Master Thesis Word Count: 14784

1

The Impact of Cultural Values and Bounded Rationality on Investment Decisions: Examining the Link between Dimensions of Project GLOBE and Home Bias

Master Thesis

Word Count: 14784

Philipp Heilmann

Student Number: 2093573 (Groningen), 109207111 (Newcastle)

University of Groningen Faculty of Economics and Business

ewcastle University

Business School

Dual Master in Advanced International Business and Management Supervisors:

2

3

Table of Content

1.1 Purpose of the Study ... 10

1.2. Audience ... 10

1.3. Methods of Analysis ... 11

1.4. Plan of Work ... 11

1.5. Data Sources ... 12

1.6. Findings of Dissertation ... 12

2. The Influence of Culture and Bounded Rationality on Economic Choices ... 13

2.1. Culture and Economic Decision Making ... 13

2.1.1. Definition of the Term Culture ... 14

2.1.2. The Link between Culture and Economic Choices: Two Perspectives ... 16

2.1.2.1. The Macro Perspective ... 18

2.1.2.2. The Micro Perspective ... 19

2.2. Bounded Rationality and Economic Decision Making ... 23

2.2.1. The Efficient Market Hypothesis as Proxy Model... 23

2.2.2. From Utility Maximisation to Prospect Theory ... 25

2.2.2.1. Expected Utility and Utility Maximisation Theory ... 25

2.2.2.2. Cognitive Biases ... 26

2.2.2.3. Prospect Theory ... 27

2.3. Linking Culture and Bounded Rationality ... 28

3. Conceptual Framework ... 31

3.1. Home Bias and Diversification Theory ... 33

3.2. Dimensions of Culture and Economic preferences: Borrowing from Project GLOBE ... 35

4. Data and Method ... 43

4.1. Cross- Country Outbound FDI Regression ... 44

4

4.2.1. Survey Design ... 46

4.2.2. Risk Preferences ... 47

4.2.3. Investment Preferences or Extent of Home Bias ... 48

4.2.4. Cultural Dimensions ... 48

5. Empirical Results ... 49

5.1. FDI Regression Results ... 49

5.2. Survey Regression Results ... 50

6. Discussion and Conclusion ... 52

6.1. Culture and Home Bias at the Macro Level ... 52

6.2. Culture and Home Bias at the Individual Level ... 54

6.3. Joint Discussion ... 58

6.4. Summary and Conclusion ... 59

6.4.1. Limitations ... 61

6.4.2. Theoretical and Practical Implications ... 61

List of References ... 62

5

List of Tables

Table 1: Overview of Hypotheses ... 42

Table 2: Required Survey Regression Variables ... 46

Table 3: Results of the Multi Model FDI Regression ... 49

Table 4: Results of the Single Model FDI Regression ... 50

Table 5: Adjusted Coefficients of Determination of Control Variables ... 50

Table 6: Results of Multi Model Survey Regression ... 51

Table 7: Additional Results of the Survey Answers ... 51

Table 8: Relation of Cultural Dimensions and Home Bias from FDI Regression ... 53

Table 9: Relations of Cultural Dimensions and Home Bias from Survey Regression ... 56

6

List of Figures

Figure 1: Value Function of Prospect Theory ... 28

Figure 2: Cultural Impacts on Financial Decision Making (Bontempo, Bottom and Weber) ... 29

Figure 3: Cultural Impacts on Financial Decision Making (Weber and Hsee) ... 30

Figure 4: Conceptual Framework ... 32

7

List of Abbreviations

CIA Central Intelligence Agency EMH Efficient Market Hypotheses FDI Foreign Direct Investments GDP Gross Domestic Product

GLOBE Global Leadership and Organizational Behaviour Effectiveness Research Project GNI Gross National Income

8

Abstract

9

1. Introduction

In a modern business environment managers are facing an increasing pressure to internationalise their business in order to keep up with competition that is no longer local but global. The servicing of home markets alone is often not enough to satisfy stakeholders and achieve efficiency gains (Levis et al, 2010). Business managers are increasingly confronted with the task of which product to market internationally and where to market it best. A lot of international diversifications have failed because the target market and especially the customers were not understood well enough. An example in this context was Wal-Mart’s failure to enter the South Korean and German market. Gandolfi and Strach (2009) have examined Wal-Marts attempt to internationalise to South Korea and attributed the failure to the misjudgement of consumer preferences of the Korean people. A different culture can be one of the main obstacles for international business, but if understood and conceptualised can provide a competitive advantage over other companies (Tse et al., 1988). An attempt to better serve the consumption requirements of humans is to understand and predict their cognitive behaviour to situations as well as the preferences they have developed in light of their cultural background. Human choices or the question “Why do people want what they want” is fundamental to International Business Managers and marks the basis for cultural theory (Thompson et al, 1990).

10

literature on Rational Choice such as Kahneman and Tversky (1986) by adding culture to the descriptive model of choice.

1.1 Purpose of the Study

The purpose of this study is to theoretically establish and practically test a link between culturally based preferences, Bounded Rationality and resulting investment choices. Through theorising this link and testing an empirical connection, the thesis shows that culture has an influence specifically on investment choices, but in a wider sense on all economic choices humans undertake. This concept is underlined by the theory of Bounded Rationality to show that the view on preferences should be altered in mainstream economic theory, as it involves cognitive limitations. In this regard, the underlying research provokes a revision of the concept of Rational Choice. Additionally, culture should no longer be treated as exogenous to economic models but be integrated in the concept of decision making, in order to encourage the development of more accurate estimations of consumer behaviour.

1.2. Audience

11 1.3. Methods of Analysis

The method utilised to analyse the theorised correlations of the hypotheses is a quantitative approach. Data was collected by the means of a survey to test for a link between cultural dimensions and extent of Home Bias on an individual level. On a national level, macro data is regressed with the cultural dimensions as indicated by Project GLOBE. For this purpose Home Bias is substituted by the ratio of outbound Foreign Direct investments (OFDI) as percentage of Gross National Product (GDP) per country. Regression analyses were conducted through Microsoft Excel.

1.4. Plan of Work

12 1.5. Data Sources

Primary data was collected through the means of a survey that was developed for the purpose of this research and conducted online. The data for the regression of FDI with cultural dimensions is based on the year 2009 due to availability of data and stems from three sources. OFDIs per country and the population figures per country were provided by the CIA World Factbook. Gross National Income (GNI) per capita and GDP figures were provided by the World Bank. Dimensions of Culture and the relevant national results were provided by Project GLOBE.

