynamics: 53–67.

Lau, F., Sarker, S., Sahay, S., 2000, On Managing Virtual Teams, Healthcare Information Management, Communications Canada, vol. 2nd Quarter: 46–52.

D. Kauffmann, How team leaders can improve virtual team collaboration

73

Leitch Peters, L.M., 2003, Now You See Them, Now You Don’t: Toward a Greater Understanding of Virtual Team Effectiveness, Amherst, University of Massachusetts.

Lewicki, R.J., Bunker, B.B., 1995, Trust in Relationships: A Model of Trust Development and Decline, in: Conflicts, Cooperation, and Justice, Jossey-Bass, San Fransisco.

Lynch, R.P., 2007, The Architecture of Collaborative Innovation Using Cross-Boundary Alliances as Innovation Engines to Unleash the Power of the Value Chain, in: The Architecture of Collaborative Innovation, s. l, The Warren Company.

Mankin, D., Cohen, S.G., Bikson, T.K., 1996, Teams and Technology: Fulfilling the Promise of the New Organization, s. l, Harvard Business Press Books.

Marguin, J.A., 2010, A Meta-Analysis of Interpersonal Trust and Team Performance, Long Beach, s.n.

Martins, L.L., Gilson, L.L., Maynard, M.T., 2004, Virtual Teams: What Do We Know and Where Do We Go from Here?, Joumal of Management, vol. 30, no. 6: 805–835.

Martínez-Miranda, J., Pavón, J., 2012, Modeling the Influence of Trust on Work Team Performance, Simulation, vol. 88, no. 4.

Mattessich, P.W., Murray-Close, M., Monsey, B.R., 2001, Collaboration: What Makes It Work, 2nd ed., Fieldstone Alliance, Minnesota.

Mayer, R.C., Davis, J.H., Schoorman, F.D., 1995, An Integrative Model of Organizational Trust, Academy of Management Review, vol. 20: 709–734.

McAllister, D., 1995, Affect- and Cognition-based Trust as Foundations for Interpersonal Co-operation in Organizations, Academy of Management Journal, vol. 38, no. 1: 24–59.

McGrath, J.E., 1962, Leadership Behavior: Some Requirements for Leadership Training, U.S. Civil Service Commission, Office of Career Development, Washington, DC. McKnight, D.H., Cummings, L.L., Chervany, N.L., 1998, Initial Trust Formation in New Organizational Relationships, Academy of Management Review, vol. 23, no. 3: 473–490.

Meyerson, D., Weick, K.E., Kramer, R.M., 1996, Swift Trust and Temporary Groups, in: Trust in Organizations, s. l, Sage Publications, Thousand Oaks, CA: 166–195.

Michie, S.G., Dooley, R.S., Fryxell, G.E., 2006, Unified Diversity in Top-level Teams: Enhancing Collaboration and Quality in Strategic Decision-making, International Journal of Organizational Analysis, vol. 14, no. 2: 130–149.

Mishra, A.K., 1996, Organizational Responses to Crisis: The Centrality of Trust, Trust in Organizations: 261–287.

Morgeson, F.P., DeRue, D.S., Karam, E.P., 2010, Leadership in Teams: A Functional Approach to Understanding Leadership Structures and Processes, Journal of Management, vol. 36, no. 1: 5–39.

Mourtos, N.J., DeJong Okamoto, N., Rhee, J., 2004, Defining, Teaching, and Assessing Problem Solving Skills, 7th UICEE Annual Conference on Engineering Education, Mumbai.

Mumbi, C., 2007, An Investigation of the Role of Trust in Virtual Project Management Success, Technology of Murdoch University.

O’Neil, H.F., 1999, Perspectives on Computer-based Performance Assessment of Problem Solving, Computers in Human Behavior, vol. 15: 225–268.

Osman, B., 2004, Antecedents to Effective Collaboration to Innovate, York University, Toronto.

74

Economics and Business Review, Vol. 1(15), No. 2, 2015

Pazos, P., Ustun, A., DelAguila, R., 2011, The Role of Conflict Management on Virtual Team Performance and Satisfaction, Institute of Industrial Engineers-Publisher, Norcross.

Peters, L.M., Manz, C.C., 2007, Identifying Antecedents of Virtual Team Collaboration, Isenberg School of Management, University of Massachusetts, Amherst Massachusetts, USA. Powell, A., Piccoli, G., Ives, B., 2004, Virtual Teams: A Review of Current Literature and Directions for Future Research, The DATABASE for Advances in Information Systems, vol. 35, no. 1: 6–36.

Pulnam, L.L., 1992, Rethinking the Nature of Groups in Organizations, in: Small Group Communication: A Reader, s. l, William C Brown, Dubuque, IA: 57–66.

Quinn, R.E. et al., 2010, Becoming a Master Manager: A Competing Values Approach, s. l, John Wiley & Sons Inc. Qureshi, S., Liu, M., Vogel, D., 2006, The Effects of Collaboration in Distributed Project Management, Group Decision and Negotiation, vol. 15: 55–75.

Raghu, T., Ramesh, R., Chang, A.-M., Whinston, A.B., 2001, Collaborative Decision Making: A Connectionist Paradigm for Dialectical Support, Information Systems Research, vol. 12, no. 4: 363–383.

Roth, I., 2010, Virtual Teams Effectiveness as a Function of Using CMC, Recanati Business School, Tel-Aviv University.

Salas, E., Dickinson, T.L., Converse, S., Tannenbaum, S.I., 1992, Toward an Understanding of Team Performance and Training, Teams: Their Training and Performance: 3–29.

Sarker, S., Valacich, J., 2003, Virtual Team Trust: Instrument Development and Validation in an IS Educational Environment, Information Resource Management Journal, vol. 16, no. 2: 35–55.

Saunder, C.S., Ahuja, M.K., 2006, Are All Distributed Teams the Same? Differentiating Between Temporary and Ongoing Distributed Teams, Small Group Research, vol. 37: 622–700.

Schaubroeck, J., Lam, S.S.K., Peng, A.C., 2011, Cognition-Based and Affect-Based Trust as Mediators of Leader Behavior Influences on Team Performance, Journal of Applied Psychology, vol. 96, no. 4: 863–871.

Sivunen, A., 2008, The Communication of Leaders in Virtual Teams: Expectations and Their Realization in Leaders’ Computer Mediated Communication, The Journal of E-working, vol. 2: 47–60.

Sørensen, E., Torfing, J., 2012, Collaborative Innovation in the Public Sector, The Innovation Journal: The Public Sector Innovation Journal, vol. 17, no. 1: 1–14.

Thomas, D., Bostrom, R., 2008, Building Trust and Cooperation through Technology Adaptation in Virtual Teams: Empirical Field Evidence, Information Systems Management, vol. 25: 45–56.

Thomas, V.B., 2010, Virtual Team Effectiveness: An Empirical Examination of the Use of Communication Technologies on Trust and Virtual Team Performance, s. l, UMI Dissertations Publishing.

Thomson, A.M., Perry, J.L., 2006, Collaboration Processes: Inside the Black Box, Public Administration Review, vol. 66, no. 6: 20–31.

Thomson, A.M., Perry, J.L., Miller, T.K., 2010, Linking Collaboration Processes and Outcomes: Foundations for Advancing Empirical Theory, Big Ideas in Collaborative Public Management: 97–120.

D. Kauffmann, How team leaders can improve virtual team collaboration

75

Trainer, E.H., 2012, Supporting Trust in Globally Distributed Software Teams: The Impact of Visualized Collaborative Traces on Perceived Trustworthiness, University Of California, Irvine, UMI Dissertation Publishing.

Turban, E., Leidner, D., McLean, E., Wetherbe, J., 2006, Information Technology for Management: Transforming Organizations in the Digital Economy, John Wiley & Sons, Hoboken.

Turban, E., Liang, T.-P., Wu, S.P., 2011, A Framework for Adopting Collaboration 2.0 Tools for Virtual Group Decision Making, Group Decision and Negotiation, vol. 20, no. 2: 137–154.

Ustun, A., Pazos, P., 2012, The Impact of Online Collaboration Spaces on Virtual Team Outcomes, s. l, s.n.

Van Gelder, S., 2011, The Effectiveness of Knowledge Sharing and Collaboration in Creating High Performance Work Teams, Pepperdine University: UMI Dissertations Publishing.

Wakefield, R.L., Leidner, D.E., Garrison, G., 2008, A Model of Conflict, Leadership, and Performance in Virtual Teams, Information Systems Research, vol. 19, no. 4: 434–455.

Warkentin, M.E., Sayeed, L., Hightower, R., 1997, Virtual Teams versus Face-to-Face Teams: An Exploratory Study of a Web-based Conference System, Decision Sciences, vol. 28, no. 4.

Webber, S.M., 2002, Leadership and Trust Facilitating Cross-functional Team Success, Journal of Management Development, vol. 21, no. 3–4: 201.

Williams, M., 2001, In Whom We Trust: Group Membership as an Affective Context for, Academy of Management Review, vol. 26: 377–396.

Yukl, G.A., 2012, Leadership in Organizations, 8th ed., Prentice Hall, NJ. Zaccaro, S.J., Rittman, A.L., Marks, M.A., 2001, Team Leadership, Leadership Quarterly, vol. 12: 451–483.

Zofi, Y., 2012, A Manager’s Guide to Virtual Teams, AMACOM, New York.

Economics and Business Review, Vol. 1 (15), No. 2, 2015: 76–94

International trade in differentiated goods, financial crisis and the gravity equation1 Udo Broll, Julia Jauer2

Abstract: The study examines the effect of financial crises on international trade with a gravity equation approach. We use a large data set covering almost 70 importing and 200 exporting countries from 1950 to 2009. Thus it is possible to put the ‘Great Trade Collapse’ witnessed during the financial crisis 2008–2009 into a historical perspective.

Both the period for which the crisis is observed, and the level of the trading partners’ economic development constitute important factors in explaining the negative effects of a banking crisis on international trade. As the analysis indicates financial crises have a stronger negative effect on differentiated goods compared to overall export flows. In addition the negative effects of financial crises persist even after the income effect is accounted for. The study therefore suggests that the increasing share of differentiated goods in international trade might be one possible reason for the comparatively large effect of the recent financial crisis on international trade relative to previous financial turmoil in post-war economic history.

Keywords: international trade, differentiated goods, globalization, financial crisis, gravity equation.

JEL codes: F13, F14, F33, G15.

Introduction In the first quarter of 2009 international trade declined nearly 30 per cent compared to the same period one year before. From the 1950’s onwards, both global economic capacity and complexity were growing with expanding international trade and as complex global interdependencies developed. As a consequence national economies became interlinked globally. This interde1

Article received 02 September 2014, accepted 19 February 2015.

We would like to thank our anonymous referee for very helpful comments and suggestions.

2 School of International Studies (ZIS), Technische Universität Dresden, Faculty of Business and Economics, Helmholtzstrasse 10, 01062 Dresden, Germany, corresponding author: udo. broll@tu-dresden.de.

U. Broll, J. Jauer, International trade in differentiated goods, financial crisis

77

pendence did not result in positive gains only; it can also lead the entire world economy into turmoil if one economy malfunctions. This happened when the initial US housing market crisis in 2007 became a world financial crisis in 2008/2009 with effects in the real economy [Didier, Hevia, and Schmukler 2010; Admati and Hellwig 2013]. The world economy experienced one of the broadest, deepest, and most complex crises since the Great Depression and it lead to a severe decline in trade relative to gross domestic product (around twenty per cent) unobserved since 1929, that came to be called “the great trade collapse” [Baldwin 2009].

The link between the economic crisis and the decline in international trade is a complex one. Previous studies suggest a negative relationship between the financial crisis and international trade, but some contradictions remain [Baldwin 2009]. Whilst some studies emphasize the role that declining overall demand had on decreasing trade flows downplaying or rejecting the effect trade finance might have had other contributors s point out the particular importance that trade finance has for international trade especially in times of financial turmoil. And yet other studies bring forth a compositional argument highlighting the fact that trade is composed of very different commodities and sectors, which might react differently to a financial crisis. The internationalisation of production chains, so called vertical linkages, was mentioned as one key factor in the massive trade decline.

Our study tries to provide additional insight into the question of the impact of the financial crisis on international trade. However, it treats a financial crisis as an exogenous event. The paper is related to the empirical so-called gravity approach trade literature. It sets out to examine the underlying factors driving the trade slump in 2008–2009 using a gravity type trade flow model incorporating country specific characteristics and including external shocks. The analytical framework is based on a recent study by Berman et al. [2012] and by adding the element of disaggregated trade flows. The study provides new evidence on the way the financial crisis affected trade this time around.

There are other approaches in the literature to investigate the question why trade collapsed during the crisis. A prominent example is, Eaton et al. 2013.

They use a framework that includes recent development in general equilibrium models of bilateral international trade into a multi-country real business cycle model. The goal of this approach is to provide a mapping between observables and an underlying set of shocks in order to separate the shocks from observables and counterfactual shocks.

In the following Section 1, the related literature is reviewed. In Section 2 the effects of financial crises on international trade are discussed and causes and causalities are addressed leading to some testable hypotheses. Section 3 describes the theoretical approach of the standard gravity model, the estimation method applied and the data used. Section 4 presents and evaluates the empirical results. The next Section concludes.

78

Economics and Business Review, Vol. 1(15), No. 2, 2015

1. Related literature: trade and financial crises The financial crisis inspired a series of studies on financial, banking and economic crises [see, for example, Admati and Hellwig 2013]. An extraordinary example of providing empirical research on cycles of debt, financial, currency and sovereign debt crises was done by Reinhart and Rogoff [2011], who also provide a publicly accessible data set dating back to the 19th century. Before the economic and financial crisis in 2008–2009 the examination of the relationship between financial crises and trade as a whole was sparse. A collection of essays, edited by Baldwin [2009], offers a good overview on the subject of trade decline in the recent financial crisis. According to Baldwin’s calculations the decrease in trading volume for the second quarter of 2008 to second quarter of 2009 was 20 percent and for some countries even 30 percent.

Frictions in trade finance and the drying up of trade credit during the financial crisis are expected to have an effect on trade [see Chor and Manova 2012].

Studies that focus particularly on trade finance during financial crises usually have a strong regional focus on global banking centres or are country specific [see Amiti and Weinstein 2011].

Bricongne et al. [2012] also examine the compositional effect of external finance on trade with French firm level data and find that the firms that are more dependent on external finance are more affected by the crisis. In general, all studies find strong support that vertical linkages are quantitatively important in understanding the global trade collapse. Global production patterns can thus be expected to explain part of the massive decline in international trade this time around, because the international supply chain has intensified over the last couple of decades. The sensitivity of trade towards output has increased over time and an underlying reason for this could be the growing share of certain goods which react in a more volatile way to economic frictions than total output [Engel and Wang 2011]. International production sharing or vertical specialisation means that trade reactions are increasingly sensitive to changes in trade costs.

Furthermore, empirical studies distinguish between differentiated and nondifferentiated goods and find that this distinction is crucial for understanding the extent to which price declines contributed to the decline in trade values.

Only very few studies applied the gravity approach to examine the trade collapse during 2008–2009. An exceptional study Berman et al. [2012] analyzes the effect of the recent financial crisis on international trade covering the whole post-war era on a global scale and using a gravity-based approach. The fall in trade caused by financial crises is magnified by the time-to-ship goods between the origin and the destination country. The authors strongly suggest that financial crises affect trade not only through demand but also through financial frictions that are specific to international trade.

Globalisation and the internationalisation of production patterns, however, have not been fully addressed by this study although previous studies suggest

79

U. Broll, J. Jauer, International trade in differentiated goods, financial crisis

an important role for these in trade. A study of Eaton et al. [2013] includes also an element of the gravity model to calculate an indicator of trade frictions between individual countries. They conclude that the bulk of the decline in trade relative to GDP may be explained by shocks in the industrial demand for goods and it is only in some countries like China and Japan that trade decline can be explained to a large extend by increased trade frictions.

2. The effects of the financial crisis on trade The years directly after World War II were remarkably tranquil and marked by the quasi-absence of banking crises. If financial crises emerged at all, they were strictly currency crises. The Bretton Woods Agreements and the gold exchange standard stabilized global economic frictions. The fixing of countries exchange rates relative to the US-dollar was abandoned in 1971. Since then banking crises were more frequent and the share of countries experiencing banking crises was rapidly increasing until recently (see Figure 1 below).

Share of banking crises

% 40 30 20 10

2007 2009

2004

2001

1998

1995

1992

1989

1986

1983

1980

1977

1974

1971

1968

1965

1962

1959

1956

1953

1950

0

Figure 1. Share of banking crisis in the post-war period (1950–2009) Source: Authors calculation [based on: Reinhart and Rogoff 2011]

The crisis in Latin America of the 1970’s and 1980’s, the Japanese banking crisis in the early 1990’s, the European and the Asian financial crises are well visible as peaks. The impetus of the share of countries experiencing banking crises in 2008/2009 came after a period that was relatively calm compared to the last ten years. The crisis had a tremendous effect on international trade.

Even though the global economy has seen financial crises before 2008/2009, international trade declined for the first time after fifty years of more or less continuously rising trade volumes. In 2009 both developed, developing and emerging countries were experiencing trade declines, but developed countries

80

Economics and Business Review, Vol. 1(15), No. 2, 2015

Trillions

% 18 16 14 44% 12

56%

10 8 Share of total trade decline in 2009 in %

6 4 2

Developing and emerging countries

Developed countries

2007 2009

2004

2001

1998

1995

1992

1989

1986

1983

1980

1977

1974

1971

1968

1965

1962

1959

1956

1953

1950

0

World

Figure 2. Import flows 1950–2009 for developing, emerging and developed countries in trillion US$ and their share in the trade collapse in 2009 Source: Authors calculation (based on International Monetary Fund DOTS)

had a relatively higher share in the decline of the total global trade decrease (see Figure 2).

The decline in trade flows was significant and on a global scale. Almost all countries experienced declines in exports and imports. Even though developed countries accounted for the larger share of the total trade decline in 2009 some developing and emerging countries’ exports were also hit hard during the financial crisis.

Previous literature suggests that the internationalization of production chains could account for the increased volatility of international trade in crisis. The importance of vertical linkages and the global production chain is well visible when looking at the share of differentiated goods3 in total trade. The share of differentiated goods has increased dramatically in the last fifty years. In the early 1960’s their share in total imports was just over 45 percent and reached a peak in the early 2000’s with over 72 percent of all imports (see Figure 3).

Besides GDP there are other indicators for trade promotion and trade disruption which need to be considered when analysing the trade decline of 2008/2009. Taking into account the behaviour of trade flows it is possible to identify the main possible factors by which international trade was affected by financial crisis. The OECD [2010] describes three direct ways how the financial crisis affected international trade. International trade is affected by the financial crisis through: (i) global demand and income, (ii) international trade finance 3

The definition of differentiated goods follows Rauch’s [1999] conservative definition.

81

U. Broll, J. Jauer, International trade in differentiated goods, financial crisis % 75 70 65 60 55 50 45

2010

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

1968

1966

1964

1962

40

Figure 3. Share of differentiated goods of all imports (1962–2010) Source: Authors calculation [based on: UN COMTRADE; Rauch 1999]

and (iii) composition of internationally traded goods. In light of the literature review a fourth channel of how financial crisis affects trade can be added. The way in which goods traded react to financial crisis also depends on the degree of economic development in exporting countries.

2.1. The income effect The financial crisis affects international trade indirectly through reduced consumption and therefore through the decline in demand for goods [Eaton et al. 2013]. With a declining demand for foreign goods fewer imports are purchased and fewer exports are sold. The drop in demand has significantly contributed to the drop in trade but it cannot explain it fully. Thus the decrease of income due to the financial crisis is only one factor in explaining the decline in international trade.

2.2. The international trade finance effect The availability of financial services is important for firms engaging in international business. During the financial crisis the sensitive cooperation of international financial service was severely disturbed and this affected international trade. Thus the price increase in trade financing or the absence of it has led to a decrease in global trade flows. This applies especially to r developing countries which might have suffered from increased risk perception and therefore more expensive trade finance [Broll et al. 2001; Berman et al. 2012]. Information on detailed trade finance on a global scale is very difficult to obtain, especially for emerging and developing countries with less integrated and less developed banking and financial systems. In response to the change in market conditions

82

Economics and Business Review, Vol. 1(15), No. 2, 2015

on trade finance, the International Monetary Fund (IMF) has undertaken a survey of major developed countries and emerging markets’ banks. According to the IMF [2009a, b] several banks reported sharp increases in the cost of trade finance; seventy percent of the surveyed banks reported that the price for trade finance services has increased.

2.3. The trade compositional effect A World Bank survey indicates that the biggest financing constraint particularly for firms operating in global supply chains is not access to trade credit (e.g., letter of credit) per se, but rather pre-export finance. Differentiated goods, such as investment and consumer goods, are therefore more demanding in terms of finance structures making them particularly vulnerable to a financial crisis. This observation indicates that a crisis might affect exports within global supply chains or with vertical linkages in a more severe way than other goods because in times of crises they require specific financial provisions which they otherwise would not need and which are even harder to obtain in times of financial turmoil. Thus, the composition of international trade has led to a distinctive decline of trade flows during the financial crisis.

2.4. The economic development effect The way international trade reacts to financial crises depends also on the level of economic development of a country. On the one hand developing countries can be more dependent on trade exports relative to their GDP than developed economies. A trade slump therefore can have an amplified effect for developing countries. Available data indicate that trade in some regions – Asia, the Middle East, Northern Africa and South America – was more severely impacted by changes in short-term trade finance than other regions (Europe and North America). This may be due to the fact that countries in these regions were considered to be more exposed to risk. Therefore trade finance prices became less affordable for those countries. On the other hand, the lack of integration with the international financial system could have been a blessing in protecting developing and emerging countries against negative chain reactions [see for example Didier, Hevia, and Schmukler 2011].

