SURVEY AND REVIEW ON BEHAVIORAL FINANCE. HyoYoun Park

Transcription

1 SURVEY AND REVIEW ON BEHAVIORAL FINANCE By HyoYoun Park THESIS Submitted to KDI School of Public Policy and Management in partial fulfillment of the requirements for the degree of MASTER OF PUBLIC POLICY 2011

2 SURVEY AND REVIEW ON BEHAVIORAL FINANCE By HyoYoun Park THESIS Submitted to KDI School of Public Policy and Management in partial fulfillment of the requirements for the degree of MASTER OF PUBLIC POLICY 2011 Professor Wook SOHN

3

4 ABSTRACT SURVEY AND REVIEW ON BEHAVIORAL FINANCE By HyoYoun Park Even though Modern Portfolio Theory (MPT) and Efficient Market Hypothesis (EMH), which represent standard finance, are successful, the alternative approach of behavioral finance brings psychological and sociological issues for investigating market anomalies and individual investors behavior. There are two big topics discussed in behavioral finance: Behavioral Finance Macro that is the recognition of anomalies in the efficient market hypothesis that behavioral models could explain, and Behavioral Finance Micro that is the recognition of individual investors behavior or biases that is not explained by traditional models of rational behavior. This paper provides a summary of two big topics in behavioral finance by reviewing latest studies and surveys. iv

5 Copyright by HyoYoun Park 2011 v

6 Dedicated to my Lord, my parents, and my sister. vi

7 TABLE OF CONTENTS I. INTRODUCTION... 1 II. TWO TOPICS IN BEHAVIORAL FINANCE... 3 A. COGNITIVE BIASES... 4 Three Bounds of Human Behavior... 8 Prospect Theory B. LIMITS OF ARBITRAGE Fundamental Risk Noise Trader Risk Evidences III. APPLICATIONS EVIDENCE FROM KOREA IV. CHALLENGES TO BEHAVIORAL FINANCE V. CONCLUSIONS VI. BIBLIOGRAPHY vii

8 LIST OF TABLES 1. Preferences between Positive and Negative Prospects 12 viii

9 LIST OF FIGURES 1. The Underpinning of Behavioral Finance 3 2. The Behavioral Finance Checklist 5 3. Kahneman and Tversky s (1979) proposed value function υ 12 ix

10 INTRODUCTION The recently acknowledged theories in academic finance are as called as standard or traditional finance. Based on the standard finance paradigm, scholars seek to understand financial markets using models which suppose that investors are rational, and the basis of traditional finance is founded in Modern Portfolio Theory (MPT) and Efficient Market Hypothesis (EMH). Harry Markowitz introduces Modern Portfolio Theory in 1952 and he illustrates relations between choices of portfolio and beliefs in line with the "expected returns variance of returns" rule. Ricciardi and Simon (2000) define that Modern Portfolio Theory is an expected return and standard deviation of particular securities or portfolios are correlated with the other securities or mutual funds held within one portfolio. Another main concept is known as Efficient Market Hypothesis (EMH). This concept states that financial markets are efficient regarding the level of information. Anyone cannot consistently achieve excessive return over market returns on a risk-adjusted basis, since all publicly available information has already been reflected in a security s market price, and the current price of security is its fair value. With these Modern Portfolio Theory (MPT) and Efficient Market Hypothesis (EMH) framework, individual investors behavior is not easily understood. Behavioral finance is a new concept in a financial market. Behavioral finance is not one of several branches of standard finance, but it replaces traditional finance and offers a better model of humanity. Even though Modern Portfolio Theory (MPT) and Efficient Market Hypothesis (EMH) theories have been considered successful in a financial market, behavioral finance has developed to stand out as one of the alternative theories of standard finance. 1

11 Behavioral finance is a research of the impact of psychology on the market participants behavior and the following effects or outcomes on markets. Behavioral finance focuses on the way how individual investors make decisions, in particular, how people interpret and act on specific information. Investors do not always show rational and predictable reactions supported by the quantitative models, and this means that the decision-making process of investors considers cognitive biases and the affective (emotional) aspects. Behavioral finance emphasizes on investors behavior leading to various market anomalies and inefficiencies. This new concept of finance explains individual behavior, group behavior by integrating the field of sociology, psychology, and other behavioral sciences and predicts financial markets. Ricciardi and Simon (2000) define behavioral finance in a following manner: Behavioral finance attempts to explain and increase understanding of the reasoning patterns of investors, including the emotional processes involved and the degree to which they influence the decision-making process. Essentially, behavioral finance attempts to explain the what, why, and how of finance and investing, from a human perspective. (See figure 1). Shefrin (2000), on the other hand, mentioned the difference between cognitive and affective (emotional) factors, cognitive aspects concern the way people organize their information, while the emotional aspects deal with the way people feel as they register information. This review paper summarizes and rearranges recent works in behavioral finance, particularly regarding market anomalies and investors behavior which cannot be explained by the traditional paradigm of finance. 2