1.6. Findings of Dissertation

13

2. The Influence of Culture and Bounded Rationality on Economic

Choices

A well- known concept on the goals of conducting business is profit maximisation, which has its origin in classical economic theory of Rational Choice and Utility Maximisation. Business managers need to conduct actions that have Profit Maximisation as their ultimate goal. Profit Maximisation converges with Utility Maximisation in the long- term (Formby and Millner, 1985). This is where the concept falls short of practical evidence that shows that businessmen rather satisfice than maximise (Simon, 1978). In organisational terms the “garbage can model” offers critique to rational decision models. It states that different streams of influence, such as decision opportunities and participants, enhance the probability that certain decisions are being made by intuition rather than approaching the decision fully rational (Cohen et al., 1972). The same principle holds true for consumers who do not distribute their purchases to maximise utility as evident from negatively sloping demand curves. To account for these deviations from theory, research has begun to focus on the ways in which decisions are made, not merely the outcomes of decisions. Experiments on human behaviour have added proof that Expected Utility Theory does not provide a reasonable approximation of behaviour (Simon, 1978). From a business point of view, the question of how humans arrive at decisions is central to not only managing an organisation, but to better understand consumer behaviour. Concepts that add to our understanding of this decision process are Bounded Rationality and preferences formed through culture. These concepts shall be clarified and linked in the extent of this section.

2.1. Culture and Economic Decision Making

14

fairness and reciprocity, even at the expense of their own payoffs. Fundamental questions in this regard are raised by Henrich et al. (2001):

“Are the deviations from the canonical model evidence of universal patterns of behavior, or do the individual’s economic and social environments shape behavior? If the latter, which economic and social conditions are involved? Is reciprocal behavior better explained statistically by individuals’ attributes such as their sex, age, or relative wealth, or by the attributes of the group to which the individuals belong? Are there cultures that approximate the canonical account of self-regarding behavior?” (p. 73)

To answer these questions that provide the basic theoretical foundation for this paper, Henrich et al. (2001) examine small scale societies. They find that differences in economic behaviour between societies stem from different preferences or expectations between groups. These inter- group differences are due to different social institutions or cultural fairness norms. From their work the assumption can be drawn, that culture influences economic decision making. The extent of this influence however, is the significant question. Will an incorporation of culture aid in rendering prediction models more precise or only add complexity with little accuracy benefits? To shed light on this question, this thesis further develops the concept by linking economic choices to specific cultural value traits and examining the link both theoretically and empirical.

The following section provides a brief definition of the term culture, to be followed by theories on how culture impacts economics by drawing on recent research.

2.1.1. Definition of the Term Culture

15

how groups and members of a society such as a nation can be distinguished. In this regard Hofstede (2011) sees it as the way people think, feel and act or the “unwritten rules of the social game”. To specify, national cultures have been found to differ through mostly unconscious values that a majority of a population share and that are different to the values of other populations. Values again can be described as “broad preferences for one state of affairs over others” (Hofstede, 1984, p. 389).These preferences are very stable over time as they are based on what we learn in childhood, requiring generations for them to change. Hofstede (1980, 2001), as one of the pioneers in this field of research, has developed an approach by establishing dimensions of key cultural differences between nations. Different nations scored differently on these dimensions on a scale from 0-100. Other researchers have picked up on these dimensions such as Schwartz (1990) and Triandis (1995). Project GLOBE is another important framework and the most comprehensive model up to date. The dimensions of Project GLOBE are an underlying part of this research and will be introduced in more detail in the process of this paper.

Hofstede’s work however, especially his methodology, has also received considerable critique. Some aspects of this critique need to be highlighted, as limitations of his approach to quantify values are also subject to this work, with regard to the utilised Project GLOBE framework:

Representativeness of questionnaires. From certain countries only a small amount of questionnaires were obtained, leaving the representativeness of the responses for the whole country under research in question.

16

independent of country and time of subsidiary existence, shares exactly the same organisational culture as all other IBM employees.

Steadiness over time. Occupational and national culture are imprinted into the mind in childhood and are from then on continuous and do not change. People who work in marketing in different countries still share the same occupational culture, no matter what and where they have studied or how they ended up in marketing.

Representativeness of respondents. In- country differences between responses lead to a national average, as no individual inhibits the one national culture. It is therefore hard if not impossible to find the one representative group for a national average. Hofstede claims he has a representative national average by drawing on samples from micro-environments.

Alternative explanation. Hofstede does not include variables that may additionally be responsible for answers, except for culture. Variables could have been age, race, gender and education. Instead Hofstede attributes all answer differences to national culture (McSweeney, 2002).

The above listed points reflect the critique on Hofstede’s work, but also on value quantification in general, which is deemed important for the method and results of this paper. The critique is also relevant to Project GLOBE which uses similar methodology and will be incorporated and further discussed in the methodology and discussion section.

2.1.2. The Link between Culture and Economic Choices: Two Perspectives

17

“For economists to rest a large part of their theory of choice on differences in tastes is disturbing since they admittedly have no useful theory of the formation of tastes, nor can they rely on a well-developed theory of tastes from any other discipline in the social sciences, since none exists. (…) The weakness in the received theory of choice, then, is the extent to which it relies on differences in tastes to "explain" behavior when it can neither explain how tastes are formed nor predict their effects (p. 380).”

18 2.1.2.1. The Macro Perspective

A prominent measure to test the influence of culture on economics has been to compare different nations’ economic indicators with scores on the dimensions of culture that were identified for these countries through surveys (Dodor and Rana, 2007; Papamarcos and Watson; 2006; Gorodnichenco and Roland, 2011). While a common concept for explaining differences between countries’ economic performance are different stages of development in time, where eventually all poorer nations catch up to developed nations, structuralists have critiqued this approach. Structuralists claim that some nations are not simply in a different stage of development, but feature distinct structural differences that distinguish them from other nations in the long term (Dodor and Rana, 2007). Part of these different structures stem from different formal and informal institutions that are rooted within national culture and are steady over time.

