POLITICAL RISK IN EMERGING MARKETS

Rasmus Lund Breddam

MSc. Economics and Business Administration – International Business

POLITICAL RISK IN EMERGING MARKETS

- A study into the mitigation strategies of multinational corporations

Copenhagen Business School 2016

Master thesis

Hand in: 17

of May, 2016 STU: 169.163

Supervisor: Torben Juul Andersen

1

Abstract

Political risk is a factor multinational companies need to consider when operating in emerging markets. There host governments can with the swing of a pen transform a multimillion dollar investment into a huge loss. With the opening of more emerging markets to foreign direct investment, the traditional approach of political risk studies to focus on the bargain over entry becomes obsolete. As a result of this, this thesis seeks to investigate the mitigation strategies of multinational corporations in emerging markets post-entry.

In illuminating this issue, this thesis firstly draws upon the literature on political risk to form a basis of understanding from which the following analysis draws. Secondly, this thesis presents a multiple case study of the experiences of two large Danish companies, Novo Nordisk, and Arla Foods, to illustrate possible strategies for mitigating political risk in emerging markets. Lastly, an expert interview is presented to extract new knowledge as well as validate the findings of the case studies.

This thesis concludes that companies are able to manage political risk. Local partnerships and engagement with the host society are valid strategies to decrease one’s political risk exposure.

One cavity with the conclusion is the preciseness of managing any risk until it materializes, i.e. if the risk does not materialize what you did was either correct or unrelated, but if the risk does materialize, your actions were either wrong or they soften the blow.

1. WHY POLITICAL RISK MATTERS 5

1.1PROBLEM STATEMENT 7

1.2STRUCTURE OF THESIS 8

2. METHODOLOGY 9

2.1RESEARCH PHILOSOPHY 9

2.2RESEARCH APPROACH 10

2.3RESEARCH DESIGN 11

2.3.1CRITICALLY REVIEWING THE LITERATURE 11

2.3.2CONFIRMATORY CASE STUDY 11

2.3.3SELECTION OF CASES 12

2.3.4COMPARATIVE METHOD 13

2.3.5EXPERT INTERVIEW 14

2.4DATA CHOICE 14

2.5TIME ASPECT OF STUDY 15

2.6VALIDITY AND RELIABILITY 15

2.6.1VALIDITY OF STUDY 15

2.6.2RELIABILITY 16

2.7DEFINITIONS 16

2.7.1POLITICAL RISK 16

2.7.2EMERGING MARKETS 16

3. POLITICAL RISK IN THE LITERATURE 17

3.1RISK MANAGEMENT 18

3.2THE BEGINNING OF POLITICAL RISK: THE CONFLICTUAL SCHOOL 20

3.3THE MACRO LEVEL RISK; THE INSTITUTIONALIST 22

3.4THE RISE OF THE ENVIRONMENTAL SCHOOL 25

3.5SUMMARY 29

3

4. MODELS FOR POLITICAL RISK MANAGEMENT 31

4.1THE LEGITIMACY-BASED VIEW OF POLITICAL RISK 32

4.2THE POLITICAL BARGAINING MODEL 35

4.3COMPLEMENTARITY AND APPLICATION 38

5. CASE 1: INSULIN FOR BANGLADESH 39

5.1BANGLADESHI SITUATION 40

5.2NOVO NORDISK 42

5.3THE POWER OF THE BANGLADESHI STATE THROUGH THE LENSES OF THE POLITICAL BARGAINING MODEL 44

5.3.1GOALS AND OBJECTIVES 45

5.3.2RESOURCES AND CONSTRAINTS 45

5.3.3BARGAINING POWER 47

5.4GAINING LEGITIMACY THROUGH SOCIAL PARTNERSHIPS 47

5.4.1PRAGMATIC PERCEPTION ON DANISH PHARMA 47

5.4.2MORALITY IN A PLACE CALLED BANGLADESH 49

5.4.3COGNITIVE OBSERVATIONS 50

5.4.4IS THE STATE MOTIVATED TO INTERVENE? 51

5.5SUB-CONCLUSION 52

6. CASE 2: NO MILK FOR THE RUSSIAN BEAR 52

6.1THE RUSSIAN CONTEXT 54

6.2ARLA FOODS 56

6.3BARGAINING WITH THE KREMLIN 58

6.3.1GOALS AND OBJECTIVES 58

6.3.2RESOURCES AND CONSTRAINTS 59

6.3.3BARGAINING POWER 60

6.4LEGITIMIZING AN EMBARGO 61

6.4.1ARLA’S LEGITIMACY 61

6.4.2DENMARK’S LEGITIMACY 62

6.4.3RUSSIAN MOTIVATION 63

6.5SUB-CONCLUSION 64

7. CASE COMPARISON 65

8. EXERTS FROM AN EXPERT 68

9. DISCUSSION 70

9.1THE LEGITIMACY-BASED VIEW REFINED 70

9.2MARKET STRUCTURE AND POLITICAL RISK – SUGGESTED CORRELATION AND POSSIBLE TOPIC FOR FUTURE RESEARCH 72

10. CONCLUSION 72

REFERENCES 75

APPENDIX 87

APPENDIX 1:INTERVIEW GUIDE FOR INTERVIEW WITH SØREN ROBENHAGEN 87 APPENDIX 2:INTERVIEW CONDUCTED WITH SØREN ROBENHAGEN ON MAY 9^TH2016 87

5 1. Why Political Risk Matters

“In no area is political risk more relevant than in analysis of globalization and the rapid growth of emerging markets.” (Bremmer & Keat, 2009, p. 68)

Since the turn of the millennia, a lot of attention has been paid to emerging markets. Especially the grouping of four large fast growing economies, commonly known as the BRIC economies, namely Brazil, Russia, India, and China, have gained significant attention in the academic literature (Biggemann & Fam, 2011; Goldstein, 2013; Holtbrügge & Baron, 2013). These economies have become a prototype for understanding the diverse field of countries labeled emerging markets. The attention is well deserved.

Since 2005, the BRIC countries share of the world foreign direct investment inflow has increased from 9.4% to 28.4% in 2014 (World Bank, 2016). But it is not just companies from developed economies that have benefitted from moving into these fast-growing markets. Simultaneously, the number of large companies from these four countries have become a bigger force on the global business scene, illustrated by the increase of companies from BRIC countries’ listed on Fortune’s Global 500 list; in 2005 they held 5.2% of the list, today they hold 23.6% (Fortune, 2016). With more than 40% of the world’s population and a growing middle class, these markets must be part of the strategic discussions in most multinational companies.