12 Figure 1. The Underpinning of Behavioral Finance 1) TWO TOPICS IN BEHAVIORAL FINANCE Behavioral finance is a study that combines psychology and economics, and it tries to explain various kinds of events happening in financial markets. For example, in terms of behavioral finance perspectives, we can see some individuals limitations and problems in expected utility theory and arbitrage assumptions. In particular, we can say there are two representative topics in behavioral finance: cognitive psychology and the limits of arbitrage. 1) Ricciardi, Victor, and Simon, Helen K. What is Behavioral Finance?, The Business, Education and Technology Journal 2 (2000): 1,

13 Cognitive psychology is the scientific study of human beings cognition or the mental processes that are considered to form human behavior. Limits of arbitrage explain a prediction of effectiveness of arbitrage forces in any circumstances. Behavioral finance investigates that some individual investors are not completely rational due to specific preferences or mistaken beliefs, and then Efficient Market Hypothesis (EMH) cannot explain all circumstances. Behavioral finance assumes that financial markets are not efficient in some particular circumstances with regard to information, and this inefficiency can be explained by psychological biases of arbitrageurs. COGNITIVE BIASES Under the traditional and standard financial theory, decision makers investors are rational. Basically, rational economic person is an individual who tries to achieve discretely specified goals to the most comprehensive, and consistent way while minimizing any economic costs. Rational economic person s choices are determined by his utility function. In contrast, modern theory in behavioral finance suggests that investors decision is subject to several cognitive illusions. Many scholars of contemporary behavioral finance feel that the field s most direct roots can be founded in cognitive psychology. Research in cognitive psychology investigates various topics such as perception, attention, creativity, memory, reasoning, knowledge representation, and problem solving. Cognitive psychology also explains human thought in terms of input, representation, processing, and output. There are many topics within the arena of behavioral finance relating to cognitive psychology, and Ricciardi and Simon (2000) introduced the checklist of behavioral finance topics (See figure 2). These topics are initiations for the various subjects in behavioral finance literature that have been studied for over the last 30 years. The validity of all of these topics 4

14 will be conducted continuously as the behavioral finance scholars eventually investigate and implement concepts or, as other practices start to diminish or to be discarded. Figure 2. The Behavioral Finance Checklist 2) Anchoring Financial Psychology Cascades Chaos Theory Cognitive Dissonance Fear Cognitive Errors Contrarian Investing Crashes Loss Aversion Herd Behavior Greed Anomalies Market Inefficiency Fads Overreaction Under-reaction Framing Mental Accounting Irrational Behavior Heuristics Risk Perception Behavioral Economics Gender Bias Overconfidence Hindsight Bias Preferences Regret Theory Economic Psychology Manias Groupthink Theory Group Polarization Risky Shift Prospect Theory Behavioral Economics Panics Affect (Emotions) Behavioral Accounting Issues of Trust Illusions of Control Cognitive Psychology Issues of Knowledge Downside Risk Experimental Psychology Familiarity Bias Below Target Returns View of Experts vs. Novices Information Overload Many research papers of cognitive psychologists show several patterns of individual investors behavior among the above checklist. Ritter (2003) describes some of these topics in a following fashion. 2) Ricciardi, Victor, and Simon, Helen K. What is Behavioral Finance?, The Business, Education and Technology Journal 2 (2000): 1,

15 Heuristics Heuristics, in other words known as a rule of thumb, an instructed guess, common sense, or an intuitive judgment, let individuals make decision easier and faster. Heuristics use experience-based techniques to discover, learn, and solve problems. In this decision making process, however, heuristic sometimes generates biases, in particular, when people face any changes in specific circumstances, these rules of thumb can be the second best investment decision. Benartzi and Thaler (2001) explain this human behavior using 1/N rule. When people have N choices in a way to invest their retirement money, many people separate identically their capital across the funds proposed in the plan. For example, if investors have five funds to choose, each probability (one-fifth) is allocated identically. If three of total five funds are stock funds, they (three-fifths) are allocated identically to equities. Overconfidence Most people are overconfident about their abilities. Barberis and Thaler (2003) classified the overconfidence into two aspects. First of all, people assign the far too narrow confidence intervals to their quantities estimation the level of the Dow index in a year. Alpert and Raiffa (1982) give an example that 98% of those confidence intervals include only about 60% of the time for the true quantity. Second, people are wrong too often when they are sure that they are right when estimating probabilities. According to the research of Fischhoff, Slovic, and Lichtenstein (1977), any events which are sure to be happened actually occur only around 80% of the time, and other events that are considered not occurring almost 20% of the time. Men tend to be more overconfident than women, and a concept forecasts that men will execute excessive trades compared to women. Barber and Odean (2001) analyze the pattern of investors trading activities having discount brokerage accounts. They find that when 6

