Do signal providers on social trading platforms exhibit behavioral biases?

Table of Content

List of Figures

List of Appendices

List of Abbreviations

1. Introduction

2. Literature Review
2.1 Review on the Disposition Effect
2.1.1 The Four Key Elements of the Disposition Effect
2.1.2 Cognitive Dissonance
2.1.3 What Factors influence the Disposition Effect?
2.2 Review on the Home Bias

3. Social Trading

4. Methodology and Empirical Results
4.1 Disposition Effect
4.1.1 Hypotheses
4.1.2 Data and Empirical Approach
4.1.3 Empirical Results
4.1.4 Odean Method
4.2 Home Bias
4.2.1 Hypothesis
4.2.2 Data and Empirical Approach
4.2.3 Empirical Results

5. Limitations and Further Research

6. Conclusion

List of Literature

Appendix

List of Figures

Figure 1: Value function in prospect theory

Figure 2: Odean’s original ratios

Figure 3: Evolution of Wikifolios

Figure 4: Trader compensation

Figure 5: Examples of gain calculation

Figure 6: Descriptive statistic of the data collection

Figure 7: Histogram of the dependent variable

Figure 8: F-test for fixed effect model

Figure 9: Breusch-Pagan test for the random effect model

Figure 10: Hausman test

Figure 11: Regression output for ln (days, gain)

Figure 12: Regression output for ln (days, loss)

Figure 13: Compendium of both regression outputs

Figure 14: Sub-sample results

Figure 15: Histogram of invested capital vs. ln of invested capital

Figure 16: Hausman test for home bias regression

Figure 17: Home bias regression output

List of Appendices

Appendix 1: Illustration of the Endowment Effect

Appendix 2: Example of an account statement

Appendix 3: Disposition effect gain do file

Appendix 4: Disposition effect loss do file

Appendix 5: Sub-sample results with trader names

Appendix 6: Home bias do file

Appendix 7: Home bias f-test

Appendix 8: Home bias Breusch and Pagan test

Appendix 9: Second home bias regression

List of Abbreviations

Abbildung in dieser Leseprobe nicht enthalten

1. Introduction

This paper examines the effect of transparency on the trading behavior in the context of social trading. The focus lies on the effect of transparency on the behavioral bias known as the disposition effect. The disposition effect indicates that trader tend to hold their losing stocks for a too long amount of time and sell their profitable stocks too soon. Additionally, the effect of transparency on the home bias will be analyzed. The home bias is the trader’s preference for domestic stocks at the expense of portfolio diversification.

The underlying question is: "Can transparency mitigate the extent of the disposition effect and home bias?” In the past there has been a wide extent of competing theories about which factors influence these biases. However, a factor which has not been in the center of attention is transparency. In other words, a permanent visible statistic about the trader’s performance, trades, positions and their success. The results of this paper may help to draw comparisons to the field of traditional fund managers.

To begin with, this thesis will start by providing information about the status quo of the disposition effect and the home bias. This chapter will highlight the factors that has been in the center of empirical research. Furthermore, it will show why it is important to look at transparency. Next, the idea behind social trading will be explained and the platform Wikifolio will be introduced. The empirical part will be divided into two segments. The first segment will cover the disposition effect. It will start with the underlying hypothesis, followed by the data sampling and empirical results. The segment about the empirical results will contain two panel data regressions that will analyze the trader’s holding time of losing and gaining positions separately. The analysis of the disposition effect will be concluded by a different approach to cross check the findings of the panel data regressions. Therefore, the original method of Odean¹ will be used. The second segment will cover the hypothesis, data and empirical results of the home bias. The final chapter illustrates the difficulties of this thesis and some approaches for further research. The thesis is recapped by the conclusion.

2. Literature Review

This chapter will cover the most important findings to the disposition effect and the home bias. The literature review chapter will start with the disposition effect, containing the four key elements, cognitive dissonance and the most important factors that influence the disposition effect. The second part will be a short survey of the home bias.