Yet, the environment that companies face in emerging markets varies greatly from what they have been accustomed to in developed economies. On March 4^th, 2016, Luiz Inácio Lula da Silva, Brazil’s former president from 2003-10 and the mentor of its current one, Dilma Rousseff, was held for questioning by the Brazilian police concerning corruption charges (The Economist, 2016c). Two years earlier, in March 2014, Russia annexed the Ukrainian peninsula of Crimea, which resulted in economic, political and financial sanctions from the West. These sanctions, combined with a currency halved since the beginning of the crisis, have placed the Russian economy in a dire condition, not to talk about the affect it has had on foreign companies operating there (The Economist, 2016b). India, possibly the politically most stable of the four, also pose a troubling environment, arising from an unclear tax code, and in the future, the world’s largest, but fragmented, market (The Economist, 2016a). Lastly, China has begun aggravating its neighbors by claiming disputed islands in the South China Sea, not to address the

elephant in the room, the possible consequences of China’s slowing economy on the world economy (Gordon, 2016). As mentioned in the previous paragraph, the emerging markets, here exemplified by the BRIC economies, are showing great potential and has become a global factor. However, these economies are also prime examples of unstable political environments, which pose great dangers on multinational company’s (MNC) investments.

History is filled with examples of companies that have been hit hard by political risk, i.e. political events that interfere with business operations. The predecessor to British Petroleum, Anglo-Iranian Oil Company, lost their entire Iranian operations as part of the nationalization of the Iranian oil sector in 1951 (Bremmer & Keat, 2009). In 2004, the two pharmaceutical giants, Pfizer, and GlaxoSmithKline had patents of key ingredients for respectively Viagra and Avandia, a diabetes drug, revoked in China, leading to a boost for the local infant pharmaceutical industry (Einhorn et al., 2004). Or even more recent, the Russian trade embargo on EU food imports has hurt many farmers and dairy companies in Europe, among which Arla Foods has estimated the cost for the company close to DKK 134mn (Arla Foods, 2015). Arla is neither the only costly example. The Danish brewer Carlsberg estimated a loss of 155mn DKK in the first quarter of 2015 on its eastern European markets, and ECCO, a Danish shoemaker, reported it lost 50% of its turnover in Russia in 2014 and 2015 (Ferslev, 2016; Nymark, 2015). History, both past and recent, has shown that political risk is a factor firm’s need to analyze when considering operations in emerging markets.

The above cases are just the tip of the iceberg when discussing political tension and the impact it can have on companies operating in emerging markets. Companies are constantly exposed to a variety of risks. These risk can be differentiated as either hazards, economic risk, operational risk or strategic risk (Andersen & Schrøder, 2010). The risks that political events impact a company’s business are categorized as political risk, and dealing with it is known as political risk management. Many companies do not have a proper understanding of how to deal with it and has thus proceeded to account for the impact of political risk as part of the statistical error term (Fägersten, 2015). A proper understanding and process of dealing with political risk in emerging markets, cannot just limit the downside risk faced

7

by a company, but also become a competitive advantage (Henisz & Zelner, 2010). Therefore, this thesis sets out to investigate possible mitigation strategies available to MNC’s operating in emerging markets.

One of the reasons why most companies do not have an adequate understanding of political risk management might stem from the lack of agreement in the academic literature, and of established theories: “The literature on political risk is diverse and remains at the pre-theory stage of its evolution”

(Jarvis & Griffiths, 2007, p. 10). Political risk straddles many different academic disciplines, such as international relations from political science, the traditional risk management methods, and the internationalization literature from the international business literature, which adds to the complexity of the field.

1.1 Problem Statement

This thesis will investigate the nature of political risk management in emerging markets. Much of the literature has been devoted to the study of political risk surrounding the entry decision, and the political bargain that occurs between firm and host country government (Boddewyn, 2005; Kobrin, 1987;

Ramamurti, 2001; Vernon, 1981). However, as most emerging markets have opened their doors to foreign direct investment (FDI), these bargains are no longer as critical, as when the literature on political risk began (Eden, Lenway, & Schuler, 2004). Therefore, this thesis will move forward as an exploratory study of political risk in emerging markets post-entry, which leads to the research question for this thesis:

How may MNC’s be able to mitigate political risk in emerging markets post-entry?

Three aspects appear from this. Firstly, the need for a thorough understanding of the literature, as to deduce possible mitigation options and a framework to view political risk through, and secondly, the applicability of said theoretical framework when operating in emerging markets. Thirdly, the materialization of political risk and its underlying reasons may also provide necessary knowledge about possible mitigation strategies. Therefore, the following three sub-questions will also be answered in support of the main research question:

o How does the literature discuss political risk and mitigation strategies?

o How well do models of political risk perform in practice?

o What strategies are available to MNC’s concerning mitigation post-entry?

1.2 Structure of thesis

The following chapter expands on the methodological consideration taken as part of this project, as well as a discussion of the validity and reliability of the findings. The subsequent two chapters are devoted to an investigation of the literature, and a presentation of the models that will be used to examine the underlying reasons for political risk. It begins with a literature review that seeks to establish three distinct schools of thought within the literature and the respective mitigation strategies each emphasize. The fourth chapter then investigates further the two models that form the backbone of the following case studies, namely Eden et al.’s (2004) political bargaining model and Stevens et al.’s (2015) legitimacy-based view on political risk. Chapter three and four will form the foundation for answering the first sub-question. The following chapters five and six investigate the cases of Novo Nordisk in Bangladesh and Arla Foods in Russia in the light of the presented models. The findings from the two cases will then be analyzed and compared in chapter seven. Then to strengthen the findings of the thesis, chapter eight takes a more inductive approach and presents the views expressed by an expert from the Danish Export Council and Danida Business in Bangladesh. Chapter nine goes in depth with a discussion of the findings of this thesis, and propose an adjustment to the legitimacy-based view on political risk, before ending with a conclusion of the findings in chapter ten.