19

environment, quicker reaction due to the predominant hierarchical structure and an ethic of wealth maximisation by continuous improvements. Furthermore, societies that take advantages of work possibilities and inhibit a climate of material consumption will increase overall production. Institutions that aid free trade and penalise free riders and corruption will also benefit economic wealth. Another point that overlaps partly with the individual level is trust and cooperation development of societies, familiar under the concept of “social capital” (Fukuyama, 2001). Societies that encourage trust and cooperation will be more likely to be able to govern themselves without central hierarchy and encourage and facilitate economic activity.

The macro perspective measures societal or national attributes and relates them to economic activity. For example, a society that has institutions which punish free riders is more successful in economic terms over a nation or society without these institutions. In order to determine how individuals that share cultural value traits arrive at these national or cultural attributes, i.e. institutions that punish free riders, a look on the individual level is necessary.

2.1.2.2. The Micro Perspective

By using the concept of micro perspective, this section reasons how culture influences decisions on the individual level that result in economic outcomes. For this purpose two important factors, namely trust as result of prior beliefs and preferences are theorised. Trust and prior beliefs are influenced by culture within a society and impact business activity as Arrow (1972) stated:

“Virtually every commercial transaction has within itself an element of trust, certainly any transaction conducted over a period of time” (p. 357).

20 Trust and Prior Beliefs

Trust is one of the key variables in the link between culture and economics on an individual level. Different cultures have developed different levels of trust towards members of their own society and members of other societies and can be defined as “the subjective probability with which an agent assesses that another agent or group of agents will perform a particular action” (Gambetta, 2000; p. 217) or “Trust is the expectation that arises within a community of regular, honest, and cooperative behavior, based on commonly shared norms, on the part of other members of that community” (Fukuyama, 1995; p. 26). Trust in this sense is of sociological origin as it is necessary and beneficial only in social interactions as alternative to rational prediction in order to minimise complexity. When rational predictions are too complex to construct, trust offers a way around predictions by living as if certain rational futures will not occur, hence reducing complexity in the decision making process. From an economic perspective, trust needs to be differentiated between interpersonal and systems trust. In a small scale society, interpersonal trust is based on personal knowledge and experience of the opposite, rather guided by emotions. Trusting the trading partner to act in a specific manner can serve as a basis of conducting business. When societies grow and increase in complexity, interpersonal trust is often no longer a sufficient means to conduct business, as the pace of business interaction increases and new business partners are unfamiliar on a personal level. Common regulations and ultimately institutions are set up to enforce regulation and facilitate impersonal trading with the means to prosecute unfavourable behaviour. People then develop trust towards that system, based on rather cognitive calculations on the effectiveness of the system (Lewis and Weigert, 1985).

21

In “The Moral basis of a backward society”, Banfield (1958) relates a strong form of self-interest in southern Italy as a reason for its economic underdevelopment. Northern Italy on the other hand had a deep routed tradition of free city trade since the middle-ages that required the evolution of impersonal trust in order to trade with strangers. Southern- Italy was not largely involved in such a system and lacked the evolution of trust that extended over the near social community (Putnam, 1993; Fukuyama, 1995). The cultural differences between the North and the South show how different levels of trust are formed. To extent, culture led to the foundations of institutions that enhance trust towards the success of individual trades. Collective punishment systems between trade intensive cities in the North are examples of these institutions (Greif, 2006).

Trust is very much influenced by acquired prejudices. Parents, amongst others, share learned knowledge with their children, which also includes lessons on who to trust. Taking these prior beliefs to a national level, deep rooted prejudices against other countries exist, often due to historical events that influence the level of trust between cultures and are visible in the actions of individuals. Guiso et al. (2004) have examined this phenomenon and find that culture largely impacts trust and trust again impacts economic exchange between countries. Knack and Keefer (1996) confirm and find that countries that encourage trust between its inhabitants tend to have an increased level of inter- trade and therefore economic performance. In order to link the concept of systems trust to capital markets, Guiso et al. (2008) contribute important findings. They examine trust as a cultural based factor in relation to stock investment and find that the level of trust is an explanation why certain people participate in the stock market and others choose not to.

Preferences

22

nutrition. These basic needs, although expressed in different eating habits, are the same to all humans, simply due to the need for survival. Other preferences differ between different groups of humans, because these preferences have developed within their culture and may be derived from biological aspects. Our preferences or interests are therefore due to social relations we take part in (Wildavsky, 1987). A debate on preferences exists in the question if these should be exogenous or endogenous to economics which again is at the basis of this research. Often economists regard preferences as exogenous or external to the system that is under research as these cannot be analysed. Stigler and Becker (1977), who debate this point, clearly express the economic opinion on preferences or tastes:

“(...) one does not argue about tastes for the same reason that one does not argue over the Rocky Mountains – both are there, will be there next year, too, and are the same to all men” (p. 76).

23

2.2. Bounded Rationality and Economic Decision Making

The previous section emphasised cultural influences on human decision making and linked them to economic choices. Another aspect that has only recently been introduced to economics is human cognition. Not only culture, but psychology aids in explaining how humans arrive at choices. This is of special relevance in an international business context: To differentiate between the choices we make, because individuals are homogenous through cognitive capabilities, or the choices we make because we are part of a collective that has common preferences and prior beliefs. The latter part aims at cultural influences on our economic decisions, the first part at human cognitive capabilities. Both concepts are exogenous to mainstream economic choice theory, resulting in poor prediction of actual outcomes.

The following section is dedicated to human cognition to depict how cognitive limitations are responsible for the decisions we make. I will begin by looking at the Efficient Market Hypothesis (EMH) as part of financial theory to arrive at the concept of cognitive limitations and resulting bias towards economic choices. I will then link the described fields of culture and cognitive limitations by drawing on Prospect Theory.