2.1 Review on the Disposition Effect

2.1.1 The Four Key Elements of the Disposition Effect

The disposition effect is a behavioral bias that has been widely analyzed by researchers during the last three decades. It describes the effect, that investors often tend to sell assets that have gained in value, while holding losing assets for longer periods.² The term "disposition effect” which is named by Shefrin and Statman (1985) consists of four key elements: prospect theory, mental accounting, regret aversion, and self-control. These will be briefly discussed in the following.³

In 1979 Kahneman and Tversky published an article, in which they explained their "prospect theory” for decision under uncertainty. The theory describes the decision processes in two stages, an editing phase and a subsequent phase of evaluation. During the editing phase the subjects start by defining a reference point as mental starting point for the following evaluation. They asses possible actions or prospects against this reference point. Subjects then decide which outcomes they consider equivalent, define outcomes greater than the reference point as gains and outcomes smaller than the reference point as losses. In the evaluation phase subjects dedicate a specific utility, based on the outcome and the occurrence probability, and then choose the alternative with the highest utility.⁴

When people have to choose between risky prospects, they often act risk averse when the possible outcome is framed as gains and risk seeking if the outcome is framed as losses. This phenomenon in prospect theory is defined as the "Reflection Effect”. Additionally, the well-known "Endowment Effect” can be also explained by the choice of reference point. The endowment effect is the hypothesis that people ascribe more value to things just because they own them. Owners incorporate the item into their status quo and therefore choose a reference point including the item. Buyers will set their willingness to pay lower and choose a reference point without the item. Buying will be interpreted as positive prospect and selling as a negative prospect.⁵ Kahneman et al. (1990) displayed the "Endowment Effect” in an experiment where the participants were given a mug and were then offered the chance to sell it or trade it for an equally valued item. They found that, once the mug was in the possession of the participants, the required compensation for the mug was approximately twice as high as the amount they were willing to pay to acquire the mug.^{6 7}

In addition to that Kahneman and Tversky developed a value function, which is "... (i) defined by deviations from the reference point, (ii) concave for gains and in general convex for losses and (iii) less steep for gains than for losses’’.⁸ This shows that people value a gain less than an equal loss and are loss averse. Figure 1 shows a value function by Weber and Camerer (1998).

Abbildung in dieser Leseprobe nicht enthalten

Figure 1: Value function in prospect theory⁹

Here you can see that the value (V(G)) of a gain (P+G) is less than the value (V(- L)) of an equal loss (P-L). In this framework Weber and Camerer show that the reference point and the risk attitudes lead to the disposition effect. In the example of Weber and Camerer a stock has the same chances to gain or lose an equal amount in each period. The intersection of the axes is the purchase price (P) and reference point. If the stock gains in the first period (P+G), the subject will sell the stock in the next period because the value (V(2G)), if the stock gains again, is lower than the value (V(G)) which the subject will lose if the stock falls. Vice versa, if the stock falls in the first period (P-L) the subject will hold the stock in the next period because the value, if the stock falls again (V(-2L) - V(-L)), is lower than the value (V(-L)) which the subject will regain if the stock gains. Here the investor will hold the stock because of the possibility of breaking even with the reference point/purchase price (P). This behavior leads to the disposition effect.¹⁰

One of the first real money experiments, with regard to prospect theory were performed by Thaler and Johnson in 1990. They showed how risk taking is affected by prior gains and losses and entitled the "house money effect” and the "break even effect”. Individuals tend to take on greater risks when investing with profits attained ("House Money Effect”). Individuals who have already lost some amount of money are more risk seeking, when given the chance to break even ("Break Even Effect”).¹¹

Despite the vast majority of studies confirming the link between prospect theory and the disposition effect, there is a growing number of studies which doubt that "Prospect Theory” can explain the disposition effect. List (2003) questioned the robustness of the exchange asymmetries within the "Endowment Effect”. He performed an experiment similar to Kahneman et al. (1990) at a sports card market. His experiment included people who often trade sports souvenirs and people who do not. List found strong evidence for exchange asymmetries in the second group, but not in the first group containing the experienced traders. He suggested that "Prospect Theory” might be less applicable for experienced economic actors.¹²