16 people trade more frequently, their performance will be worse in general. In addition, male traders have poorer results by executing more transactions than female traders. Mental Accounting Mental accounting tries to explain the process of the way people code, categorize, and evaluate economic outcomes. Many people sometimes divide decisions which should be combined in principle. For example, since people think separately about home cooking and restaurant foods, they choose to limit their home cooking as they order expensive meals at restaurant. Representativeness Because many people attempt to give too much weight to recent experiences, they relatively underweight on long-term averages. For example, when stock returns have been low over a long time, most people perceive that low equity returns are general or normal situations. We can call this phenomenon as the law of small numbers. Conservatism When things change, many people attempt to under-react and to be slow to aware of the changes. Then, they consider these changes as permanent ways that things have normally been. However, if there is a continuous and long pattern that is enough to believe, people finally adopt this pattern and sometimes overreact, underestimating the long-term average. 7

17 Three Bounds of Human Behavior Widely accepted traditional economic model that explains human behaviors has at least three unrealistic traits, and Mullainathan and Thaler (2000) document past research works in terms of bounded rationality, bounded willpower, and bounded selfishness. Bounded Rationality Traditional economic theory assumes that an economic man is rational. This man has clear, ample, and complete information and well-organized, stable system of preferences as well. However, recent developments in economics have raised great doubts that this classical model of economic man can offers a proper basis of economics. Simon (1955) assumes that the concept of economic man or administrative man needs to be revised fairly, and proposes some other suggestions as to the direction that the revision might take. He applies this bounded rationality concept to explain a more realistic aspect of human capabilities in problem solving. Conlisk (1996) insists that there are four words explaining bounded rationality: evidence, success, methodology, and scarcity. He indicates the reason why we fail to incorporate a concept of bounded rationality into traditional economic models is because those models are bad. He emphasizes throughout that we cannot apply an appropriate rationality assumption to all contexts at once because this assumption is not something to decide. He, in principle, assumes that we can refer to one encompassing single theory which appears as various kinds of forms as special cases in bounded and unbounded rationality. However, the evidence and models which he studies suggest that we can also find that there are various kinds of integrated rationality assumptions by context. Because we have enough time and enormous brainpower, we are not expected to solve difficult problems effectively and optimally. People are considered as rational economic 8

18 men to adopt rules of thumb in economizing cognitive abilities: because the heuristics are commonly considered as the standard economic model which disregards any bounds. One article of Kahneman and Tversky (1974) shows that people tend to depend on a limited number of heuristic principles which can curtail several steps of the complex tasks for probability assessment and value prediction to make judgmental operations simpler. They say, in general, these rules of thumb are fairly useful, but they sometimes make systematic and critical errors. Therefore, we can improve judgments and decisions in uncertain situations when we are aware of several types of heuristics and biases. Bounded rationality emerges both in beliefs and in preferences. According to Kahneman, Slovic, and Tversky (1982), many people make decisions with their beliefs in uncertain situations, and the differences between judgment based on belief and rationality are huge and extensive. Concepts such as overconfidence, extrapolation, anchoring, optimism, and frequency or likelihood of judgments are some examples of availability heuristics or representative heuristics. Many cases of the bounded rationality are handled by Kahneman and Tversky (1979) in prospect theory. They describe how people make decisions under uncertainty. Bounded Willpower Secondly, economic persons tend to choose the optimum in terms of bounded willpower. Human beings sometimes fail to make decisions for self-control reasons, even though they know what the best choice is. Most people eat or drink too much even on a diet or spend too much time and work too little. People can at least somewhat realize and are aware of their self-control problems, and they also even try to control themselves despite having these problems. However, they sometimes fail to maintain their continuous willpower finally. Jolls, Sunstein, and Thaler (1998) summarize bounded willpower as follows: Bounded willpower refers to the fact that human beings often take actions that they 9

19 know to be in conflict with their own long-term interests. Most smokers say they would prefer not to smoke, and many pay money to join a program or obtain a drug that will help them quit. As with bounded rationality, many people recognize that they have bounded willpower and take steps to mitigate its effects. They join a pension plan to prevent under-saving, and they do not keep tempting desserts around the house when trying to diet. In some cases they may vote for or support governmental policies, such as social security, to eliminate any temptation to succumb to the desire for immediate rewards. Thus, the demand for and supply of lay may reflect people s understanding of their own (or others ) bounded willpower; consider cooling off periods for certain sales and programs that facilitate or even require saving. Bounded Selfishness Lastly, people are bounded selfish. Economists emphasize self-interest as one of the primary motive, even though existing economic theory, in a practical matter, accept altruism. The free-rider problem can be one of the good examples, because it is likely to occur generally in economic aspects, and we expect individuals not to contribute goods to the public unless their private welfare is likely to be improved by considering this free-rider problem. We also can find similar altruistic behavior in controlled laboratory experiments. Mullainathan and Thaler (2000) describe that subjects in experiments often cooperate systematically each other in any situations relating to public goods and in prisoners dilemma games, and reject unfair suggestions in ultimatum games. Jolls, Sunstein, and Thaler (1998) call bounded selfishness as a bounded self-interest and describe it as below: We use the term bounded self-interest to refer to an important fact about the utility function of most people: They card, or act as if they care, about others, even strangers, in some circumstances. Thus, we are not questioning here the idea of utility maximization, but rather the common assumptions about what that entails. Our notion is distinct from simple altruism, which conventional economics has emphasized in areas such as bequest decisions. Self-interest is bounded in a much broader range of settings than conventional economics assumes, and the bound operates in ways different from what the conventional understanding suggests. In 10