2.2.1. The Efficient Market Hypothesis as Proxy Model

24

The EMH explains stages of market efficiency and is used as the basis for many financial models. The theory states that in efficient markets all available information is incorporated into share prices and shares will always trade at their equilibrium in relation to available information. When new information arrives, investors will immediately process this information and the share price will move into a new equilibrium. Therefore, for every piece of information, one state of equilibrium exists, ergo one price (Fama, 1970). Henceforth, no investor is able to outperform the market in the long term, as he does not have access to superior information in relation to other investors. All financial actors always have the same access to information regarding their stocks and interpret this information similar. However, this field of study yields a big gap to practice and is subject to widespread criticism. If all actors on financial markets accepted the EMH to hold true, there would be no actively managed investment funds, as no fund can achieve more return for their investors than the investors themselves could. Still, there are thousands of investment funds, earning billions of dollars every year in management fees because investors do believe that fund managers can outperform markets due to superior access and a more thorough evaluation of information. This gap has been acknowledged in the academic world and over the last twenty years research has begun to focus on the inefficiency of markets and behavioural finance (Malkiel, 2003). In this regard, Robert Shiller (2002) can be considered a pioneer in the financial academic world by stating that share prices are determined by their discounted future dividend payments which have been relatively constant in the past century. Yet, share prices are very volatile, even without new information available on the markets.

25

between utility maximising humans, as foundation for models such as the EMH and psychological boundaries in decision making.

2.2.2. From Utility Maximisation to Prospect Theory

In this section I briefly describe the economic reasoning on how choices are formed, namely Expected Utility and Utility Maximisation Theory. The shortcomings of these theories are explained by drawing on cognitive biases that describe why humans do not always act economically rational. Prospect Theory takes these limitations into account and describes that humans are all subject to limitations when confronted with specific dilemmas involving risk and that it is possible to incorporate these findings into models. The section on Prospect Theory then aids in linking the factors together: Bounded Rationality and culture.

2.2.2.1. Expected Utility and Utility Maximisation Theory

Human Rationality is based on the axioms of von Neumann and Morgenstern (1944) which provide the basis for the Expected Utility Theory of microeconomics. This theory implies that humans act rational and have individual risk preferences. Their overall goal, no matter what their personal risk preference is, is to maximise their utility. Individuals attach different personal utilities to specific goods and therefore people will inhibit different preferences for different goods. Expected Utility Theory adds uncertainty in choice situations that can be regarded as lotteries. Here, people are confronted with different probabilities of outcomes and have to assign utilities to those outcomes. In general, people will have differing utility functions, but all still act rational according to their individual risk preferences, weighing all alternative options and choosing the option that maximises personal utility. Relating this theory to the above described EMH, a rational investor will know his1 risk ability, as in how much risk he is able to be

1

26

exposed to. On the basis of this knowledge, the investor adjusts his portfolio to optimize it for his given risk. As new information arrives on the market, an investor will adjust his portfolio according to new information on risk and return with his given level of risk aversion (Hens and Wang, 2007).

Utility Maximisation Theory is part of Rational Choice Theory. To this theory the questions of “Why we want what we want”, meaning the origin of human motivation for our action, is extrinsic. Instead it focuses on costs of our actions and their payoffs, making humans payoff maximisers. This normative approach describes the model of an idealised decision maker and stands in opposition to describing the real decision making behaviour of humans (Kahneman and Tversky, 1986). The critique of this theory is nicely pictured in the explanation of Utility Maximisation Theory by Grant and van Zandt (2007), who use the following sentence:

“We ignore much of the information processing that Anna must do in order to make a decision.” (p. 3)

When looking at the decision the example person Anna can make, we have to know her utility function in order to understand or predict her decision making process. However the decision making process that Anna undergoes to arrive at her perceived utility does not necessarily reflect rational behaviour. Her choices are influenced by cognitive reasoning, which limit the economic rationality with which people can respond to economic situations. This cognitive reasoning is subject to cognitive biases.

2.2.2.2. Cognitive Biases

27

confirm rather than disconfirm their inferences. Also, they over-judge the accuracy of their judgements as memories of positive judgements overrule those of negative ones (Engel, 2010). These limitations are needed to deal with complexity and ambiguity by simplifying, in order to arrive at quick, sometimes life- saving decisions similar to the concept of trust which was explained in Chapter 2.1.2.2. However those cognitive processes often result in faulty decision making, known as cognitive biases. Subconscious mental procedures and not emotional or intellectual predisposition are responsible for these biases. Important aspects of these mental errors are their consistency and predictability (Heuer, 1999). This systematic of mental errors provide important implications for their utilisation in decision making models. In academic terms, the concept is known as Bounded Rationality, where humans tend to arrive at a satisfactory result given their cognitive restrictions rather than at a maximising one, as described in Rational Choice Theory (Simon, 1956). The implications of these findings are stated by Kahneman and Tversky (1986):

“(…) the deviations of actual behavior from the normative model are too widespread to be ignored, too systematic to be dismissed as random error, and too fundamental to be accommodated by relaxing the normative system” ( p. S252).

2.2.2.3. Prospect Theory

28

humans are indifferent to the reference point. Therefore, this example shows that humans are biased when making decisions.

Figure 1: Value Function of Prospect Theory

2.3. Linking Culture and Bounded Rationality

Cognitive biases are common to all humans but humans are segmented from each other by values, norms and beliefs that are rooted within their culture. However, culture can also impact on cognitive biases, like loss aversion. A logical conclusion therefore constitutes a link between cognitive biases and culture. A promising field of research to demonstrate this is Prospect Theory. If cognitive limitations shift or create preferences similar to cultural value traits, both must influence decisions.

29

Additionally, Bontempo, Bottom and Weber (1997) conducted a study on the risk aversion of students from Hong Kong, Taiwan, the USA and the Netherlands. They found that the risk levels of Asian students depend on the level of the losses and not on the gains, as compared to the western participants. Prospect Theory yields a graphic image of these cultural differences:

Figure 2: Cultural Impacts on Financial Decision Making (Bontempo, Bottom and Weber)

30

Figure 3: Cultural Impacts on Financial Decision Making (Weber and Hsee)

31

3. Conceptual Framework

In the theoretical background I argued that there are influences on the decision making process that are not part of Utility Maximisation Theory. Culture, as an explanation for how humans arrive at preferences and why they differ between social groups is regarded as an extrinsic variable and Bounded Rationality is considered contradictory to mainstream economic theory. I linked both factors utilising Prospect Theory, showing how they can be combined to increase our understanding of human choices. As described, understanding human choices increases the justification and efficiency of management and business decisions, especially in an international and culturally diverse context.

The theoretical background was constructed twofold. The first section, Chapter 2.1., aimed at explaining the influence of culture on economic decision making. The second section, Chapter 2.2., illustrated the concept of Bounded Rationality, as adaptation to the view on how decisions are made by drawing on the assumption of fully rational investors and the EMH. This Bounded Rationality stems from cognitive limitations that lead to behavioural biases.