Another study which questions "Prospect Theory” is Hens and Vlcek (2011). They argue, that prospect theory cannot explain the disposition effect as mentioned by Kahneman and Tversky. Investors who sell winning stocks and hold losing stocks would not have invested in the stock market in the first place. The "prospect Theory” is correct in an ex-post point of view, when the investment has occurred. But not ex-ante, when it is required that the investment has to be made.¹³

The second key element of the disposition effect is mental accounting expressed by Thaler (1980). Consumer behavior often deviates from the standard economic prediction. Mental accounting is a framework which can explain these deviations and explain why consumers might violate the standard economic principles. Thaler mentions that individual expenses are not considered as a reduction in value of wealth. Instead, they are considered in two accounts, the current budgetary period, and the category of expenses.¹⁴ An example therefore is credit card and cash payments. A credit card payment delays the payment to a later date, when the monthly bill is paid, and it is added to a larger existing sum. This deferred payment is less anchored in the memory. E.g. an additional €5 to an existing €120 credit card bill is more justifiable than an out of pocket €5 cost.¹⁵ This is also consistent to the value function of figure 1, where the step from (P) to (P-L) is a bigger value reduction than the step from (P-L) to (P-2L).

The third key element of the disposition effect is regret aversion. Ex post having made a bad decision, people may feel ashamed to admit this mistake because they tend to avoid regret. People with regret aversion might prevent situations where they face regret afterwards. This leads to the disposition to realize gains and delay losses. In regard to the disposition effect, a loser stock will be held for too long, in order to avoid the feeling of regret when the loser stock is sold, that means the loss realized.¹⁶ There are several experiments which confirm regret aversion.

Filiz-Ozbay and Ozbay conducted a first price auction aimed to regret aversion. Here a loser’s regret can be induced by revealing the winning bid of an auction, and so revealing to the loser whether they were able to make a profit out of the auction. E.g. a participant who would have valued an item €80 and lost with a bid of €60 sees that the winning bid was €65 and that he would have made a profit by bidding anything between €65 and €80. If participants anticipate the possibility of regret, they will bid more in an auction where the winning bid is revealed, than an auction with no feedback about the winning bid.¹⁷ Decision over lotteries also provide evidence for regret aversion. E.g. when confronted with the choice of certain €40 or a lottery where a correct guessed coin toss is worth €100 and a false guess is worth €0, the majority will opt for the certain €40. Not only does this choice minimize the risk but also eliminates the possibility of regret, when the coin toss is falsely guessed.¹⁸ In the field of investor behavior this might provide an incentive to stay inactive and neither realize gain nor losses. An experiment by Shefrin and Statman (1985) in which investors sell at a small gain and observe the following stock price, indicates that the initial pride of selling might be reduced by regret of selling too early if the price rises.¹⁹

The problem of self-control could also be an explanation for the disposition effect. Thaler and Shefrin (1981) formulated a framework, where an individual is both, a farsighted planner and a myopic doer. They said, that individuals want to increase their feeling of pride and try to postpone the feeling of regret by not selling loser stocks. The myopic doer does approve such behavior, while the farsighted planner does not have enough power to prevent this, which leads to the disposition effect.²⁰

2.1.2 Cognitive Dissonance

This chapter will provide a different approach of explaining the disposition effect based on the notion that investors hold loser stocks for too long in order to justify the initial purchase decision. Barber et al. state: “For some investors, the tendency to hold losers may be driven on a more basic level than probabilities of gains and losses. We live in a world in which most decisions are judged ex post and most people find it psychologically painful to acknowledge their mistakes. When a stock is sold for a loss, it becomes, irrevocably, an (ex post) mistake.”²¹ Usually investors want to maintain a positive attitude about their ability to make good investment decisions and fit their actions with their self-evaluation. When faced with a bad investment, the realization of the loss would be in conflict with the initial purchase decision. In regard to the disposition effect, investors hold loser stocks for too long because they do not want to admit that the initial purchase was, with the benefit of hindsight, a mistake.²² This need to justify the initial investment can be explained through cognitive dissonance theory. The initial cognition like “I have good investment skills” or “I will make a profit out of my investment” stands in dissonance with the cognition “This stock I purchased is losing”. To protect the positive image of oneself, a desire arises to explain the action ex post. Therefore, new cognitions are created such as “The investment will regain” and "This is only a short-run loss”. These new cognitions are biased expectations about the future. The theory says that a discrepancy between the actions of an investor and attitudes creates discomfort, and changing an attitude requires psychological cost.²³