The compositional effect of international trade is also quite different regarding the level of economic development. In general developing countries’ exports differ from the exports of developed countries. If differentiated goods have a higher elasticity then other exports, developing countries might react differently compared with developed countries in times of crisis.

To sum up the arguments can be expressed in the following hypotheses.

H0: Financial crises have, in general, a negative impact on international trade.

H1: Financial crises reduce trade due to income effects.

H2: Trade finance has played a role in trade disruption.

83

U. Broll, J. Jauer, International trade in differentiated goods, financial crisis

H3: Differentiated goods are more sensitive to financial frictions during financial crises.

H4: Emerging and developing countries differ in trade composition and access to trade finance during financial crises.

3. Gravity trade equation, estimation method and data The trade gravity approach has long been one of the successful models in international economics [Anderson and van Wincoop 2003, 2004]. The gravity equation is fundamentally about inferring trade costs in a setting where much of what impedes trade is not per se observable to the econometrician because there is only limited information on direct measures of trade costs.4 However, trade flows and proxies for different types of trade costs are observable.

3.1. Gravity equation The gravity model states that export and import flows between countries (or regions) depend positively on the GDP of the trading partners (as a measure of economic size) and negatively on geographical distance (as a proxy for transaction costs): force of trade

º G ª¬YY i j¼ dist i , j

elasticity 1 , with

§ 1 G {¨ © Ωi

·§ 1 ¸ ¨¨ 1 elasticity ¹ © Pj

· ¸. ¸ ¹

(1)

GDP of country i (j) is denoted by Yi (Yj); Pj is country j’s price index and Ωi a proxy to what is called market potential in the economic geography literature (often measured by the sum of its trade partners GDPs divided by bilateral distance). It is an indicator for openness of a country i. The term G is the ‘gravital un-constant’, because it varies over time (as prices and GDPs change). This is what Anderson and van Wincoop called the multilateral trade resistance. Two countries will often exchange more goods and services the bigger they are and also the less further apart that they are. The average elasticity of international trade is estimated close to unity to around 0.9 says a meta-study by Disdier and Head [2008] on standard gravity estimations.

A structural gravity model has the following general form: Xij

4

§ t ji YY i j ¨ Yw ¨© Pi Pj

1 σ

· ¸ . ¸ ¹

(2)

For a comprehensive and up to date introduction of the theories behind the gravity model see: [Chaney 2008; Head and Mayer 2013].

84

Economics and Business Review, Vol. 1(15), No. 2, 2015

The volume of exports Xij of a country i to country j in equation (2) is explained by the relative size of the exporter (measured as a proportion of income Yi), the importer (measured as a proportion of income Yj) and of the world GDP YW. In addition exports depend on the bilateral trade cost tji, which are all trade barriers. The elasticity of substitution between different types of goods is recognized by Ď&#x192;.

The unobservable trade cost factor tij can be written as a log-linear function of observable characteristics, namely as the bilateral distance dij and whether there is an international border bij between i and j. After some rearrangements we obtain the theoretical gravity equation:

ln Xij

β ln P β ln P   ξ ,

Îą β1 ln Yi β2 ln Y j β3 Ď ln dij β4 ln bij 5

i

6

j

(3)

ij

where Îą is a constant, Ď is (1 â&#x20AC;&#x201C; Ď&#x192;) and Îľ represents the normally distributed error term. Equation (3) is expanded by adding other factors to the trade cost as dummy variables such as common language, common currency, free trade union member, currency unions and common border. The fact that history played a role in shaping trade relationships can also be accounted for by including colony dummies, controlled by the same colonizer, similarities in religion, legal system and military conflicts [see Martin, Mayer, and Thoenig 2008; Head, Mayer, and Ries 2010].

3.2. Estimation method A consensus has emerged in gravity literature on the use of fixed effects (FE) for panel trade flows. According to the literature replacing the remoteness variable with exporter and importer dummies to proxy for multilateral resistance is the â&#x20AC;&#x2DC;correct specificationâ&#x20AC;&#x2122; of the gravity model. Exporter and importer dummies will be added to equation (3). Estimation will be made for country pair dummies. To measure the effect that banking crises have on trade flows, a further adjustment has to be made. A dummy variable BCjt for banking crises is introduced, which takes one when a banking crisis has occurred in the importing country at time t.

ln Xijt

β developing  country  Ο

Îą β1 ln Yit β2 ln Y jt β3 Ď ln dij β4 ln bij β5ln RERijt

β6 Tijt β7 BC jt

8

ij

ij

 ξijt ,

(4)

where Tijt stands for a set of time-varying bilateral controls (like regional trade agreements, colonial relationships, currency unions, etc.). Because for a large sample of countries representative price indexes are not available, the best al-

U. Broll, J. Jauer, International trade in differentiated goods, financial crisis

85

ternative is to use real exchange rate (RER) indexes. RERijt is the bilateral real exchange rate between country i and j. Fixed effects are included with μij. This allows control of all time independent country specific characteristics which might influence the bilateral trade relation.

The second dummy introduced in the equation (4) is the level of development for the partner’s economy. To assess whether the financial crisis in the importing country has a different effect on developing and emerging countries the level of economic development dummy is included. The effect of a banking crisis on trade is expected to be negative. It is not clear however how exports from developing and emerging countries are affected by the banking crises because one can argue both ways. It is however tested that they have a different reaction to banking crises than developed countries.

It is assumed that differentiated goods are more prone to the effects of financial crises (our hypotheses H3). Equation (4) estimates separately for exports of differentiated goods only. The negative effect of a banking crisis on trade is expected to be stronger for differentiated goods. Looking at the banking crises before the global financial crises in 2008 and the recent crises separately will allow an inference to be made about the different impact the financial crisis had on trade in this period.

3.3. Data Trade data come from two sources. Aggregated trade flows are obtained from the International Monetary Fund Direction of Trade Statistic (IMF DOTS) dataset which covers trade flows of almost all country pairs of the world from 1949 onwards. Reported bilateral imports by the importing countries will be used. Disaggregated trade flows come from the UN COMTRADE. The data are transformed in a series of steps to match the classification of disaggregated goods provided by Rauch [1999].5 The data for banking crisis comes from Reinhart and Rogoff [2011], period 1950 until 2009. The occurrence of a banking crisis for a given year is marked by a binary variable for up to 69 different countries.

Banking crisis are defined by two types of events: (i) bank runs that lead to the closure, merging, or takeover by the public sector of one or more financial institutions; or (ii) if there are no runs, the closure, merging, takeover, or large-scale government assistance of an important financial institution (or group of institutions) that marks the start of a string of similar outcomes for other financial institutions.

Gravity relevant variables come from the Centre d’Etudes Prospectives et d’Informations Internationales (CEPII) gravity data set which is publicly available and covers the period from 1948–2006. The updated version up to 2009 will be used for the estimations below. The definition for developed countries versus developing and emerging countries comes from the United Nations 5

On request the authors can provide more detailed information about this procedure.

[86]

cursib

curcol

comcol

comlang_off

rta

comcur

contig

lnexchange_rate_o_d

lndistw

lngdp_d

lngdp_o

UN COMTRADE 0.856c (0.000) 0.837c (0.000) –1.307c (0.000) –0.002 (0.407) 0.480c (0.000) 0.101 (0.126) 0.092c (0.000) 0.658c (0.000) 0.267c (0.000) 1.157c (0.000) 0.405 (0.454)

IMF DOTS

0.923c (0.000) 0.816c (0.000) –1.046c (0.000) 0.007c (0.000) 0.532c (0.000) 0.303c (0.000) 0.494c (0.000) 0.557c (0.000) 0.359c (0.000) 1.040c (0.000) 1.520c (0.000)

Table 1. Gravity trade model for export flows

1.085c (0.000) 0.835c (0.000) –1.094c (0.000) 0.005c (0.006) 0.479c (0.000) 0.050 (0.322) 0.343c (0.000) 0.555c (0.000) 0.353c (0.000) 0.886c (0.000) 0.822c (0.001)

IMF DOTS same

UN COMTRADE same 0.929c (0.000) 0.890c (0.000) –1.199c (0.000) –0.000 (0.907) 0.454c (0.000) 0.062 (0.345) 0.219c (0.000) 0.620c (0.000) 0.100b (0.033) 1.645c (0.000) 1.560c (0.000) 0.236b (0.028) 1.509c (0.004)

0.442c (0.000) 0.575c (0.000)

–0.005c (0.000)

0.852c (0.000) 0.777c (0.000)

IMF DOTS

0.923c (0.000) 1.283c (0.003)

–0.019 (0.614) 0.342c (0.000)

–0.002 (0.297)

0.844c (0.000) 0.780c (0.000)

UN COMTRADE

0.474b (0.016) 1.057c (0.000)

0.181c (0.000) 0.423c (0.000)

–0.006c (0.000)

0.961c (0.000) 0.844c (0.000)

IMF DOTS same

1.382c (0.000) 2.002c (0.000)

–0.021 (0.539) 0.357c (0.000)

–0.001 (0.455)

UN COMTRADE same 0.888c (0.000) 0.811c (0.000)

[87]

0.390c (0.000) 0.159c (0.000) 271,210 0.766 Yes No

0.406c (0.000) 0.224c (0.000) 317,045 0.703

Yes

No

No

Yes

0.435c (0.000) 0.215c (0.000) 218,652 0.734

a b

c

No

Yes

0.351c (0.000) 0.208c (0.000) 218,652 0.761

Yes

No

317,114 0.836

Yes

No

271,210 0.879

Yes

No

218,652 0.869

Yes

No

218,652 0.887

Source: [IMF DOTS; UN COMTRADE; Reinhart and Rogoff 2011; CEPII Gravity dataset (update)]; author’s estimation.

Exogenous Variables (table 1–3) gdp_o gdp origin (exporting country) gdp_d gdp destination (importing country) distw distance weighted contig contiguity comcur common currency rta member of the same regional trade agreement comlang_off common official language comcol common colonizer post 1945 curcol current colonial relationship cursib currently the same colonize comrelig common religion comleg_posttrans common legal origin BKdDevel banking crisis and developing countries (interaction effect) globalgdpslowdown global GDP slow down (when GDP grew less than 3 percent a year)

Robust standard errors in parentheses, clustered by destination-year, with , , and respectively denoting significance at the 1%, 5% and 10% levels. Year dummies are included in all estimations.

Observations R2 Exporter and importer fixed effects Country-pair fixed effects

comleg_posttrans

comrelig

88

Economics and Business Review, Vol. 1(15), No. 2, 2015

Statistics Division [UN STATS].6 There are two final data sets used for the following estimations. One is based on the IMF DOTS trade flows and consists of 69 importing countries and 206 exporting countries from 1950 to 2009 and the second data set consists of 68 importing countries and 198 destination countries from 1962 to 2009. The lower number of importing countries for both datasets is due to the limited data of banking crises.

4. Empirical results Estimation results, implied by equation (3) for two different data sets, are shown in Table 1. The overall fit is promising with R2 ranging from 0.70 to 0.89. All the variables have the expected sign and plausible values. As suggested by the theory the elasticity of trade with respect to income is significant and close to unity.

Column 1 and 5 display the estimation results for the full available sample of IMF DOTS bilateral export flows. The first column is estimated with exporter and importer fixed effects and column 5 is estimated with country pair fixed effects. The time invariant variables are therefore omitted in the country-pair fixed effects estimations in columns 5 to 8. Column 2 and 6 show the estimates for the full available sample for UN COMTRADE bilateral export data. Column 3, 7, 4 and 8 present the results for the sample available for both IMF DOTS and UN COMTRADE respectively. The estimated coefficient values for the same years of observation and trading partners slightly differ for IMF DOTS and UN COMTRADE bilateral export data but they correspond in sign and magnitude.

The following estimations will exploit the richness of the full samples available keeping in mind that IMF DOTS covers a longer time span – 12 years more – than the UN COMTRADE data.

Table 2 displays estimation results for equation (4). The estimated coefficient of banking crisis dummy variable is only significant for columns 5 and 6; the coefficient is negative as expected and its magnitude is in line with previous studies [Berman et al. 2012]. A dummy variable included for developing and emerging countries reveals a significant positive effect. If estimations include an interaction effect of the dummy for developing or emerging countries and financial crisis an interesting effect becomes visible. Exports involving only developed countries are more negatively affected by banking crises than exports for trading partners in developing and emerging countries.

The interaction effect of banking crisis and developing countries is statistically significant and strongly positive both for the estimations with IMF DOTS data and with UN COMTRADE data. The coefficient for banking crises when controlled for the level of economic development becomes statistically significant with a negative impact on exports between 27.1 and 41.3 per cent.

6

http://unstats.un.org/unsd/methods/m49/m49regin.htm.

U. Broll, J. Jauer, International trade in differentiated goods, financial crisis

89

When the estimations of Table 2 are replicated only for differentiated goods (results not displayed but available upon request) the negative effect of the banking crises gets even larger (32.4 compared to 27.1 per cent). International trade in differentiated goods are more vulnerable to financial frictions than homogenous goods.

When the banking crises dummy is split into two variables, a dummy for the recent crisis and a dummy for the previous banking crises, the results support the hypothesis that the effect on trade of the recent crisis was different compared to previous crises. The financial crises prior to 2008 had no statistically significant effect. For aggregated export flows a statistically significant negative effect of the recent financial crisis on exports between 25.2 and 26.8 per cent is estimated. The effect on differentiated goods of the recent financial crisis is even 10 per cent higher. When controlled for the level of economic development for the trading partner, this effect is more pronounced. For differentiated goods however both effects of the financial crises are magnified. For the previous financial crises a significant negative effect on trade of 31.3 per cent is estimated. The effect of the recent financial crisis is 47.7 per cent.

Table 3 shows the result from testing the income effect on trade during a financial crisis. We use an additional dummy variable for the income effect. The dummy variable is equal to 1 when the global GDP grew less than 3 percent in one year following the classification of a recession by the IMF. The coefficient for the slowdown variable is significantly negative using the UN COMTRADE export aggregate and disaggregated data even when controlling GDP. Thus exporting responds more negatively than the GDP effect alone would imply. This result also holds good when including interaction effects of the level of economic development and a banking crisis. In addition to the negative effect of GDP slowdown on exports the effect of financial crisis continues to be significantly negative. This suggests that the statistically negative coefficient of a banking crisis holds beyond a recession effect. Other components of financial crises, such as the disruption of trade finance, play an independent role. Since there is no valid data available on international trade finance the specific factors other than GDP slowdown remain elusive.

The strong negative effect of the recent financial crisis and the higher volatility of trade for developed countries, generally trading more differentiated goods, can be interpreted as a support of the sector compositional hypothesis (Hypothesis H3). Differentiated goods are more demanding in terms of finance structures making them particularly vulnerable to a financial crisis. This observation brings to mind that a crisis might affect exports within global supply chains or with vertical linkages in a more severe way than other commodities.

The increased elasticity of international trade flows due to vertical linkages and therefore higher share of traded differentiated goods offers an explanation as to why international trade declined in response to a financial crisis. However, a gravity model capturing vertical linkages and the increased vulnerability of

[90]

curcol

comcol

comlang_off

rta

comcur

contig

lnexchange_ rate_o_d

lndistw

lngdp_d

lngdp_o

0.923c (0.000) 0.816c (0.000) –1.046c (0.000) 0.007c (0.000) 0.532c (0.000) 0.303c (0.000) 0.494c (0.000) 0.556c (0.000) 0.359c (0.000) 1.040c (0.000)

IMF DOTS

UN COMTRADE 0.856c (0.000) 0.837c (0.000) –1.307c (0.000) –0.002 (0.406) 0.480c (0.000) 0.101 (0.126) 0.092c (0.000) 0.658c (0.000) 0.267c (0.000) 1.157c (0.000)

0.907c (0.000) 0.815c (0.000) –1.148c (0.000) 0.005c (0.003) 0.495c (0.000) 0.347c (0.000) 0.639c (0.000) 0.519c (0.000) 0.470c (0.000) 1.034c (0.000)

IMF DOTS

UN COMTRADE 0.844c (0.000) 0.841c (0.000) –1.410c (0.000) –0.003 (0.170) 0.435c (0.000) 0.180c (0.002) 0.289c (0.000) 0.611c (0.000) 0.390c (0.000) 1.141c (0.000)

Table 2. The effect of banking crises on export flows

0.905c (0.000) 0.809c (0.000) –1.163c (0.000) 0.005c (0.003) 0.474c (0.000) 0.373c (0.000) 0.667c (0.000) 0.511c (0.000) 0.472c (0.000) 1.057c (0.000)

IMF DOTS

UN COMTRADE 0.844c (0.000) 0.840c (0.000) –1.414c (0.000) –0.003 (0.166) 0.428c (0.000) 0.195c (0.001) 0.305c (0.000) 0.608c (0.000) 0.391c (0.000) 1.151c (0.000) 0.236b (0.028)

0.442c (0.000) 0.575c (0.000)

–0.005c (0.000)

0.852c (0.000) 0.777c (0.000)

IMF DOTS

0.923c (0.000)

–0.019 (0.614) 0.342c (0.000)

–0.002 (0.297)

UN COMTRADE 0.844c (0.000) 0.780c (0.000)

0.236b (0.028)

0.442c (0.000) 0.575c (0.000)

–0.005c (0.000)

0.852c (0.000) 0.777c (0.000)

IMF DOTS

0.923c (0.000)

–0.019 (0.614) 0.342c (0.000)

–0.002 (0.297)

UN COMTRADE 0.844c (0.000) 0.780c (0.000)

0.236b (0.028)

0.441c (0.000) 0.575c (0.000)

–0.005c (0.000)

0.852c (0.000) 0.777c (0.000)

IMF DOTS

0.923c (0.000)

–0.019 (0.613) 0.342c (0.000)

–0.002 (0.297)

UN COMTRADE 0.844c (0.000) 0.780c (0.000)

[91]

271,210 0.766

Yes

No

Yes

No

0.405 (0.453) 0.390c (0.000) 0.159c (0.000) –0.098 (0.111)

317,045 0.703

1.520c (0.000) 0.406c (0.000) 0.224c (0.000) –0.026 (0.728)

No

Yes No

Yes

271,210 0.771

(0.000)

(0.000)

317,045 0.707

1.169c

0.411 (0.467) 0.460c (0.000) 0.168c (0.000) –0.094 (0.129)

1.049c

1.564c (0.000) 0.477c (0.000) 0.227c (0.000) –0.030 (0.686)

No

Yes

(0.000) 0.621c (0.000) 317,045 0.707

0.704c

1.569c (0.000) 0.476c (0.000) 0.220c (0.000) –0.532c (0.000)

No

Yes

(0.000) 0.274c (0.000) 271,210 0.771

1.021c

0.413 (0.464) 0.459c (0.000) 0.165c (0.000) –0.316c (0.000)

Yes

No

317,114 0.836

–0.010 (0.851)

1.509c (0.004)

Yes

No

271,210 0.879

–0.013 (0.776)

1.283c (0.003)

Yes

No

317,114 0.836

–0.010 (0.851)

1.509c (0.004)

Yes

No

271,210 0.879

–0.013 (0.776)

1.283c (0.003)

Yes

No

0.092 (0.275) 317,114 0.836

–0.085 (0.132)

1.509c (0.004)

Yes

No

0.027 (0.722) 271,210 0.879

–0.035 (0.533)

1.283c (0.003)

Source: [IMF DOTS; UN COMTRADE; Reinhart and Rogoff 2011; CEPII Gravity dataset (update)]; author’s estimation.

Robust standard errors in parentheses, clustered by destination-year, with a, b, and c respectively denoting significance at the 1%, 5% and 10% levels. Year dummies are included in all estimations.

Observations R2 Exporter and importer fixed effects Country-pair fixed effects

BKdDevel

developingcountries

banking_ crisis_d

comleg_posttrans

comrelig

cursib

92

Economics and Business Review, Vol. 1(15), No. 2, 2015

Table 3. Banking crises, GDP slowdown and level of economic development IMF DOTS lngdp_o lngdp_d lndistw lnexchange_rate_o_d contig comcur rta comlang_off comcol curcol cursib comrelig comleg_posttrans banking_crisis_d globalgdpslowdownd developingcountries

0.906c (0.000) 0.815c (0.000) –1.150c (0.000) 0.005c (0.003) 0.501c (0.000) 0.342c (0.000) 0.635c (0.000) 0.519c (0.000) 0.469c (0.000) 1.034c (0.000) 1.565c (0.000) 0.478c (0.000) 0.228c (0.000) –0.031 (0.669) –0.121 (0.195) 1.046c (0.000)

BKdDevel Observations R2

315,890 0.707

0.904c (0.000) 0.809c (0.000) –1.164c (0.000) 0.005c (0.004) 0.481c (0.000) 0.368c (0.000) 0.663c (0.000) 0.511c (0.000) 0.471c (0.000) 1.057c (0.000) 1.569c (0.000) 0.477c (0.000) 0.221c (0.000) –0.534c (0.000) –0.122 (0.192) 0.702c (0.000) 0.620c (0.000) 315,890 0.707

UN COMTRADE 0.844c (0.000) 0.841c (0.000) –1.410c (0.000) –0.003 (0.170) 0.435c (0.000) 0.180c (0.002) 0.289c (0.000) 0.611c (0.000) 0.390c (0.000) 1.141c (0.000) 0.411 (0.467) 0.460c (0.000) 0.168c (0.000) –0.094 (0.129) –1.874c (0.000) 1.169c (0.000)

271,210 0.771

0.844c (0.000) 0.840c (0.000) –1.414c (0.000) –0.003 (0.166) 0.428c (0.000) 0.195c (0.001) 0.305c (0.000) 0.608c (0.000) 0.391c (0.000) 1.151c (0.000) 0.413 (0.464) 0.459c (0.000) 0.165c (0.000) –0.316c (0.000) –1.871c (0.000) 1.021c (0.000) 0.274c (0.000) 271,210 0.771

UN COMTRADE differentiated goods 0.994c 0.993c (0.000) (0.000) 0.950c 0.949c (0.000) (0.000) –1.537c –1.543c (0.000) (0.000) –0.013c –0.013c (0.000) (0.000) c 0.485 0.475c (0.000) (0.000) 0.312c 0.336c (0.000) (0.000) 0.259c 0.283c (0.000) (0.000) 0.704c 0.699c (0.000) (0.000) 0.485c 0.486c (0.000) (0.000) c 1.227 1.242c (0.000) (0.000) –0.070 –0.067 (0.920) (0.923) 0.538c 0.537c (0.000) (0.000) 0.165c 0.161c (0.000) (0.000) –0.051 –0.391c (0.436) (0.000) –2.148c –2.144c (0.000) (0.000) 0.766c 0.539c (0.000) (0.000) 0.419c (0.000) 271,210 271,210 0.798 0.798

Standard errors in parentheses, clustered by destination-year, with a, b, and c respectively denoting significance at the 1%, 5% and 10% levels. Year dummies are included in all estimations.

