Some investors are curious about whether there is a disposition effect in the real world. The matter has been settled, since there is obvious empirical evidence to support an occurrence of the disposition effect in the markets, including stock markets. However, underlying causes for it are still ambiguous. Some claim that the disposition effect could be explained by various theories, both traditional and behavioural theories. Another observer argues that the explanations cannot be captured and fully understood by those theories. To prove that the disposition effect occurs in the markets, Odean (1998) performed a test on 10,000 customer accounts provided by a nationwide discount brokerage house between 1987 and 1993. On each day of the study, each stock in customers’ portfolios was recorded and categorised into one of four positions when it was sold. These four stock positions are sold for gain, sold for loss, not sold and presenting a gain, and not sold and presenting a loss, called realised gain, realised loss, paper gain, and paper loss, repectively. He then used the recorded numbers to calculate the two ratios, illustrated in figure 1. The result shows strong evidence for the existence of disposition effect, as the 14.8 percent PGR ratio exceeded the 9.8 percent PLR ratio for the entire year, and the ratio of PGR to PLR is approximately 1.5. This figure indicates that more than 50 percent of stocks that appreciate in value are more likely to be sold than stocks whose value has declined. The table of results from the test is shown in figure 2. The tendency to sell winners too early and hold losers too long was first official observed by Shefrin and Statman (1985). Understanding the disposition effect and why it occurs is helpful in understanding the trading behaviour of investors. Explanations of this tendency regularly mention prospect theory, which is one of the most widely accepted theories among decision making under risk and uncertainty. This theory was pioneered by Kahneman and Tversky (1979), implying that investors usually perceive outcomes as gains or losses and the value is measured by deviations from the reference point. Investors’ initial purchase prices are supposed to be their reference points. The main feature is the S-shaped value function, which can be seen from figure 3. It displays a concave curve in the domain of gains and a convex curve in the domain of losses. A graph that is steeper for losses than for gains indicates loss aversion. This means that when investors experience gains, they tend to be more risk averse and vice versa when they experience losses. Thus, risk averters may tend to sell stock that appreciates in value, and tend to hold on to stock that declines in value. Considering the example from Shefrin and Statman (1985), an investor purchases a stock at 50 dollars and one month later the price drops to 40 dollars. The investor has two alternatives: selling the stock now to realise a 10 dollar loss, or hold on the stock for one more period. The latter will be selected over the former, as shown by the convex curve of the S-shaped value function. Consequently, the investor will hold on to the loser. Prospect theory generally plays an important role in explaining the disposition effect; hence, there are several researchers working on this article (see e.g. Weber and Camerer 1998; Barber and Odean 1999; and more recently Hens and Vlcek 2009). Barberis and Xiong (2009) were among those who studied two cases in order to determine whether the disposition effect can be predicted by prospect theory. The first case investigated prospect theory in relation to the annual profits of stock trading over a one-year period, and the other was related to the realised gains and losses when some stocks are sold. The first case used 10,000 investors with prospect preferences. Barberis and Xiong also note that they follow Odean’s methodology to calculate PGR and PLR ratios with various expected risky asset returns and trading periods. The result, however, shows that the model is unsuccessful in predicting the disposition effect when the expected return is high, and the number of trading periods is low. The investors tend to hold shares when the prices rise rather than drop, which is opposed to the disposition effect. Moving on to the second case, they perform the test regarding the optimal share holding with a specific expected return on three dates: at the date of purchase (X0), at the date of stocks doing well (Xu), and at the date of stocks doing poorly (Xd). The result is presented in figure 4. When the expected return is equal to 1.11, an investor buys 3.7 shares of a risky asset. If the stock performs well, he/she decreases the holding to 3 shares. On the other hand, if the stock performs poorly, he/she holds the same number of shares, 3.7 shares. This indicates that the disposition effect can be predicted by prospect theory only in realised gains and losses, not in paper gains and losses. Prospect theory, however, does not provide an obvious cause of this effect due to the lack of behavioural considerations. For this reason, alternative behavioural theories such as cognitive psychology and specific emotions were introduced. In Shefrin and Statman (1985)’s framework, they present a concept of mental accounting or narrow framing that was first introduced by Richard H. Thaler. According to Thaler (1984, cited in Shefrin and Statman, 1985, p. 780), investors are inclined to organise different type of gambles faced into separate accounts in their minds. For example, when an investor buys stocks, he/she establishes new mental accounts for each stock in his/her mind. The investor then uses prospect theory to make a decision on each account. In other words, an investor pays attention to single a stock instead of the whole portfolio. This way of thinking might lead to the disposition effect. Considering