Staw (1976) and Brockner (1992) delivered experimental designs which proved the self-justification hypotheses. Self-justification occurs usually by people who were responsible for the initial action. In matter of investment decisions this means that investors who are responsible for the initial purchase are more likely to fall for the disposition effect. Weber and Camerer (1998) did a laboratory experiment with two groups, where one group were allowed to trade freely at all times and the other group was forced to sell all their stocks at the end of each trading round. The subjects in the group with the selling rule were allowed to rebuy all the shares they had to sell at the end of the trading round. Because there were no transaction costs involved, the standard prediction would be, that they buy back all shares they were forced to sell. With this prediction there would have been no difference between the two groups. With the selling rule, the subjects had to actively repurchase all their losing stocks to align their portfolio with the free trading group. The subjects in the selling rule group bought back only some of the losing stocks, which shows that the selling rule helped to mitigate the effect that losing positions are held for too long.²⁴

Results on mutual funds allow an interesting interpretation. The available evidence for mutual funds does not show a disposition effect. It seems that Self­justification can be escaped by investing in mutual funds, because if the investment goes down it is easy to blame the fund manager for the mistake. This means mutual funds allow investment with no risk of suffering a blow to self- image.²⁵

2.1.3 What Factors influence the Disposition Effect?

The study of Odean (1998) introduced a method for measuring the disposition effect, which several later studies used as well. The data of his study contains the stock investments of 10.000 accounts in U.S discount brokerage. In his method he uses two measures, the "Proportion of Gains Realized” (PGR) and the "Proportion of Losses Realized” (PLR).

Abbildung in dieser Leseprobe nicht enthalten

Figure 2: Odean’s original ratios²⁶

If an investor sells a stock Odean checked if this particular stock was sold for a gain (realized gain) or for a loss (realized loss). If a stock was not sold but gained in value, it was counted as paper gain and if the stock was not sold and dropped in value, it was counted as paper loss. To calculate whether the stock made a gain or a loss, Odean compared the purchase price and the sale price, or the day’s closing price as their hypothetical sale price. From PGR and PLR, the commonly used measure of the disposition effect is derived. If PGR is significantly greater than the PLR the disposition effect is present. Odean finds a strong evidence for the disposition effect. In general, only 9,8% of the losses are realized, while 14,8% of the gains are realized. This means investors are 50% more likely to realize gains than losses.²⁷

Since Odean (1998), numerous studies tried to identify factors which stimulate and which lessens the extent of the disposition effect. There is literature to a various number of factors, e.g. gender, age, experience, education, information etc.

Age is a factor that can be seen as an indicator for trading experience, which is a factor that a person cannot influence by himself. It is expected that older traders are less affected by the disposition effect, due to increasing adeptness with age. One prominent study regarding age and the disposition effect was published by Dhar and Zhu (2006). They showed that older investors are less prone to the disposition effect. So, increasing age gives an advantage regarding the disposition effect, but age is a factor that cannot be changed.²⁸

Gender is a highly analyzed factor regarding the disposition effect as well. Barber and Odean (2001) showed that men trade 45% more than women and are more confident than women. Since the trading frequency is higher by men, they should experience less disposition effect than women.²⁹ This was confirmed by Rau (2014), who found that female investors realize less capital losses and are more loss averse than men, hence they have a significantly higher disposition effect than male investors.³⁰ Whereas, Da Costa Jr. et al (2008) showed, that women do not keep losing stocks and sell winners when the reference point shifts from the purchasing price to the previous price. This shows that the disposition effect disappears for women and remains for men. They explain their results with the speculation that male and female brains interpret changing reference points differently.³¹