Source: [IMF DOTS; UN COMTRADE; Reinhart and Rogoff 2011; CEPII Gravity dataset (update)]; author’s estimation.

U. Broll, J. Jauer, International trade in differentiated goods, financial crisis

93

international trade remains to be thoroughly developed theoretically and empirically tested.

Conclusions What happened during the crisis that led international trade to nearly collapse? Global trade decreased around 29 percent during the global recession of 2008–2009. The gravity model is one specific valuable economic instrument for studying and measuring the effect of financial crises on international trade.

The gravity approach is helpful in addressing differences in country specific trade relations and the effect of financial crises over time. One of our results in the study is that developed countries seem to be effected more by financial crises than developing and emerging economies. The higher share of differentiated goods in exports traded by developed countries seems to be one reason for this phenomenon. The trade in differentiated goods suffered more during the financial crisis 2008–2009 with a statistical and economic significance of over 17 percent compared with aggregated international trade flows. Differentiated tradable goods are particularly vulnerable to financial crises because of their complex pre-finance structures.

References Admati, A., Hellwig, M., 2013, The Bankers’ New Clothes, Princeton University Press, Princeton and Oxford.

Amiti, M, Weinstein, D.E., 2011, Exports and Financial Shocks, Quarterly Journal of Economics, vol. 126: 1841–1877.

Anderson, J.E., van Wincoop, E., 2003, Gravity with Gravitas: A Solution to the Border Puzzle, American Economic Review, vol. 93: 170–192.

Anderson, J.E., van Wincoop, E., 2004, Trade Costs, Journal of Economic Literature, vol. 42: 691–751.

Baldwin, R., 2009, The Great Trade Collapse: Causes, Consequences and Prospects, VoxEU.org Publication.

Berman, N., de Sousa, J., Martin, P., Mayer, T., 2012, Time to Ship during Financial Crises, Center for Economic Policy Research, 2012–25.

Bricongnge, J.-C., Fontagne, L., Gaulier, G., Taglioni, D., Vicard, V., 2012, Firms and the Global Crisis: French Exports in the Turmoil, Journal of International Economics, vol. 87: 134–146.

Broll, U., Rajiv, M., Pong Wong, K., 2001, International Trade and Hedging in Economies in Transition, Economic Systems, vol. 25: 149–159.

Chaney, T., 2008, Distorted Gravity: The Intensive and Extensive Margins of International Trade, American Economic Review, vol. 98: 1707–1721.

Chor, D., Manova, K., 2012, Of the Cliff and Back: Credit Conditions and International Trade during the Global Financial Crisis, Journal of International Economics, vol. 87: 117–133.

94

Economics and Business Review, Vol. 1(15), No. 2, 2015

Didier, T., Hevia, C., Schmukler, S.L., 2011, How Resilient Were Emerging Economies to the Global Crisis? World Bank Policy Research, Working Paper, no. 5637.

Disdier, A.-C., Head, K., 2008, The Puzzling Persistence of the Distance Effect on Bilateral Trade, Review of Economics and Statistics, vol. 90: 37–48.

Eaton, J., Kortum, S., Neiman, B., Romalis, J., 2013, Trade and the Global Recession, Booth School of Business, University of Chicago (and National Bureau of Economic Research, Working Paper, no. 16666; 2011).

Engel, C., Wang, J., 2011, International Trade in Durable Goods: Understanding Volatility, Cyclicality, and Elasticities, Journal of International Economics, vol. 83: 37–52.

Head K., Mayer, T., 2013, Gravity Equations. Workhorse, Toolkit, and Cookbook, Centre for Economic Policy Research, Discussion Paper, no. 9322.

Head K., Mayer, T., Ries, J., 2010, The Erosion of Colonial Trade Linkages after Independence, Journal of International Economics, vol. 81: 1–14.

IMF, 2009a, Global Financial Stability Report, Navigating the Financial Challenges Ahead, October.

IMF, 2009b, Survey of Private Sector Trade Credit Developments, February.

Kellman, M.. Shachmurove, Y., 2011, Diversification and Specialization Paradox in Developing Country Trade, Review of Development Economics, vol. 15: 212–222.

Kellman, M.. Shachmurove, Y., 2012, Evolving Sophistication of Trade Pattersn in a Transition Economy – Machinery Exports of Poland 1980–2009, Poznań University of Economics Review, vol. 12: 9–41.

Kellman, M.. Shachmurove, Y., 2013, Exports and Development Montenegro 2006–2012, Montenegrin Journal of Economics, vol. 9: 29–43.

Martin, P., Mayer, T., Thoenig, M., 2008, Make Trade Not War?, Review of Economic Studies, vol. 75: 865–900.

OECD, 2010, Trade and Economic Effects of Responses to the Economic Crisis, Trade Policy Studies, Paris.

Rauch, J.E., 1999, Networks versus Markets in International Trade, Journal of International Economics, vol. 48: 7–35.

Reinhart, C.M., Rogoff, K.S., 2011, From Financial Crash to Debt Crisis, American Economic Review, vol. 101: 1676–1706.

Internet sources for data CEPII: Gravity Data, http://www.cepii.fr/CEPII/en/bdd_modele/presentation.asp?id=8 [access: 19.02.2013].

IMF DOTS: Direction of Trade Statistics, http://elibrary-data.imf.org/FindDataReports. aspx?d=33061&e=170921 [access: 19.02.2013].

Rauch, J.E., Research on Incomplete Information and Networks in International Trade – Classification of SITC Rev. 2, http://weber.ucsd.edu/~jrauch/research_international_trade.html [access: 19.02.2013].

UN COMTRADE: United Nations Commodity Trade Statistics Database, http:// comtrade.un.org/db/default.aspx [access: 19.02.2013].

UN STATS: United Nations Statistics Division definition of Composition of macro geographical (continental) regions, geographical sub-regions, and selected economic and other groupings, http://unstats.un.org/unsd/methods/m49/m49regin.htm [access: 19.02.2013].

Economics and Business Review, Vol. 1 (15), No. 2, 2015: 95–109

Tax revenues and aging in ex-communist EU countries1 Mihai Mutascu2, Maciej Cieślukowski3

Abstract: The paper explores the relationship between tax revenues and aging in the case of 10 ex-communist EU countries, for the period 1995–2012, by using a panel model approach. The main finding shows that the ageing has a significant and positive impact on tax revenues. In these ex-communist countries, the persons are more interested in the redistribution. On the other hand, there are high revenue amounts for older persons and a low degree of meritocracy. Both aspects put pressure on public expenditures and required additional financial needs for the government.

Keywords: tax revenues, determinants, aging, ex-communist countries, aging effects.

JEL codes: H20, Q56, C23, C26.

Introduction Population ageing is one of the most important topics for the social science in the last decades with complex implications in the area of economics also. The most affected countries by this tendency are European countries, especially the ex-communist states. Ageing, as a demographic element, has significant implications for the economy with extended effects in the field of taxation. In this way, illustrating the classical determinants of tax revenues, Lago-Peñas and Lago-Peñas [2008] include the age of the population in the group of socio-demographic characteristics, jointly with gender, marital status, education, employment status, religion and social class.

Given this relevance of the ageing phenomenon to economic tax literature, the main aim of the paper is to investigate whether the ageing is associated with tax revenues in the case of ten ex-communist EU countries in the peri1

Article received 19 June 2014, accepted 19 February 2015.

West University of Timisoara, Faculty of Economics and Business Administration, Romania.

3 Poznań University of Economics, Public Finance Department, al. Niepodległości 10, 61-875 Poznań, Poland, corresponding author: m.cieslukowski@ue.poznan.pl.

2

96

Economics and Business Review, Vol. 1(15), No. 2, 2015

od 1995–2012. This group of states arouse additional interest as they passed through two different political regime periods: one autocratic and the other democratic. In this way, Farkas [2011] emphasizes that such countries have a set of common characteristics: “the lack of capital, weak civil society and the impact of the European Union and other international organisations influencing the new member states” [p. 15].

Another common feature of those countries is a similar course of the share of registered population aged 65 and above in the total population, from 1960 to 2012, as in Figure.

19 17 15 13 11 9

5

1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012

7

Bulgaria Latvia

Czech Republic Lithuania

Slovak Republic

Slovenia

Estonia Poland

Hungary Romania

The population aged 65 and above as a percentage of total population in the case of ex-communist EU countries for the period 1960–2012 Source: [World Bank 2014]

The figure above illustrates two periods with an ascending trend of population aged 65 and above as a percentage of the total population in the case of ex-communist EU countries: firstly until 1979–1980, and secondly since 1986.

The second interval of time registers a significant increase in the population aged 65 and above in total population especially after 1990, covering all democratic period of the countries considered.

The main novelty of the paper in the field literature is the fact that the results are the first ones obtained for ex-communist EU countries regarding the “tax revenues-ageing” connection. On the other hand, the study is one of the first analysis that use, as a tool of investigation, a panel model approach, testing also for non-linearity of the considered nexus between tax revenues and ageing.

M. Mutascu, M. Cieślukowski, Tax revenues and aging in ex-communist EU countries

97

The rest of the paper is structured as follows: Section 1 presents a general literature framework; Section 2 illustrates the data and methodology; whilst Section 3 describes the results. Finally, Section 4 concludes the study.

1. Literature review According to Peñas and Lago-Peñas [2008], ageing is one of the most important socio-demographic classic tax determinants. The other groups of tax factors include political and social attitudes, the fiscal parameters and the contextual determinants. The political and social attitudes are related to the trust in courts, legal system, in politicians, the level of democracy , national pride, social capital, perception of corruption and voting behaviour. The fiscal parameters include tax rates, fine rate, audit probability, risk aversion and personal income, whilst the contextual determinants are the extent of direct democracy, language fragmentation and the existence of regional divisions.

Literature regarding the “tax-ageing” nexus is not so prolific. Even so, it reveals contradictory results concerning the relationship between tax revenues and ageing. One strand of research shows that the ageing has a positive impact on tax revenues, whilst the second claims a negative correlation between them.

Adding significant findings to the first group of contributions, Mincer [1970] shows that the older labour force increases the level of tax revenues as the income level of older taxpayers is greater than the income level of younger persons. In other words, the incomes of older workers are greater than the incomes of younger people as result of professional experience and maturity. As income is one of the most important parts of the tax-base, the older people’s income tax-base and tax payments are superior to the tax-base and taxes of young workers. He also explains that the age-earning characteristics register an upward tendency over a large period of the life cycle. On the other hand, when the percentage of the old population goes down, the percentage of the elderly (65 years and above) to working people (15–64) in the total population increases, tax revenue shows a descending trend. Visco [2001] finds similar results, but with different arguments . He states that the extension of an ageing population generates more pressure on public expenditure as a result of additional social needs provided by the welfare state related to the ageing population. In this case supplementary tax revenues should finance the redistribution amongst the generations and all additional social needs. The increase of tax revenue represents a direct effect of a public expenditure rise under the pressure of the ageing population (i.e., when the percentage of elderly people in the total population increases, the corresponding public expenditure increases and generates a need for additional tax revenue).

The second group of research reveals the contrary: the increase of ageing in the population results in a decrease in tax revenue. On the one hand, as

98

Economics and Business Review, Vol. 1(15), No. 2, 2015

Goudswaard and Van de Kar [1994] note, when the ratio of elderly (65+) to working people (15–64) increases the tax revenues decrease because the number of taxpayers is reduced. More precisely the tax base contracts as results of the decrease in the number of taxpayers. Thus, if the segment of an ageing population is greater than the segment of young population, the older people will pay fewer taxes by comparison with young taxpayers, because their tax-base reduces as result of retirement.

According to Razin, Sadka, and Swagel [2002], the same effect is obtained but through the voting process. Firstly, the taxes tend to be higher because the ageing voters vote for additional financing for social needs. Furthermore, the tax revenues will collapse due to reluctant behaviour regarding the tax payments of the young taxpayers. These arguments are also related to Kirchler’s [2007] explanations, which states that different perceptions regarding the tax payments exist between old and young people. If the old persons are more responsible in respect to taxation (i.e. they have a high level of tax compliance), the younger generations are less responsible (i.e. they reveal a low level of tax compliance).

Thus, the older persons will pay “more taxes” than the young taxpayers.

Finally, Brett [2012] analyzes the effects of population ageing on optimal redistributive taxes in an overlapping generation model, having as a starting point the results of Ordover and Phelps [1979]. The optimal tax concept is related to a nonlinear approach which has as its starting points the production and utility functions. The author shows that a decrease in the rate of population growth leads to an ageing population, which puts pressure on the relative price of consumption per person in retirement. Given this pressure the price of consumption rises and tends to reduce the level of consumption. Simultaneously the scholar also shows that, when the population becomes older, wage rate increases and the implicit marginal income tax rates remain unchanged. As a consequence, the optimal marginal income tax rates are independent of the rate of population growth.

Nearly al these approaches follow either theoretical modelling or econometric tools by especially using linear estimation techniques.

2. Methodology and data Taking into account the main literature the investigation of the relationship between tax revenues and ageing in the case of the ten ex-communist EU countries for the period 1995–2012 (data available period) is based on the hypothesis below: H: Ageing has a significant impact on government tax revenues, being an important determinant of macroeconomic policies, especially in taxation policy.

Our balanced panel sample includes, for the period 1995–2012, the following ex-communist EU countries: Bulgaria, the Czech Republic, Estonia, Hungary,

M. Mutascu, M. Cieślukowski, Tax revenues and aging in ex-communist EU countries

99

Latvia, Lithuania, Poland, Romania, the Slovak Republic and Slovenia (Croatia has not been targeted as this country integrated in EU later).

The dependent variable is represented by collected tax revenues per capita (τ), which denotes the level of tax revenues collected by central government in US dollars. This variable includes all type of taxes and better captures the source of the whole financial needs of ageing, as these are divided per capita and covered by both personal and company taxpayers.

The independent interest variable is the ageing (λ), which is measured as the population aged 65 and above as a percentage of the total population.

In order to isolate the effect of the interest variable, several control variables have also been considered, following the methodology of Mutascu [2014].

The first group includes variables inspired by classic tax literature, such as: the gross domestic product (GDP), the size of the industrial sector, the size of the agricultural sector and the size of the service sector. GDP controls the size of the economy, being expressed in millions of US dollars. It is expected to have a positive effect on tax revenues per capita [Katircioglu 2010]. The size of the industrial sector, the size of the agricultural sector and the size of the service sector are measured as a percentage of GDP and capture the structure of the economy, with significant impact on the dependent variable [Agbeyegbe, Stotsky, and WoldeMariam 2006]. We expect that the industrial and service sectors have a positive impact on tax revenues, whilst the agricultural sector has a negative one.

The second group of control variables is related to the appropriate macroeconomic policy area and is represented by the government final consumption expenditures and the net inflow of foreign direct investment (FDI). The first variable quantifies the government final consumption expenditures as a percentage of GDP, being expected to have a positive influence on tax inputs [Taha and Loganathan 2008]. FDI measures the difference between inward foreign direct investment and outward foreign direct investment as a percentage of GDP. Many studies point-out its positive impact on tax revenues [e.g., Wildasin 2003].

The last set of control determinants concerns robustness. We consider here the following determinants: the government effectiveness, level of corruption and political stability. The government effectiveness denotes the government quality and “captures perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation and the credibility of the government’s commitment to such policies” [World Bank 2014]. The value –2.5 shows a weak governance performance, whilst 2.5 – a strong one. The impact of government effectiveness on tax revenues is expected to be positive, as Lisi [2011] notes. The level of corruption is expressed in percentile rank and reveals the intensity of corruption, from 0 (high level) to 100 (low level). This dimension does not offer any information about tax compliance. According

100

Economics and Business Review, Vol. 1(15), No. 2, 2015

to Imam and Jacobs [2007], a negative impact of corruption on tax revenues is expected. The political stability denotes the number of years since the change of the most recent regime or the end of the transition period defined by the lack of stable political institutions. This dimension has an ambiguous impact on tax revenues [Estrada, Mutascu, and Tiwari 2013].

The tax revenues per capita and GDP are treated as elasticities, with natural log form, excepting the regressors already expressed in percentage and the political stability, which does not have strict positive values. The sources of data and the descriptive statistics of variables are illustrated in Tables 1 and 2 in the Appendix.

The main function is: ln(Ď&#x201E;) = f (Îť),

(1)

where: Ď&#x201E; â&#x20AC;&#x201C; the amount of tax revenues per capita in U.S. dollars, Îť â&#x20AC;&#x201C; the population aged 65 and above as percentage of the total population.

The extended panel-model is as follows: n

ln(Ď&#x201E; )it

ι ι1 Νit Œ βk X k , it Οi νt ξit,

(2)

k 1

where: Îą â&#x20AC;&#x201C; the intercept, Îą1 â&#x20AC;&#x201C; the slope of interest variable, βk â&#x20AC;&#x201C; coefficient of control independent variable k by n type, X â&#x20AC;&#x201C; control variables, Îźi â&#x20AC;&#x201C; stands for country fixed effects, Ď&#x2026;t â&#x20AC;&#x201C; time-specific effect that checks for unaccounted common time-varying factors, i â&#x20AC;&#x201C; country, t â&#x20AC;&#x201C; time, Îľit â&#x20AC;&#x201C; the error term.

The main function is also tested for polynomial non-linearity by using the wald-test, partial F-test and taking into account the significance of the resultant coefficients. Furthermore, the first estimations are performed in several scenarios, with different factorial determinants, by using the OLS panel regressions. At the same time, the multi-coliniarity implications of independent variables are also investigated.

Secondly, we deal with the homogeneity of the panel. As the panel-data model may have heterogeneity in the data, we test this property by analyzing both cases of fixed and random panel-model effects, through the F-test and

M. Mutascu, M. Cieślukowski, Tax revenues and aging in ex-communist EU countries

101

Hausman-test. The random effects estimation requires the number of crosssections to be greater than the number of coefficients. As in our case the number of cross-sections is equal to the number of coefficients The random model is not employed.

The last potential issue in estimations is the endogeneity, especially due to the reverse causality of the interest variable. We deal with this issue by following the IV model (Instrumental Variables regression, also known as Two-Stage Least-Squares estimator). Unfortunately, this estimation is not consistent under heteroskedasticity disturbance. In this case, as Baum, Schaffer, and Stillman [2003] note, “if heteroskedasticity is present, the GMM estimator is more efficient than the simple IV estimator, whereas if heteroskedasticity is not present, the GMM estimator is no worse asymptotically than the IV estimator” [p. 11].

The Pagan-Hall general test statistic is calculated to illustrate evidence of the heteroskedasticity in IV regression, whilst the Wu-Hausman F-test shows if the regressors are or not exogenous in IV regression.

3. Results The first outputs in Table 3 in the Appendix, reveal that in the case of the naive model 1, the interest variable (λ) is significant and positively correlated with ln of the dependent variable (τ). Further, the wald-test for non-linearity of the main function, entering the square of the interest variable, shows that we cannot reject the result that the coefficient of the square interest regressor is zero.

Additionally, the same output is obtained by performing the partial F-test for the coefficients of the block “interest variable and the square of interest variable”. Based on the value 0.04 of partial F-test (p-value = 0.842), in Table 3 and in the Appendix, we cannot reject the null hypothesis (H0 = the coefficients of both interest variables are 0).

Thus, for the whole analysis, we assume that the main function is linear.

This conclusion is reinforced by the insignificance of all coefficients of the non-linear model 2.

Entering the control variables progressively, the main output shows that, in the case of OLS models 3 and 4, the coefficients of the interest variable are still significant, having the same positive impact on ln of tax revenues per capita as in the naïve scenarios. Between the control variables, ln of GDP, size of the industrial sector, size of the agricultural sector, control of corruption and political stability are significant. Only the size of the agricultural sector is negative , whilst the rest of the determinants are positively correlated with the dependent variable. The rest of the control determinants remain insignificant.

In the case of the complete OLS model 4, the variance inflation factor test results (VIF-test) in Table 4 in the Appendix, show there is not any significant

102

Economics and Business Review, Vol. 1(15), No. 2, 2015

multicoliniarity issue between the regressors, as all test levels are less that the critical level 4 [O’Brien, 2007], at limit for the size of service sector and government effectiveness.

Furthermore, in the fixed-effects model 5 the homogeneity of the considered panel is tested. The F-test illustrates that this fixed-model model 5 is more appropriate for our analysis than the OLS model 4. In this case, the main finding shows that the interest variable is still significant and positive as in the previous OLS estimations.

The last step of our panel investigation is the endogeneity issue. This can be caused either by omitted variables and measurement error or by the reverse causality, especially of the interest variable. In order to deal with this, a TSLS model 6 is performed, using as instruments the lags of the interest-instrumented variable. The Wu-Hausman F-test for endogeneity reveals that the null of “the regressor is exogenous” and cannot be rejected. The last test is for the presence of heteroskedasticity. The output of the Pagan-Hall general test statistic, in model 6, clearly shows that the null of “disturbance is homoskedastic” and cannot be rejected. Thus, the disturbance of model 6 is homoskedastic and the TSLS model 6 is more efficient than any GMM estimations [Baum, Schaffer, and Stillman 2003]. As a consequence, the TSLS model 6 is selected for the final analysis of the relationship between ln of tax revenues per capita and the population aged 65 and above.