20 many markets and bargaining settings as opposed to nonmarket settings such as bequest decisions, people care about being treated fairly and want to treat others fairly if those others are themselves behaving fairly. As a result of these concerns, the agents in a behavioral economic model are both nicer and (when they are not treated fairly) more spiteful than the agents postulated by neoclassical theory. Prospect Theory One of the important concepts in behavioral finance is prospect theory initiated by Kahneman and Tversky (1979) which helps individual investors to make decisions. Prospect theory essentially describes the way how individual investors assess gains and losses differently. This theory designates two representative processes in thinking: editing and evaluation. During the editing phase, individuals rank many alternatives according to heuristics or rules of thumb. Then, during the evaluation phase, individuals assign a specific reference point which explains a relative basis for assessing gains and losses. A value function which passes over this reference point and assigns a value to each positive and negative result is asymmetrical S shaped curve. This curve reflects loss aversion which is a tendency to assess the impact of losses much larger than that of gains in domain gains situation, while this can be considered as risk seeking in domain losses which is named reflection effect. Figure 3 describes a value function which is a representative diagram in prospect theory. 11

21 Figure 3. Kahneman and Tversky s (1979) proposed value function υ 3) Kahneman and Tversky discuss preferences between positive and negative prospects. In each of the four cases in Table 1, the preference in negative prospects is the mirror image of the preference in positive prospects and they name this pattern as the reflection effect. Table 1. Preferences between Positive and Negative Prospects 4) Positive prospects Negative prospects A: (4,000,.80) < (3,000) A : (-4,000,.80) > (-3,000) N=95 [20] [80]* N=95 [92]* [8] B: (4,000,.20) > (3,000,.25) B : (-4,000,.20) < (-3,000,.25) N=95 [65]* [35] N=95 [42] [58] C: (3,000,.90) > (6,000,.45) C : (-3,000,.90) < (-6,000,.45) N=66 [86]* [14] N=66 [8] [92]* D: (3,000,.002) < (6,000,.001) D : (-3,000,.002) > (-6,000,.001) N=66 [27] [73]* [70]* [30] 3) Kahneman, Daniel, and Tversky, Amos. Prospect Theory: An Analysis of Decision under Risk, Econometrica 47 (1979): 2, ) Kahneman, Daniel, and Tversky, Amos. Prospect Theory: An Analysis of Decision under Risk, Econometrica 47 (1979): 2,

22 Kahneman and Tversky find out three implications in their results. First of all, the reflection effect shows that people tend to show risk-averse behavior in the positive domain while they attempt to show risk-seeking behavior in the negative domain. For example, in case A, most of the participants tend to choose a risk of.80 to lose 4,000 rather than to choose a sure loss of 3,000, even though they recognize that the first option has a lower expected value. Secondly, the preferences among choices in positive prospects indicated in Table 1 are not in line with expected utility theory. In the positive domain, the certainty effect is applied and individuals prefer a sure gain to a larger gain with lower probability. In the negative domain, on the other hand, risk seeking preference is working and individuals prefer a larger loss with lower probability to a certain smaller loss, even though the same certainty effect is applied. Because people tend to overweight on certainty in gaining situation, they show risk-averse behavior in the positive domain. On the other hand, as people tend to overweight on uncertain results in losing situation, they show risk-seeking behavior in the negative domain. Lastly, we can say that the reflection effect mitigates aversion for uncertainty or volatility because of the certainty effect. For example, if individuals generally prefer (3,000) to (4,000,.80) or (4,000,.20) to (3,000,.25), people tend to resolve this obvious inconsistency in preferences and make the assumption that they prefer prospects with high expected return and small variance. Since (3,000) has no variance while (4,000,.80) has relatively big volatility, individuals choose the former prospect in spite of its lower expected value. This can explain the phenomenon of loss aversion and show that the marginal utility of gains decreases faster than the marginal disutility of losses. This experimental result shows that we can find some evidences which explain people s behavior not to follow the assumptions and predictions of expected utility theory. We can initially find considerable evidence that individuals are subject to changes more easily in asset values than in net asset levels, to gains and losses from a reference point rather than to levels of wealth and welfare, and this idea is first proposed by Markowitz (1952). Tversky 13