The two sections were combined by using Prospect Theory in Chapter 2.3., to emphasise that cognitive biases are universal to humans, but their extent may differ according to the cultural background of decision makers. The aim of this paper, after having linked the concepts theoretically, is to show that the assumption of influence on economic decisions not only holds true for specific relations of biases and culture, such as loss aversion, but also on a generalizable level. While an influence of cognitive limitations on investment decisions has been researched in behavioural finance, the influence of culture remains under-researched. To incorporate Bounded Rationality into the conceptual model of this thesis, the influence of cultural value traits on investment preferences are researched on an investment phenomenon, which is subject to Bounded Rationality.

The Main Research Question is therefore:

32

Providing an answer to the research question is attempted by statistical regressions on the micro and macro level. Home Bias is selected for both regressions as an example investment phenomenon for cognitive biases. The impact of cultural dimensions upon Home Bias is measured, utilising the framework of Project GLOBE for both regressions. Research in this regard has been conducted by Beugelsdijk and Frijns (2010) and Anderson et al. (2010) who find an effect of culture on Home Bias. However, it remains unclear if this influence is generally measurable outside of fund level research and how strong and persistent this influence is, given different dimensions of culture. Furthermore another attempt of this analysis is to answer if the relation is measurable both at the individual and at the macroeconomic level. This closed research question is analysed by reasoning deductively what dimensions of culture influence investment decisions, specifically how the Project GLOBE dimensions relate to the extent of Home Bias. An overview on the aim of this paper is depicted below in the conceptual model, which was theorised in the literature review by linking recent research. In the further process the link between culture and preferences is theorised in detail with the help of cultural dimensions of Project GLOBE. This relation as well as cognitive limitations is tested empirically on the micro and on the macro level to arrive at findings that answer the research question.

Cognitive Limitations

Culture

Preferences Investment Decisions

• Values • Norms • Prior Beliefs Dimensions of Culture Bounded Rationality

33 3.1. Home Bias and Diversification Theory

The example of Home Bias was chosen as it is a well- known financial phenomenon associated with a cognitive bias. It is well researched and can be explained with the help of financial theory as to why it remains a common bias even though its inherent logic is well known to international investors. This section is dedicated to explaining the phenomenon of Home Bias in the field of Bounded Rationality in order to provide a theoretical foundation for the development of the hypotheses and their further testing.

“Don’t put all the eggs in the same basket” translates to Asset Diversification or Portfolio Theory; a concept of great importance in finance. Diversification reduces the risk in a portfolio by spreading assets over industries and countries. If assets in a portfolio are not perfectly positively correlated, the standard deviation or risk will be lower, than each asset held seperatly. Ergo any investor should seek to minimise standard deviation given the same expected return (Stulz, 1995), in line with Utility Maximisation Theory and the EMH. As international markets are only correlated to a certain extent, cross- country diversification is vital. However, studies have shown that most investors do not take advantage of risk reduction and exclusively or invest in their home country or have a major percentage of assets there. This phenomenon is known as Home Bias (Chan et al., 2005). According to Huchzermeier (2005), there are several explanations for the occurrence of this bias:

Home country information advantage

34

Relative Optimism

Investors will perceive local shares with greater potential gains than foreign shares. This is often associated with patriotism, where investors believe in their countries supreme capabilities (Huchzermeier, 2005). In this regard relative optimism must also be attributable to cultural dimensions, as patriotism is a form of in-group collectivism.

Psychological Approach

In conjunction with relative optimism, investors have a loss aversion as depicted by the graphs in the recent part on Prospect Theory. As humans have a cognitive flaw that leads them to herd2, they are especially loss averse in relation to other humans. When they lose but others lose even more, the loss is considered to be less strong. This leads to benchmark thinking: Trading decisions are no longer based on fixed gains or losses but on relative gains and losses. This approach comes closest to the argumentation of this paper, as it involves cognitive limitations.

Institutional Approach

Investing in foreign countries requires dealing with different host institutions. This involves political- and exchange rate exposure, increased transaction complexity resulting in higher costs, and therefore increases the risk compared to home investments (Huchzermeier, 2005).

In the following part a detailed causal relationship is established theoretically by reasoning on the link between cultural dimensions and their effect on Home Bias. For this purpose a hypothesis for each of the nine cultural dimensions of Project GLOBE is derived, which is tested with the help of statistical measures.

2

35

3.2. Dimensions of Culture and Economic preferences: Borrowing from Project GLOBE

GLOBE (Global Leadership and Organisational Behaviour Effectiveness) is a multi- phase international research program incorporating different methods to study cross- cultural effectiveness. It focusses on culture and leadership in 62 nations and its main result is that leadership effectiveness is contextual and embedded within the value, norms and believes of a society or organisation. The study is build up on prior work by Hofstede (1980), Schwartz (1994), Smith and Peterson (1995) and Inglehart (1997) amongst others and marks the most comprehensive framework regarding the quantitative measures of cultural value perceptions up to date. GLOBE has identified distinct value traits in the extent of their research that provide a quantification of different value traits between nations worldwide. Therefore, I utilise the questions and results of the GLOBE framework for the quantitative research and theorise the channels of influence between cultural value traits and Home Bias in the following section. The limitations associated with the GLOBE framework impact the underlying research in this paper. Due to the lack of a better model and for practical reasons this paper will utilise the GLOBE framework and draw on the limitations in the discussion.

GLOBE has identified nine dimensions of national cultural differences as the results of a survey on thousands of middle managers from different industries (House et al., 2004). The dimensions and findings form the basis to theorise the effects on economic preferences. The nine dimensions consist of:

36 8. Uncertainty Avoidance

9. Gender Egalitarianism

Performance Orientation

Performance orientation depicts to what extent members of a society reward performance and excellence of individuals within that society (House et al., 2004). This relates strongly to individualism, as members of a society are encouraged to provide results that are outstanding compared to other members. In this regard, the competition within a social group is encouraged. Performance can be attributed to dedicating more work and resources towards an objective than others. In an economic perspective this translates to efficiency. The more efficiency is achieved at a constant level of input, the higher the level of output. Specifically in financial terms, better performance can often be achieved by taking more risks than others, as more risk should be rewarded with increased reward. Members of a collective cannot be financially outstanding by buying the same assets as other members of that group. They will need to buy assets of a higher risk class or invest in assets with the same reward but less risk. This can be achieved by selecting higher risk stocks from the same market, or on the other hand investors can choose to minimise risk with a similar reward by diversification, ultimately resulting in higher performance. This diversification can be achieved by spreading the investments across different countries. On the example of Home Bias, investors of a high performance oriented culture will be more likely to hold assets from other countries. To include behavioural aspects, people with a high performance orientation will hold assets from countries other than their own, not for reasons of diversification, but because they believe foreign stock are related to higher risk and therefore higher reward. Concluding, I theorise that performance orientation will have a large impact on the level of Home Bias.