that the investor faces a paper loss he/she sells off the stock to realise that loss. The mental account is now closed and the investor is pained as the paper loss becomes a certain loss. However, it is difficult to close mental accounts at losses, since that action proves that the investor made a wrong decision in buying the stocks. Hens and Vlcek (2009) also provide an alternative explanation using two mental accounts: realised gains and losses, and paper gains and losses. The paper account has less weight than the realised account. An investor who experiences paper losses is less pained than one who experiences realised losses, and realised gains provide greater happiness than paper gains. That is why investors keep the losers and sell the winners. What concerns an investor is not only an outcome of a decision, but also the feeling about that outcome. To be more obvious, regret theory could be used to explain the disposition effect. When the level of wealth would have been higher with an alternative decision that the investor did not make in the first place, he/she feels regret. As for pride, when the decision the investor made turns out to be better than an alternative decision, he/she feels pride. (Muermann and Volkman, 2006) Shefrin and Statman state that an investor feels regret when a stock account is closed at a loss and feels pride when a stock account is closed at a gain. The circumstance of seeking pride and avoiding regret can lead to the disposition effect. In Summer and Duxbury (2012), the balance of emotions resulting from decision making was investigated. They claim that in the specific decision, experiencing the balance of positive and negative emotions can predict the disposition effect. For instance, when an investor purchases and holds a stock, he/she has to make a decision to hold it or sell it at the end of the first period. The future outcome is uncertain; both currently winning and losing stock can be winning or losing stock in the future. This can create both positive and negative emotions. If an individual owns a loser and sells it to realise a loss, he/she feels regret about buying this stock. To hold on to this loser might be the optimal choice, as it offers a path to avoid negative emotion, namely regret. Summer and Duxbury note that “Thus behavior predicted for regret is consistent with the behavior observed for losers in the disposition effect (continue to hold in an attempt to put things right)” (p. 4). What if an individual owns a winner? Sell the winner creates rejoicing, since a paper gain is turned into a realised gain. The emotions that individuals feel vary depending on the responsibility for outcomes or whether they own stocks. They feel elation or disappointment due to the non-responsibility. An individual, who owns stocks, will also feel joy and regret. Another psychological examination of emotions is Muermann and Volkman’s paper, which introduced the model to measure the amount of regret and pride for investors. An individual focuses only on the realised return of stocks he/she holds, and feels regret or pride for all decisions he/she made. They conclude that the disposition effect can be explained by individuals’ feelings of regret and pride. The reason that investors decide to sell winners and ride losers might not be that they are reluctant to realise losses or avoid regret, but that they believe in price mean reversion. According to Odean (1998), investors hope the losing stocks will outperform currently winning stocks in the future. They might sell winners, as they believe the prices are now reflected in their information. Conversely, they decide to hold losers because they believe the prices have not been reflected in their information yet. The test was carried out by calculating the excess return after selling winners and the paper loss on riding losers. This belief is, however, regularly inaccurate, since the excess return for the sale of winning is higher than paper losses. This point of view is also supported by Hens and Vlcek (2009). Such irrational belief that losers will outperform winners and today’s winners will be losers in the future could lead to the disposition effect. In behavioural explanations, self-control is an essential part to make investors overcome a reluctance to realise their losses. In Thaler and Shefrin (1981)’s framework, an individual is considered as an organisation, which has a conflict between a planner and a doer, called the planner-doer model. The planner has lifetime utility and it is a rational part. The doer, in contrast, relates to only one period of time and it is an emotional part; therefore, it is assumed to hold power to influence individuals’ actions. Selling winners quickly to feel pride and continuing to hold losers to defer regret thus comes from the doer part. The planner might not be powerful enough to stop this action of the doer. The disposition effect then occurs. Alternatively, tax considerations and portfolio rebalancing are investigated to find whether they can be used to explain the disposition effect. Lakonishok and Smidt (1986) studied the turnover of stocks related to taxation. They find that losers have lower turnover than winners. Moreover, the wash-sale rule, which prevents investors from repurchasing the same stocks at least 30 days after selling, is also a good explanation. Investors have a tradeoff between realising tax benefits and not trying to decrease equity exposure. However, this rule does not explain why investors realise gains. (Shefrin and Statman 1985) When investors are faced with a major price movement of a stock in an upward direction, they may sell a part of that stock in order to rebalance their portfolios. Odean (1998), therefore, performs the test by excluding the traders who desire to restructure portfolios and tax considerations, and then calculates PGR and PLR again. The