The more information investors gather regarding an investment decision, the less they experience the disposition effect. This was observed by Shapira and Venezia (2001), who analyzed Israeli investors and found evidence for the disposition effect. They had two groups in their experiment. One group of independent investors and one group of investors who have been professionally advised. Their results show that the disposition effect exists in both groups, but it is significantly stronger in the group without the additional information from professional advisors.³²

As shown above, there has been a variety of studies that are searching the factors which influence the disposition effect. A factor which is not much studied is transparency. On the platform Wikifolio the portfolio of the trader goes through different stages. The starting stage is the test phase. Here the trader can test his trading strategies without any observers. The following stage is the published phase, where the trader has to convince investors on Wikifolio to fund his portfolio. In this phase the trader can be observed by every user on this platform. The last stage is the investable phase with full transparency.³³ Therefore, this thesis will perform a similar analysis compared to the one conducted by Lukas et al. (2017), who studied traders on the social trading platform Wikifolio and analyzed individual trades as portfolios become publicly visible in multiple stages. Since the traders on social trading platforms are compensated by asset under management (AUM), it is the traders’ main goal to convince other users to fund their portfolio.³⁴ Hence, the traders want to give the best impression of their portfolio, without any biases. Consequently, the key argument for the mitigating effect of transparency on the disposition effect is that the cognitive unease of realizing losses (loss aversion) gets canceled out by the cognitive unease of publicly holding losing stocks in one’s portfolio.³⁵

2.2 Review on the Home Bias

The term "Home Bias” describes the strong preference for domestic equities exhibited by investors in international markets, even though the benefits from international diversification are well documented. Economic theory indicates, that each investor selects his portfolio to maximize the expected utility of the portfolio’s payoff. The investors risk aversion predicts that their portfolio should be diversified. Especially the major benefits of international diversified portfolios stand against the home bias.

"Home Bias” was first stated by French and Poterba in 1991. They found a difference between the amount invested in U.S. stocks by U.S. equity traders and the share of the U.S. equity market in the global equity market. Furthermore, U.S. traders invested 94% of their funds in domestic stocks. However, the U.S. equity market represented less than 48% of the global market. This mismatch was later known as "Home Bias” and exists also in other countries than the U.S., where investors are focusing too much on investments in their home country and are neglecting foreign opportunities.³⁶

This behavior is highly inefficient through a lack of diversification within the investor’s portfolio. Domestic barriers to foreign trades are the primary reason for this bias, such as high transaction costs and foreign taxes.³⁷ However, while these obstacles declined over time, the preference to invest in one’s home country stayed. Tesar and Werner (1995) made three significant conclusions concerning "Home Bias”. First, they said there is strong evidence of "Home Bias” in national portfolios despite the potential gains from international diversification. Secondly, they found that to the extent investors hold international securities, the composition of the portfolio of foreign securities seems to reflect factors other than diversification of risk. Thirdly, they concluded that the high turnover rate on foreign equity investments relative to domestic equity markets suggest that transaction costs are unlikely to be an explanation for "Home Bias”.³⁸ The statement of Tesar and Werner (1995) that transaction costs do not help to explain "Hhome Bias” is confirmed by Warnock (2002), by using data regarding transaction costs in 41 markets.³⁹

The information asymmetries may be one of the best explanations currently discussed. There is no need for national boundaries or foreign taxes to foster the "Home Bias”. High geographical distance between the investor and the potential investment may be enough to discourage the trade. The access to information about companies located near the investor may be easier, resulting in a preference to hold the stocks of local firms for which they have an information advantage. Investors may obtain information through local media, talk to employees of the firm or might be friends with managers of the firm. All of this might result in an information advantage for local stocks. Another reason to invest locally could be the desire to support the local economy. Investors could just feel comfortable about companies they hear a lot about. There might be even the possibility that local brokerage firms have an incentive to advertise local stocks if there are close ties between brokers and local corporate managers.⁴⁰ In this context Huberman (2001) found that shareholders of a Regional Bell Operating Company, a voice carrier for a specific area, tend to live in the area which it serves. Huberman argued that this is closely related to the tendency of employees to own their employers’ stocks in their retirement account. He attributed such behavior to a cognitive bias to invest in the familiar.⁴¹