9 2. Methodology

“We think that it is of practical benefit to understand the taken-for-granted assumptions that we have about the way the world works. Only if we have such an understanding can we examine these assumptions, challenge them if we think it appropriate, and behave in a

different way” (Saunders, Lewis, & Thornhill, 2009, p. 109)

This section will be structured along the lines of the ‘research onion’ presented by Saunders et al.

(2009). The model is presented in figure 1. It is their argument that before elaborating on the methods used to answer the research question, important layers are needed to be understood first, i.e. the ontological and epistemological justifications that are underlying your research. Therefore, this chapter starts with placing this thesis into a philosophical tradition, before moving on to the applied approach of gathering knowledge. Following this, the research design is discussed, before elaborating on the validity

and reliability of the research design.

2.1 Research philosophy

How do we perceive the world? This fundamental question is the traditional beginning of any research, as it leads the methodological choice that follows in the research process and the understanding of what is under investigation (Moses & Knutsen, 2012; Saunders et al., 2009). That is to say, our ontological position are the glasses through which we view the world. I subscribe to the realism strain, more precisely the critical realism philosophy. “Critical realists argue that what we experience are sensations, the images of the things in the real world, not the things directly” (Saunders et al., 2009, p.

115). This position is particularly useful when studying the complexity of political risk. Political risk exists, it is here taken to be something that occurs in the real world and is not socially constructed; the

Figure 1 Research Onion, adapted and modified from Saunders et al. (2009)

actions of political actors interfere with business operation. However, how it is perceived, the changes that occur because of political risk, and the underlying consequences of these actions are more complex, and cannot reflect in full the reality of the events, thereby a layer of social construction appear on top of the real events.

In accepting a research position of realism lies an acceptance of data, and also the misinterpretation of data (Saunders et al., 2009). What the data may show, are subject to misinterpretation of the underlying social phenomena, as well as a bias of the research being value-laden. How I perceive a context and my views on a political situation are inherent in how I understand and view the world. This research bias is attempted mitigated to the best of my abilities, through a use of triangulation, however, it lies at the core of the critical realism’s axiology, and will impact the research.

2.2 Research approach

Throughout the main part of this thesis, the deductive method will be used. This stems from the initial query that sparked the interest in this thesis, the search for knowledge on political risk. The deductive approach is also evident in the problem statement, as the sub-question of ‘how the literature discusses political risk’ illustrates. This question is in nature deductive, as it relies on an understanding of theory, and from this understanding, attempt to form new knowledge. Yet, the method of case study, which this thesis relies upon, includes elements of both the deductive and inductive approaches. The main part will be focused on examining the research question through theory and the use of models from the literature. Even so, when aspects diverge from the literature, inductive reasoning will guide the enlightenment of the discrepancy. Further, the interview that has been conducted also introduces an element of inductiveness, as Søren Robenhagen’s views represent his own experience of advising Danish companies in politically unstable environments. Following a rigid deductive approach would blind the research from other possible conclusions. Saunders et al. (2009) argue that it is beneficial to combine the two, to benefit from the structure of the deductive, but also be open to new interpretations. This thesis thus applies a hybrid research approach.

11 2.3 Research design

Forming an understanding of political risk is a complex matter, one the literature itself has struggled with (Fägersten, 2015). This thesis will mainly take a theoretical approach to the concept of political risk, due to this complex nature. Thus, to investigate and arrive at a conclusion to the research question, a threefold approach will be used. A literature review will form an understanding of existing knowledge on political risk, and frame the first sub-question. Then two case studies will be used to test models of political risk, i.e. a confirmatory multiple case study, and the findings will be compared through a method of agreement. The third approach will be to combine the learnings from the case, with the views expressed in an expert interview, thereby introducing an inductive element. These elements of the research design are presented in this section, with its justification.

2.3.1 Critically reviewing the literature

Academic work is not done in isolation. Authors build on each other’s experiments and learnings to create new knowledge. This is part of the scientific method. As was mentioned in the introduction, the literature on political risk is vast, and some of the reasons for this is that it straddles several academic fields, from political science to strategic management literature and the risk management literature. To answer to what extent firms are able to mitigate political risk in emerging markets, and the understanding of the subject matter is imperative. The literature review has been conducted with a deductive approach, to help identify theories and find relevant methods for testing it, and has thus been instrumental in forming the latter study. It has been critical in identifying where the field is moving, and consequently, it places the following case studies in a better perspective. By evaluating and grouping the literature, as well as investigating the assumptions behind established theories, a better understanding can be presented (Saunders et al., 2009). A critical survey of the literature allows for the subsequent analysis to be understood within the broader framework in which it fits.

2.3.2 Confirmatory case study

Based on the literature review, the analysis continues with a case study. This approach has been chosen as it allows for an investigation of a phenomena in its real life context, where multiple sources influence the object at hand, and the context may provide valuable understanding (Saunders et al., 2009). “The

case study is a research strategy which focuses on understanding the dynamics present within single settings” (Eisenhardt, 1989, p. 534). When considering the variables ‘political risk’ and ‘emerging markets’, the context becomes especially crucial, as politics is difficult to study in isolation. Often there exist various formal and informal institutional structures that are able to affect the political process, and therefore, it is necessary to observe it in its natural environment (Fuchs, 2007).

As the case studies will take its onset in existing theory, i.e. following the deductive approach and building on the literature review, it will take the form of a confirmatory case study. The cases will be structured on a theoretical foundation, and seek to answer why political risk has appeared or not. Two cases have been chosen to represent theory confirming cases, as to later compare the benefits and flaws of the presented models (Moses & Knutsen, 2012). Yin (2003, as quoted in Saunders et al., 2009) argues that the multiple case study is preferred to the single case study as the findings become more generalizable when tested in two different settings. For the case of political risk across a broad and diverse category of countries such as that of emerging markets, this is certainly true.

The approach taken is similar to that of Conway (2013) in his study of political risk management in Kazakhstan’s uranium industry. His study approached the case study as a combination of theory testing and heuristic development. Theory testing evaluates the validity of a theory while the heuristic case studies seek out new development to the theory (Eckstein, 1991; George & Bennett, 2005).

The case study method also has its critics. The general critique of the case study method is the many biases inherent in the method, as well as the issue that what might be true in one case is not necessarily true in another (Moses & Knutsen, 2012; Saunders et al., 2009). With regard to the lack of generalizability, it is to some extent mitigated by the use of multiple cases. It is not to be understood that by applying the same framework to a second case all these issues disappear, but it does move the research towards a more general state.