23 and Kahneman (1991) say that this dependence on reference is the key analytic assumption in prospect theory. It shows the opposite side of individuals expected utility function which commonly defines levels of assets. Prospect theory also analyzes how individuals mentally frame the predicted results, often in very subjective terms; this can affect expected utility accordingly. Tversky and Kahneman (1991) assert that the reference point is generally consistent with the current position of decision maker, but they acknowledge that this is not always happened. The framing of the reference point can be also influenced by following concepts: aspirations, expectations, social comparisons, and social norms, etc. Barberis and Thaler (2003) consider this theory as the most successful tool to capture the experimental results. They argue that prospect theory does not need to be considered as a normative theory in finance: It just simply tries to capture people s behavior to prefer risky gambles as begrudgingly as possible. Mullainathan and Thaler (2000) also regard that prospect theory can be an excellent example of behavioral economic theories because it integrates the theoretical components of finance and several important features of psychology. They also recommend three key concepts of the prospect theory in terms of a value function as follows: overconfidence, loss aversion, and mental accounting. Overconfidence In its most basic form, overconfidence can be regarded as unguaranteed faith in human being s intuitive reasoning, judgments, and cognitive abilities. We can obtain evidences of overconfidence from a large pool of experiments and surveys in cognitive psychology participants in experiment tend to overestimate their own predictive capabilities and the accuracy of the given information. Many people are poor at probability estimation, and in particular, events they are sure to be occurred are often happen with far less than 100 percent probabilities. To sum up, people tend to believe and think they are having better information and are smarter than they actually are. For example, if investors are 14

24 overconfident and overestimate their abilities, they will tend to execute trades even they do not have correct and fruitful information. One of the most popular researches in social psychology describes, For nearly any subjective and socially desirable dimension most people see themselves as better than average (Myers, 1998). For instance, people think that they are above average in driving ability, their sociability, and their job-seeking ability that they can enter anywhere they want to go (College Board, ; Svenson, 1981; Weinstein, 1980). Some people assert that we are too overconfident and biased to make proper business decisions such as foundation of a new company, entering to an existing market, or introduction of a new product. (Cooper, Woo, and Dunkelberg, 1988; Dunning, Heath, and Suls, 2004; Hayward and Hambrick, 1997; Malmendier and Tate, 2004; Odean, 1998; Zajac and Bazerman, 1991). Overconfidence is one of the popular traits studied recently in finance, in particular for trading behavior and corporate finance such as excessive trading volumes on financial markets, or incorrect valuation of securities, or unfavorable acquisitions undertaken by some companies. Skala (2008) describes that overconfidence is considered as a well-developed psychological theory with main facts including an incorrect calibration (miscalibration), illusion of control, better-than-average effect, and unrealistic optimism. Following studies are associated with overconfidence in corporate finance. First of all, Ben-David, Graham, and Harvey (2006) show empirical evidence in their study that overconfidence in management environment is related to aggressive corporate policies such as investment, financing, financial reporting, and executive compensation. They collect 5-year forecast data regarding stock market returns and confidence intervals of company s Chief Financial Officers (CFOs). Authors find that overconfidence is correlated with personal characteristics. Companies having overconfident CFOs show excessive trading pattern, pay out smaller dividends, prefer higher leverage to use debt aggressively, engage in market timing, offer more managerial estimations, and tilt executive compensation 15

25 towards performance. Bernardo and Welch (2001) argue that overconfidence can be continued as overconfident behavior provides and circulates valuable private information to the group because most individuals follow the herding behavior, information integration is difficult in groups. A group with few entrepreneurs is often doomed to make incorrect decisions and whole group cannot help following the incorrect way due to an inferior way to integrate information among individual members. On the other hand, when there are many entrepreneurs in one group, they are likely to waste too much time (high attrition) because all members depend only on their own information and they eventually make mistakes too often. Authors also identify a set of influences such as degree of overconfidence, information precision, group size, and type of decision which affects degree of overconfidence in group level and the optimal numbers of entrepreneurs. Malmendier and Tate (2006) study whether CEO s overconfidence helps to explain merger decisions. Overconfident CEOs often overestimate their ability to outperform the expected returns. They find in the model that overconfident CEOs are overpaying for target companies and undertaking value-destroying mergers. In addition, when mergers do not require external financing, the numbers of making acquisition can be increased. For merger announcement, the overconfident CEOs react more negatively than non-overconfident CEOs. In this situation, overconfident CEOs refer to several alternative interpretations such as risk tolerance level, signals, and inside information. Some papers are studying overconfidence in terms of trading behavior. Biais, Hilton, Mazurier, and Pouget (2005) measure the degree of overconfidence in judgment. They investigate some evidences regarding incorrect calibration (miscalibration) and selfmonitoring of 245 participants of their experiment, and observe each participant s behavior under asymmetric information in a designed financial market. Mis-calibrated traders tend to underestimate the conditional uncertainty of the asset value and they are vulnerable, in particular, to the winner s curse. Strategic self-monitoring is likely to be happened 16

26 frequently and participants can achieve better results. Authors suggest that mis-calibration decreases trading performance while self-monitoring improves trading performance. They also find that the impact of the psychological variable is stronger for men but is not existed for women. Glaser and Weber (2007) measure overconfidence of online broker investors in a group level in various dimensions such as mis-calibration, the better than average effect, illusion of control, or unrealistic optimism, and the authors investigate that a degree of overconfidence are significantly associated with trading volume of individual investors. Authors try to specify what kind of overconfidence might influence trading behavior, and they find that overconfident investors execute excessive trades. However, overconfidence is not related to trading volume. Jaimovich and Rebelo (2007) focus on the two specific biases indicated in behavioral finance: optimism and overconfidence. Optimistic investors are biased to expect better outcomes while overconfident investors overestimate the accuracy of information that they have. Their study is based on the model which generates co-movement between consumption and working by investigating changes in expectation. They find that overconfidence can be a potentially useful mechanism as overall volatility is higher when investors are overconfident, while optimism is not a significant source of volatility in their model. Autonomous shocks to expectations also can be a useful source of volatility. They confirm that both overconfidence and shocks to expectations can be useful sources of volatility, but those sources are not enough to be foundations of a theory in the business cycle. Loss Aversion In prospect theory, loss aversion refers to the tendency for people to highly prefer loss avoidance to gain acquirement. Some studies suggest that losses are psychologically much more powerful than gains about twice. Loss aversion is firstly tested and verified by Kahneman and Tversky (1979). Shefrin and Statman (1985) refer to disposition effect 17