The main output in model 6 indicates that the interest variable λ is significant and has a positive impact on ln of tax revenues per capita. In other words, when the level of the older population increases a as percentage of the total population, the level of collected tax revenues increases also. The same positive sign has several significant control variables, such as: ln of GDP, size of the industrial sector, control of corruption and political stability. The size of the agricultural sector remains significant, but negatively correlated with dependent variable τ. The rest of control variables are non-conclusive.

Our findings confirm the results of Mincer [1970] and Visco [2001], but are in contradiction to the outputs obtained by Goudswaard and Van de Kar [1994], Razin, Sadka, and Swagel [2002] and Brett [2012]. The explanation is that the considered sample includes an ex-communist area, more interested in redistribution, with strong economic disequilibrium, high revenue amounts for older persons and a low degree of meritocracy. Moreover, these countries have a common and special “model” of governance, as Farkas [2011] notes. The author highlights three main common characteristics: (i) a low level of capital inflow, (ii) a weak civil society, and (iii) the strong impact of the European Union institutions on economic and social environment. All these factors, plus the international migration to other countries , generate over time a strong disparity between the number of older and the number of younger people, with a negative impact on government inputs (i.e. the number of taxpayers becomes lower than the number of retired people).

M. Mutascu, M. Cieślukowski, Tax revenues and aging in ex-communist EU countries

103

Conclusions Ageing is a new and important determinant of tax revenues, even if the literature in the field is not so prolific. In the case of the ten ex-communist EU countries, for the period 1995–2012, the main findings show that ageing has a significant and positive influence on tax revenues.

In ex-communist countries the people are more interested in the redistribution of revenues, the memories of the “left political period” are still fresh in the minds, especially in the older persons. On the other hand, many of these countries are still contributary to gerontocracy, the older persons having a high level of revenue. Both issues put pressure on public expenditure and the additional financial needs imposed on the government.

Regarding the policy implications in these countries the governments should follow a mixed policy strategy. In the tax policy area it is a requirement that the maximizing of tax revenue should be achieved only through a coherent demographic policy by improving the birth rate.

This investigation focuses exclusively on the impact of ageing on the area of taxation but can be easily extended to the government spending component of fiscal policy. This topic would be a good topic for further research.

104

Economics and Business Review, Vol. 1(15), No. 2, 2015

Appendix

Table 1. The sources of data Variable

Source

Tax revenues per capita (US dollars)

Calculation based on the level of tax revenues as percentage of GDP, and GDP per capita, offered by World Bank online database, 2014

Population of age 65 and above (percentage of total population)

World Bank online database, 2014

GDP (US dollars)

World Bank online database, 2014

Size of the industrial sector (percentage of GDP)

World Bank online database, 2014

Size of the agricultural sector (percentage of GDP)

World Bank online database, 2014

Size of the service sector (percentage of GDP)

World Bank online database, 2014

Government final consumption expenditure (percentage of GDP)

World Bank online database, 2014

FDI (percentage of GDP)

World Bank online database, 2014

Government effectiveness (percentile rank) World Bank online database, 2014 Control of corruption (percentile rank)

World Bank online database, 2014

Political stability (years)

Polityâ&#x201E;˘ IV Project Political Regime Characteristics and Transitions, 1800â&#x20AC;&#x201C;2012 Dataset

M. Mutascu, M. CieĹ&#x203A;lukowski, Tax revenues and aging in ex-communist EU countries

105

Table 2. Descriptive statistics of variables Mean

Median

Maximum

Minimum

1832.84

1447.437

6294.514

242.3813

1247.799

180

Population of age 65 and above (percentage of total population)

0.148633

0.148479

0.194275

0.111419

0.019492

180

GDP (US dollars)

7.75E+10 4.21E+10 5.29E+11 3.78E+09 9.84E+10

180

Variable Tax revenues per capita (US dollars)

Std. Dev. Obs.

Size of theindustrial sector 0.329469 (percentage of GDP)

0.328603

0.499482

0.215758

0.048456

180

Size of theagricultural sec0.060759 tor (percentage of GDP)

0.046032

0.249472

0.021786

0.040252

180

Size of the service sector (percentage of GDP)

0.619971

0.628083

0.765597

0.319488

0.072221

180

Government final consumption expenditure (percentage of GDP)

0.226797

0.196361

0.987734

0.111273

0.127115

180

FDI (percentage of GDP)

0.054652

0.041118

0.518958

â&#x20AC;&#x201C;0.16418 0.066903

180

Government effectiveness (percentile rank)

0.691311

0.732174

0.897874

0.282927

0.13443

180

Control of corruption (percentile rank)

0.638965

0.648752

0.89757

0.239244

0.121784

180

Political stability (years)

11.13889

11

0

5.603036

180

22

[106]

9.097*** (2.821)

Population of age 65 and above (λ)

2.893** (0.831) –4.486*** (0.943) 0.746 (0.605)

3.668*** (1.168) –11.684*** (1.055) 0.751 (0.838)

Size of the industrial sector

Size of the agricultural sector

0.037 (0.045) 0.185** (0.081) –0.071 (0.094)

0.066 (0.184) 0.093 (0.329) 0.451 (0.321)

FDI

Government effectiveness

–0.044 (0.142)

–0.853*** (0.241)

0.286 (0.211)

1.014*** (0.025)

2.110*** (0.503)

–17.739** (0.599)

(5)

Government final consumption expenditure

Size of the service sector

0.061** (0.026)

3.623** (1.467)

1.601* (0.906)

(4)

0.193*** (0.030)

56.223 (129.59)

12.255*** (1.648)

–0.263*** (0.038)

(3)

Model

Ln GDP

Square of population of age 65 and above (λ2)

7.143** (2.876)

5.909*** (0.422)

Constant –7.705 (38.835)

(2)

(1)

Independent variables

Dependent variable: ln tax revenues per capita – ln(τ)

Table 3. Empirical results of panel regressions

0.487 (0.336)

0.072 (0.337)

0.106 (0.192)

0.733 (0.621)

–4.544*** (0.985)

2.935*** (0.852)

0.066** (0.028)

4.733*** (1.769)

1.292 (0.988)

(6)

[107]

0.859

TSLS

(a) (…) denotes the standard error; (b) PLS represents panel least squares; (c) FE means fixed effects; (d) TSLS denotes Two-Stage Least-Squares regression; (e) ***, **, and * denote significance at 1, 5 and 10 % level of significance, respectively.

0.334 Prob.= 0.564

350.25 Prob.= 0.000

0.993

FE

Wu–Hausman F–test

0.865

PLS

9.646 Prob.= 0.472

0.720

Model summary

PLS

Pagan–Hall general test statistic

F–test for fixed effects

0.04 Prob.= 0.842

Partial F–test

0.056 0.433 Prob.= 0.664

0.055

R–squared

PLS naive

T–statistic of wald–test for square of interest variable “λ”

PLS naive

0.050*** (0.006)

–0.009*** (0.599)

0.051*** (0.005)

Political stability

Type of estimation

2.432*** (0.343)

–0.023 (0.094)

2.451*** (0.325)

Control of corruption

108

Economics and Business Review, Vol. 1(15), No. 2, 2015

Table 4. Variance inflation factor test results Variable

VIF

1/VIF

Population of age 65 and above

1.8

0.55412

Ln GDP

1.99

0.50282

Size of the industrial sector

3.58

0.27932

Size of the agricultural sector

3.18

0.31428

Size of the service sector

4.22

0.23687

Government final consumption expenditure

1.21

0.82457

FDI

1.07

0.93197

Government effectiveness

4.13

0.24239

Control of corruption

3.46

0.28918

Political stability

2.41

0.41492

Mean VIF

2.71

References Agbeyegbe, T., Stotsky, J., WoldeMariam, A., 2006, Trade Liberalization, Exchange Rate Changes, and Tax Revenue in Sub-Saharan Africa, Journal of Asian Economics, 17: 261–284.

Baum, C., Schaffer M., Stillman S., 2003, Instrumental Variables and GMM: Estimation and Testing, The Stata Journal, 3: 1–31.

Brett, C., 2012, The Effects of Population Aging on Optimal Redistributive Taxes in an Overlapping Generations Model, International Journal of Public Finance, 19: 777–799.

Estrada, F., Mutascu, M., Tiwari, A., 2013, Political Stability and Taxation, Análisis Político, 77: 133–152.

Farkas, B., 2011, The Central and Eastern European Model of Capitalism, Post-Communist Economies, 23: 15–34.

Goudswaard, K., Van de Kar, H., 1994, The Impact of Demographic Change on Tax Revenue, Atlantic Economic Journal, 22: 52–60.

Imam, P., Jacobs, D., 2007, Effect of Corruption on Tax Revenues in the Middle East, IMF Working Paper, no. 270.

Katircioglu, S., 2010, Is There a Long-Run Relationship between Taxation and Growth: The Case of Turkey, Romanian Journal of Economic Forecasting, 13: 99–106.

Kirchler, E., 2007, The Economic Psychology of Tax Behavior, Cambridge University Press.

Lago-Peñas, I., Lago-Peñas, S., 2008, The Determinants of Tax Morale in Comparative Perspective: Evidence from a Multilevel Analysis, Instituto de Estudios Fiscales Working Paper, no. 2.

Lisi, G., 2011, Job Search Theory and the Slippery Slope Framework: An Attempt to Integration, Department of Economics, University of Cassino, Working Paper, no. 2.

M. Mutascu, M. Cieślukowski, Tax revenues and aging in ex-communist EU countries

109

Mincer, J., 1970, The Distribution of Labor Incomes: A Survey With Special Reference to the Human Capital Approach, Journal of Economic Literature, 8: 1–26.

Mutascu, M., 2014, Influence of Clime Conditions on Tax Revenues, Contemporary Economics, 8: 315–328.

O’Brien, R., 2007, A Caution Regarding Rules of Thumb for Variance Inflation Factors, Quality & Quantity, 41: 673–690.

Ordover, J., Phelps, E., 1979, The Concept of Optimal Taxation in the Overlappinggenerations Model of Capital and Wealth, Journal of Public Economics, 12: 1–26.

Razin, A., Sadka, E., Swagel, P., 2002, The Aging Population and the Size of the Welfare State, Journal of Political Economy, 110: 900–918.

Taha, R., Loganathan, N., 2008, Causality between Tax Revenue and Government Spending in Malaysia, The International Journal of Business and Finance Research, 2: 63–73.

Visco, I., 2001, The Fiscal Implications of Ageing Populations in OECD Countries, Organization for Economic Cooperation and Development, Presented at the Oxford Centre on Population Ageing Pensions Symposium, June.

Wildasin, D., 2003, Fiscal Competition in Space and Time, Journal of Public Economics, 87: 2571–2588.

World Bank, 2014, online database. *** Polity™ IV Project Political Regime Characteristics and Transitions, 1800–2012 Dataset. *** World Bank online dataset 2014.

Economics and Business Review, Vol. 1 (15), No. 2, 2015: 110–129

The analytics of the New Keynesian 3-equation Model1 Jean-Christophe Poutineau2, Karolina Sobczak3, Gauthier Vermandel4

Abstract: This paper aims at providing a self contained presentation of the ideas and solution procedure of New Keynesian Macroeconomics models. Using the benchmark “3 equation model”, we introduce the reader to an intuitive, static version of the model before incorporating more technical aspects associated with the dynamic nature of the model. We then discuss the relative contribution of supply, demand and policy shocks to the fluctuations of activity, inflation and interest rate, depending on the key underlying parameters of the economy.

Keywords: dynamic IS curve, impulse response analysis, New Keynesian Macroeconomics, New Keynesian Phillips Curve, output gap, Taylor rule.

JEL codes: C63, E12, E32, E52.

Introduction Keynesian ideas returned to the forefront of academic research in the mid 90’s in new clothes to address questions related to unemployment, economic fluctuations and inflation. This followed a twenty year period that witnessed the domination of new classical ideas on both monetary and real macroeconomics questions. Before contributing to the building of what is now considered as the workhouse of modern macroeconomics [Carlin and Soskice 2014], the New Keynesian School proposed in the 80’s a series of models aimed at providing microeconomic foundations to price and/or wage rigidity5 and at showing 1

Article received 15 April 2014, accepted 19 February 2015.

CREM-CNRS, University of Rennes 1, Faculty of Economics, 7 place Hoche, 35065 Rennes cedex, France, corresponding author: jean-christophe.poutineau@univ-rennes1.fr.

3 Poznań University of Economics, Department of Mathematical Economy, Poznań, Poland.

4 CREM-CNRS, University of Rennes 1, Faculty of Economics, Rennes, France.

5 On New Keynesianism, its history, development and significance for modern economics, see for example Bludnik [2009] or Romer [1993].

2

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

111

that this key feature of the real world can be explained in a setting with optimizing agents with market power. An important breakthrough was about 15 years ago, with the papers of Goodfriend and King [1997] and Clarida, Gali, and Gertler [1999]. These contributions introduced a framework mixing Real Business Cycle features with nominal rigidities. This setting now forms the basic analytical structure of contemporaneous macroeconomic models as exemplified by Woodford [2003] or Gali [2008].

Besides new ideas and a new modelling strategy this New Keynesian Synthesis (NKS) has adopted new solution procedures that may appear cumbersome to non-specialists. Because of their recursive structure NKS models do not admit a closed form solution but should be solved by borrowing procedures developed for the analysis of stochastic discrete time dynamics systems.6 The aim of this paper is to provide a compact and self contained presentation of the structure and of the standard solution procedure of the basic NKS framework known as the “three equation model”. We particularly separate the main ideas conveyed by this model, using a static version of the reference framework, from the technical aspects of the solution procedure. In the presentation we emphasise the qualitative similarities between the simple graphical analysis of the static model and the Impulse Response Functions (IRFs) of the model following the occurrence of exogenous shocks. We then illustrate the key features of this model regarding the analysis of business cycles characteristics.

The paper is organized as follows: In the first section we introduce the general structure of a benchmark NKS model that combines (the log linear versions of) a Philips curve, an Euler equation and a monetary policy (Taylor) rule.7 In the second section we set a simple static version of the model to obtain closed form solutions for the key macroeconomic variables and to provide the reader with a graphical analysis of the consequences of demand and supply shocks.

In the third section we introduce the Blanchard-Kahn solution procedure to get IRFs and dynamic reactions of the model around a stable steady state following exogenous supply demand and policy shocks. This third section is also devoted to a discussion of business cycles characteristics of the model. Section four concludes.

1. The 3 equation new Keynesian model The New Keynesian Synthesis (NKS) mixes the methodology of Real Business Cycles (RBC) with nominal and real rigidities to characterise short run macroeconomic developments. More particularly the NKS seeks to explain the macroeconomic short run evolution of an economy subject to real and monetary 6 7

For an up to date exhaustive introduction to this literature see Miao [2014].

In the appendix we provide the micro foundations of the framework used in this paper.

112

Economics and Business Review, Vol. 1(15), No. 2, 2015

shocks and to replicate business cycle statistics. The core representation of this synthesis has given rise to what is called the â&#x20AC;&#x153;3-equation modelâ&#x20AC;? as the basic NKS setting reduces to a system of three equations corresponding to an ASAD model. First, the AS curve is represented by the New Keynesian Phillips curve that relates inflation to the output gap. Second, the AD component of the model combines a dynamic IS curve (that relates the evolution of the output gap to the interest rate) and a MP (Monetary Policy) schedule (that describes how the nominal interest rate is set by the central bank following fluctuations in the output gap and in the inflation rate. This model is based on agentsâ&#x20AC;&#x2DC; micro founded decision rules where consumers maximize their welfare subject to an intertemporal budget constraint and where firms maximize their profit, subject to nominal rigidities, characterising the imperfect adjustment of prices on the goods market. For convenience the micro foundations of this model and the derivation of the log-linear system are presented in appendix. These equations in turn determine three main variables of interest in a closed economy, namely the output gap ( yË&#x2020;t) which is the gap between the effective output and potential output, the inflation rate (Ď&#x20AC;Ë&#x2020;t) and the nominal interest rate (rË&#x2020;t).

Formally, the model is defined as follows: The New Keynesian Philipsâ&#x20AC;&#x2122; Curve (PC) links current inflation (Ď&#x20AC;Ë&#x2020;t) to expected future inflation (Et ^Ď&#x20AC;Ë&#x2020;t 1`), to the current output gap ( yË&#x2020;t) and to an exogenous supply shock that takes the form of a cost push shock (ÎľtS):8 Ď&#x20AC;Ë&#x2020;t

βEt ^Ď&#x20AC;Ë&#x2020;t 1` ÎşyË&#x2020; t ÎľtS.

(1)

As shown in the appendix this relationship comes from the aggregation of the supply decision of firms that have market power and can re-optimize their selling price with discontinuities (i.e. nominal rigidities â&#x20AC;&#x201C; they cannot modify their selling price at any point in time). Thus they set the selling price of their product depending on three main criteria. (i) The first criterion is anticipated inflation: as firms cannot re-optimize their price, they take into account future inflation to set their price today. (ii) The second term is the output gap: when firms set their price they take into account the difference between supply and demand so that inflation reflects stresses on the goods market: firms increase their prices during periods of expansion ( yË&#x2020;t > 0) whilst they decrease it during recessions ( yË&#x2020;t < 0). (iii) Finally, this relation incorporates a cost push term ÎľtS (such that ÎľtS > 0 may indicate an increase in raw materials or energy price in the economy). In a standard way we assume that ÎľtS is an AR(1) process:9 ÎľtS Ď S ÎľtS 1 ΡtS with ΡtS ~ N (0, Ď&#x192; S2 ) and iid. The New Keynesian Phillips Curve is 8

In the paper all parameters are positive.

This assumption is commonly adopted in the literature to characterize exogenous shocks [see for example, Gali 2008 for a discussion].

9

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

113

derived from the Calvo model [1983] which combines staggered price-setting by imperfectly competitive firms. As presented in the appendix, the Calvo approach assumes that in each period, only a fraction θ of firms, randomly chosen, can reset their selling prices10. Using this assumption, Clarida, Gali, and Gertler [1999] show that the Phillips curve then takes a particularly simple form in which inflation depends on the current gap between actual and equilibrium output as in the standard Phillips curve but on expected future inflation rather than on past inflation.

The dynamic IS curve is a log linearization of the Euler bond equation that describes the intertemporal allocation of consumption of agents in the economy:

yË&#x2020; t

Et ^ yË&#x2020; t 1`

1 rË&#x2020; Et ^Ď&#x20AC;Ë&#x2020;t 1` ÎľtD. Ď&#x192; t

(2)

This relation plays the same role as the IS curve in the IS-LM model. As shown in the appendix it comes from the intertemporal optimization of the welfare index of a representative consumer subject to its budget constraint.

Once aggregated over consumers and log-linearized around the steady state this relation can be expressed in terms of the output gap ( yË&#x2020;t). The dynamic IS curve links the current output gap to the difference between the real interest rate (rË&#x2020;t Et ^Ď&#x20AC;Ë&#x2020;t 1`), to the expected future output gap (Et ^ yË&#x2020; t 1`) and to an exogenous preference shock ÎľtD (that represents a demand shock henceforth). The demand shock is described in a standard way by AR(1) process of the form: ÎľtD Ď D ÎľtD 1 ΡtD with ΡtD ~ N (0, Ď&#x192; D2 ) and is iid.

The Monetary Policy schedule (MP) is based on the Taylor rule. It links the nominal interest rate (that is controlled by monetary authorities) to the inflation rate and to the output gap:11 rË&#x2020;t

Ď&#x2022;Ď&#x20AC; Ď&#x20AC;Ë&#x2020;t Ď&#x2022; y yË&#x2020; t ÎľtR.

(3)

In this equation variable ÎľtR denotes a monetary policy shock that follows an AR(1) process of the form: ÎľtR Ď R ÎľtR 1 ΡtR with ΡtR ~ N (0, Ď&#x192; R2 ) and iid. This shock identifies monetary policy decisions which imply deviations from the standard Taylor rule such as unconventional measures or to reshape the inflation expectations in the medium run. This MP schedule aims at replacing the standard LM curve commonly found in the standard AS-AD model. It proposes an up-to-date description of the behaviour of central banks that control a short run nominal interest rate instead of a monetary aggregate [Clarida, Gali, and Gertler 1999].

10

Baranowski et al. [2013] propose an endogenous mechanism.

In appendices, we provide an interest rate smoothing with smoothing parameter Ď . In this section we neglect these features such that Ď = 0.

11

114

Economics and Business Review, Vol. 1(15), No. 2, 2015

This 3-equation model is a stylised shortcut that encompasses supply and demand relations to determine how the three main macroeconomic variables of interest (the output gap, the inflation rate and the nominal interest rate) react to exogenous supply and demand shocks. In this short presentation we ignore more recent developments associated with the introduction of financial frictions that give rise to an acceleration phenomenon [see for example Poutineau and Vermandel 2015a, b].

2. The solution to a static version of the model This second section simplifies the previous system (1)â&#x20AC;&#x201C;(3) to convey the main ideas of the NKS model. Following Bofinger, Mayer, and Wollmershäuser [2006] and Poutineau and Vermandel [2015b] we neglect the dynamic aspects of the model and we concentrate on a static version of the framework.12 This is helpful to obtain the reduced form for the main variables of interest and to understand intuitions regarding the working of the model using tools similar to the IS-LM and AD-AS frameworks. To obtain the static version of the model we firstly assume that the monetary authorities are perfectly credible in the conduct of monetary policy so that the private sector expects that they reach the targeted inflation rate in future, namely that Et ^Ď&#x20AC;Ë&#x2020;t 1` = Ď&#x20AC;0, where Ď&#x20AC;0 is the long-run targeted rate of inflation. Secondly, we assume that the economy is very close to full employment so that the authorities are able to close the output gap in the future, namely that Et ^Ď&#x20AC;Ë&#x2020;t 1` = y0. Thus the gap between the real interest rate and the natural interest rate disappears. In this case we can express the monetary policy rule in terms of the real interest rate. Imposing these restrictions, the simplified static framework gives: Ď&#x20AC; = Ď&#x20AC;0 + Îşy + ÎľS,

(4)

y = y0 â&#x20AC;&#x201C; Ď&#x192;r + ÎľD,

(5)

r = Ď&#x2022;Ď&#x20AC;(Ď&#x20AC; â&#x20AC;&#x201C; Ď&#x20AC;0) + Ď&#x2022;yy + ÎľR.