Another stand of the home bias literature finds supportive aspects of this behavior. One example is Coval and Moskowitz (2001). They found that managers earn substantial abnormal returns in nearby investments. Especially the returns of small funds that focus on few holdings are particularly strong. Moreover, while the average fund only shows a mild bias towards local stocks, some funds excessively focus on local stocks and exhibit a great local performance. Lastly, they hinted that the extent to which a firm is held by nearby investors is positively related to their performance. The explanation of the authors was that these investors are trading at an information advantage.⁴² Hau (2001) conducted a study where he tested whether German speaking investors earn excessive returns on German investments. As a result, he found that traders in non-German-speaking cities show lower trading profit than locals.⁴³

3. Social Trading

The financial industry faces an ongoing transformation. The rise of “fintechs” (financial technology), that allow investors to bypass financial intermediaries, is a big competition to the traditional providers. One of these “fintech innovations” are social trading platforms, which allow users to make financial decisions based on information gathered in online communities similar to Facebook, Linkedin, Myspace etc.

Since the last financial crisis, the willingness to gather information about investment opportunities got a lot higher. As a result of a loss of trust in the traditional financial markets, investment ideas got shared on sites like Facebook. Later on, startups like Ayondo, eToro, and ZuluTrade created the first independent social trading platforms. Since then, a lot of brokerage firms developed their own social trading platform for their users.

A social trading platform consists of two groups of users. On the one hand, there are the signal providers, which make their investment decisions available on the platform. On the other hand, there are signal followers or just followers, which can observe every trade made by the signal provider and replicate those trades. This procedure is known as so-called “copy trading” that allows users to become signal followers by subscribing to one or more signal providers. All signals subscribed by a follower will be automatically processed into the follower’s brokerage account. This will proceed proportionally to the invested amount of money and in real time. Signal providers will be compensated by a platform- specific performance fee for their service based on the success of their investment strategy.⁴⁴

The initial service offered by the first social trading platforms was that incoming trading signals can be manually approved or declined by the follower. This is still offered by some platforms and known as manual replication. Yet, copy trading established itself as the predominant method on social trading platforms.⁴⁵

Social trading networks provide a different approach for delegated portfolio management. Signal providers on these networks act as portfolio managers, even though there are no funds transferred from the followers to the signal providers account. In 2015 the "Federal Financial Supervisory Authority” BaFin made a statement about the regulations of social trading platforms. In their annual report of 2015 the BaFin says:

” In social trading, the operators of special platforms manage public portfolios for signal providers in order to make their trading activities visible. Customers can link their own portfolios to these reference portfolios: In this way, trade decisions taken by the signal provider are automatically also implemented for their own account. The signal provider is, as a rule, attributed to the platform, which complements the provider’s activity by executing concrete customer orders and acting as the promoter of the business model. This means that the platform normally provides at least portfolio management within the meaning of the Banking Act.’’⁴⁶

According to their statement, the BaFin sees copy trading as portfolio management and therefore, the platform operators require equivalent authorization.

The agency problem and renumeration are two big challenges a social trading platform has to face. To solve this, the platform establishes standardized real time track records for every signal provider, based on historical trades. Instruments like search functions and rankings are added as a feature. When a follower subscribes to a signal provider he can track trading activities in real time, as they are mirrored into his account. Compared to classical portfolio management social trading leads to higher transparency. However, it is not visible whether a good performance is based on pure skill or only by chance. Given that every internet user can join a social trading platform, the differentiation between amateurs and experts is complicated. Huddart (1991) noted, that charlatans that bet on pure luck to build a solid track record are more likely on platforms with a low entry barrier.⁴⁷ Therefore, it is unknown whether a social trading platform is a better way of mitigating the agency problem than the classical portfolio delegation. However, the transparency setting, social trading provides, may diminish the risk of moral hazard by the signal provider.⁴⁸

In order to reduce moral hazard, the platform tries to align the interests of the signal provider and follower by the selection of the renumeration scheme. Several studies underline the impact of a compensation structure on the risk a signal provider is willing to take and the effort he is willing to invest.⁴⁹ Typical renumeration schemes for social trading platforms are profit-based models, follower-based models and volume-based models.