2.3.3 Selection of cases

For this thesis the case of Novo Nordisk in Bangladesh, and Arla Foods in Russia have been selected as confirming case studies, to test the theory in. Both cases have been chosen due to their ability to test

13

theoretical sampling is to choose cases which are likely to replicate or extend the emergent theory”

(Eisenhardt, 1989, p. 537). Novo Nordisk operations in Bangladesh, a country categorized as least developed by the World Bank, has been chosen for its ability to depict the importance of local partnerships, whereas the case of Arla in Russia was chosen to test whether the legitimacy-based view on political risk holds in a global political crisis. Thus, both cases provide an opportunity to test the theories and investigate the validity of these, which in turn will help in answering the research question, as their actions shed light on possible mitigation strategies.

However, a danger lies in this method. As I have explicitly chosen these cases, instead of other cases, a deliberate choice has been made, i.e. a selection bias. The two cases may possibly not represent the best cases for this study, as there is a danger that a lurking variable is present that is not identified during the study. Furthermore, the cases have to be chosen as either confirming, deviant, didactical, or generalizing (Moses & Knutsen, 2012). By choosing confirming case studies less emphasis is given to theory building, or educational aspects, which would have been found in the inductive methods of the generalizing case study, or the educational style of the didactical. However, the cases have not been chosen to fit perfectly with the theory, but rather to test the validity of the theories, and attempt to gain insights into their mitigation strategies, when faced with politically instable environments.

2.3.4 Comparative method

By comparing two cases of Danish multinationals’ experience with management of politically unstable environments, it allows us to evaluate whether there are elements that are similar between two otherwise different contexts. Even though both host countries in this study, Bangladesh, and Russia, are considered under the heading of emerging markets, it does not mean that they are completely similar in all respects. For one, Russia’s economy is close to ten times the size of Bangladesh, even though their populations are close to equal in size (World Bank, 2016). Therefore, the method of agreement by John Stuart Mills is applied. This comparative method allows for an investigation of two different settings with regards to a similar circumstance, i.e. different institutional infrastructures, but similar potential for political risk. “The Method of Agreement is simple in that the investigator merely collects cases of a particular phenomenon in an attempt to find common factors in these cases that are

otherwise quite different” (Moses & Knutsen, 2012, p. 102). The major issues with applying the Method of Agreement is the possibility of hidden lurking variables (Moses & Knutsen, 2012). However, per definition, a hidden lurking variable, is close to impossible to identify.

2.3.5 Expert interview

For this thesis, an expert interview has also been conducted. The respondent, Søren Robenhagen, works as a Commercial Counsellor and as Danida Business Program Coordinator at the Danish embassy in Dhaka. His past experience also includes serving as Consul General and Trade Commissioner at the Danish Trade Council in Istanbul. This makes him knowledgeable about businesses needs in politically instable markets. The interview was conducted as a semi-structured phone interview via Skype on the 9^th of May 2016. Semi-structured interviews allow for a possibility to avert from the prepared questions and ask elaborative questions by the interviewer (Saunders et al., 2009). The interview guide prepared for the interview can be found in appendix 1 and the interview itself is transcribed, translated into English and placed in appendix 2.

2.4 Data choice

A mix of quantitative and qualitative data will be used throughout this thesis. The validity and reliability of these sources are discussed in a subsequent section, whereas this section illustrates the importance of a multiple methods data approach. This approach of applying multiple types of data and analysis is being increasingly encouraged in the management literature, for its ability to provide more than one method of answering the research question (Saunders et al., 2009). Quantitative data, such as descriptive statistics, GDP figures, and demographic data is used to underline arguments regarding the context in which the case firms operate. A mix of primary and secondary qualitative data is then used to present information on the chosen cases. The data takes the form of company and government reports and statements (when available), news articles (triangulated across borders whenever possible), academic journals, as well as an expert interview. This ensures a more holistic approach to the research question, increasing the validity of the conclusion.

15 2.5 Time aspect of study

The case studies in this thesis both apply the longitudinal aspect. If political risk is dealt with only at the time it materializes, important aspects of why it develops, and how it could be mitigated are left out of the study. Many aspects are able to influence the decision to exert political risk onto a foreign company, and therefore, it should be studied as a series of events, i.e. a longitudinal study (Simon, 1982; Stevens et al., 2015). Both cases involve companies that have had some sort of business operations in their respective host countries for more than ten years, and this historical aspect also plays a part.

2.6 Validity and reliability

One of the objects of writing a master’s thesis is to demonstrate sound academic research. Still, how do we know that the presented research is sound? The short answer is we do not know. As mentioned above, hidden lurking variables may be at play, interfering with the conclusions. Instead, what is possible is to reduce the probability that the conclusions are incorrect. This is done through a continuous process of reliability and validity checks (Saunders et al., 2009).

2.6.1 Validity of study

How accurate are the conclusions drawn from this study? In relying mainly on qualitative data, therein lies a certain risk. To ensure that this risk is diminished the used data is placed under scrutiny through a triangulation: “In this sense the validity in quantitative research is very specific to the test to which it is applied – where triangulation methods are used in qualitative research” (Golafshani, 2003, p. 603).

Triangulation, i.e. the use of at least two independent sources, has been used throughout the case studies to ensure correct data. Information published in a national newspaper have been cross- referenced with news agencies from different countries, which has been imperative when dealing with Russian media, where the media is not considered free to press what they want (Freedom House, 2016).

Likewise, the different sources have been critically evaluated.

Also, the use of a multiple case study ensures corroboration on the case findings. By applying a multiple mixed model approach to a multiple case study and drawing on an expert interview, the validity of the findings are strengthened (Saunders et al., 2009). The validity could be strengthened further through additional interviews, and possibly also interviews with people from the case companies.

2.6.2 Reliability

Reliability concerns the degree to which the conclusions drawn from this study will be the same, given another observer, as well as the inferences drawn from the data are transparent. For this study, the main concern has been that of observer bias, i.e. my interpretation of the data (Saunders et al., 2009).

To counter this bias, a logical structure has been applied to the case analysis, where the case data is presented firstly, and the following analysis builds on this data, to ensure that others may draw the same conclusions.

2.7 Definitions

This section serves the purpose of aligning the reader with the author's understanding of key terms that are used throughout the thesis. This ensures better coherence between what is written and what is read.