37

Future Orientation

This dimension describes how individuals of a society engage in future orientated actions, like savings and expenses for education (House et al., 2004). In this context investments are a way to enhance future outcomes by delaying gratification. College funds are an example of securing future education by dedicating resources today. In this regard future oriented societies are more likely to participate in financial markets to save for future consumption, also increasing the probability of members investing in stocks other than local ones. Nevertheless, this dimension gives no indication of the risk perception of a society. Thus it cannot be stated what kind of assets these societies are likely to buy. Future orientation may be attributable to conservative investments to have more certainty over future outcomes. From a rational choice point of view, investors will be more likely to spread assets across the globe to minimise risks. From my line of argumentation the opposite may also hold true. According to behavioural finance, people will not spread the assets due to a false perception of risk or information asymmetries. That would mean that a future oriented society that invests conservatively will rather invest in home stocks, as they believe that less risk is associated with these investments. On the contrary, a future oriented society that is risk seeking will invest in global stock, because they believe they can achieve more returns on the global market due to the tolerance of additional risk, even though they will ultimately decrease risk by these actions through diversification. In this line of thought, future orientation yields a larger probability of investments participation but provides no clear indication of the risk preferences of that society. Hence no link to the extent of Home Bias can be drawn.

H2: Future Orientation is unrelated to the level of Home Bias

Assertiveness

38

culture but can be regarded problematical by members of other societies, especially consensus based cultures. However, I argue that assertiveness will have little effect on the investment behaviour of individuals. A possible influence would be the behaviour towards ill performance of an asset and may result in emotional based decisions rather than rational choice. Meaning the investors may be more likely to continuously switch ill performing assets. However, this does not influence diversification choices between countries and has no effect on Home Bias.

H3: The level of Assertiveness of a culture is unrelated to the level of Home Bias.

Power Distance

Power distance measures the extent to which individuals of a society expect and agree that power should be distributed unequally (House et al., 2004). In economic terms, power is not only of political nature but includes power over resources. Therefore, a large percentage of wealth may be in the hands of a few in societies high on Power Distance. However, this does not have immediate implications for diversification choices of individuals. It also does not tell if individuals are less likely to engage in financial markets. Therefore no conclusions can be drawn regarding Home Bias through this dimension.

H4: The extent of Power Distance is unrelated to the level of Home Bias.

Humane Orientation

39

the third world, therefore reducing Home Bias, although not out of economic rationality. The products for which this argument holds true are very specific and since the purchase of these products has little personal gain motivation, the impact will hardly be measurable, as they are largely independent of other financial products. In terms of investment funds, this may relate to a decreased level of assets regarded as unethical private portfolios, compared to nations scoring low on humane orientation. However, ethical investing will have no measurable effect on diversification.

H5: The level of humane orientation is unrelated to the level of Home Bias.

Institutional Collectivism

Institutional or societal collectivism measures how organisational or societal institutions reward an even distribution of resources within the collective (House et al., 2004). The more even resources are spread within a society, the higher the percentage of members of a society to potentially generate excess returns. These excess returns leave the possibility to invest some or all of the returns that are not spend on consumption in available investment vehicles. This will increase the general percentage of nationals that invest in stocks and therefore also results in a higher probability that nationals will make decisions to invest in foreign stock. However, this does not shift preferences of individuals and is only a question of large numbers. The effect may be visible in certain macro analysis, but not on a group level.

H6: The level of institutional collectivism is unrelated to Home Bias

In-group Collectivism

40

part of, being loyal to their interest (Ball, 2001). Expressing loyalty to the in- group also refers to an attitude of supporting that in- group economically. This translates to collectivists primarily buying products that are produced locally and supports the own economy, because indirectly this supports the own family and the larger collective. With strengthening the collective, eventually the individual is strengthened, compared to other groups. With regard to Home Bias, nations scoring high on in-group collectivism will also primarily buy stocks from the home market, resembling an economic and financial patriotism. This assumption is supported by Morse and Shiver (2003) who, by examining Home Bias, arrive at patriotism as one explaining factor. They state that ethnocentrism, for example buying local goods, changes worldwide production systems and equilibrium. This equilibrium, which is according to Adam Smith created by the invisible hands of the market, is influenced by culture; and not exclusively by rational decision making. Morse and Shiver’s research therefore “contributes to the growing evidence that investment decisions depend on societal characteristics other than aggregate expectations and risk preferences” (p.5). In this line of argumentation the following hypothesis is derived:

H7: The higher the degree of in-group collectivism, the higher the extent of Home Bias.

Uncertainty Avoidance

41

terms of economic implications, members of a group scoring low on uncertainty avoidance are more willing to take risks and may therefore be more progressive in economic terms. This is aided by the reasoning that the economy in a low uncertainty avoidance environment has more opportunities to unfold, as competition is appreciated and fewer rules apply to regulate it. Societies with higher uncertainty avoidance will be more risk averse and the economy will evolve slower as the focus is set on maintaining a secure lifestyle on the basis of the status quo. Papamarcos and Watson (2006) have conducted empirical research into the above described relation between a nation’s uncertainty avoidance and economic performance and have found a negative relation, meaning the higher the uncertainty avoidance of a country, the lower their economic performance. In terms of investment decisions, risk tolerance is one of the key concepts. From a behavioural economics perspective, individuals with a low risk tolerance will have a conservative investment style, meaning they will invest more in products they trust, i.e. familiar local products. However, from a rational economics point of view they would minimise their risk by diversification. Vice versa, individuals of risk tolerant cultures are more likely to diversify as they expect more performance from accepting more risk that is associated with investing abroad. Since Bounded Rationality seems to offer a more adequate theory on investment decisions, this thesis expects a strong positive relation.