A follower-based model compensates the signal provider, based on the number of users following his portfolio. This scheme is used by one of the largest platforms, measured by the amount of users, “eToro”. They pay a fixed sum depending on the number of followers, which assign at least $100 and $20,000 at the most to the signal provider. The number of followers can be seen as a proxy for asset assigned to the signal provider, which is comparable to the asset- based fee typically charged by mutual funds.

[...]

---

¹ Cf. Odean (1998), pp. 1781-1782.

² Cf. Pleßner (2017), p. 2.

³ Cf. Shefrin and Statman (1985), pp. 777-790.

⁴ Cf. Kahneman and Tversky (1979), pp. 263-291.

⁵ Cf. Kahneman and Tversky (1979), p. 288.

⁶ Cf. Kahneman et al. (1990), pp. 1325-1348.

⁷ There is an illustration for the mug experiment in the appendix.

⁸ Kahneman and Tversky (1979), p. 279.

⁹ Weber and Camerer (1998), p.173.

¹⁰ Cf. Weber and Camerer (1998), pp. 172-174.

¹¹ Cf. Thaler and Johnson (1990), pp. 648-651.

¹² Cf. List (2003), pp. 41-71.

¹³ Cf. Hens and Vlcek (2011) pp. 141-142.

¹⁴ Cf. Thaler (1980), pp. 39-60.

¹⁵ Cf. Prelec et al. (2001), pp. 5-12.

¹⁶ Cf. Loomes and Sugden (1982), pp. 805-824.

¹⁷ Cf. Filiz-Ozbay and Ozbay (2007), pp. 1407-1418.

¹⁸ Cf. Zeelenberg et al. (1996), pp. 148-158.

¹⁹ Cf. Shefrin and Statman (1985), pp. 777-790.

²⁰ Cf. Thaler and Shefrin (1981), pp. 392-406.

²¹ Barber et al. (2007), p. 425.

²² Cf. Zuchel (2001), pp. 15-18.

²³ Cf. Festinger (1957), p. 21-24.

²⁴ Cf. Weber and Camerer (1998), pp. 167-184.

²⁵ Cf. Kaustia (2010), pp. 18-19.

²⁶ Cf. Odean (1998), p. 1782.

²⁷ Cf. Odean (1998), pp. 1781-1783.

²⁸ Cf. Dhar and Zhu (2006), pp. 726-729.

²⁹ Cf. Barber and Odean (2001), pp. 261-263.

³⁰ Cf. Rau (2014), pp. 33-34.

³¹ Cf. Costa Jr. et al (2008), pp. 411-413.

³² Cf. Shapira and Venezia (2001), pp. 1573-1575.

³³ The different stages will be explained more detailed in chapter 3.0.

³⁴ The specific compensation systems will be illustrated in chapter 3.0.

³⁵ Cf. Lukas et al. (2017), pp. 1-2.

³⁶ Cf. French and Poterba (1991), p. 223.

³⁷ Cf. Black (1974), p. 337.

³⁸ Cf. Tesar and Werner (1995), p. 467.

³⁹ Cf. Warnock (2002), p. 795.

⁴⁰ Cf. Coval and Moskowitz (1991), p. 2046.

⁴¹ Cf. Huberman (2001), pp. 659-661.

⁴² Cf. Coval and Moskowitz (2001), pp. 811-813

⁴³ Cf. Hau (2001), pp. 1959-1960.

⁴⁴ Cf. Oehler et al. (2016), pp. 202-205.

⁴⁵ Cf. Doering and Neumann (2015), pp. 1-2.

⁴⁶ BaFin Annual Report (2015), p. 39

⁴⁷ Cf. Huddart (1999), p. 260.

⁴⁸ Cf. Doering and Neumann (2015), p. 4.

⁴⁹ Cf. Baker et al. (1988), p. 593.