27 meaning that investors are associated with past winners differently from past losers. Past winners want to keep winners in their portfolio while past losers tend to sell. According to Odean (1998) s study, investors in a discount brokerage firm, they are more likely to sell a value-increasing stock than a value-decreasing stock, and this reluctance for gain realization finally leads a certain amount of cost. Odean also shows that the portfolio which holds loser securities underperformed the portfolio with winner securities that was sold. Abdellaoui, Bleichrodt, and Paraschiv (2007) say that loss aversion can explain various fields and experimental data in behavioral finance. Authors propose a quantitative method and use it to measure loss aversion in their experiments without making any parametric assumptions. This paper is the first study to provide non-parametric results in utility function under prospect theory, and this study also offers an efficient way to draw out utility midpoints. In addition, the degree of loss aversion is different depending on several definitions used in various ways, and authors find strong evidence of loss aversion in individual level as well as aggregate level in line with the definitions. Kouwenberg (2009) study the impact of loss aversion in asset prices. Berkelaar and Authors derive closed-form solutions for the market price, mean, and volatility in an economy with heterogeneous agents. They find that some agents are normal while other agents show lossaverse tendency. In a favorable situation, loss-averse investors tend to be momentum traders and they push prices far above the level and stimulate a boom in the market. In a moderate bad situation, on a contrary, loss-averse investors follow a contrarian strategy and equilibrium prices are set relatively high and stable. In an extremely bad situation, lossaverse investors are compelled to withdraw their money from the stock market owing to avoid bankruptcy, and then they will result in a significant price drop. Finally, they find that loss-averse investors generally make the stocks mean return and the volatility to increase, asset returns are predictable, and volatility is non-constant. Tovar (2009) studies the consequences of loss aversion when people determine trade policies, and shows the way how 18

28 to make people to understand and explain a number of important and puzzling features of trade policy. This paper shows that if we can find a tendency of loss aversion in individual preferences, protection will be higher in certain sectors in which profitability is decreasing. Moreover, the author shows that an industry will become organized and lobbied for protection when it has a loss. This study also finds that if the coefficient of loss aversion is large enough, we can find an anti-trade bias in trade policy. The anti-trade bias is trade policy which restricts trade by favoring import-competing sectors. The author also tests some expected outcomes on the lobbying side and he finds evidence of loss aversion in lobby formation. The predictions of lobby formation strengthen the anti-trade bias because of loss aversion. In addition, this study finds the data favoring the author s model over the current leading political economy in trade protection. Kobberling and Wakker (2005) say that risk aversion is caused by loss aversion, and they introduce several indexes of loss aversion in their paper. One of the suggested indexes can lead people to discompose their risk attitude into three discrete components such as basic utility, loss aversion, and probability weighting. They study how these indexes of different decision makers can be compared with each other through observed choices. Mental Accounting A concept first named by Thaler (1999), mental accounting is the set of cognitive operations applied to individuals and households when they organize, evaluate, and keep track of financial activities. Mental accounting tries to explain the process of the way people code, categorize, and evaluate economic outcomes. Thaler (1999) and many researchers try to extend the scope of mental accounting and he considers the scope of questions that mental accounting supports people to answer as follows: Why do firms pay dividends? Why do people buy time-share vacation properties? 19

29 Why are flat-rate pricing plans so popular? Why do sales contests have luxuries (instead of cash) as prizes? Why do 401(k) plans increase savings? Why do stocks earn so much higher a return than bonds? Why do people decline small-stakes attractive bets? Why can't you get a cab on a rainy day? (Hint: cab drivers earn more per hour on rainy days). One of the detailed examples of mental accounting, named behavioral life cycle hypothesis (Shefrin & Thaler, 1988), assumes that households treat components of their assets as being under their current income or under their current wealth or under their future income. If anyone is attracted to spend now, it would be account for the greatest part of the current income and for the least part of the future income. The key assumption of the behavioral life-cycle theory is that households consider components of their income or wealth as nonfungible, even though they are not restricted by a credit rationing. The mental accounting bias is also applicable to investment. For example, some investors separate their investments into two parts: a safe investment portfolio and a speculative portfolio. This is because they tend to avoid the negative return which is affected by speculative investment. Even with these separate portfolios, investors net wealth will not have any differences from an assumption that investors hold one large portfolio. Prelec and Loewenstein (1998) find financial behavior of consumers by introducing mental accounting of savings and debts. They show that consumers have pleasure coming from consumption itself, magnitude of payments, and timing of payments, and there are hedonic interactions which cannot be explained by traditional economic model. Authors bring a theoretical evidence of these interactions between consumption and payment which can explain the patterns of behavior and help to understand in payment mechanisms designed for products and services. On the contrary to the traditional concept that assumes people prefer to consume goods now and pay for them later, the authors model estimates strong debt 20