H8: Higher levels of uncertainty avoidance will result in higher levels of Home Bias.

Gender Egalitarianism

42

percentage of home country investments will remain unchanged. In this paper it is not expected that there are significant preference differences between the genders concerning diversification decisions.

H9: Gender Egalitarianism is unrelated to levels of Home Bias.

Empirically testing the hypotheses intends to answer the call of Ivkovic and Weisbenner (2007) to explore “the role of societal characteristics in portfolio decisions” (p. 1356), by not only establishing a statistical relationship but determining the magnitude of effects. The following table provides an overview of the nine hypotheses and the theorised relation towards Home Bias.

Table 1: Overview of Hypotheses

GLOBE Dimension of Culture Influence on extent of Home Bias

1. Performance Orientation - - -

2. Future Orientation Unrelated

3. Assertiveness Unrelated

4. Power Distance Unrelated

5. Humane Orientation Unrelated

6. Institutional Collectivism Unrelated

7. In- Group Collectivism +

8. Uncertainty Avoidance +++

43

4. Data and Method

The above theorised hypotheses are tested through quantitative methods that constitute two separate sections. The first section contains a macro approach to the hypotheses, testing whether these hold up on a cross country regression of Outbound Foreign Direct Investments (OFDI) per country as proxy for Home Bias. The approach utilises the Project GLOBE responses per country, as found through the original Project GLOBE surveys. The second section aims at testing the hypotheses at the individual level, through a multiple regression analysis of the answers to a conducted survey. This approach utilises the questions from the Project GLOBE survey aimed at quantifying cultural dimensions, to identify cultural value traits of the respondents. The two approaches are compared and analysed for differences to provide an answer to the research question. In the figure below, the utilised methods are summarised and the following section provides a detailed overview on the structure and purpose of the utilised methods.

44

4.1. Cross- Country Outbound FDI Regression

I elaborated that Bounded Rationality impacts investment decisions and chosen the phenomenon of Home Bias to demonstrate this. The extent of Home Bias was theorised to be influenced by cultural value traits as depicted by the dimensions of Project GLOBE in the hypotheses. These hypotheses are tested through macro data on Home Bias, namely OFDI as percentage of Gross Domestic Product (GDP). Whereas Home Bias is often associated with limited global diversification of publicly listed companies’ shares, an investment Home Bias can be regarded in a wider sense. If Home Bias is associated with a limited purchase of shares from foreign companies, another aspect of Home Bias related investments is the purchase of stock in non- publicly listed companies (Levis et al., 2010). These direct investments in at least 10% of the voting stock, according to standard definition, are mostly conducted by companies due to the size of the deals and the adjoining transaction costs (The World Bank, 2011). However, no difference between individual and institutional investors exists regarding Home Bias (Coval and Moskowitz, 1999; Ivkovic and Weisbrenner, 2005). Therefore, Outbound FDI is chosen as an equal measure to show different levels of overseas investments on the country level. Kogut and Singh (1988) have theorised on the link between culture and FDI and confirm a relationship. On these grounds this thesis takes a unique approach testing the influence of cultural value traits by drawing on OFDI as investment proxy for Home Bias.

45

the possibility that rising wealth does influence the preferences and therefore the investment style. It is assumed that individuals with greater wealth have more funds to spare for savings, but that this does not change with the percentage of excess money for saving (Dunning, 1981; 1991; 1998). There are other variables that influence the level of OFDI such as tax rates, market integration, availability of investment products and standard of institutions and education (OECD, 2000). However, culture must have an impact on all of these variables, as culture impacts the economic variables on a large scale. This was described in the literature section by drawing on trust and institutions that form, depending on the prevailing cultural values. The analysis of an indirect impact of culture on OFDI via the described variables is beyond the feasibility of this work and theorised as given.

Internal validity is of an issue in this regression, as it does not seem feasible to establish a control group that is independent of cultural value traits. To increase internal validity, countries were included that have low levels of OFDI and are not part of the developed world. The amount of countries is limited in general through the availability of data from GLOBE and availability of macro data, especially GNI. A total of 47 countries were included in the regression. The standard formula for multiple linear regressions was utilised, as depicted below, where γ represents the dependent variable of OFDI in relation to GDP of a country. The assumption holds that the larger the percentage of OFDI to GDP, the lower the prevailing Home Bias in a specific nation. X represents the explanatory variables, which are dimensions of culture from Project GLOBE, as well as the above mentioned control variables. The β’s are the regression coefficients, determining the relation between the explanatory variables and the dependent variable. The error term ε capturers all other factors apart from x, that influence γ.

46

4.2. Survey for the Individual Level Regression

The second section of the analysis tests for a link between cultural value traits and Home Bias on an individual level through collected survey data. The main purpose of the survey is to find statistical significance between the responses to culture related questions from the GLOBE framework and investment preferences. In order to test the validity of loss aversion from Prospect Theory, questions regarding the individual risk perceptions were also included. To control for internal validity, age, gender, education and nationality were required answers. The following section provides an overview of the questionnaire design, the three sections of the survey and explains how the questions aid in finding a statistical pattern between the researched variables and therefore answering the research question.

4.2.1. Survey Design

The survey is designed as an analytical survey. The purpose is to collect data as input for the multiple regression analysis on the individual level to compare it to the larger population of the first section (OECD, 2003). Therefore, the questions are set up in a closed manner to scale the answers on scores, as in how much the participants agree to a certain statement. The required data for the regression are the following variables.

Table 2: Required Survey Regression Variables

Variable Name

Explanatory Variables Cultural Value Traits

Dependent variable Level of Home Bias

Controlled Variables Age, Gender, Education and Nationality

47

The survey was set up according to general survey standards (Wimmer and Dominick, 2011; Oppenheim, 1996), taking ethical considerations into account. The participation was voluntary for all participants and set up in English. No duplicate answers were possible through software mechanisms that are not under the control of the questionnaire designer. The survey was conducted online via a standard questionnaire program, so answers could not be altered and no influence on the answers could be undertaken by any party. The questions were set up using examples to avoid misunderstandings and the examples were provided two fold, to not influence respondents by example numbers. The survey was conducted anonymously and respondents were selected randomly, although there is a strong cluster regarding certain nationalities, age and level of education that do limit the assumption of a random sample. All 247 survey responses were included in the analysis.