(6)

In equilibrium the values of the output gap y*, the inflation rate Ď&#x20AC;* and the interest rate r* solution to the model (4)â&#x20AC;&#x201C;(6) are a linear combination of exogenous shocks: y* 12

y0 Ď&#x192;Ď&#x2022; Ď&#x20AC; S Ď&#x192; R 1 D Îľ Îľ Îľ ,

Ί Ί Ί Ί

Dynamic aspects will be reintroduced in Section 3.

115

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

Ί ÎşĎ&#x192;Ď&#x2022; Ď&#x20AC; S Îş Ď&#x192;Îş Îş y0 Îľ R Îľ D Îľ , Ί Ί Ί Ί Ď&#x2022; Ď&#x20AC;Îş Ď&#x2022; y y0 Ď&#x2022; Ď&#x20AC;Îş Ď&#x2022; y D Ď&#x2022; Ď&#x20AC; S 1 R Îľ Îľ Îľ ,

Ί Ί Ί Ί

Ď&#x20AC; * Ď&#x20AC;0 r*

where Ί = 1 + Ď&#x192;(Ď&#x2022;Ď&#x20AC;Îş + Ď&#x2022; y).

The adjustment of the output gap, the inflation rate and the nominal interest rate following alternative shocks is summarized in Table 1. As shown in the first column a supply shock leads to a decrease in the output gap, (activity decreases below its natural level), and to an increase in the inflation rate and in the interest rate. As shown in the second column a demand shock leads to an increase in the output gap, (activity increases), in the inflation rate and the interest rate. As observed, the reactions of the variables of interest to exogenous shocks are clearly affected by the value of the parameters of the interest rate rule of the authorities (Ď&#x2022;Ď&#x20AC; and Ď&#x2022; y).

Table 1. Reduced form of the static model

Output gap wy / wÎľ

Inflation wĎ&#x20AC; / wÎľ

Interest rate wr / wÎľ

Supply Shock ÎľS

Demand Shock ÎľD

Monetary Shock ÎľR

Ď&#x192;Ď&#x2022; Ď&#x20AC; 0 Ί

1 !0 Ί

Ď&#x192; 0 Ί

1 Ď&#x192;Ď&#x2022; y !0 Ί

κ !0 Ί

Ď&#x192;Îş 0 Ί

Ď&#x2022;Ď&#x20AC; >0 Ί

Ď&#x2022; Ď&#x20AC;Îş Ď&#x2022; y !0 Ί

1 !0 Ί

To understand more clearly the reaction of the economy to supply and demand shocks we refer the reader to figures 1 and 2. Graphically the model can be represented as consisting of two panels: in the lower panel of each figure, the IS-MP block (equations (5) and (6)) presented in the (y, r) space focuses on demand side aspects and can be treated as a New Keynesian representation of the IS-LM framework; in the upper panel the AD-PC block presented in the (y, Ď&#x20AC;) space determines the global equilibrium of the economy and can be treated as a New Keynesian representation of the AD-AS framework. The PC curve is given by equation (4) and the AD curve is obtained by combining equations (5) and (6) and is defined in equation (7), y

y0 Ď&#x192;Ď&#x2022;Ď&#x20AC; 1

ÎľtD. (Ď&#x20AC; Ď&#x20AC;0 ) 1 Ď&#x192;Ď&#x2022; y 1 Ď&#x192; Ď&#x2022; y 1 Ď&#x192;Ď&#x2022; y

(7)

116

Economics and Business Review, Vol. 1(15), No. 2, 2015

The consequences of the demand shock are presented in Figure 1. The first panel displays the adjustment of the inflation rate and the output gap. The second panel displays the adjustment of the demand side, accounting for the reaction of the central bank to the shock. π PC

B

πC π0

C

A AD’ AD y0

yB

yC

r

y

MP’ MP

C’

rC

B’

IS’ r0

A’ y 0 yC

IS

yB

y

Figure 1. Demand shock

To understand the main differences between the two panels one has just to remember that the IS curve (5) moves one for one with a demand shock whilst the demand curve moves by less than one. Thus, taking point A as the initial equilibrium of the model a positive demand shock moves the IS curve from IS to IS’ in the lower panel, which in turn, ignoring the reaction of the central bank, moves the demand schedule to the dotted line. As the temporarily equilibrium B implies an increase in the inflation rate the central bank reacts by increasing the interest rate for any value of the output gap. Thus, the MP curve in the lower panel moves left from MP to MP’. This, in turn, leads the aggre-

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

117

gate demand curve to move to the left from the dotted line to AD’. In the final equilibrium C, the evolution of aggregate demand from AD to AD’, that combines both the initial demand shock and the monetary reaction, is less than proportional to the demand shock. Furthermore, with the reaction of the central bank the increase of inflation is dampened. Finally the positive demand shock leads to an increase in the output gap, an increase in inflation and a rise in interest rate, as summarized in Table 1. π

PC’ PC

πB π0

B A

AD yB

y0

y

r MP’ MP

rB r0

B’ A’ IS

yB

y0

y

Figure 2. Supply shock

The consequences of the supply shock are presented in Figure 2. In this example the supply shock is a positive inflation shock (that corresponds to a decrease in the supply of goods ). Following this supply shock the Phillips curve moves upwards to the left in the (y, π) space. This shock leads to an increase in the rate of inflation and the central bank reacts by raising the interest rate.

Graphically the reaction of the central bank means increasing the interest rate for any value of the output gap so that the MP curve moves left to MP’ in the

118

Economics and Business Review, Vol. 1(15), No. 2, 2015

lower panel of Figure 2. Once all the adjustments have been implemented the final equilibrium lies at point B which is characterized by a negative output gap (namely activity falls below its natural value), an increase of the inflation rate over its targeted value and at point B’ an increase in the interest rate (needed to dampen part of the inflation consequences of the supply shock).

Finally, the balance between the consequences of the shocks on activity and inflation depends on the slope of the demand curve which, in turn, is affected by the reaction of the central bank to inflation rate and output gap developments.

A more conservative central bank (namely a central bank that puts a higher weight on inflation and a lower weight on the output gap) makes the slope of the demand curve of the economy flatter in the upper panel of figures 1 and 2, which translates into lower fluctuations in the interest rate but to a higher variability of the output gap. Conversely if the stance of the central bank reaction is more sensitive to the output gap and less sensitive to inflation then the MP and AD curves become steeper and shocks have a lower impact on activity and a higher impact on inflation.

3. The fully-fledged model In the dynamic version of the model (1)–(3), each period t corresponds to a quarter. As the fully fledged model does not have a closed form solution it must be simulated around a stable steady state. The solution procedure, based on the Blanchard-Kahn [1980] approach,13 requires the choice of numerical values for the parameters of the model in order to compute Impulse Response Functions (IRFs hereafter ) and the corresponding variance decomposition of the three variables of interest of the model.

3.1. The solution procedure The solution procedure introduced by Blanchard and Kahn [1980] is based on matrix calculus and is aimed at selecting a unique stable dynamic path to describe the reaction of the variables following the occurrence of exogenous shocks. The Blanchard-Kahn condition defines a necessary criterion to get this result through the equality between the number of forward variables and the number of unstable eigenvalues. Practically the problem of the eigenvalues translates into the problem of appropriate values of the structural parameters of the model or their combinations. To be solved the model first has to be 13 In this paper we adopted the Blanchard-Kahn approach for solving the model, given its anteriority and popularity in literature. However, the reader should be aware of the existence of other methods introduced by Klein [2000] and Sims [2000]. Miao [2014] offers a nice comparison between these three approaches.

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

119

written in a state-space representation. For our linear model (1)â&#x20AC;&#x201C;(3), defining Î&#x17E; (Ď&#x192; Ď&#x2022; y Îş Ď&#x2022; Ď&#x20AC; ) 1, this representation is:

ÂŞ yË&#x2020; t Âş ÂŤË&#x2020; Âť ÂŹĎ&#x20AC;t Âź

Î&#x17E; βĎ&#x192;

ÂŞĎ&#x192; Âş 1 βĎ&#x2022;Ď&#x20AC; Âş ÂŞ Et ^ yË&#x2020; t 1`Âş Î&#x17E; ÂŞ Ď&#x192;ÎľtS

ÂŤ Âť ÂŤ Âť . (8) ÂŤ Âť y D S R ÂŤÂŹ Ď&#x192;Îş Îş β(Ď&#x192; Ď&#x2022; )Ÿ ÂŹ Et ^Ď&#x20AC;Ë&#x2020;t 1`Âź βĎ&#x192; ÂŤÂŹ Ď&#x192;βξ t βξt βξt Ÿ

The Blanchard-Kahn condition states that there are as many eigenvalues of ÂŞĎ&#x192; 1 βĎ&#x2022; Ď&#x20AC; Âş the matrix ZT ÂŤ greater than one in modulus as there are y Âť ÂŹÂŤ Ď&#x192;Îş Îş β(Ď&#x192; Ď&#x2022; )Ÿ non-predetermined variables. Since there are two forward-looking variables in the model (1)â&#x20AC;&#x201C;(3) ( yË&#x2020;t and Ď&#x20AC;Ë&#x2020;t), we know that there should be exactly two eigenvalues outside the unit circle to get one unique stable trajectory of each of the modelâ&#x20AC;&#x2122;s variable around the steady state. Given the form of the matrix ZT , the Blanchard-Kahn condition for the model (1)â&#x20AC;&#x201C;(3) reduces to the following relation: Îş(Ď&#x2022;Ď&#x20AC; â&#x20AC;&#x201C; 1) + (1 â&#x20AC;&#x201C; β)Ď&#x2022;y > 0.

Table 2. Calibration of parameters Parameter

Value

β

0.99

Ď&#x192;

1

relative risk aversion

Îľ

6

elasticity of substitution amongst goods

Ď&#x2020;

1

elasticity of marginal disutility with respect to labour

Ď&#x2022;Ď&#x20AC;

1.5

Ď&#x2022;y

0.5/4

influence of output gap in the interest rate rule

Ď S

0.90

persistency of supply shock

D

0.90

persistency of demand shock

R

Ď

0.40

persistency of monetary policy shock

θ

3/4

probability of retaining old price

Ď

Description discount factor

influence of inflation rate in the interest rate rule

Source: Authorsâ&#x20AC;&#x2122; synthesis.

This condition reduces to the choice of appropriate values for the parameters of the model. A sufficiently relevant condition for the previous one to be hold is that the monetary authorities should respond more than proportionally to inflation developments (namely, Ď&#x2022;Ď&#x20AC; > 1) according to the Taylor principle. In this case a rise in inflation leads to a more than proportional rise in nominal interest causing an increase in real interest rates that affects agentsâ&#x20AC;&#x2122; economic decisions and thus the real macroeconomic equilibrium of the model. The

120

Economics and Business Review, Vol. 1(15), No. 2, 2015

choice of parameters is therefore a main feature of the analysis as it must both represent economic features and contribute to the Blanchard-Kahn condition. As presented in Table 2, following Galí [2008], we use a calibration of the model parameters that is commonly selected in the literature. The intra-temporal elasticity between intermediate goods is set at 6 which implies a steady state mark-up of 20 % in the goods’ market corresponding to what is observed in main developed economies. The sensitivity of the inflation rate to changes in the marginal cost is equal to 0.13 roughly. The value of the discount factor set at 0.99 implies the steady state quarterly interest rate equal to one and the steady state real return on financial assets of about 4 percent per year. Average price duration amounts to three quarters which is consistent with empirical evidence.14 The values of coefficients in the interest rate rule (3) are consistent with variations observed in the data on inflation and the interest rate given in the annual rates.15 Because in our model periods are interpreted as quarters the output gap coefficient has to be divided by 4.

3.2. Impulse-response analysis The mechanisms by which random innovations change into endogenous variables fluctuations may be illustrated by impulse response functions (IRFs).

Each IRF isolates the impact of a particular shock throughout the economy. To document the response of activity, inflation and nominal interest we sequentially describe the consequences of a supply, demand and interest rate shock.

The demand shock: Figure 3 documents the consequences of a 1% positive demand shock. As observed the increase in goods demand for leads to an in-

Benchmark regime is: ϕπ = 1.5, ϕ y = 0.5/4, inflation target regime: ϕπ = 1.7, output gap regime: ϕ y = 0.8/4

Figure 3. Effects of a 1% demand shock 14

Galí, Gertler, López-Salido [2001] and Sbordone [2002] provide estimations based on aggregate data. Galí [2008] points also to some micro evidence.

15 These values were originally proposed by Taylor [1999] as a good approximation of the monetary policy conducted by the Federal Reserve in years 1986–1999 when the head of the USA central banking system was Alan Greenspan. His monetary policy decisions largely followed standard Taylor rule recommendations.

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

121

crease in activity so that the output gap becomes positive. However, as production overshoots its natural value this rise in activity increases the inflation rate.

Since both the output gap and inflation rate increase the central bank should react by raising the nominal interest rate.

According to the Taylor principle the nominal interest rate increases more than proportionally to inflation developments to affect real exchange rates. This policy however is not sufficient to close the positive output gap immediately or to dampen the inflation rate. The effect of monetary policy should be assessed over time on the output gap (activity goes back to its natural value as time passes) and on the rate of inflation (that converges towards its natural value). The adjustment time path is affected by the parameter value of the Taylor rule. As presented in Figure 3 a higher concern for inflation or output gap reduces the volatility of both activity and inflation. Thus, stricter monetary policy leads to more moderate responses of variables to the demand shock.

The supply shock: Figure 4 represents the consequences of a 1% increase in inflation (i.e. the negative supply shock acts as an increase in the price of raw materials or energy that increases the real marginal cost of production).

This shock has a direct impact on inflation that rises and overshoots its targeted value. As a consequence monetary authorities should react according to the Taylor principle by raising the interest rate. Since the increase in the nominal interest rate is higher than the rate of inflation, the real rate rises. This, in turn, negatively affects output that decreases under its natural value. However, as time passes, the increase in the interest rate dampens inflation. Finally, the output gap goes back to its steady state value whilst the inflation rate reaches its targeted value. As previously for the demand shock, the time path of variables is affected by the parameter values of the Taylor rule. A higher concern for output gap (as represented with ‘inflation target’ IRF) leads to weaker responses of real variables and stronger responses of nominal variables. Inversely a higher concern for inflation leads to stronger responses of real variables and weaker responses of inflation and nominal interest rate.

Benchmark regime is: ϕπ = 1.5, ϕ y = 0.5/4, inflation target regime: ϕπ = 1.7, output gap regime: ϕ y = 0.8/4

Figure 4. Effects of a 1% supply shock

122

Economics and Business Review, Vol. 1(15), No. 2, 2015

Benchmark regime is: ϕπ = 1.5, ϕ y = 0.5/4, inflation target regime: ϕπ = 1.7, output gap regime: ϕ y = 0.8/4

Figure 5. Effects of 1% a monetary policy shock

The monetary policy shock: Figure 5 documents the consequences of a 1% increase in the nominal interest rate (corresponding to a 25 basis point increase in the exogenous shock measured in quarterly terms as presented in the figure). Because of sticky prices the initial increase in the nominal interest rate implies a corresponding increase in the real interest rate at the initial period.

This depresses demand in the economy as it leads households to delay their consumption through intertemporal consumption smoothing as reported in the Euler condition. Since activity is demand determined, firms’ production decreases. In the meanwhile the drop in demand generates deflation. The economy recovers overtime, since, according to the Taylor rule, a decrease in both activity and in the inflation rate leads to a reduction in the nominal interest rate after the initial period.

3.3. Business cycle statistics IRF analysis aims at isolating the effect of a particular shock on the dynamics of endogenous variables. However, in real life situations, shocks occur both randomly and jointly to affect the macroeconomic equilibrium. The combined effect of supply and demand shocks over time is captured by historical variance analysis. The aim of this exercise is both to evaluate the relative contribution of each type of shock on the motion of macroeconomic variables over time and to appreciate how a particular design for economic policy may dampen the effect of one particular type of shock. Table 3 shows the variance decomposition of activity, inflation and the nominal interest rate under the benchmark calibration of Table 2 and evaluates the sensitivity of the benchmark results to alternative values of key behavioural and policy parameters of the model.

In the first panel of Table 3 (Benchmark calibration), supply side shocks (namely price mark- up shocks) explain most of the output variability leaving only a marginal contribution (around 4%) to demand and interest rate shocks.

In contrast the variability of the inflation rate is mainly explained by demand and monetary policy innovations. Finally, around 2/3 of interest rate variability is explained by real supply side shocks.

123

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

Table 3. Variance decomposition (in %) Supply

Demand

Monetary Policy

Production

95.93

3.16

0.91

Inflation

48.13

51.31

0.56

Interest rate

63.00

36.65

0.34

Production

96.72

3.16

0.09

Inflation

99.07

0.76

0.17

Interest rate

99.02

0.08

0.00

1 – Benchmark

2 – Sticky economy θ = 0.95

3 – Quasi-flexible economy θ = 0.01 Production

90.84

2.99

6.17

Inflation

0.00

99.53

0.47

Interest rate

0.00

93.72

6.28

π

4 – Aggressive Monetary Policy ϕ = 2.5 Production

99.09

0.46

0.45

Inflation

39.81

59.36

0.83

Interest rate

62.49

36.36

1.15

y

5 – Output-oriented monetary policy ϕ = 1 Production

96.02

3.16

0.82

Inflation

89.10

10.85

0.06

Interest rate

97.60

2.38

0.02

In panel 2 (sticky economy) and panel 3 (quasi flexible economy) we evaluate the sensitivity of the benchmark results to alternative assumptions regarding nominal rigidities. In the sticky economy only 5% of the total number of firms can reset their price each period. Whilst in the quasi flexible situation 99% of the total number of firms reset their prices each quarter. The main consequences can be assessed with regard to the contribution of supply side shocks to inflation and interest rates. Remarkably supply side shocks have no effect on either inflation or interest rates when prices are flexible. In contrast the fluctuations of the output gap are more sensitive to interest rate shocks whilst the effect of demand shocks on activity is almost unobsevable.

In panel 4 and 5 we evaluate the sensitivity of the benchmark results to alternative assumption regarding the conduct of monetary policy. When a monetary policy is more aggressive in terms of inflation (panel 4) it dampens the effect of demand shocks on activity (and in contrast makes output development

124

Economics and Business Review, Vol. 1(15), No. 2, 2015

more sensitive to supply shocks) and reinforces the impact of demand shocks on inflation (whilst , conversely, it dampens the impact of supply side shocks on this variable). Finally, this policy has almost no noticeable effect on the relative contribution of shocks on interest rate developments. In panel 5 an output oriented monetary policy increases the effect of supply shocks on inflation and interest rate whilst leaving the relative contribution of shocks on activity almost unchanged. The results obtained in these last two panels may serve as simple guideline to determine the nature of monetary policy depending on both its objective and the origin of shocks. If an economy is mainly affected by price mark-up shocks monetary policy should be more closely oriented towards output developments. As this policy is able to dampen the effect of supply shocks on inflation, whilst having no noticeable effect on activity, monetary authorities are able to stabilise prices more easily. In contrast if the economy is affected by demand shocks the authorities have to use arbitrage because a more aggressive policy against inflation dampens the impact of demand shocks on activity whilst it increases the impact of demand shocks on inflation.

Conclusions In this paper we have described in a concise way the main ideas conveyed by the 3 equation New Keynesian model and the main elements of the solution procedure required to analyse the dynamics of the model. To introduce the reader to this class of models we have presented a simple static version of the model that gives both direct reduced forms and provides the basis for a simple graphical analysis of the macroeconomic equilibrium. We have then introduced the Blanchard-Kahn solution procedure and report IRFs to describe the dynamic adjustment of the economy over periods. Finally we have used the historical variance analysis to evaluate how a modification of values of the key parameters of the model affect the relative contribution of supply side and demand side shocks. Our aim was not to provide the reader with a comprehensive and up to date catalogue of all the results obtained by this New Keynesian literature but rather to offer a clear and simple presentation of the basic ideas and the required technical tools needed to solve this class of models that have become the conventional workhorse of today’s macroeconomics.

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

125

Appendix A. Micro-foundations A.1. Households There is a continuum of households j ď&#x192;&#x17D; [0; 1] with a utility function Ct1 Ď&#x192; H 1 Ď&#x2020; Ď&#x2021; t , the representative household maximizes its welU (Ct , H t ) 1 Ď&#x192; 1 Ď&#x2020; fare, defined as the expected stream of utilities discounted by β ď&#x192;&#x17D; (0, 1): f

max

Ct ( j ), H t ( j ), Bt ( j )

Et ÂŚ β Ď&#x201E;U Ct Ď&#x201E; ( j ), H t Ď&#x201E; ( j ) .

(A.1)

Ď&#x201E; 0

Under the budget constraint: D

Pt Ct ( j) e Ď&#x192;Îľt Bt ( j) Rt 1Bt 1 ( j) Wt H t ( j),

(A.2)

where Ď&#x192; > 0 and Ď&#x2020; > 0 are shape parameters of the utility function with respect to consumption and to labour supply whilst Ď&#x2021; is a shift parameter which scales the steady state labour supply to realistic values. As in Smets and Wouters [2005] we introduce an AR(1) demand shock process in the budget constraint of the D representative household denoted by Îľt .