30 aversion, which confirms a concept that people actually prefer to pay for consumption beforehand. In addition, on the contrary to concepts of economic efficiency which describe people should pay for whatever they consume at the margin, their model estimates a preference for flat-rate pricing. They also show other kinds of predictions such as differences in spending patterns between cash and credit purchases, and preferences for savings to specific purchases. Rockenbach (2001) reports an experiment on behavior in investment decisions particularly for financial options, and the author suggests that mental accounting is a very important factor for making financial decision especially in an option pricing environment. We can test an arbitrage-free option pricing against three hypotheses in terms of mental accounting. The author finds the continuously unexploited arbitrage opportunities in his experimental data even with considerable experience. Even though we need to recognize the connections among the different investment possibilities for arbitragefree pricing, participants in experiment do not seem to make any connections. They act, on a contrary, as if they combine risky assets with the same mental account while they treat bonds separately. Grinblatt and Han (2004) write the paper regarding the tendency of some investors behavior contrary to disposition effect. According to disposition effect, people tend to hold loser securities driven by prospect theory and mental accounting, even though they see a divergence between a security s fundamental value and its equilibrium price. We can also see a price under-reaction to information. Authors develop a model of equilibrium asset pricing referred to prospect theory, mental accounting, and disposition effect. In this model, the differences between the market value of a stock and the aggregate cost are positively related to the stock s expected return. They find that past 1-year returns are not good predictors to expect future returns. In addition, momentum strategies are profitable but the profits are depending on the path. Results in this paper are robust and cannot be explained by the interaction of past returns and turnover. According to the authors, completely rational arbitrageurs cannot mitigate the impact of the capital gain on equilibrium 21

31 prices. It is risky that arbitrageurs try to arbitrage away the under-reaction although prices always under-react to news. Soman (2004) studies framing in investors decision making. When individuals face a decision making problem, they form a mental representation of that problem using a particular frame, and they solve this representation. Therefore, the author thinks the study of decision making should be completed with studying the process of individuals framing and some factors influencing framing. Shafir and Thaler (2006) consider mental accounting of delayed consumption. They investigate some accounting rules and find that many people treat advanced purchases as investment activities rather than simple spending. In addition, earlier consumption of goods and planned usage of them is often considered as free or savings. On a contrary, when consumption is occurred not as planned, this consumption is recognized as the cost of goods replacement, and in particular, people consider this replacement is expected to be happened. Authors also discuss related phenomena and several implications. Lim (2006) investigates the effectiveness of mental accounting in multiple outcomes to the way that investors selling pattern of stocks. This study shows that investors tend to sell multiple stocks when realized losses dominate gains meaning that individuals prefer to integrate losses and to segregate gains. Moreover, investors are likely to select stocks which gains and losses to realize together and integrate sales of winners and losers, so this integrated outcome is more profitable than segregated outcome. In this situation, mental accounting plays a significant role in individuals trading patterns and decisions. She develops and tests a prediction from psychological theories regarding the market participants behavior, and investigates how selling decisions for multiple stocks associate with each other, even in the absence of common fundamental factors. Empirical findings in this study also have further implications on the research of equilibrium stock prices. As investors make biased selling decisions between winner stocks and loser stocks, these asymmetric decisions can affect the stock market to be skewed. In addition, asset prices in financial market will be affected since investors selectively adopt 22

32 only one among several different mental accounting systems. LIMITS OF ARBITRAGE In terms of traditional and standard finance, discussions on the limits of arbitrage are not widely accepted, and so called Efficient Market Hypothesis (EMH) is considered as the best established model in economics. According to one of the major principles under the Efficient Market Hypothesis (EMH), prices are correct in a sense that asset prices reflect the fair value of the security. In a rational and efficient market, all individuals are not allowed to have any free lunch, so no investment strategy can make excessive risk-adjusted average returns. Behavior finance argues that, however, we can see many deviations in asset prices and these deviations are brought by the irrational investors. According to Friedman (1953), there are two common situations in a financial market. First, when there is a mispricing and the current asset price is deviated from fundamental value, the arbitrage opportunity is created as an attractive investment opportunity. Second, rational investors will immediately catch this opportunity, and then the mispricing will be corrected. Behavioral finance criticizes for the second step. Even if attractive opportunities become revealed and recognized by traders, we cannot confirm that these opportunities will be immediately exploited, and as a result, the mispricing can remain unchanged. At the first step, arbitrage opportunity is considered as riskless strategy. However, behavioral finance argues that this is very risky and investors considered as arbitrageurs actually can be noise traders. Another well-known example is the case of the same stock traded in two different places described by Froot and Dabora (1999). Those two shares have same characteristics and should be traded in a certain ratio in line with a concept assuming rational model. However, they are actually not: the actual 23