2.7.1 Political Risk

The most central term in this thesis is ‘political risk’. In the introduction, it was referred to as the risk that political events impact a company’s business. This broad definition covers the essential understanding of the concept. However, as the literature on political risk is divided a further refinement is needed to avoid misinterpretation (Fägersten, 2015). The definitions range from the very broad:

“Political risk can be understood as the probability and impact of a politically motivated action or system failure that negatively affects a given company’s people, reputation of performance” (Fägersten, 2015, p. 31), to the inclusive: “Political decisions, events, and conditions, including of social nature, that change the expected outcome or the expected value of an economic activity, and thus affects a company’s realization of its goals” (Iversen, 2006, p. 1). For this paper the definition by Stevens et al.

will be used: “Government actions that have a negative impact on firm performance” (Stevens et al., 2015, p. 315). It thus narrows the scope to government’s actions, and does not include social unrest, which can be more difficult to assess, and thereby it provides a clearer definition of the subject at hand.

2.7.2 Emerging Markets

As the research question highlights emerging markets, a brief mentioning of what is understood by the

17

i.e. countries with high GDP growth rates through a longer time period, or countries lacking the regulatory and institutional environment that provides stability. It thus follows Bremmer’s (2005) definition of stability: “A nation’s stability is determined by two things: political leaders’ capacity to implement the policies they want even amidst shocks and their ability to avoid generating shocks of their own” (pg. 5). Governments that are able to govern smoothly, without creating shocks of their own, are stable. The opposite case in point is Russia, where the government is able to pass regulation, but have also created external shocks. While being a high-income country, this instability places it in the bracket of emerging markets for this thesis.

3. Political risk in the literature

“If I have seen further it is by standing on the shoulders of giants” – Sir. Isaac Newton

Political risk management is not a new phenomenon. People have been attempting to influence those in power to provide themselves with preferential treatment for centuries, and thus, manage their political risk exposure. Be it nobility trying to influence the king, or modern day lobbyist persuading politicians to support their cause. However, the modern academic literature is still attempting to find a common standpoint with regards to the nature, and definition of political risk management (Fägersten, 2015). One reason behind this difficulty might stem from the fact that political risk occurs in both strategic management literature, political science, international business literature, and risk management (Jarvis & Griffiths, 2007).

Through researching the political risk literature three distinguishable schools have emerged; the conflictual school, the environmental school, and the institutional school. The latter two can also be understood as a split in research design between micro and macro studies of political risk, i.e. those studies that investigate the political risk exposure at the firm level, and those that focus on a cross- section of countries’ political risk levels. Both draw inspiration and originate from the conflictual school.

The literature has been classified as such because they present different approaches to political risk and are based on diverse assumptions of business-government relations. In the conflictual school, state and business are seen as opponents in a zero-sum game, battling for higher rents. From the

environmental school, business no longer sees host country governments as a counterpart, but rather as a part of the greater business environment, and are therefore based on the assumption that politics and policy can be shaped. The institutional school focuses its analysis of political risk as a cross-country study in variation in political and social institutions and is mainly focused on identification.

This chapter will go through selected writings in these three schools, and thereby present a broad understanding of the field, as well as introduce models used to examine political risk through the years.

The two main models for this thesis will be further elaborated on in the next chapter while in this chapter they will only be presented briefly. To frame the understanding of political risk, this chapter starts with an introduction into corporate risk management processes, from which a great deal of the literature refers to their understanding of risk.

3.1 Risk management

At the core of any work related to risk management is an understanding of risk itself; albeit, that in itself can be a difficult task. There is no common definition of risk (Roggi & Altman, 2013). Generally, when speaking of risk, it is understood to be the result of an event that may happen, and often associated with a negative impact. This need not be the case, as argued by Andersen, Garvey, and Roggi (2014).

They argue that only focusing on risk management as a tool for limiting downside risk, miss the point that risk management can also be used as a value generating process within a company. Thus, risk encompasses both a risk of a loss, as well as an opportunity for gains. Or, put into a political risk perspective, governments, as well as social actors and other political players are part of the business environment, when and where their actions can affect a company’s performance, lies a risk for companies.

Risk also holds an element of measurability. Often the literature refers to Frank Knight’s (1921) distinction of risk and uncertainty. Risk, he defines, as events that have a measurable probability, whereas uncertainty is the events that cannot be measured. In regards to political risk, some parts of the literature have placed political risk in the box of uncertainty (Fägersten, 2015).

When firms conventionally deal with risk management, they go through a four-step process. Firstly, a

19

accordance with the organizations’ risk appetite, before lastly responding to these risk exposures (Andersen & Schrøder, 2010). How does one measure a risk? One method that is commonly used is that of assessing the value of the risk to the firm, often done by multiplying the likelihood of the risk materializing with the impact it will have on the company. Some risks can be difficult to quantify; this is also why some companies tend to use a subjective measure of the potential risk. Some use a six-step process, where a score of 1 represent very likely and 6 represent highly unlikely. These are then often plotted on a risk map that indicates whether a risk should be given more attention. The visualization of the identified risk can be helpful when deciding how to deal with them. Others try to measure the risk by more quantifiable variables, or use measures such as the Value at Risk formula.

Most companies have established risk management processes in place, often in the form of an enterprise risk management (ERM) system, which has risen in popularity following the financial crises of the past decades. Although many variations of ERM frameworks exist, e.g. FERMA, AS/NZS 430, ISO 31000, Turnbull Guidance, COSO etc. (Hayne & Free, 2014; Power, 2009). The ERM system advocated by COSO is being advertised as a strategic framework that attempts to instill a risk culture throughout the organization (COSO, 2004). Critics believe it to be closer to an advanced control system (Andersen

& Schrøder, 2010). An ERM system is able to help mitigate both hazardous events, economic risks, as well as operational risk, but it is not well placed to help firm’s deal with risks that are strategic in nature, such as political events. To this end, Andersen and Schrøder (Andersen & Schrøder, 2010) have argued for an extension to the ERM system, to ensure that risk management is integrated into the strategy framing and objective setting process of the company. Key to a good Strategic Risk Management approach to risk is the implementation of methods from strategy planning into the framework.