After replacing the Lagrange multiplier the first order conditions are defined by the Euler bond condition: § C ( j) ¡ Et ¨ t 1 ¸ Š Ct ( j) š

where Ď&#x20AC;t 1 mined by:

Ď&#x192;

§ R ¡ Et ¨ t ¸ , e Š Ď&#x20AC;t 1 š β

ÎľtD

(A.3)

Pt 1 is the inflation rate and the labour supply equation is deterPt Ď&#x2021;Ct ( j)Ď&#x192; H t ( j)Ď&#x2020;

Wt .

Pt

(A.4)

These equations define the optimal paths of labour and consumption and maximize the welfare index of the representative household.

A.2. Firms The representative firm i maximizes its profits: max

H t (i ), Yt (i )

^Pt (i)Yt (i) Wt H t (i)`,

(A.5)

126

Economics and Business Review, Vol. 1(15), No. 2, 2015

under the supply constraint: Yt(i) = Ht(i).

(A.6)

We suppose that firms solve a two-stage problem. In the first stage, firms choose labour demand in a perfectly competitive market. The first order condition is: MCt (i) MCt

Wt , Pt

(A.7)

where MCt denotes the nominal marginal cost of producing one unit of goods.

In the second stage problem the firms cannot optimally set prices. There is a fraction of firms θ that are not allowed to reset prices. Prices then evolve according to Pt(i) = Pt â&#x20AC;&#x201C; 1(i). The remaining share of firms 1 â&#x20AC;&#x201C; θ can set their selling price such that Pt(i) = Pt*(i), where Pt*(i) denotes the optimal price set by the representative firm given the nominal rigidity. The maximization programme is thus defined as: Îťct Ď&#x201E; (βθ )Ď&#x201E; ÂŹÂŞPt*(i) MCt Ď&#x201E; (i)ټYt Ď&#x201E; (i), c 0 Îťt

f

max Et ÂŚ Pt*(i )

Ď&#x201E;

(A.8)

under the downward sloping constraint from goodsâ&#x20AC;&#x2122; packers: Et ^Yt Ď&#x201E; (i)`

Îź t Ď&#x201E; ­ ½ Îźt Ď&#x201E; 1 § ¡ P * ( i ) ° t ° Et Ž¨ Yt Ď&#x201E; ž ,  Ď&#x201E; ! 0, ¸ P °Š t Ď&#x201E; š ° ÂŻ Âż

(A.9)

where: Îźt

Â&#x2039; Â&#x2039; 1

S

e γξt â&#x20AC;&#x201C; the time-varying mark-up,

ď&#x192;˛

â&#x20AC;&#x201C; denotes the imperfect substitutability between different goods varieties, ÎľtS â&#x20AC;&#x201C; denotes the mark-up shock, Îł â&#x20AC;&#x201C; a shift parameter that normalizes the shock to unity in the log-linear form of the model as in Smets and Wouters [2005].

Since firms are owned by households they discount the expected profits usc ing the same discount factor as households (βĎ&#x201E;Îťt+Ď&#x201E; /Îťtc). The first order condition is thus: Îťct Ď&#x201E; ( βθ )Ď&#x201E; ÂŞÂŹ Pt*(i) Îźt Ď&#x201E; MCt Ď&#x201E; (i)ºŸ Yt Ď&#x201E; (i) 0. c 0 Îťt Îźt Ď&#x201E; 1

f

Et ÂŚ Ď&#x201E;

(A.10)

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

127

A.3. Authorities To close the model the monetary policy authority sets its interest rate according to a standard Taylor Rule: Ď&#x20AC;

y

1 Ď

Ď Ď&#x2022; Ď&#x2022; § Rt ¡ §¨ § Ď&#x20AC;t ¡ § Yt ¡ ¡¸ ¨ ¸ ¨¨ Ď&#x20AC; ¸ ¨ ¸ ¸ Š R š ŠŠ š Š Y š š

Rt R

R

e Îľt ,

(A.11)

where: Rt â&#x20AC;&#x201C; the nominal interest rate, Ď&#x20AC;t â&#x20AC;&#x201C; the inflation rate, Yt â&#x20AC;&#x201C; the level of output, ÎľRt â&#x20AC;&#x201C; an AR(1) monetary policy shock.

Finally, parameters R, Ď&#x20AC; and Y are steady state values for the interest rate, the inflation rate and GDP16. The central bank reacts to the deviation of the inflation rate and the GDP from their steady state values in a proportion of Ď&#x2022;Ď&#x20AC; and Ď&#x2022; y, the central bank also smoothes its rate in a proportion of degree Ď .

A.4. Equilibrium conditions After aggregating all the supplies by firms the resource constraint for the economy is defined by: Yt = Ct.

(A.12)

Whilst the aggregation between constrained firms and non-constrained firms leads to the following equation for aggregate prices: 1 1 Îźt

(Pt )

1 1 Îźt

θ (Pt 1 )

1 1 Îźt

(1 θ )(Pt*)

.

(A.13)

B. Linearization To obtain the steady state of the model, we normalize prices i.e. P = 1 whilst we assume that households work one third of their time H = 1/3. Then we find: C Y W

H,

MC

1 , Îź

Ď&#x2021; WC Ď&#x192; H Ď&#x2020; .Â

16

Under a credible central bank, Ď&#x20AC; and Y also can be interpreted as the targets of the central bank in terms of inflation rate and GDP.

128

Economics and Business Review, Vol. 1(15), No. 2, 2015

First, combining the Euler bond equation (A.3) and the resources constraint (A.12), i.e. yË&#x2020; t cË&#x2020;t, we get production determined by: yË&#x2020; t

Et yË&#x2020; t 1

1 rË&#x2020; Et Ď&#x20AC;Ë&#x2020;t 1 ÎľtD. Ď&#x192; t

(A.14)

The labour supply equation (A.4) in log-deviation is: wË&#x2020; t

Ď&#x192;cË&#x2020;t Ď&#x2020;hË&#x2020;t,

(A.15)

where wË&#x2020; t denotes the variations of the real wage. Up to a first order approximation of the firm price optimization solution (A.10) and the aggregate price equation (A.13), the linearized new Keynesian Phillips curve is: Ď&#x20AC;Ë&#x2020;t

βEt Ď&#x20AC;Ë&#x2020;t 1

(1 θ )(1 θβ) Ë&#x2020; t ÎľtS. mc θ

(A.16)

Ë&#x2020; t wË&#x2020; t and the production function yt = ht, Thus the real marginal cost is: mc then from the labour supply equation, the marginal cost can be simplified as: Ë&#x2020; t (Ď&#x192; Ď&#x2020;) yË&#x2020;t. Then the Philips curve is: mc Ď&#x20AC;Ë&#x2020;t

βEt Ď&#x20AC;Ë&#x2020;t 1

(1 θ )(1 θβ) (Ď&#x192; Ď&#x2020;) yË&#x2020; t ÎľtS. θ

(A.17)

Finally, the monetary policy is determined by: rË&#x2020;t

Ď rË&#x2020;t 1 (1 Ď ) Ď&#x2022; Ď&#x20AC; Ď&#x20AC;Ë&#x2020;t Ď&#x2022;y yË&#x2020; t Îľ tR.

(A.18)

To summarize, our model is determined by the following set of three equations: ­ ° ° ° ÂŽ ° °rË&#x2020;t °¯

1 rË&#x2020;t Et Ď&#x20AC;Ë&#x2020;t 1 ÎľtD , Ď&#x192; (1 θ )(1 θβ) Ď&#x20AC;Ë&#x2020;t βEt Ď&#x20AC;Ë&#x2020;t 1 (Ď&#x192; Ď&#x2020;) yË&#x2020; t ÎľtS , θ Ď rË&#x2020;t 1 (1 Ď ) Ď&#x2022;Ď&#x20AC; Ď&#x20AC;Ë&#x2020;t Ď&#x2022;y Ë&#x2020;yt Ď&#x2022;Î&#x201D;y Ë&#x2020;yt Ë&#x2020;yt 1 ÎľtR . yË&#x2020; t

Et yË&#x2020; t 1

Where shock processes are determined by: Îľti

Ď i Îľti 1 Ρti , i D, S, R.

(A.19)

J.-C. Poutineau, K. Sobczak, G. Vermandel, The analytics of the New Keynesian

129

References Baranowski, P., Gałecka-Burdziak, E., Górajski, M., Malaczewski, M., Szafrański G., 2013, Inflacja a mechanizmy aktualizacji cen. Studium dla Polski, Łódź, Wydawnictwo Uniwersytetu Łódzkiego, Wydawnictwo Naukowe PWN. Blanchard, O., Kahn, C.M., 1980, The Solution of Linear Difference Models under Rational Expectations, Econometrica, vol. 48, no. 5: 1305–1311.

Bludnik, I., 2009, The New Keynesianism – Proclamation of a Consensus?, Poznań University of Economics Review, vol. 9, no. 1: 5–24.

Bofinger, P., Mayer, E., Wollmershäuser, T., 2006, The BMW Model: A New Framework for Teaching Monetary Economics, Journal of Economic Education, vol. 37, no. 1: 98–117.

Carlin, W., Soskice, D., 2014, Macroeconomics: Institutions, Instability, and the Financial System, Oxford University Press.

Clarida, R., Galí, J., Gertler, M., 1999, The Science of Monetary Policy: A New Keynesian Perspective, Journal of Economic Literature, vol. 37, no. 4: 1661–1707.

Galí, J., 2008, Monetary Policy, Inflation, and the Business Cycle: An Introduction to the New Keynesian Framework, Princeton University Press, Princeton.

Galí, J., Gertler, M., López-Salido D., 2001, European Inflation Dynamics, European Economic Review, vol. 45, no. 7: 1237–1270.

Goodfriend, M., King, R., 1997, The New Neoclassical Synthesis and the Role of Monetary Policy, NBER Macroeconomics Annual, vol. 12: 231–296.

Klein P., 2000, Using the Generalized Schur Form to Solve a Multivariate Linear Rational Expectations Model, Journal of Economic Dynamics and Control 24(10), September: 1405–1423.

Miao, J., 2014, Economic Dynamics in Discrete Time, MIT Press.

Poutineau, J.C., Vermandel, G., 2015a, Cross-border Banking Flows Spillovers in the Eurozone: Evidence from an Estimated DSGE Model, Journal of Economic Dynamics and Control, no. 2: 378–403.

Poutineau, J.C., Vermandel, G., 2015b, A Primer on Macroprudential Policy, The Journal of Economic Education, forthcoming.

Romer, D., 1993, The New Keynesian Synthesis, Journal of Economic Perspectives, vol. 7, no. 1: 5–22.

Sbordone, A.M., 2002, Prices and Unit Labor Costs: Testing Models of Pricing Behavior, Journal of Monetary Economics, vol. 45, no. 2: 265–292.

Sims, C., 2000, Solving Linear Rational Expectation Models, Working Paper, Department of Economics, Princeton University.

Smets, F., Wouters, R., 2005, Comparing Shocks and Frictions in US and Euro Area Business Cycles: a Bayesian DSGE Approach, Journal of Applied Econometrics, 20(1).

Taylor, J. B., 1999, A Historical Analysis of Monetary Policy Rules, in: Taylor, J.B. (ed.), Monetary Policy Rules, University of Chicago Press , Chicago.

Woodford, M, 2003, Interest and Prices: Foundations of a Theory of Monetary Policy, Princeton University Press, Princeton.

Economics and Business Review, Vol. 1 (15), No. 2, 2015: 130–148

Investments and long-term real interest rate in Poland. Study of investment structure, current account and their correlation with long-term real interest rates1 Jakub Krawczyk, Szymon Filipczak2

Abstract: One of the effects of the globalization of financial markets is unhampered international capital flows that can be freely reallocated amongst different countries.

Considering that the investor’s aim is to look for the most profitable investment opportunities, they have natural inducement to invest in emerging economies. Taking into account differences in the level of interest rates between Poland and developed economies we claim that the contribution of foreign capital in domestic investments decreases the cost of capital in Poland. The two main aims of this paper are to identify crucial channels of foreign capital flows to Poland and to find out whether co-financing by foreign investment in Poland influences the domestic cost of capital (long-term real interest rate). Our findings, based on empirical analysis, appear to confirm that there is a connection between the structure of investors and the cost of capital.

Keywords: international finance, current account, investments, interest rate, capital flows.

JEL codes: E22, E43, E44, F21, F30.

Introduction Long term economic success of every country is inseparable from growth of effective investments, especially financed by private capital. They create fundamentals for other components of the economy, such as private consumption or industrial production. Significant role of investments for economy was a matter of interest for many academics. They investigated it from a perspective of 1

Article received 20 May 2014, accepted 19 February 2015.

Poznań University of Economics, Department of Capital Investments and Financial Strategies, al. Niepodległości 10, 61-875 Poznań, Poland, corresponding author: kuba-krawczyk@wp.pl.

2

131

J. Krawczyk, S. Filipczak, Investments and long-term real interest rate in Poland

real economy as well as from financial perspective. Overtime second issue has gained on importance due to, among others, global competition for capital.

The main focus in this paper are investments, as a crucial component of Gross Domestic Product (GDP) growth (see Figure 1). It played a vital part in the transformation of developing economies towards those more marketoriented. We would like to analyze it in the Polish context as one of the biggest European economies, that used to be a part of communist block and now is a European Union (EU) member and benefited from a high increase of investment flows over years (see Section 2).

Investment was one of the GDP growth component in recent years, especially positively contributing in 2011, what is illustrated in Figure 1. On the other hand, in last available quarters (third quarter 2012 – second quarter 2013) investment contributed negatively and GDP growth was maintained mostly by net export.

It is obvious that there is a link between investment and GDP growth, however our aim was not to prove such a relationship, as it has already been analyzed by others. For example Rydarowska-Kurzbauer [2012] calculated that the correlation coefficient of GDP growth rate versus changes in investments amounted to 0.955 for Poland in the period 2000–2007. Our goal has been to investigate the historical investment patterns, relationships between foreign in% 8 6 4 2 0 –2

Private Consumption [C] Investments [I] net Export [NX]

Public Consumption [G] Inventories (change) Gross Domestic Product (%, yoy)

Figure 1. GDP real growth and the contribution of main components, 1Q’2009 – 2Q’ 2013 Source: Eurostat

2Q'13

1Q'13

4Q'12

3Q'12

2Q'12

1Q'12

4Q'11

3Q'11

2Q'11

1Q'11

4Q'10

3Q'10

2Q'10

1Q'10

4Q'09

3Q'09

2Q'09

–6

1Q'09

–4

132

Economics and Business Review, Vol. 1(15), No. 2, 2015

vestment flows and the real interest rate as well as the identification of the most important channels of cash flows that end up as an increase in the fixed capital formation in Poland. What is more, we believe that findings from this research may help a better understanding of Polish future economy developments.

The paper is divided into three sections and a conclusion. The first section is devoted to a literature overview. In the second section we make a short overview of historical data associated with our research field. The third section deals with empirical analysis substantiating our claim.

1. Literature overview Investments in the Central and Eastern Europe (CEE) region and particularly in Poland have been the subject of numerous studies most of which, however, did not approach it in a broader context (we would like to emphasise that we have focused on studies which concern the CEE region). Most authors dealt with one specific channel of investment and investigated determinants and effects for the economy resulting from it (e.g. foreign direct investments (FDI)).

Sometimes in analyses of flows to the CEE region authors put more stress on regional factors and did not go into specific reasons for certain situations in Poland, which is the centre of our interest. What is more, usually interest rates were not at the forefront (in the context of foreign investment) and frequently were not deeply investigated. Additionally, some of the analyses were prepared some time ago; therefore, do not cover recent developments.

A wide-ranging study of foreign capital flows to the region that we are interested in was conducted by Claessens, Oks & Polastri [2000]. They investigated capital flows to the whole of Central and Eastern Europe and the former Soviet Union in the period from 1990 till 1997. Their scope of interest was the whole region rather than one particular country; however, some of their findings were relevant for major countries in the region, including Poland. They brought out the fact that in the beginning of Polish transition investment flows were dominated by official flows, which has changed towards a more balanced official versus private split over the years (parallel to an increase in investment over that time). The bulk of increasing private flows took the form of FDI driven by the Polish good macro performance, the convertibility of the currency, moderate fiscal deficit and favourable prospects for EU membership. About 20% of them were accounted for by privatization. Of course the liberalization of the capital account was gradual in the early 90’s with preference for longterm investments at the beginning which had its impact on the flow structure [Arvai 2005]. Claessens, Oks & Polastri also tried to approach the interest rate influence on investments (especially portfolio investments), by calculating interest rate differentials (domestic interest rate corrected for the devaluation and London Interbank Offered Rate (LIBOR)) and comparing it with non-equity

J. Krawczyk, S. Filipczak, Investments and long-term real interest rate in Poland

133

portfolio investments which led them to the conclusion that there is a positive association. Claessens, Oks & Polastri found on the basis of output analysis that private flows, and FDI in particular, were determined first by reform efforts and then by prospective EU membership and credit worthiness. It is of interest that the sustainability of capital flows was associated with the continuation of reform efforts. It implies a source of financing (public versus private) and type of flows (reforms support FDI investments whilst they have a negative impact on short-term debt flows). Foreign Direct Investments flows were further investigated by others. Dunning [1993] focused on determinants and divided motives for FDI into four groups: resource seeking, market seeking, efficiency seeking and strategic asset seeking. It had been investigated in the Polish context by Gorynia, Nowak, and Wolniak [2007], who have analyzed seven cases of FDI in Poland in the 90’s with respect to the first three motives pointed out

by Dunning and concluded that foreign investments were motivated by market seeking (e.g. looking for access to the new market) and efficiency seeking (e.g. low labour cost). In any case resource seeking was identified as an important factor. Number of theories explaining internationalization and FDIs were revived and categorized by Trąpczyński [2014, p. 33] in his study of determinants FDI performance in the internationalisation process of Polish companies. Nevertheless he focused on outward FDI from Poland.

Following Claessens, Oks, and Polastri [2000] study, Garibaldi, Mora, Sahay, and Zettelmeyer [2001] looked at 25 economies in the region (transition economies in Europe and the former Soviet Union) and analyzed in greater detail the heterogeneity of capital flows to transition economies in quantitative terms.

What is more, their study covered three more years (1991–1999) and was primarily focused on direct and portfolio investments for which Garibaldi, Mora, Sahay, and Zettelmeyer ran regression. Their studies show that direct investment patterns can be explained by a set of standard economic fundamentals.

Inward direct investments are attracted mainly by a macroeconomic environment (measured by economic growth and a high fiscal balance as well as state of economic liberalization), economic reforms (using EBRD Trade and Foreign Exchange Index), the privatization methods (insider privatization discourages direct investments whilst equal access for foreign investors encourages same) and the presence of natural resources. What is of interest is that their sample does not reveal that wages have a stimulating impact on direct investments. In terms of portfolio investment a relatively small number or explanatory variables have been noted (at least in the model in question). The presence of market infrastructure and the protection of property rights seem to be the only significant factors.

Ancypowicz [2009] looks into FDI’s evolution in the context of Polish accession to the European Union. She points out that the influence of accession should not be overestimated and states that the amount of greenfield investments after accession was lower than the amount of reinvested earnings and

134

Economics and Business Review, Vol. 1(15), No. 2, 2015

bank borrowings of companies that were already present in Poland before EU accession; therefore their impact on GDP growth was not that significant (based on the data from the years 1995–2008 she calculated the correlation between GDP growth and investments to be 0.89, however GDP growth versus FDIs was only 0.25). In our opinion, investments that were made before EU accession were probably somehow anticipating Polish accession; therefore, they should be taken into account as well. Nevertheless, foreign investments provided a major stimulus for domestic entrepreneurs to develop their businesses (improve organization, logistics, technology) so as to be able to compete successfully).

Bijsterbosch and Kolasa [2009] draw attention to the positive impact of FDIs as well, concluding that FDI inflows played an important role in accounting for productivity growth in the CEE region (with critical dependence on the capacity to absorb).

International capital flows and their impact on economy growth in CEE countries that joined the European Union in 2004 and 2007 were also investigated by Śliwiński [2011]. He performed an analysis of determinants for capital and financial accounts. Based on the research of others [e.g. Calvo, Leiderman, and Reinhardt 1993; Hernandez, Mellado, and Valdes 2001; Carlson and Hernandex 2002; Ralhan 2006; Culha 2006; Schmitz 2009 and others] he created a list of potential determinants for CEE the region and tried to verify them using GLS, OLS and LSDV methods. He found out that there is a limited effect of international interest rates and GDP growth in the EU on flows to CEE countries that were chosen for the study. On the other hand, there was a positive relationship between domestic GDP growth and capital flows. A similar positive relationship has been established with the real currency rate. What is not surprising is that he found out that there is a negative relationship between budget deficit and portfolio flows. Changes in the overall current account deficit (which we treat as a foreign capital inflow – it is explained later) were investigated by Sobański [2010]. He carried out a signal analysis and logistic regression for developing countries (including some of the CEE countries). He found out that e.g. low export/foreign debt ratio or low export dynamics may predict some current account adjustments that result in changes in their value.

What is interesting from our perspective, Sobański concluded that the higher the difference between domestic and foreign interest rate, the lower the risk of rapid current account adjustments.

Going beyond the analyses discussed above the consequences of such flows are interesting as well. In a somewhat broader context they were looked at by Janicka [2008]. She points out that the liberalization of capital flows has a number of advantages such as better capital allocation, efficiency improvement of domestic companies, the possibility of creating more diversified portfolios of investments and a decrease in cost of capital for companies. The latter point was confirmed by Henry [2000] who investigated 11 developing countries (unfortunately none of them from the CEE region) and proved that there are good

J. Krawczyk, S. Filipczak, Investments and long-term real interest rate in Poland

135

grounds for accepting a supporting thesis that stock market liberalization positively influences investments (it is worth mentioning that Henry’s contribution focused on private investments). On the other hand, Janicka [2008] mentions that liberalization might create some threats such as the outflow of domestic capital, difficulties in macroeconomic and monetary policies, a possible decrease in tax inflows (due to capital outflow). According to Stiglitz, Janicka mentioned also that especially short-term capital might generate a risk of financial crisis and destabilization that might not be compensated for by advantages of liberalization. In a narrower context (only FDI, not capital flows in general) the effects of foreign investments on the Polish economy were investigated by Weresa [2009]. She reviewed literature on that topic and found out that from a theoretical point of view it was proven that FDIs might have a positive or negative influence on specific area of an economy (e.g. trade) depending on circumstances and countries. She claimed that FDIs have two opposite effects on investment activity in Poland. First, they increase the level of available financing, thereby support investment by entities with foreign capital. On the other hand, it indirectly limits the investment potential of domestic competitors. According to Weresa, FDI has a positive impact on the labour market by creating new jobs directly (especially through greenfield investments) as well as indirectly. What is more, they have a positive impact on labour efficiency.

