國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
2. Literature Review
We review the relevant literature regarding the relationship between trading activity and prior returns in this chapter. Gervais and Odean (2001) examine the relationship between past returns and trading volume, thus, they develop a multiperiod market model describing both the process by which traders learn about their ability. In the paper, they assume a trader initially does not know his own ability, and he infers this ability from his successes and failures. They fine that trading volume is greater after market gains and lower after market losses. Moreover, the degree of overconfidence is greater after market gains and lower after market losses.
Statman et al. (2006) examine whether the trading volume could predict formal overconfidence models and the market-wide and the positive relationship between individual security turnover and lagged returns for many months. Using monthly observations on all NYSE/AMEX common stocks, excluding closed-end funds, REITs, and ADRs, from August 1962 to December 2002, the span of the daily CRSP files, they find that both market-wide and individual security turnover are responsive to past market returns, and overconfidence effect that market returns is positively related to volume is more pronounced in small-cap stocks and in earlier periods where individual investors hold a greater proportion of shares. Kim and Nofsinger (2007) test this hypothesis using Japanese market level data. They identify stocks with
‧
institutional investors. Moreover, the relation is stronger for developing countries than developed ones; the result shows that turnover is more sensitive to adverse return shocks in markets with short-sale restrictions and high levels of corruption.
Odean (1999), Barber and Odean (2002) use a dataset from a U.S. discount broker to analyze individuals. Odean (1999) finds that the individual investors reduce their returns through trading; thus, they trade excessively. Their result is consistent with overconfidence hypothesis. Barber and Odean (2002) analyze 1,607 investors who switched from phone-based to online trading during the 1990s. They find that those who switch to online trading perform well prior to going online and beat the market, after going online individual investors trade more; however, they reduce their profitably. Besides, Glaser and Weber (2009) use 3,000 individual investors who have discount broker accounts in German over a 51 month period to examine whether prior market or portfolio returns are following subsequent stock trading volume, and other trading activities. They find that past market returns and past portfolio returns affect trading activity of individual investors, such as the stock portfolio turnover, the number of stock transactions, and having a tendency to trade stocks in a given month.
Moreover, their results also show that individual investors will buy high risk stocks and reduce the number of stocks in their portfolio after experiencing high portfolio returns. However, high past market returns do not have the same effect as high portfolio returns; individual investors do not increase their risk taking or underdiversification. In addition, the results approve overconfidence is based on self-attribution and the explanation variable in trading activity.
O’Connell and Teo (2009) using data on mutual fund managers’ trading decisions test the effects of trading gains and losses on risk-taking among large institutional investors. Their results show that large institutions take on more risk following gains in both long position and short position gains.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
While a dataset in Taiwan Futures Exchange, Liu et al. (2010) find that the three different types of market makers the relationship between prior outcomes and subsequent risk taking is positive. Chou and Wang (2011) examine overconfidence biases by directly testing the signed trading volume and order aggressiveness of traders following trading gains from their prior positions in different trader types - foreign institutions, domestic institutions, futures proprietary firms, and individual traders. They find that investors tend to buy (sell) more after they have previously hold a long (short) position and gains; moreover, the overconfidence biases in individual traders are the strongest in all trader types.
Barberis and Xiong (2009) use a two-period model to study the trading behavior of an investor who derives utility from realized gains and losses with a utility function that is concave over gains and convex over losses. They find that investors may tend to increase their risk taking after prior gains.
There are other researches that focus on whether overconfidence bias affects the trading activities. There are a lot of researches find that the more overconfidence bias , the more trading activities, such as, Odean (1998), Barber and Odean (2001), Campbell et al. (2004), Malmendier and Tate (2005), Menkhoff et al. (2006), Grinblatt and Keloharju (2009), and Li and Tang (2010).
Odean (1998) develops models to test whether overconfidence investors
‧
with overconfidence models that gender is a substantial number of other attributes that might affect trading.
Campbell et al. (2004) examine the relationship between narcissists and overconfident, and the relationship between overconfident and decision making. They find that narcissism is positive related to overconfidence. Besides, the willing to accept risk is following increasing confidence though the value of bets is decline.
Menkhoff et al. (2006) examine the impact of experience on overconfidence and risk taking of 117 German fund managers. They find that the returns of inexperienced fund managers are greater than of experienced ones. However, their results are not able to explain the relation between experience and overconfidence and risk taking.
Although the results are signification positive, they might be based on the importance of learning.
Grinblatt and Keloharju (2009) use equity trading data from Finland to examine that two psychological attributes, sensation seeking and overconfidence, are related to the tendency of investors to trade stocks. After controlling a number of variables that might explain trading activity, such as wealth, income, age, number of stocks owned, marital status, and occupation by using cross-sectional regressions, they find that both overconfident investors as well as those investors who are prone to sensation seeking trade more frequently.
Using a sample of firms, Malmendier and Tate (2005), and Li and Tang (2010) find that overconfident CEOs will increase companies’ and themselves risk taking.
Malmendier and Tate (2005) use a dataset consist of 477 large publicly listed on Forbes 500 and traded U.S. firms from the years 1980 to 1994. They examine the relationship between managerial overconfidence and corporate investment decisions.
They define that CEOs are overconfident if they do not reduce their personal portfolio exposure to company-specific risk. Their results show that the sensitivity of
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
investment to overconfident CEOs is significantly positive related to cash flow, especially for firms have abundant internal funds. Li and Tang (2010) use a dataset from 2,790 CEOs of diverse manufacturing firms in China to examine whether CEO hubris lead to firm risk taking. Their results show that the relationship between CEO hubris and firm risk taking is significantly positive.
There are other studies that regarding the relationship between prior outcomes and subsequent risk taking. Barberis and Xiong (2009) examine whether subsequent investment decisions is related to prior trading outcomes. They find that when preferences are defined as realized gains and losses not annual gains and losses, investors may tend to increase their risk taking after prior gains.
Liu et al. (2010) use a dataset from Taiwan Futures Exchange to examine whether the risk-taking in the afternoon is related to morning profitably. They find that market maker take above-average risks in afternoon trading after morning gains. Their results are consistent with the controversy that prior outcomes affect subsequent risk taking through a relationship that is sensitive to the model parameters.
While investors are professional traders, Coval and Shumway (2005), and Locke and Mann (2005) examine whether professional traders are able to avoid irrational behaviors. Coval and Shumway (2005) find that widespread losses in the morning lead to increases in short-run afternoon volatility but no increase in volatility measured over longer intervals. Their result is consistent with the argument that any