result, however, still exhibits the disposition effect. Odean concludes that tax considerations and portfolio rebalancing do not cause the disposition effect. All of the previously mentioned explanations appearing in various publications are possible causes of the disposition effect. Since they do not fully explain the disposition effect or do not have enough explanatory power in some authors’ viewpoints, a number of studies are proposed to contradict these explanations. Hens and Vlcek (2009) argue that prospect theory cannot explain the disposition effect, as this theory considers ex-post disposition behaviour, not ex-ante. They study individuals’ risk-taking behaviours according to an increase and decrease in stock prices. The ex-ante disposition occurs over two periods of time, whereas the ex-post disposition involves only one period. The behaviour in the second period relates to the movement of stock prices in the first period. The investors are inclined to ex-post disposition behaviour more than ex-ante disposition. This is because “Those investors who sell winning stocks too early and keep losing stocks too long would not have invested in stocks in the first place.”(p. 5). In Zuchel (2001)’s paper, it is also claimed that prospect theory involves a one-period decision making. It does not go along with the disposition effect, which account for sequential decision making. The explanation that the degree of risk aversion falls after losses and rises after gains seems to not have its origins in the prospect theory. Consequently, the disposition effect cannot be explained by the S-shaped value function. Furthermore, Weber and Camerer (1998) introduced the opposite view of the mean reversion. The disposition effect exists only if winners would fall and losers would rise. According to their framework, the belief in mean reversion is wrong. This is because stocks increasing in value are more likely to indicate a positive trend and further rise. This is similar to losers, which are inclined to fall. The explanatory power of regret theory is limited. This is to say, it cannot apply to every situations. For example, it cannot apply to a situation that has more than two alternatives. Zuchel’s framework testing regret theory to find the connection between a current decision and a past price found no such connection, which should exist with the disposition effect. Again in relation to tax considerations, the turnover of selling losers seems to be higher than normal in December and exhibits a smaller disposition effect due to tax benefits. Taxes are generally calculated on a tax-year basis, thus December is the last chance for investors to realise their losses and receive the tax benefits. (Lakonishok and Smidt 1986) Barber and Odean (1998) also support this point of view. As for the self-control issue, the main problem is that investors do not have enough self-control to end their investments at losses. However, various techniques can be used to control investors’ doers. One of the techniques that some professional traders adopt is iron-clad rules. These rules are used to compel traders to realise losses when they reache a predetermined level. For example, traders will sell stocks when losses stretch to ten percent of the purchase prices. The findings of Dhar and Zhu (2002) confirm that the disposition effect, on average, exists among individual investors. Despite this, they discover that approximately twenty percent of their sample investors present the reverse disposition effect. They summarise that the size of the disposition effect can be affected by investor characteristics and sophistication. The wealthier, more professional, more experienced or more frequently trading investors tend to exhibit a smaller disposition effect. Although the disposition effect among individuals exists, there is no strong evidence that it occurs among mutual funds and foreign investors. In Taiwan Stock Exchange (TSE), Barber, Lee, Liu and Odean (2007) find that investors are more likely to realise gains than losses. They, however, do not find the disposition effect in mutual funds and foreign investors. Jin and Scherbina (2011) also note that US equity mutual funds, on average, tend to realise capital losses rather than capital gains because the new managers are more likely to sell off losers from their portfolios. Since mutual funds have more experience obtained from continuous trading in the markets, this makes them more skilled and they can keep away from disposition effect behaviour. In the financial market, investors do not always act rationally. An investor who sells winners to realise gains and holds losers to defer losses exhibits the disposition effect. The persistence of this effect in trading patterns in the financial markets is confirmed by many researchers such as Odean (1998). This paper is, hence, concerned with the review of the disposition effect and the evaluation of its explanations. It documents the possible underlying factors studied by a number of researchers. Those factors, explored in the traditional theory and behavioural theories, could be used to explain why the disposition effect exists. The traditional theory is prospect theory which is the widely accepted model in financial studies. Since the disposition effect is not fully explained by this theory, the behavioural theories, which are mental accounting, regret aversion, self-control and mean reversion, were introduced. Alternatively, tax considerations and portfolio rebalancing are tested to determine whether they are the causes of the disposition effect. However, they are not the underlying causes. Furthermore, some empirical studies and frameworks disagree and argue that the disposition effect cannot be explained by the previously mentioned theories. Therefore, there is still a lack of full understanding of the underlying causes of the disposition effect.