EMPIRICAL EVALUATION OF OVERCONFIDENCE HYPOTHESIS AMONG INVESTORS THE EVIDENCE IN VIETNAM STOCK MARKET Phan Nguyen Ngoc Xuan My Huynh Luu Duc Toan and Nguyen Thi Kim Cuong June 2016 EMPIRICAL EVALUATION OF OVERCONFIDENCE HYPOTHESIS AMONG INVESTORS - THE EVIDENCE IN VIETNAM STOCK MARKET Phan Nguyen Ngoc Xuan My, Huynh Luu Duc Toan, and Nguyen Thi Kim Cuong June 2016 ABSTRACT. This paper highlights the role played by overconfidence bias in investors’ behaviors of finance. Using Vietnam stock market data sets during the period 2008 — 2015, this paper provides the quantitative research of the overconfidence hypothesis in Vietnam: market gains (losses) increase (decrease) investors’ confidence, and consequently they trade more (less) in subsequent periods. Overall, we find empirical evidence that an increase in stock returns (VN- Index) is always followed by an increase in trading volume, as well as an increase in the Vietnam Investor Confidence Index ® (VICI), as a proxy for investors’ confidence.
We further investigate the contemporaneous relations between the three variables. The analysis shows that the more confident investors are, the more trading volume they exercise, and unfortunately the less return they can gain. Key words: behavioral finance, overconfidence bias. Phan, Nguyen Ngoc Xuan My Huynh, Luu Duc Toan Nguyen, Thi Kim Cuong Massey University Toulouse 1 Capitole University Graduate, Foreign Trade University 1.
Introduction The important assumption that all investors are rational underlies the conventional asset pricing models. However, empirical literature consistently illustrate that those models do not explain some of stylized facts observed in securities markets’. There is currently a growing concern among researchers who argue that the failure of the conventional asset-pricing model is critically due to the inappropriateness of the rationality assumption. There are developing research lines to explain such phenomenon, including models based on special trading strategies taken by irrational investors”, models of investors’ cognitive bias’, models on “limit to ot arbitrage”, the momentum effect models”, and the negative long-term autocorrelations in many asset and securities markets”.
Recently, behavioral finance models have been motivated by offering a unified explanation of short-run underreaction and long-run overreaction. For example, Daniel, Hirishleifer, and Subrahmanyam (1998) (hereafter, DHS) state that trading volume in speculative market is too large, and volatility of asset prices relative to fundamentals is also too high. Trading motivated from hedging and liquidity purposes is likely to explain only a small fraction of the observed trading activity and fails to support a large amount o f informational trade. Overconfidence has been advanced as an explanation for the observed trading volume and volatility.
In short, the overconfidence hypothesis, among other things, offers the following testable empirical hypothesizes. First, overconfident investors have a tendency to overreact to private information and underreact to public information. Second, an increase in market gains (losses) ' Fama (1998) and Daniel, Hirshleifer, and Subrahmanyam (1998) review the literature on those anomalies. Moreover, Daniel, Hirshleifer, and Teoh (2002) and Heaton and Korajezyk (2002) discuss those anomalies.
? Cutler, Poterba, and Summers (1990, 1991) and De Long et al. (1990b) indicate that some irrational traders do not take negative feedback trading strategy which can help to explain short-term momentum and long-term reversal. Bange (2000),Choe et al. (1999) and Grinblatt and Keloharju (2000) show evidence that certain classes o f investors engage in positive feedback trading.
> Barberis, Shleifer, and Vishny (1998) provide a model measuring investor sentiment based on two assumptions of cognitive bias: conservatism and representative heuristic; while Daniel, Hirshleifer, and Subrahmanyam (1998) develop a theory based on alternative assumptions: overconfidence and self-attribution bias. Gervais and Odean (2001) propose a muti-period market model showing how a learning bias impacts on overconfident level of traders 4 Delong et al. (1990) state that noise traders can create price risks on risky asset which deters rational arbitrageurs from actively hedging against them. Black (1986) and Barberis and Thaler (2002) discuss about “limits to arbitrage”.
> Possible explanations for momentum include data mining, risk, and behavioral patterns. However, in some empirical tests, risk and data mining finds it difficult to explain the effect (e., Jegadeesh and Titman (1993, 2001, 2002), Fama and French (1996), Conrad and Kaul (1998), and Rouwenhorst (1998, 1999)). © DeBondt and Thaler (1985, 1987), Fama and French (1988), Poterba and Summer (1988), Culter et al. (1991), and Richards (1995,1997) Page 1 leads to an increase (decrease) in investors’ overconfidence, and consequently they trade more (less) aggressively in subsequent periods.
Third, as overconfident investors, they fail to estimate risk appropriately and, as such, trade riskier securities. Fourth, excessive trading by overconfident investors in securities markets makes a contribution to the observed excessive volatility. Previous empirical studies have found evidence on various implications of the overconfidence hypothesis. Odean (1998b), and Gervais and Odean (2001) develop their models, which show evidence of the second hypothesis that implies a positive causality running from stock return to trading volume.
In Vietnam, there are some studies which indicate the impacts of behavioral finance on Vietnam Stock Market. Tran Thi Hai Ly (2011)’, and Nguyen Duc Hien (2012)Ẻ shows the model to measure which factors contribute to investors’ behaviors. Therefore, on a stock market in general and on the Vietnamese one in particular, investment decisions are not only affected by conventional financial theories, but also driven by various factors, among of which is behavioral finance. In other words, investment decisions or investors’ behaviors rely on psychological factors.