Weresa’s analysis of FDI suggests that in the case of Poland it has rather positive consequences on trade (due to the fact that foreign capital usually flows to sectors in which Poland has comparative advantages). Impact on innovation of an economy depends on the sector; sometimes FDIs have positive influence (e.g. car industry) or there is no correlation (e.g. high tech). On the other hand, Weresa investigated portfolio investments which had positively affected the development of the Polish capital market (however a high level of foreign portfolio flows might create risks as well). Nevertheless FDIs have a bigger impact on the economy than portfolio investments.

The importance of the structure of capital flows – one of our main objects in this paper – was also brought up by Mitra [2011]. However, Mitra went a step further and tried to investigate flows distinguishing those that end up in the real estate sector. According to her research flows to this sector have a greater impact on GDP (surges and collapses) than other sectors. Nevertheless it shows that the destination and channel of capital flow matter, especially in the context of possible financial crises and swings in GDP.

2. Historical data overview At this point we would like to make a short overview of historical data associated with our research field. Based on Eurostat data we divided investments into the following groups: households, financial corporations, non-financial

[mln PLN, current prices]

136

Economics and Business Review, Vol. 1(15), No. 2, 2015 %

350 000 300 000

30 25

250 000

20

200 000 15 150 000 10

100 000

5

50 000

Households

Financial Corporations Central Govterment

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

0

Non-Financial Corporations Pub. Inv. Share (rhs)

Figure 2. Structure by investor of the gross fixed capital formation in Poland, 1995–2011 Source: Eurostat % 100 80 60 40 20 0

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

–20 –40 Households

Financial Corporations

Non-Financial Corporations

Central Govterment

Figure 3. Nominal growth rates of gross fixed capital formation made by the following groups, 1996–2011 Source: Eurostat

137

J. Krawczyk, S. Filipczak, Investments and long-term real interest rate in Poland

%

350 000

35

300 000

30

250 000

25

200 000

20

150 000

15

100 000

10

50 000

5

Domestic Capital

Foreign Capital

For. Cap. Share (rhs)

Figure 4. Structure of financial investment in Poland, 1995–2011 Source: Eurostat

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

0 1995

[mln PLN, current prices]

corporations and central government. Absolute data and the nominal growth of those groups are illustrated in Figure 2 and Figure 3.

Over the years the main source of investments in Poland were non-financial corporations (see Figure 2), however with a negative growth rate observed in the period from 2000 to 2004 and from 2008 to 2011 (see Figure 3). It coincided with the economic situation at that time. Investments by financial corporations were more volatile but their value was not that important for the economy (as illustrated in Figure 2). Much more stable were household investments which were mainly associated with the purchase of real estate. The government of the country made a significant contribution to investments as well (see Figure 2).

What is of interest in times of economic uncertainty public investments were characterized by a positive nominal growth rate which was probably an effect of an expansionary fiscal policy. Moreover, assuming that the current account balance equals investment and domestic savings in the economy (CA = S – I), so the current account deficit can be treated as foreign capital inflow (foreign investment) and the current account surplus can be treated as capital outflow and we define relationship of current account to investments as a share of investment financed by foreign investors. Given this assumption, Figure 4 illustrates the structure of financial investments in Poland over the years.

As illustrated in Figure 4 domestic capital was the main source of investment financing in Poland in recent years. This is of course not surprising tak-

138

Economics and Business Review, Vol. 1(15), No. 2, 2015

ing into account that liberalization of capital flows started in 90’s (as mentioned in Section 1) and therefore foreign capital started to gain importance around 1995. In longer perspective share of foreign capital amounted from ca. 13% to 30% of overall investments.

3. Analyses of investment’s relationship with the long-term interest rate One of the two goals of this paper is to look for connections between foreign capital flows and the long-term real interest rate in Poland. In our study we consider the real long-term interest rate as an approximation of cost of capital for the Polish economy. We would like to leave aside discussion about the best definition of a real interest rate. Taking into account that investments are characterized by their long-term nature, in our calculation we define the real rate as the yield of the tenyear Polish government bonds denominated in polish zloty adjusted by current consumer price index (CPI). Moreover the relationship of the current account to investments tells us what share of investments was financed by foreign capital. Considering that the level of interest rates in Poland is more attractive for investors from developed economies (because it is higher) it creates an inducement to invest in Poland. Clearly there are some additional risks associated with investment abroad and especially risks associated with investment in emerging market which should be taken into consid%

%

35

8

30

7 6

25

5

20

4 15

3

10

2

CA/I (lhs)

10Y real rate (rhs)

Figure 5. Fourth quarter averages of the current account relative to investments and the ten-year real rate, 4Q’2000–4Q’2012 Source: Eurostat, NBP, Bloomberg

4Q'12

2Q'12

4Q'11

2Q'11

4Q'10

2Q'10

4Q'09

2Q'09

4Q'08

2Q'08

4Q'07

2Q'07

4Q'06

2Q'06

4Q'05

2Q'05

4Q'04

2Q'04

4Q'03

2Q'03

4Q'02

2Q'02

0 4Q'01

0 2Q'01

1

4Q'00

5

J. Krawczyk, S. Filipczak, Investments and long-term real interest rate in Poland

139

eration. Under that assumption the level of interest rates should be influenced by the presence of foreign investors because they create additional demand for domestic investment and above all they are benchmarked to their local rates.

To test that hypothesis we collected quarterly data for the period 2000–2012.

To minimize the influence of seasonal factors we decided to work on fourth quarter average real rate (data shown in Figure 5).

In general, relation of current account to investments vary from 13% to 30%.

Interest rate, after 2002/2003, was in downward trend.

The results basically confirm our expectations and that which Janicka [2008] points out after Henry [2000] in terms of the decreasing cost of capital after stock market liberalization. The share of investments financed by foreign investors is negatively correlated with the long-term real rate, Pearson’s r coefficient is –0.61. Based on 49 observations we carried out regression analysis, that is statistically significant. P-value is 0.0000038 that is less than assumed level of significant α = 0.05 (Table 1).

Table 1. Regression summary output Regression Statistics Multiple R

0.607

R Square

0.368

Adjusted R Square

0.355

Standard Error

0.012

Observations

49

ANOVA df

SS

MS

F

Significance F

Regression

1

0.004

0.004

27.388

0.000

Residual

47

0.007

0.000

Total

Intercept CA/I

48

0.011

Coefficients

Standard Error

t Stat

P-value

Lower 95%

0.071

0.007

9.935

0.000

0.057

0.085

–0.235

–0.104

–0.170

0.032

–5.233

0.000

Upper 95%

It is worth mentioning that R square is 0.36 therefore there are some other factors determining the interest rate. This finding is consistent with Claessens, Oks, and Polastri conclusion that there is a positive link between interest rate differentials and capital flows. Additionally we decided to illustrate the path of data relations which can be helpful in identifying crucial moments (Figure 6).

%

10Y real interest rate

4Q'02 6

1Q'03

2Q'02

3Q'02

2Q'03 5

1Q'02

3Q'03

4Q'01 1Q'04 4Q'03 3Q'06

4

2Q'04

3Q'01

4Q'06

3Q'04

y = –0.1696x + 0.071

1Q'07

R² = 0.3682

2Q'06

3

1Q'06 4Q'05

2Q'01 2Q'10

2

3Q'07

3Q'10

1Q'10

4Q'09

2Q'07

4Q'04 4Q'10

1Q'01

4Q'07 1Q'11 2Q'11

1Q'08 1Q'09 2Q'08 3Q'11 2Q'09 4Q'11 4Q'08 4Q'00 3Q'08 2Q'12 1Q'05

3Q'05

2Q'05

3Q'09

1 4Q'12

1Q'12

3Q'12

%

0 10

15

20

25

30

CA/I

Figure 6. Path of relations of the current account relative to investments and the ten-year real interest rates, 4Q’2000–4Q’2012

10Y real interest rate

% 4Q'02 6

2Q'02 2Q'03

5

3Q'03

1Q'03 3Q'02 1Q'02

4Q'01

4

y = –0.3998x + 0.1179 R² = 0.8283

1Q'04

4Q'03

3Q'01

3 2Q'01 1Q'01

2

1 12

14

16

18

20

22

24

CA/I

Figure 7. Path of relations of the current account relative to investments and the ten-year real interest rate, 4Q’2000–1Q’2004

4Q'00 % 26

J. Krawczyk, S. Filipczak, Investments and long-term real interest rate in Poland

141

What is interesting is that since Poland’s accession to the European Union (2Q’04) the relationship started to change and became much more volatile (data in Figure 7 versus data in Figure 6) which might have been the result of the capital flow liberalization that is required in the EU accession process [Arvai 2005]. Figure 7 illustrates the path of the relationship up to 1Q’04, once Figure 6 gives broader perspective. The above conclusions give us a background for further analysis, however at this point it seems reasonable to sum up critical views about our results. First of all, the interest of foreign investors in capital allocation in emerging markets is partly affected by market sentiment in a broad sense. One may say that a negative sentiment is also the reason for the rising yield of government bonds and therefore data would be falsely correlated. Secondly, the assumption that the long-term real rate is considered a cost of capital for Polish economy it should be influenced mostly by the level of investment, regardless of its origin (domestic or foreign). Last but not least, our study makes use of relative data (the share of foreign financed investment to overall investment) but a perspective on absolute data would also be worthy of note. We opted for relative data to neutralize the effect of inflation and changes in the level of investment.

We next attempted to establish which channels of capital flow are most significant. The current account is financed mostly by the financial account but what is characteristic for Poland (and similar CEE economies) is that the capital account is crucial, mostly caused by European Union co-funding. On the basis of data provided by the National Bank of Poland, we classified specific channels into categories: capital account, net FDI, long-term capital flows, short-term capital flows, others, errors and omissions and change in official reserves. Additionally, long-term capital flows were subdivided into government debt flows and credit flows, whereas, short-term flows were divided into equity, government debt and credit. In order to identify the most important channels we used the Principal Component Analysis (PCA), which helps to limit the number of variables to the most important. PCA procedure starts with the creation of correlations matrix (Table 2). The next step requires the calculation of eigenvectors and eigenvalues (Table 3). To confirm calculation correctness Aw must be equal to λw (Table 4). At the end PCA tells us how much each factor contributes to the total variance of the data (Table 5). Results show that the capital account, net FDI and long-term capital flows are responsible for over 70% of the total variance. Long term credit, equity, short term debt and short term credit had definitely lower contribution. It is compatible with previous research, that short-term flow has a speculative character and should not be taken into consideration in the fundamental analysis (see Section 1).

Taking into account only long-term flows in relation to all investments (Figure 8) and repeating the above procedure we find out that that the correlations with the long-term real rate is still strong and statistically significant, r = –0.59 (Table 6). Based on 49 observations we carried out regression analy-

[142]

–0.041

–0.420

–0.151

Errors and omissions

Reserve

0.163

ST Gov. debt

Others

0.397

Equity

–0.151

0.376

LT Credit

ST Credit

0.178

–0.051

net FDI

LT Gov. debt

1.000

CapAc

CapAc

Correlation matrix

–0.189

–0.065

–0.156

–0.105

–0.178

–0.263

0.164

0.039

1.000

–0.051

net FDI

–0.498

–0.134

–0.167

–0.161

–0.260

0.235

–0.341

1.000

0.039

0.178

LT Gov. debt

Table 2. Correlation matrix of channels of capital flow

0.110

–0.500

0.068

0.113

0.216

–0.085

1.000

–0.341

0.164

0.376

LT Credit

0.027

–0.092

–0.259

–0.170

–0.104

1.000

–0.085

0.235

–0.263

0.397

Equity

0.123

–0.220

0.144

0.070

1.000

–0.104

0.216

–0.260

–0.178

0.163

ST Gov. debt

0.259

–0.004

–0.103

1.000

0.070

–0.170

0.113

–0.161

–0.105

–0.151

ST Credit

–0.547

–0.306

1.000

–0.103

0.144

–0.259

0.068

–0.167

–0.156

–0.041

Others

0.070

1.000

–0.306

–0.004

–0.220

–0.092

–0.500

–0.134

–0.065

–0.420

Errors and omissions

1.000

0.070

–0.547

0.259

0.123

0.027

0.110

–0.498

–0.189

–0.151

Reserve

[143]

0.2439

ST Gov. debt

–0.5624

–0.2467

Reserve

0.2897

Errors and omissions

Others

–0.1275

0.1308

Equity

ST Credit

0.4313

LT Credit

0.0198

net FDI

0.0750

0.5040

CapAc

LT Gov. debt

2.0846

Eigenvalues

CapAc

–0.4764

0.0803

0.0241

–0.3428

–0.3441

0.2468

–0.3702

0.5652

0.0690

0.0884

2.0082

net FDI

0.4053

–0.0109

–0.5640

0.0306

–0.0200

0.5750

0.0261

0.1127

–0.2257

0.3482

1.6559

LT Gov. debt

–0.0509

0.0890

0.3924

–0.0042

0.3203

0.1860

–0.3184

–0.0693

–0.7683

–0.0424

1.2805

LT Credit

0.1061

0.2483

0.0023

–0.8430

0.1536

0.0830

0.0279

–0.4059

0.1340

0.0499

0.9169

Equity

0.0474

–0.0561

0.3045

0.0143

–0.7860

0.2474

0.2494

–0.3486

–0.1913

–0.0486

0.7359

ST Gov. debt

Table 3. Eigenvectors and Eigenvalues of the correlation matrix of channels of capital flow

0.4144

–0.6559

0.0055

–0.3112

–0.0586

–0.0705

–0.1574

0.2261

–0.0702

–0.4603

0.5515

ST Credit

0.1045

0.1299

–0.1678

–0.2222

–0.2231

–0.6543

0.2267

0.2495

–0.4609

0.3011

0.4260

Others

0.2583

–0.1767

0.1720

0.0886

–0.1477

–0.2095

–0.6358

–0.2351

0.1918

0.5501

0.3035

Errors and omissions

–0.5314

–0.3551

–0.5413

–0.0201

–0.0516

–0.1082

–0.1772

–0.4499

–0.2120

–0.0703

0.0370

Reserve

[144]

0.495676

–0.691062

–0.688484

0.048355

0.161343

0.272733

0.508532

–0.265741

0.603821

–1.17231

0.161343

TRUE

TRUE

0.603821

–0.956727

0.048355

–0.265741

–0.51421

–0.688484

0.508532

–1.17231

0.495676

–0.691062

0.272733

1.135109

–0.74352

0.899187

0.138497

0.041367

0.156437

0.17752

1.050754

λw

–0.956727

–0.033142

–0.74352

0.899187

–0.51421

0.952087

1.135109

0.156437

TRUE

0.671122

–0.018007

–0.933892

0.050632

–0.033142

0.952087

0.04327

0.186535

–0.373697

0.576639

0.671122

–0.018007

–0.933892

0.050632

0.04327

0.186535

–0.373697

0.138497

0.041367

0.576639

0.17752

1.050754

Aw

Table 4. Testing if Aw = λw

TRUE

–0.065235

0.114006

0.502439

–0.005392

0.41017

0.238203

–0.407705

–0.088688

–0.983813

–0.054329

–0.065235

0.114006

0.502439

–0.005392

0.41017

0.238203

–0.407705

–0.088688

–0.983813

–0.054329

TRUE

0.097247

0.227694

0.002123

–0.772906

0.140798

0.076105

0.025554

–0.372201

0.122821

0.045747

0.097247

0.227694

0.002123

–0.772906

0.140798

0.076105

0.025554

–0.372201

0.122821

0.045747

TRUE

0.034889

–0.041271

0.224093

0.010532

–0.57843

0.182065

0.183556

–0.256545

–0.140752

–0.035742

0.034889

–0.041271

0.224093

0.010532

–0.57843

0.182065

0.183556

–0.256545

–0.140752

–0.035742

TRUE

0.228547

–0.361698

0.003042

–0.171604

–0.032336

–0.038891

–0.086806

0.12469

–0.03872

–0.253855

0.228547

–0.361698

0.003042

–0.171604

–0.032336

–0.038891

–0.086806

0.12469

–0.03872

–0.253855

TRUE

0.044527

0.055353

–0.071496

–0.094655

–0.095062

–0.278732

0.096592

0.106301

–0.196349

0.128293

0.044527

0.055353

–0.071497

–0.094655

–0.095062

–0.278732

0.096592

0.106301

–0.196349

0.128293

TRUE

0.0784

–0.053648

0.052212

0.02689

–0.044821

–0.063598

–0.192988

–0.071354

0.058205

0.166991

0.0784

–0.053648

0.052212

0.02689

–0.044821

–0.063598

–0.192988

–0.071354

0.058205

0.166991

TRUE

–0.019661

–0.013138

–0.020024

–0.000742

–0.001908

–0.004002

–0.006556

–0.016646

–0.007845

–0.002601

–0.019661

–0.013138

–0.020024

–0.000742

–0.001908

–0.004002

–0.006556

–0.016646

–0.007845

–0.002601

145

J. Krawczyk, S. Filipczak, Investments and long-term real interest rate in Poland

Table 5. Share of total variance Factor

Eigenvalue

[%]

[Σ %]

CapAc

2.084642

20.85

20.85

net FDI

2.008222

20.08

40.93

LT Gov. debt

1.655852

16.56

57.49

LT Credit

1.280483

12.80

70.29

Equity

0.916881

9.17

79.46

ST Gov. debt

0.735918

7.36

86.82

ST Credit

0.551451

5.51

92.33

Others

0.426014

4.26

96.59

Errors and omissions

0.303541

3.04

99.63

Reserve

0.036996

0.37

100.00

%

% 8

60

7

50

6 40

5 4

30

3

20

2 10

1

LT flows/I

4Q'12

2Q'12

4Q'11

2Q'11

4Q'10

2Q'10

4Q'09

2Q'09

4Q'08

2Q'08

4Q'07

2Q'07

4Q'06

2Q'06

4Q'05

2Q'05

4Q'04

2Q'04

4Q'03

2Q'03

4Q'02

2Q'02

4Q'01

2Q'01

0 4Q'00

0

10Y real rate (rhs)

Figure 8. Fourth quarter averages of long-term flows relative to investments and the ten-year real rate, 4Q’2000–4Q’2012

sis, that is statistically significant. P-value is 0.000010 that is less than assumed level of significant α = 0.05, R square amounted to 0.3435. We think that this analysis can be useful for further studies of the foreign capital flows influence on the domestic economy.

146

Economics and Business Review, Vol. 1(15), No. 2, 2015

Table 6. Regression summary output Regression Statistics Multiple R

0.5861

R Square

0.3435

Adjusted R Square

0.3295

Standard Error

0.0122

Observations

49

ANOVA df

SS

MS

F

Significance F

24.587

0.000

Regression

1

0.004

0.004

Residual

47

0.007

0.000

Total

48

0.011

Coefficients

Standard Error

t Stat

P-value

Lower 95%

Upper 95%

Intercept

0.062

0.006

10.735

0.000

0.050

0.074

LT flows/I

–0.087

0.018

–4.959

0.000

–0.122

–0.052

Conclusions A short overview of historical statistics for investment in Poland shows us that the most important source of them were companies whose activity coincided with the economic situation. On the other hand public investment were characterized by a positive nominal growth rate at the time of economic uncertainty. Households investment were much more stable but were mainly associated with the purchase of real estate. This paper aimed to identify crucial channels of foreign capital flow to Poland and find out whether co-financing by foreign investment influences the domestic cost of capital in Poland. To test our hypothesis we analyzed empirical data for the period 2000–2012. Based on our analysis we state that the cost of capital, measured as a long-term real interest rate, is influenced by foreign capital flows. Share of investments financed by foreign investors is negatively correlated with the long-term real rate (r = –0.61).

In our opinion, the reason for it is that foreign investors are willing to accept lower interest rate because it is still higher than on developed markets. Moreover, the results of the Principal Component Analysis suggest that the most important channels of capital flow are capital account (mainly effect of co-funding investment by the EU), net FDI and long-term flows. They are responsible for over 70% of the total variance. It provides background for further analyses.

J. Krawczyk, S. Filipczak, Investments and long-term real interest rate in Poland

147

One might carry out similar research in respect of a wider range of countries, especially from Eastern Europe. Another question which our paper raises is how the level of disparity of the interest rate between developed and emerging economies influences the domestic cost of capital.

References Ancypowicz, G., 2009, Wpływ bezpośrednich inwestycji zagranicznych na wzrost polskiej gospodarki w okresie poakcesyjnym, Główny Urząd Statystyczny.

Arvai, Z., 2005, Capital Account Liberalization, Capital Flow Patterns, and Policy Responses in the EU’s New Member States, IMF Working Paper, no. 213.

Bijsterbosch, M., Kolasa, M., 2009, FDI and Productivity Convergence in Central and Eastern Europe: An Industry-level Investigation, ECB Working Paper Series, no. 992.

Calvo, G.A., Leiderman, L., Reinhart, C., 1993, Capital Inflows and Real Exchange Rate Appreciation in Latin America: The Role of External Factors, IMF Staff Paper, no. 40.

Carlson, M., Hernandez, L., 2002, Determinants and Repercussions of the Composition of Capital Inflows, Board of Governors of the Federal Reserve System International Finance Discussion Paper, no. 717.