Examples include SWOT analysis, scenarios planning, and a real options approach (Andersen and Schrøder 2010). Scenarios planning especially can be valuable as it allows risk personnel to decide and evaluate risks before they materialize, and the real options approach allow firms to have a detailed and thought-through plan of action to a changing risk landscape. These methods hold great promise for firms operating in politically volatile markets.

For political risk, many argue that it is difficult to assign probabilities to political events, such as a change in regulation, or that social unrest will erupt in a country, but there are indicators that can be watched (Bremmer & Keat, 2009). A strong risk management setup may allow firms to be on the forefront of emerging political risks. Political risk exposure is more difficult, but it is not impossible to assess in the same manner as in the traditional risk process. The following sections investigate the different variations in political risk management literature.

3.2 The beginning of political risk: the conflictual school

As mentioned at the start of this chapter, political risk has been a presence for firms and individuals throughout history. The first academic conceptualization of political risk was by Franklin Root in the early 1960s, where he defined it as: “… threats to foreign investments that can arise from attitudes and behavior of host governments and significant political and social groups.” (as quoted by Simon, 1982, p. 62). The business literature at the time was concerned with the risk of expropriation, seeing this as the biggest risk to MNCs’ international operations. This came in the light of several expropriation events in the 1950s and 1960s, such as the nationalization of Anglo-Iranian Oil Company by the Iranian government in 1951, the 1956 nationalization of the Suez Canal by the Egyptian government, and the nationalization of many American companies’ assets after the Cuban revolution. Root (1968) surveyed large American MNCs and found that very few performed any political risk analysis prior to foreign investments, and he thus encouraged multinationals to appraise the political risks, including risk of expropriation in a systematic fashion before the investment decision.

The period around 1960 to 1975 was characterized by a focus on the two actors, business, and government, as opponents. Root contributed to this by his focus on expropriation, but other authors also weighed in on the discussion. John Fayerweather called it “an era of confrontation between governments and MNEs” (as quoted in Boddewyn, 2005, p. 29). This confrontational view of business and politics led to a period where authors were classifying possible negative effects of political actions on business operations in host countries (Jarvis & Griffiths, 2007; Kennedy, 1987). The power of governments was seen as a decisive element in international business operations. Because of this, business managers were warned to be aware of the dangers in the political environment: “Because of

21

the worldwide trend for governments to play an increasing role in the shape and direction of the business environment, it is essential that both a company’s exposure to political risks, and its subsequent attitude, is the result of a thorough analysis of these risks” (Lloyd, 1974, p. 31). Kobrin (1979) found in his literature review that a large part of the field had been dealing with the interference of governments into business affairs, and saw political risk as the unfavorable consequence of exogenous political events on businesses. Others have added to this view, acknowledging that the pioneers in the field contributed to a list of non-economic conditions that might affect foreign business activity (Simon, 1982). Political risk was in this period seen as an element of the government, and exogenous to business as political risk is represented as market interventions by political actors.

However, the resulting conclusion, that government should stay out of market activity, does not fit with economic historians’ findings that states’ involvement actually decrease political risk (Jarvis & Griffiths, 2007). Thus, the first writers on political risk were mainly attempting to decompose the phenomenon and investigate how widespread the consequences were.

The model that has contributed the most to the study of political risk analysis that emerged from this period is known as the Obsolescing Bargaining Model, originally proposed by Ray Vernon in 1971 (Eden et al., 2004; Kobrin, 1987; Vernon, 1981). The Obsolescing Bargaining Model proposes that host countries and MNCs go through a negotiation upon the MNC’s entry into the host country’s market.

Originally developed with a focus on the extractive industries, it has also been applied to manufacturing industries and lately service industry with less predictability (Kobrin, 1987; Stevens et al., 2015). The MNC’s desire to move abroad has been investigated plenty in the field of international business, with market seeking, resource seeking, efficiency seeking and knowledge seeking activities, as commonly stated objectives for the MNC (Dunning, 2000; Ietto-Gillies, 2012). The host country, on the other hand, is seeking foreign knowledge and capital. Thus, both have something the other party desires. At the beginning of the process, pre-entry, the MNC is in a position to make a better bargain, as it can move elsewhere while the host country needs the foreign expertise. However, the position changes once the firm has entered the host country, as large sunk costs occur for the MNC, and knowledge spillover into the local economy resulting in the emergence of local players, able to lessen the host country’s dependency on the MNC (Eden et al., 2004). This places the host country with the upper hand and

causes the original bargain to obsolete. Therefore, firms are advised to seek highest possible rents after entry, as their market position will erode. The model is based on three elements, the resources held by each party, the constraints placed on both, and the underlying goals of the two. The model has been used to explain how MNCs can mitigate political risk, by either establishing constraints on the host country through bilateral agreements through home country governments (Ramamurti, 2001), or limitation of technological spillovers from MNC to host country as a barrier to host country bargaining power (Kobrin, 1987). In essence, the model can be used to establish whether or not a host country government is able to exercise power over (foreign) corporations, not focusing much on whether or not the government wants to use this power.

From Root to Vernon, many of the early writers on the topic saw political risk as market intervention by political actors. This initial focus sparked the academic interest in the topic. The period was focused on classifying and analyzing possible outcomes that could result from political risk, though marred by its assumptions on business and government relations. From this conflictual view came the first model on political risk, the Obsolescing Bargaining Model, which laid the groundwork for political risk analysis.

3.3 The macro level risk; the institutionalist

Out of the earlier school’s focus on identification and assessment of political risk, grew a greater desire to create large-scale macro indexes, to aid decision-makers about the political risk level in various countries. This school has been defined as the institutionalist, as the common thread in their work, has been a focus on the macro level analysis, as well as a view that variation in countries institutional infrastructure can be used as a proxy for political risk. Works in this school are mostly focused on developing or using macro index calculations of political risk to make inference about the future;

moving towards a more qualitative approach (Simon, 1982). This school had its beginning in the 1980s inspired by the Latin American debt crisis, and the failure of the checklist systems of the conflictual school: “At the same time, several concerned bankers were somewhat disillusioned with checklist systems because of their failure to help predict the major crises of 1982, particularly Mexico and Argentina… As a result, there has been some visible movement recently among banks toward short-

23

term, econometric models.” (Kennedy, 1987, p. 5). However, it has also been receiving more attention in the new millennium, especially after Henisz published his political constraints index in 2000.