As mentioned in the literature review, Hofstede’s method of research is subject to criticism in terms of representativeness. In order to limit representativeness issues and ecological fallacy, nationality was not included as variable in the regression. The reason is mainly that certain nationalities are dominant in the total percentage of provided answers, whereas only few answers from certain other countries are available. Cultural value traits and their link to Home Bias were regressed including the responses from all countries, but without differentiating between nationalities. Other issues regarding the selection of the peer group and non- randomness are addressed in the discussion and limitations section.

4.2.2. Risk Preferences

48

yields results on the general validity of behavioural bias towards investment decisions, apart from the specific example of Home Bias.

4.2.3. Investment Preferences or Extent of Home Bias

The participants in the questionnaire were confronted with a question demanding to make a choice on the percentage of preferred home country investment of a fund. The available choices are limited to the respondents’ home country and the rest of the world and only percentages of diversification can be chosen. The fund question does not yield information such as fund size, emitting bank, name, investment style, etc. and excludes these choices from the survey. The results of these answers aim at providing a pattern between the participants’ cultural value traits and their choice of diversification or Home Bias. The results also show if respondents are subject to Bounded Rationality in investment choices, pertaining the specific example of Home Bias.

4.2.4. Cultural Dimensions

49

5. Empirical Results

The results obtained from both regressions are presented in the following section. They constitute information on the coefficient of determination (R2), as more variables were introduced. The regressions were first conducted without control variables, introducing all cultural dimensions at once to the selected Home Bias. Then control variable were added, testing for the change in R2. Ultimately, multiple regressions were conducted introducing one cultural dimension at the time including the control variables that offered the highest R2.

5.1. FDI Regression Results

The regression model of GLOBE dimensions and their influence on Home Bias offers an R2 of 0.194. The introduction of the control variable GDP per capita increases R2 to 0.238, whereas population decreases R2 to 0.218. The best approximation constitutes a model of only GNI per capita as control variable with a R2 of 0.249. The regression of each cultural variable with the control variables results as depicted below. Using only GNI as control variable did not alter the results significantly:

Table 3: Results of the Multi Model FDI Regression

Cultural Dimension Coefficient Standard Error

50

A regression including all variables in one model with GNI as single control variable yields the following results:

Table 4: Results of the Single Model FDI Regression

Cultural Dimension Coefficient Standard Error

Uncertainty Avoidance 0.066 0.288 Future Orientation -0.154 0.363 Power Distance -0.010 0.350 Collectivism I 0.009 0.284 Humane Orientation -0.069 0.257 Performance Orientation 0.643*** 0.347 Collectivism II 0.141 0.243 Gender Egalitarianism 0.543*** 0.290 Assertiveness 0.497 0.331

GNI per Capita ($ thds) 0.015*** 0.008

Number of Observations 47

Adj. Coefficient of determination

0.234

5.2. Survey Regression Results

The regression of Home Bias with all nine cultural dimensions, excluding control variables yields an R2 of -0.004 offering conclusions on the quality of the explanatory model. The inclusion of control variables increases R2 with each selected variable, as indicated in the following table.

Table 5: Adjusted Coefficients of Determination of Control Variables

R2 introducing Control Variables R2 R2 adding Age 0.027 R2 adding Age and Gender 0.030

51

A single model regression of all variables did not shift significance of cultural dimensions, compared to a multi regression model, where each dimension was introduced one at the time. The result of the multi model regression including all control variables is depicted below.

Table 6: Results of Multi Model Survey Regression Cultural Dimension

Cultural Dimension Cultural Dimension

Cultural Dimension CoefficientCoefficientCoefficientCoefficient Standard Standard Standard Standard Error Error Error Error Power Distance Power Distance Power Distance Power Distance -1.23** 0.582 Uncertainty Avoidance Uncertainty Avoidance Uncertainty Avoidance Uncertainty Avoidance -0.494 0.664 Humane Orientation Humane Orientation Humane Orientation Humane Orientation 0.806 0.651 Collectivism I Collectivism I Collectivism I Collectivism I 0.094 0.657 Collectivi Collectivi Collectivi Collectivissssm IIm IIm II m II -0.403 0.676 Assertiveness Assertiveness Assertiveness Assertiveness -0.568 0.635 Gender Egalitarianism Gender Egalitarianism Gender Egalitarianism Gender Egalitarianism 0.163 0.541 Future Orientation Future Orientation Future Orientation Future Orientation 0.252 0.556 Performance Orientation Performance Orientation Performance Orientation Performance Orientation -0.243 0.698 Number of Observations Number of Observations Number of Observations Number of Observations 247

Table 7: Additional Results of the Survey Answers Answers on additional Questions

Answers on additional QuestionsAnswers on additional Questions

Answers on additional Questions AverageAverageAverageAverage Home Bias

Home BiasHome Bias

Home Bias 62.310

Confronted with Gains Confronted with GainsConfronted with Gains

Confronted with Gains 50.818 Confronted with Losses

Confronted with LossesConfronted with Losses

Confronted with Losses 53.998 Deviation from rational 0

Deviation from rational 0Deviation from rational 0

52

6. Discussion and Conclusion

The results of both regressions yield contradicting results, which is partly expected due to ecological fallacy, but yields important findings and assumptions for further analysis. The results are discussed separately and then compared to discuss the findings. The aim and results of the thesis are then summarised in the conclusion, including limitations and practical implications.

6.1. Culture and Home Bias at the Macro Level

The FDI regression yielded a better R2 than the individual level regression and a link between performance orientation and assertiveness to the extent of Home Bias, both on a 10% significance level and both having an equally strong negative effect on Home Bias. For both dimensions, a one point increase on the dimensional scale, leads to an over 40% increase on the OFDI/GDP ratio. The macro level analysis provides strong support for H1, as performance orientation seems clearly related to Home Bias (Klijn, 2011). However, assertiveness was theorised to be unrelated. A relation to assertiveness can be explained using the concept of masculinity from Hofstede’s (1980, 2001) dimensions. In masculine societies, inhabitants are more assertive and competitive. Assertiveness is one fraction in that dimension, while competitiveness may refrain most closely to performance orientation which already showed a positive result. Anderson et al. (2010) find the same relation towards assertiveness and masculinity, where a higher level of masculinity and a higher level of assertiveness are related to less Home Bias.