Claessens, S., Oks, D., Polastri, R., 2000, Capital Flows to Central and Eastern Europe and the Former Soviet Union, National Bureau of Economic Research (Capital Flows and the Emerging Economies: Theory, Evidence and Controversies): 299.

Culha, A.A., 2006, A Structural VAR Analysis of the Determinants of Capital Flows into Turkey, Central Bank Review, no. 2: 11–35.

Dunning, J.H., 1993, Multinational Enterprises and the Global Economy, AddisonWesley Publication Company, Harlow, Essex.

Garibaldi, P., Mora, N., Sahay, R., Zettelmeyer, J., 2001, What Moves Capital to Transition Economies?, IMF Staff Papers, vol. 48: 109–145.

Gorynia, M., Nowak, J., Wolniak, R., 2007, Motives and Modes of FDI in Poland: An Exploratory Qualitative Study, Journal for East European Management Studies, vol. 12, no. 2: 132–151.

Henry, P.B., 2000, Do Stock Market Liberalization Cause Investment Booms?, Journal of Financial Economics, vol. 58, no. 1–2: 301–334.

Hernandez, L., Mellado, P., Valdes, R., 2001, Determinants of Private Capital Flows in the 1970s and 1990s: Is There Evidence of Contagion?, IMF Working Paper, no. 64.

Janicka, M., 2008, Liberalizacja przepływów kapitałowych – korzyści i zagrożenia, Bank i Kredyt, vol. 39, no. 3: 34–49.

Mitra, P., 2011, Capital Flows to EU New Member States: Does Sector Destination Matter?, IMF Working Paper, no. 11/67.

Ralhan, M., 2006, Determinants of Capital Flows: A Cross Country Analysis, Econometrics Working Papers, no. 0601, Department of Economics, University of Victoria.

Rydarowska-Kurzbauer, J., 2012, GDP Fluctuation and Changes in Consumption and Investment. Comparative Analysis of Polish Economy and Some Countries of the European Union, Equilibrium – Quarterly Journal of Economics and Economic Policy, vol. 7, iss. 3: 52.

148

Economics and Business Review, Vol. 1(15), No. 2, 2015

Schmitz, M., 2009, Financial Reforms and Capital Flows to Emerging Europe, The Institute for International Integration Studies, IIIS Discussion Paper, no. 278.

Sobański, K., 2010, Empiryczne modele oceny stabilności deficytu obrotów bieżących w krajach rozwijających się i ich aplikacja w krajach Europy Środkowo-Wschodniej, in: Najlepszy, E., Sobański, K. (red.), Niestabilność równowagi zewnętrznej krajów rozwijających się, PWE, Warszawa.

Stiglitz, J.E., 2000, Capital Market Liberalization, Economic Growth, and Instability, World Development, vol. 28, no. 6: 1075–1086. Śliwiński, P., 2011, Przepływy kapitału międzynarodowego a wzrost gospodarczy w krajach Europy Środkowo-Wschodniej w latach 1994–2008, Wydawnictwo Uniwersytetu Ekonomicznego w Poznaniu, Poznań.

Trąpczyński, P., 2014, Determinants of Foreign Direct Investment Performance in the Internationalisation Process of Polish Companies, Doctoral dissertation, Poznań University of Economics, Poznań, http://www.wbc.poznan.pl/dlibra/docmetadata? id=325673&from=publication [access: 31.10.2014].

Weresa, M.A., 2002, Skutki inwestycji zagranicznych dla gospodarki kraju przyjmującego – doświadczenia Polski, Zeszyty BRE Bank – Centrum Analiz Społeczno-Ekonomicznych, Fundacja Naukowa, nr 62.

Economics and Business Review, Vol. 1 (15), No. 1, 2015: 149–150

BOOK REVIEWS

Paweł Marszałek, Systemy pieniężne wolnej bankowości. Koncepcje cechy, zastosowanie [Free Banking Monetary Systems. Concepts, Characteristics, Application], Wydawnictwo Uniwersytetu Ekonomicznego w Poznaniu, Poznań 2014: 342, ISBN 978-83-7417-827-3 The monograph by P. Marszałek is an extensive study of the topic. The analysis devoted to evolution of monetary systems, with particular reference to free banking monetary systems, undoubtedly demonstrates a complex, multi-aspect, profound and critical, scientific approach, It will be extremely valuable for both financial theory and practice.

The main objective of the author is an in-depth study of the „characteristics of contemporary concepts of free banking and an assessment of the possibility for their possible practical application”. The hypothesis is explored as to whether “free banking systems are not an alternative to central banking systems”.

The monograph is divided into two parts. Chapters 1–3 are a systematic overview of historical issues and problems, combining a well-balanced approach to both theoretical and empirical aspects.

Chapters 4–6 are devoted to the specific concepts of free banking perceived by the author as an alternative to the present institutional arrangements. The reader can readily follow the theoretical development, the events, the nature and the processes and future trends in monetary practice.

The contents of the first chapter are a collection of determinants characterising a monetary system and factors which

shape it, as well as the classification of different monetary systems throughout history. It also presents the premises and stages of transition to monetary systems of fiduciary money. From this the reader will receive a clear picture of the essence, structure and changes that the monetary systems have undergone as well as the impact of various factors and processes upon monetary systems which have shaped their form and development.

The second chapter deals with the author’s assertion that “it is necessary to set in order and expand on the theoretical issues, merely indicated so far… it is necessary to determine what concepts constitute the theoretical background for the functioning of contemporary monetary systems… It has become even more necessary because the transition to the systems based upon independent money was accompanied by significant shifts of values in the theory of macroeconomics and economic policy”. The focus is on the concepts of K. Wicksell and J.M. Keynes, which the author regards as foundations of the theory of contemporary money.

He also presents the evolution of the approach to money and monetary policy in mainstream economics after 1970 and the basics of the so-called “New Neoclassical Synthesis”.

150

Economics and Business Review, Vol. 1(15), No. 2, 2015

This part examines the implications of the New Neoclassical Synthesis for the functioning of monetary systems and for the shaping of the “New Consensus” in the implementation of monetary policy.

The findings, based on the analysis of the elements of this consensus (amongst others, inflation targeting, relations between price stability and financial stability) and the conclusions he formulates in respect of the “New Consensus Monetary Policy”, create a systematic picture of new problems which emerged in the course of the evolution of monetary systems.

The author acknowledges that the list of issues which would then facilitate the analysis of modern free banking is incomplete, so he devotes a subsequent chapter to selected conditions for the functioning of contemporary monetary systems (and their consequences), such as the globalization of financial activities, changes in banks’ activities, financialization, changes in the forms and modes of using money and, as a result of other phenomena – the so-called monetary disorder, being the main consequence of changing circumstances. The author and the literature cited give valuable insights into factors of a global character and the microeconomic factors relating to the functioning of banks.

The final part of this chapter is devoted to criticism of the theoretical bases on the functioning of contemporary monetary systems. These concern not only the current state of affairs but also constitute relevant reasons for the subsequent reflections. These conclusions led the author to devote the subsequent three chapters to the issues concerning „features, construction and principles of the functioning of modern free banking concepts and – to the assessment of the possibilities of their practical application”.

The next three chapters deal with one of the three variants of modern free banking: Selgin-White’s and Rothbard-Mises’ models, Friedrich Hayek’s denationalization of money and the New Monetary Economics enabling a comparison of the concepts presented.

The monograph enriches the list of supportive arguments for the hypothesis.

The author shows great caution in his conclusions which closely correspond with the results found in related literature.

The final part of the monograph is devoted to an overall analysis of the research procedures which verify the hypothesis that individual modern free banking systems, although cohesive in their structure, are not an alternative to central banking systems. The author states clearly that “restructuring of monetary systems on the basis of free banking concepts would not be the best solution, even if it were possible due to political reasons”. He also points at theories of heterodox economists, particularly the concepts formulated by the Post-Keynesian School and so-called New Economic Sociology, as more suitable for analysis of changes in the contemporary monetary systems.

The monograph by P. Marszałek is characterized by a high substantive and methodological level of approach, the wide use of related literature and consideration of scientific trends, theories and concepts, taking other opinions into account, yet clearly stating his own position in seeking to verify the hypothesis. The monograph is a significant contribution to monetary theory and practice, especially with reference to banking.

Bogusław Pietrzak Warsaw School of Economics Collegium of Socio-Economics

Economics and Business Review, Vol. 1 (15), No. 2, 2015: 151–153

Ewa Mińska-Struzik, Od eksportu do innowacji. Uczenie się przez eksport polskich przedsiębiorców [From Export to Innovation – Learning by Exporting in Polish Enterprises], Difin, Warszawa 2014: 284, ISBN 978-83-7930-361-8 The fundamental factor of growth at the present stage of the globalization of the world economy (semi-globalization) is scientific-technological progress and, within its frameworks, innovation. On the basis of crucial inventions modern branches of industry and services were developed and their share in the economy of a given country, including exports, improves its competitiveness and welfare.

Awareness of this is widespread but the possibility of implementing the concept of a “knowledge-based economy” encounters huge barriers from various sources. This concerns not only developing countries in particular, but also many well developed countries, which is proved, e.g. by a failure in the implementation of the European Union’s Lisbon Strategy. These problems are particularly acute in Poland’s economy which, due to its exceptional weakness in the sphere of innovative development, is threatened by a fall into the middle-income growth trap, thus ruining its chance to catch up with the highly developed EU countries. Poland as is described by the author of the monograph reviewed has a distant position in the rankings of innovation and its position in Europe is one of the last.

Due to this fact one should welcome research works whose aim is to explain the conditions, state, and above all, ways to solve these key problems for Poland. The monograph reviewed conforms to this scientific trend and I regard the choice of its subject matter as relevant and useful, not only for scientific but also for practical reasons. It offers valuable recommendations concerning the sphere of innovativeness for both politicians and entrepreneurs.

The monograph itself is of an innovative character as regards the subject of research in the Polish literature. There are many publications about importing as a channel of innovation transfer, but exporting still awaits proper research and publications. Meanwhile studies on the mechanism and consequences of learning by exporting (LBE) have dynamically developed abroad for over two decades. The author has made an attempt to bridge the existing research gap. She gives a positive answer to the question as to whether LBE effects occur in Polish enterprises but nonetheless this confirmation is limited by a number of additional conditions. Her findings mainly concern a fairly narrow group of high-tech enterprises. Although this is an important group, the symptoms of learning by exporting observed cannot be simply generalized to cover all of Poland’s industrial enterprises. Since the phenomenon studied is complex, difficult to grasp and separate from other relationships, there is a probability that the results of potential subsequent studies may be different.

Therefore the subtitle of the book may also be too broad.

Despite these limitations, the book is quite interesting from the methodological point of view. The main goal of the work which is (p. 12) showing conditions and implications of learning by exporting and the verification of its occurrence in Polish manufacturing industry, with special reference to high-tech enterprises was clearly formulated. Apart from this the author constructed a series of relevant research steps which should be treated as auxiliary aims.

152

Economics and Business Review, Vol. 1(15), No. 2, 2015

The fundamental thesis of the work is as follows (p. 13): learning by exporting is not an automatic process. It is determined by a bundle of factors in each enterprise and its meso- and macroeconomic environment, so it is impossible to predetermine the final learning results in the form of the implementation of innovation and productivity growth. Moreover three hypotheses were formulated, namely: – H1: between enterprises at different levels of technological advancement there exist differences as regards the significance of exporting in the process of implementing innovation; – H2: conducting export in certain defined conditions stimulates high tech enterprises to implement innovation and increase productivity; – H3: learning begins along with the start of interaction with the foreign market entities.

The author divided the procedure she devised into three stages and each one of them was used to verify one of the previously formulated hypotheses. Doing this, she used a system of combined research methods (triangulation), due to the complexity and unclear nature of learning by exporting.

To determine the significance of exporting for innovative activities of enterprises with a different degree of technological advancement an econometric study was carried out using individual data from the Central Statistical Office (CIS 2008), comprising about 8.5 thousand manufacturing enterprises and the model of logistic regression. This stage of the research procedure showed that exporting stimulates implementation of innovation (above all product innovation) in all groups of enterprises but the strongest relation between the variables appeared in high-tech entities.

The second hypothesis was verified by the author’s nation-wide survey where the

results were interpreted using methods of descriptive statistics, inductive statistics and independence analysis. On the basis of interviews conducted at the beginning of 2012 with competent representatives of 200 high-tech enterprises the author confirmed the existence of a link between exporting and innovation of a non-determinist nature and she showed which conditions stimulate LBE in this group.

The third hypothesis was verified at the final stage and was based on a multiple case study of enterprises preparing themselves for starting exporting. For this purpose the author designated seven firms for participation in the study and she observed them for nearly two years after the survey was completed. She used such research techniques as partly structured in-depth interviews, non-participant observation, analysis of documentation and the contents of the websites of the enterprises’ investigated. The author returned many times to the persons she had interviewed in order to supplement information and determine more precisely the interdependencies she had assumed. The collected material did not confirm the third hypothesis, so she came to a conclusion that in itself the commencement of interaction with foreign markets is not sufficient to initiate the mechanism of learning by exporting.

The construction of the book is logical and subordinated to the main objective.

In accordance with general methodological principles the author passes from general and theoretical problems to detailed issues based upon empirical studies. The monograph consists of introduction, five chapters, concluding remarks, bibliography, and a list of tables and figures. The first chapter „International trade as a source of innovation” ends with an original conceptualization of the notion of “learning by exporting”. “A theoretical basis of learning by

Book reviews

exporting”, which is the subject of chapter two, was extensively described, taking into account several approaches in the theory of foreign trade, the theory of economic growth, the theory of organizational learning and the theory of internationalization.

Chapter three „Learning by exporting in the light of empirical studies” contains a critical analysis and a synthesis of the results of over ninety research works concerning the mechanism and consequences of learning by exporting. There the author’s added value consists of constructing a paradigm of learning by exporting and devising a research plan to identify the effects of learning by exporting in Polish enterprises, with special reference to high-tech firms.

The final two chapters, are of a strictly empirical character. They are devoted to the presentation of studies conducted on the impact of exports upon innovativeness of Polish enterprises and the verification of the previously postulated research hypotheses.

In chapter four entitled „The significance of exporting for innovation in Polish manufacturing enterprises” the author, used the tools of descriptive statistics and regression analysis, to present the modes of behaviour of the Polish companies, tak-

153 ing into account their branch classification, the level of technological advancement, and the division into exporters and non-exporters. Chapter five „Learning by exporting in Polish high-tech enterprises” is crucial to the whole book because it contains the results of basic studies conducted by the author, i.e. surveys and multiple case studies. The aim of the surveys was to diagnose the effects of learning by exporting and the spillover of those effects into non-exporting enterprises. The multiple case studies were aimed at establishing the moment when learning by exporting was initiated.

In my opinion the reviewed monograph is of a high scientific and cognitive value. It starts filling the gap in Polish literature on the relationship between exporting and innovative activity. The author highlights an underestimated channel of innovation transfer and potential benefits of its effective exploitation stimulated by appropriate activities at both macro- (policy) and micro- (enterprise) levels. This is all the most important because for us innovation means „to be or not to be”!

Jan Rymarczyk Wrocław University of Economics Department of International Economics

Aims and Scope Aims and Scope

Economics and Business Economics Review andisBusiness the successor Review toisthe thePoznań successor University to the Poznań of Economics University Review of Economics which wasReview which was published by the Poznań published University by the Poznań of Economics University Press ofin Economics 2001–2014.

Press TheinEconomics 2001–2014. and ThBusiness e Economics Review and Business Review is a quarterly journal is a quarterly focusing on journal theoretical focusing andon applied theoretical research andwork applied in the research fields work of economics, in the fields manof economics, management and finance. agement The Review and finance. welcomes The Review the submission welcomes of the articles submission for publication of articles dealing for publication with micro, dealing with micro, mezzo and macro mezzo issues. and All macro texts are issues. double-blind All textsassessed are double-blind by independent assessedreviewers by independent prior to reviewers acceptance. prior to acceptance.

Notes for Contributors Notes for Contributors

1. Articles submitted 1. Articles for publication submittedinfor thepublication Economicsinand theBusiness Economics Review and Business should contain Revieworiginal, should contain original, unpublished workunpublished not submitted work for not publication submitted elsewhere. for publication elsewhere.

2. Manuscripts intended 2. Manuscripts for publication intended should for publication be writtenshould in English be written and edited in English in Word andand edited sentin to:Word and sent to: review@ue.poznan.pl. review@ue.poznan.pl.

Authors should upload Authors twoshould versions upload of their twomanuscript. versions of One theirshould manuscript. be a comOne should be a complete text, while inplete the second text, while all document in the second information all document identifying information the author(s) identifying should thebe author(s) removed should be removed from files to allowfrom themfito lesbe tosent allowtothem anonymous to be sent referees. to anonymous referees.

3. The manuscripts 3. are Thetomanuscripts be typewritten arein to12’ be typewritten font in A4 paper in 12’ format font inand A4 be paper left-aligned. format and Pages be left should -aligned. Pages should be numbered. be numbered.

4. The papers submitted 4. The should papers submitted have an abstract shouldofhave not an more abstract than 100 of not words, morekeywords than 100and words, the keywords Journal and the Journal of Economic Literature of Economic classification Literature code.classification code.

5. Acknowledgements 5. Acknowledgements and references to grants, and references affiliation, to postal grants,and affiliation, e-mail addresses, postal andetc. e-mail should addresses, appearetc. should appear b, etc a, b, etc as a separate footnote as a to separate the author’s footnote name toa,the author’s name and should not be and included should innot the be main included list of footnotes. in the main list of footnotes.

6. Footnotes should 6. Footnotes be listed consecutively should be listed throughout consecutively the text throughout in Arabicthe numerals. text in Arabic Cross-references numerals. Cross-references should refer to particular should refer section to particular numbers: e.g.: section Seenumbers: Section 1.4. e.g.: See Section 1.4.

7. Quoted texts of7.more Quoted thantexts 40 words of more should thanbe 40separated words should frombe theseparated main body from by the a four-spaced main bodyindenby a four-spaced indentation of the margin tation as aof block. the margin as a block.

8. Mathematical notations 8. Mathematical should meet notations the following should meet guidelines: the following guidelines: – symbols representing – symbols variables representing should be variables italicized, should be italicized, – avoid symbols above – avoid letters symbols and use above acceptable letters and alternatives use acceptable (Y*) where alternatives possible, (Y*) where possible, – where mathematical – where formulae mathematical are set out formulae and numbered are set out these andnumbers numbered should thesebe numbers placed against should be placed against the right margin as... the (1), right margin as... (1), – before submitting – before the final submitting manuscript, the ficheck nal manuscript, the layout of check all mathematical the layout of formulae all mathematical carefullyformulae carefully (including alignments, (including centring alignments, length of centring fraction lines lengthand of fraction type, sizelines and and closure type,ofsize brackets, and closure etc.), of brackets, etc.), – where it would –assist where referees it would authors assistshould refereesprovide authorssupplementary should provide mathematical supplementary notes mathematical on the notes on the derivation of equations. derivation of equations.

9. References in the 9. text References should be in the indicated text should by thebeauthor’s indicated name, by the date author’s of publication name, date andofthe publication page num-and the page number where appropriate, ber where e.g. Acemoglu appropriate, ande.g.

Robinson Acemoglu [2012], and Robinson Hicks [1965a, [2012], 1965b].

HicksReferences [1965a, 1965b]. shouldReferences should be listed at the endbeoflisted the article at the in end theofstyle the article of the in following the styleexamples: of the following examples: Acemoglu, D., Robinson, Acemoglu, J.A.,D., 2012, Robinson, Why Nations J.A., 2012, Fail.Why The Origins Nations of Fail.

Power, The Prosperity Origins of Power, and Poverty, Prosperity and Poverty, Profile Books, London.

Profile Books, London.

Kalecki, M., 1943, Kalecki, Political M., Aspects 1943, ofPolitical Full Employment, Aspects of Th Full e Political Employment, Quarterly, The Political vol. XIV,Quarterly, no. 4: 322–331. vol. XIV, no. 4: 322–331.

Simon, H.A., 1976,Simon, From Substantive H.A., 1976,toFrom Procedural Substantive Rationality, to Procedural in: Latsis, Rationality, S.J. (ed.),in: Method Latsis,and S.J. Appraisal (ed.), Method and Appraisal in Economics, Cambridge in Economics, University Cambridge Press, Cambridge: University Press, 15–30.

Cambridge: 15–30.

10. Copyrights will 10.beCopyrights establishedwill in the be established name of theinE&BR the name publisher, of the namely E&BR publisher, the Poznań namely University the Poznań of University of Economics Press. Economics Press.

More informationMore and advice information on the and suitability advice and on the formats suitability of manuscripts and formats canofbe manuscripts obtained from: can be obtained from: Economics and Business Economics Review and Business Review al. Niepodległościal.

10Niepodległości 10 61-875 Poznań 61-875 Poznań Poland Poland e-mail: review@ue.poznan.pl e-mail: review@ue.poznan.pl www.puereview.ue.poznan.pl www.puereview.ue.poznan.pl

Subscription Economics and Business Review (E&BR) is published quarterly and is the successor to the Poznań University of Economics Review. The E&BR is published by the Poznań University of Economics Press.

E&BR is listed in ProQuest, EBSCO, and BazEkon.

Subscription rates for the print version of the E&BR: institutions: 1 year – €50.00; individuals: 1 year – €25.00. Single copies: institutions – €15.00; individuals – €10.00. The E&BR on-line edition is free of charge.

Correspondence with regard to subscriptions should be addressed to: Księgarnia Uniwersytetu Ekonomicznego w Poznaniu, ul. Powstańców Wielkopolskich 16, 61-895 Poznań, Poland, fax: +48 61 8543147; e-mail: info@ksiegarnia-ue.pl.

Payments for subscriptions or single copies should be made in Euros to Księgarnia Uniwersytetu Ekonomicznego w Poznaniu by bank transfer to account No.: 96 1090 1476 0000 0000 4703 1245.