The focus of analysis became political systems, and institutions, and from this an attempt to predict the probability of political risk events occur. Early authors based their work on the conflictual schools listing of variables and made attempts to put them into indexes able to predict the future. One example of this school can be found in Kennedy (1987), who establishes a political economy framework, based on both qualitative socio-economic variables, and quantitative economic variables. However, in mixing both qualitative and quantitative variables, Kennedy himself recognizes one of the issues, being failure to estimate qualitative variables, such as strategic goals of a states government, which makes the framework appear useless in practice. The same conclusion can be drawn to similar early attempts of model building, in which the predictability often have come into question (Jarvis & Griffiths, 2007).

Around the late 1970s and early 1980s, several agencies that focused on measuring political risk levels emerged. The Business Environment Risk Index (BERI), Business International’s Country Assessment Service (BI), and International Country Risk Guide¹ (ICRG), were all developed in the early 1980s, where the two former were indexes based on a quantification of surveys by ‘expert’ opinion-makers, the latter was and still is decided by internal editors on a predefined set of questions (Simon, 1982; The PRS Group, 2016). This was also the common practice found in some larger corporations in the early 1980s, including Bank of America and Royal Dutch Shell (Simon, 1982). These approaches had a common aim, to help guard companies against expropriation. Based on these indexes, some attempts were made within the field of international relations, to create a political risk grand theory, linking the results from these indexes, and information about state types from the political science literature (Jarvis & Griffiths, 2007). From this strain of research came the notion of using political stability as a measure of political risk. Originally used for the state type comparisons in the institutional school, it has rooted itself in the

1 Then part of Frost & Sullivan Inc. and also known as the World Risk Forecast (WPRF), today part of the PRS Group a political risk management agency.

literature and is today used generally as a definition of political risk (Bremmer & Keat, 2009; Bremmer, 2005; Henisz, 2000; Peng & Meyer, 2011).

The earlier works within the institutional school have attained some criticism. Firstly, with regards to the risk indexes, there lies a bias issue, as they have mainly relied upon so-called ‘expert’ opinions on the political and social situation within a given country. As Simon noted with regards to the BERI and BI: “The ultimate country ratings are dependent upon the “expert’s” definition of the situation and his prior experience. By presenting the findings in terms of country rankings, a pretense of objectivity is given to subjectively derived data” (Simon, 1982, p. 65). Secondly, these large country indexes, also risk a problem of being too general, and neglect that the risk environment perceived by companies differ, e.g. a politically instable country might be a horrible investment destination for a brewing company with large sunk cost, but not for a company operating within the security industry, which thrives in unstable environments. Macro indexes do not take account of variation within risk exposures on an industry basis (Simon, 1982). The indexes can help companies establish an overview of the situation, but they do not aid in pinpointing the relevant exposures for the individual firms, limiting their usefulness (Fägersten, 2015). Thirdly, they lack predictive power. 2 out of 10 models tested by Kennedy (as quoted in Yaprak & Sheldon) were able to offer proper assessment ahead of the Iranian revolution in the 1970s, and several indexes failed to alarm business and relevant policy makers ahead of the Asian Financial Crisis of the 1990s (Peng & Meyer, 2011; Yaprak & Sheldon, 1984). The earlier developments within the institutional school might offer a certain overview of potential investment destinations, though they need to be taken with a pinch of skepticism.

Though, a later development within this field is worth noting, even with regards to the criticism of the earlier attempts of inference based on indexes, political systems, and institutions. Henisz (2000) also noted the increased use of subjective indexes in the literature, especially the ICRG, and instead created an objective measure of political constraints; the results is the POLCON model. He noted that the contemporary models were suffering from four faults, 1) the models did not make a specific link between a government’s ability to commit not to interfere with private property rights, 2) they were biased due to subjective unit of measurement, 3) they did not have a large sample and/or were limited

25

to a short time frame, and 4) they were applied inconsistently with regards to theory (Henisz, 2000).

The POLCON model is based on two elements: The number of veto points in a given political system, and the distribution of the actors, i.e. legislative branch, executive, judiciary, sub-federal units, preferences in the political system. The model seeks to investigate whether a political actor, mainly considered the executive branch of government, is constrained by the political system in their future choice of policy, e.g. can a leader of a country, rapidly impose new legislation or policy that imposes extra costs on investors. The basic assumption is, that political systems with few veto points constitute’s a more politically risky environment (Peng & Meyer, 2011). He finds an expected correlation between the POLCON and ICRG, but also finds that unlike the ICRG, the POLCON does not have a lag tendency, i.e. the POLCON model is able to predict political transitions in the year they happen, unlike the experienced-based measure, that fails in times of rapid institutional change. This more objective measure has gotten much positive attention recently and has initiated a new period of institutional studies of political risk (Lee, Biglaiser, & Staats, 2014; Stevens & Dykes, 2013).

3.4 The rise of the environmental school

Root noted early on that companies did not have adequate systems in place to manage political risk.

However, by 1983, a study suggested that 74% of a sample of 61 large firms had institutionalized political risk management practices (Kennedy, 1984). The turning point for this rise in attention has by some been suggested to be the OPEC oil crisis of 1973-74 (Kennedy, 1987; Simon, 1982). Earlier, companies had been scared by expropriations but were now realizing that they needed to understand political movements in the business environment to be able to have their foreign subsidiaries operate smoothly.

Scholars began focusing on the environment and possible variables that could help explain and predict political instability and risk (Simon, 1982). At the same time, the governments in emerging markets began changing their

Figure 2: Developing countries share of World FDI and GDP. Source: World Bank 2016

attitude towards foreign firms, and instead opened their arms to FDI. Looking back at the development in FDI and GDP in the developing world since 1990, the growing importance of foreign investments becomes apparent, illustrated in figure 2 by the share of developing countries of world FDI and GDP (World Bank, 2016) To attract FDI, states became increasingly aware that a better business environment could improve their ‘bargaining’ position. For firms and international business academics, the focus changed to building and utilizing models of risk mitigation at the operational level. From this, a new school of political risk studies emerged the environmental school.