33 level of price deviates from the expected price by more than 35%. This case explains that prices can be deviated from intrinsic value due to limits of arbitrage. Malkiel (2003) also introduces the evidence that investors cannot create a portfolio trading opportunity by making extraordinary excessive risk adjusted returns even in any anomalous stock market environments. The other principle of the Efficient Market Hypothesis (EMH) is unpredictability. The Efficient Market Hypothesis is related to the idea of a random walk, which explains a price movement: prices are deviated from the previous one as a random movement. As any publicly available information is already reflected in securities price, today s price changes will be influenced only by today s news and will not be affected by the price movements in yesterday. Even though prices show volatile movement in extreme manner, they are brought back to equilibrium by mean-regression model. However, from the start of the 21st century, the dominant trend of the Efficient Market Hypothesis has become far less common. Many financial economists have become believe that securities prices are not fully predictable, but at least, partially foreseeable. Those researchers apply psychological and sociological factors to determine stock price, and they believe that price movements are not random at all and future stock prices are predictable on the basis of historical price patterns. They also insist that many investors will be able to make excessive return with those considerations. Study by De Bondt and Thaler (1985) introduces Investor Overreaction Hypothesis opposed to Efficient Market Hypothesis. Individual investors tend to overreact to new information and underweight prior information in incorporating new data. They refer to representativeness heuristics which makes investors are unduly optimistic about recent winner stocks while they are extremely pessimistic about recent loser stocks: most of investors tend to overreact to both good and bad news. Overreaction also makes the previous loser stocks to become undervalued while leads the previous winner stocks to become overvalued, and this derives a prediction about future returns: past loser stocks 24

34 should underperform while past loser stocks should outperform the market. They study two portfolios of 35 stocks traded in the New York Stock Exchange: one is composed of the extreme winner stocks over the last three years, and the other is composed of extreme loser stocks for the last three years. Consequently, past loser stocks ultimately outperform winner stocks for the next four years past losers are up 19.6% relative to the market while past winners are down 5% relative to the market. Therefore, they suggest an investment strategy to investors to buy and hold recent losers while sell recent winners. Other recent studies find other cases of unpredictability which show the opposite pattern from overreaction introduced by De Bondt and Thaler, as called under-reaction. According to Shleifer (2000) study, we can find an initial price jump in stocks on the day of the specific announcement which can influence the fundamental value such as large earnings changes, dividend payout plans, share repurchases, splits, and seasoned equity offerings followed by a slow drift in the same direction for certain period of time. Over short periods of time, less than a year, stocks are also showing display momentum stocks which have been going up fastest for the last six months attempt to keep going up. These findings regarding underreaction are a further challenge to the Efficient Market Hypothesis. Both over and underreaction rely on psychological evidences and they all explain market anomalies by explaining different time horizon between underreaction and overreaction that underreaction continues for short period of time while overreaction appears for longer term. They are now changing the nature of finance so far. The traditional concept of arbitrage in theoretical rational financial markets does not require capital and any financial risks. However, almost all arbitrage requires capital that entails risks in a real financial market. Shleifer and Vishny (1997) introduce professional arbitrage which is conducted by a relatively small number of highly specialized investors using other people s money. In particular, many researchers consider that arbitrage which requires 25

35 additional risky capital become very important issue in agency problem. Without agency problem, arbitrageurs are commonly more aggressive when prices are deviated further from fundamental values. Shleifer and Vishny (1997) give an example about Bunds: consider the simple case of two Bund futures contracts to deliver DM250,000 in face value of German bonds at time T, one traded in London on LIFFE and the other in Frankfurt on DTB. In their Bund example, an arbitrageur would generally increase positions as long as he has the capital if contract prices between London and Frankfurt move further from fundamental value. However, when arbitrageur manages other people s capital, and these customers do not have any ideas how this arbitrageur operates and do not understand exactly what the manager is doing, they only can observe results that arbitrageur is losing money when futures prices in London is varied with those in Frankfurt. They tend to refuse providing more capital or withdraw money even when the expected return of the funds is increasing. Authors in this article argue that investors cannot use arbitrage opportunity effectively to achieve market efficiency. They also analyze which markets attract arbitrage and why some markets are more attractive for arbitrage than others. Consequently, high volatility makes arbitrage less attractive if expected return does not increase identically with volatility, and in particular, this would be true when fundamental risk is a substantial part of volatility. Finally, such arbitrage activities are not effective to bring the securities price back to the fundamental value, especially in extreme circumstances. Pontiff (1996) identifies significant arbitrage costs and shows that these costs are related to large deviations of prices from fundamental values. When costs make arbitrage unprofitable, mispricing is not fully eliminated. He considers four factors which affect arbitrage profitability in a costly-arbitrage framework: systematic risk which means the security s fundamental risk that cannot be diversifiable, smaller dividends payout, lower market value with higher transaction costs, and higher interest rate. He finds that fundamental risk which cannot be hedged lowers arbitrage profits because arbitrage is risk- 26