Where the conflictual school was mainly focused on the identification and assessment of political risk and the institutional school considered variation in political risk level across markets, the environmental school’s focus was and is more on mitigation tactics for the individual firm. It follows the logic that political risk is contextual. That is, what constitutes a political risk for one company, might not be so for a company in a different industry operating in the same foreign environment (Bremmer & Keat, 2009;

Fägersten, 2015). E.g. the Muhammed Crisis, where the Danish-Swedish Dairy giant Arla was exposed to boycotts of its products, due to a cartoon published in a Danish newspaper, whereas other foreign dairy producers were not affected (The Economist, 2006). Based on this notion of induvial risk, the literature turned to focus on micro political risks.

One of the earliest divisions of political risk into a micro and a macro element can be found in the early 1970s (Robock, 1971). However, with more than four decades of research into micro political risk, the environmental school appear to lack a commonly agreed upon framework (Alon & Herbert, 2009; Jarvis

& Griffiths, 2007). Alon and Herbert (2009) inspired by the conflictual schools ideals of listing possible variables, followed suit in their model of micro political management, where instead of presenting a mitigation or prevention framework, they present another assessment framework, that builds on subjective inputs. Thus, it becomes difficult to test the validity of the framework in practice.

Others in the environmental school have focused on linking political risk management with the techniques of traditional risk management. Casson and da Silva Lopes (2013) found that even though the international business literature emphasizes the importance of political risk in high-risk environments, its prime suggestion for handling it is avoidance and withdrawal strategies. The authors

27

argue that firms are actually able to manage high-risk environments by the use of risk management strategies; considering four generic strategies of withdrawal, avoidance, mitigation and prevention (Casson & da Silva Lopes, 2013). They argue that direct political lobbying may be inefficient in emerging markets in some cases, and instead, in line with Ramamurti (2001), firms may rely on home country governments to exert political influence on behalf of the company (Casson & da Silva Lopes, 2013).

However, while the authors warn against direct lobbying activity, they also suggest that companies may benefit from creating ties with “indigenous elites” to manage public perception. Thus, some forms of clout on the political establishment is possible. Similar findings have also been found in other studies (Oliver & Holzinger, 2008; Stevens et al., 2015).

Understanding that the foreign environment is a myriad of risks (and opportunities), some authors have begun to see risk mitigation as a likely complex task. The situation might vary depending on the political event that is about to materialize, be it regulatory risk, expropriation, or social instability. Bremmer and Keat (2009) argues for a nuanced, company specific, approach to dealing with political risk. They argue for a range of different mitigation strategies, depending on the political risk companies face, and their industry characteristics. These strategies vary from lobbying, both directly and through personal networks with government officials, to reducing expropriation risk by engaging in joint ventures, or creating a network of interdependency within the subsidiary network (Bremmer & Keat, 2009). The core of their approach is, that it all depends on the environment and context in which the firm operates while drawing on traditional risk management approaches.

More recent frameworks and models for dealing with political risk from the environmental school have been attempting to move the focus away from the entry decision, and rather shifted the attention to post-entry. Some have chosen to integrate new insights from transactional economics and the resource-based view of the firm with the early models of political risk management. One example of this is Eden et al. (2004) in which the authors present an updated version of Vernon’s Obsolescing Bargaining Model, and present the Political Bargaining Model. Unlike its predecessor, the Political Bargaining Model is not focused on a negotiation over market entry conditions, but rather focused on MNCs’ and host governments continuous bargain over industry policies (Eden et al., 2004). The

theoretical lenses of Vernon’s model, i.e. the goals, resources and constraints of the two parties, as well as the bargain outcome, have been kept, but is instead used to explain public policy post-entry, and a focus on establishing continuous constraints on the host government becomes a mitigation strategy for firms.

Another example of reconfiguration of the earlier models can be found in Stevens et al. (2015), in which the authors argue for a complementary approach, i.e. utilization of three political risk frameworks. They demonstrate that the conflictual, institutional and environmental models can be used in combination to provide a better, more holistic analysis of a state’s ability to intervene, and also whether it will do so. They apply the institutional school, and conflictual school’s models, i.e. POLCON and obsolescing bargaining, to establish the power of the host country, and add the nascent legitimacy-based view as a way to analyze whether a host country will actually be motivated to intervene. The model also argues for a more holistic approach, emphasizing that political risk does not necessarily emanate exclusively from emerging markets, but can also stem from home; something that is supported by other authors (Ramamurti, 2001; Stevens & Dykes, 2013; Stevens et al., 2015). Their model will be presented further in the next chapter.

The environmental school of political risk management has been receiving increased attention recently.

The two approaches presented in the previous two paragraphs have been published respectively in 2004 and 2015. It can be argued that following the financial crisis of 2007 and China’s rise, multinational corporations have become increasingly aware of the need for an understanding of how politics influences their business, at the firm level. In such this has led to an increased attention into the environmental school, with its focus on firm-level application, and mitigation strategies. The environmental school can also be seen as building on top of the conflictual school, with the Political Bargaining Model, and working as a complementary approach to the institutionalist macro focus, illustrated in Stevens et al.’s complementary approach. These two models are also both focused on a longitudinal view on political risk, and will form the basis of the latter case studies. How they will be operationalized to answer the research question will be introduced in the following chapter. The focus

29

in the environmental school has been on application, over grand-theorizing, which remains a highly sought after price in the political risk management literature (Fägersten, 2015; Jarvis & Griffiths, 2007).

3.5 Summary

In this chapter, the literature on political risk has been inspected. Firstly, through the traditional risk management approach and its emerging strategic risk management approach, and secondly, through the political risk management literature, from which three distinct schools emerged: The conflictual school, where the emphasis was on state and business relations as a competitive zero-sum game and for firms guided towards avoiding expropriation by host country governments; The institutional school, that concentrated on developing grand indexes to predict political risk, with limited success²; and lastly the environmental school, that neglects theorizing for method, and attempts to bring the study of political risk back to the firm-level, while assuming the state might not always be an adversary, but might also be used to strengthen or protect competitive advantages. The findings are summarized in figure 3. The early literature’s mitigation strategies were focused on either avoidance or withdrawal from instable markets, but with the environmental school, it has moved towards a focus on relationships with key stakeholders, as well as a strategy of embedding oneself in the host environment to establish legitimacy, and thereby lessen the inherent political risk. From this analysis, this thesis will continue to work within the environmental view, thus acknowledging the lack of a grand theory, and instead work on post-entry micro mitigation strategies applicable to emerging markets.

2 With the exception of Henisz POLCON framework, that is able to identify veto points in a political system as a proxy for political risk, and is widely used today as a proxy for political risk