Whether or not an investor can constantly make rational decisions? According to behavioral financial theories, investment decisions are influenced by psychological factors, namely overconfidence, herd mentality, uncertainty, etc. Featuring the nature of an immature market where there are numerous individual investors and speculation frequently happens, Vietnam stock market is subject to behavioral factors, especially investors’ overconfident level. Therefore, the study of behavioral psychology proves to be reasonably necessary to the market and investors, particularly in the current period when Vietnam has finished TPP negotiation and is subject to different opportunities and challenges. The specific objective of this study is to show the empirical evidence on the second hypothesis in Vietnam Stock Market over the period 2008 - 2015 by focusing on stock returns, trading volume and investor behavior.
We follow and build upon the approach by Odean (1998b), and Gervais and Odean (2001), and analyze the link among the three factors mentioned above. This implication is tested by performing the bivariate Granger causality tests from stock 7 See Tran Thi Hai Ly (2011) — “The impacts of psychology on individual investors’ behavior in Vietnam Stock Market”. The study shows the model to measure the factors contributed to investors’ behavior. 8 See Nguyen Duc Hien (2012) — “The investor behavior on the Vietnamese stock market”.
The study uses questionare method to develop a model to measure behaviors of individual investors with five main bias: (1) Overoptimism; (2) Herding mentality; (3) Overconfidence, (4) Risk aversion, and (5) Pessimism. Page 2 retumn to trading volume. Our results show that the null hypothesis that stock returns do not Granger-cause trading volume is rejected. The Granger-causality tests for the four monthly variables of trading volume, stock returns, Vietnam Domestic Investor Confidence Index ® (VDIC), and Vietnam Foreign Investor Confidence Index ® (VFIC), which are used as proxy for investors’ confidence’, are also performed to indicate evidence that the positive causality running from stock returns to trading volume is due to investors’ overconfidence enhanced by stock returns.
Our results show that stock returns positively Granger-cause both the VDCI and trading volume, which implies that an increase in Vietnam stock returns makes only domestic investors become more confident and consequently trade more aggressively in subsequent periods. This finding is important since it provides (indirect) evidence to disentangle the overconfidence hypothesis. Furthermore, we find evidence that the VDIC slight Granger-causes stock returns. Besides, we do not find the evidence to support the Granger causality between stock returns, trading volume and foreign investors.
This finding seems to suggest that the foreign investors are fairly neutral and the market is still quite efficient in that investors’ overconfidence doesn’t drive the market. Although the results from the Granger causality tests are consistent with the prediction of the overconfidence hypothesis, care must be taken to make a conclusion that our hypothesis is supported by empirical examination before we find evidence that there exists a positive causal relation between the lagged Vietnam Confidence Index and current trading volume and stock returns. By performing the Ordinary Least Squares (OLS) regression on the Vector Autogressive Model (VAR) and the OLS with HAC — Newey West standard errors and covariance, we find that the three main following findings. First, there is the strong positive causal relation between lagged monthly stock returns and current monthly trading volume, but there is not any causal relation from trading volume to stock returns.
This implies that stock return is not driven by trading volume. We also do not find the evidence that show causal relation between confident level index of both domestic and foreign investors, and the trading volume. This implies that, the confidence level index contains no additional information to predict trading volume. Another explanation is that the influence of the ° Fisher and Statman (2002) find that there exists a positive and statistically significant relationship between changes in the American Association of Individual Investors (AAID measure o f investor sentiment and changes in the Index of Consumer Sentiment and that the Index of Consumer Sentiment goes up and down with stock returns (see also Fisher and Statman (2000)).
Page 3 investors’ confidence index on trading volume lasts faster than one month, and then the use of monthly variables may fail to capture the relation between them. Second, there is a negative relation of Vietnam Domestic Investors Confidence to stock return with the lag of 1 month and a positive relation of trading volume to Vietnam Domestic Investors Confidence. This indicates that the more confident investors are, the more trading volume they exercise, and unfortunately the less return they can gain. In previous studies, Barber and Odean (2002) find that investors who have often earned high returns are more likely to switch from phone-based to online trading.
Online investors trade more frequently and perform worse. They argue that one important reason for the switch is overconfidence. In retrospect, Vietnam has changed to launch online trading system in Vietnam stock market since 2008. Third, we find that domestic investors have a tendency to last the positive effect in the last one month when the stock return increase, and consequently, they trade more aggressive and then get loss in the next period (one month), which makes them regret after that (two months).
This is also supporting evidence for the second finding above. Due to the result of Granger causality and the OLS regression, we find that there is no relationship between Vietnam Domestic Investor Confidence Index and Vietnam Foreign Investor Confidence Index. For the purpose to robust the evidence, we perform Quantile regression. The finding shows that there exits a causal impact from foreign investors to domestic ones, which is more and more influent on the investors who show high volatility of their overconfident behavior in the Vietnam stock market.
It shows that Investors’ overconfidence is posited to be stronger in a bull or bear market (DHS (2001)). The paper is organized as follows. We briefly review related literature in Section 2. Section 3 presents the data and methodology.
In section 4, we discuss the empirical results of the tests of overconfidence hypothesis. Section 5 produces concluding remarks. The final section offers some implications. Overconfidence theory Overconfidence is the psychological state in which a person’s subjective confidence in his judgments relies on weak reasoning, evaluation and intuition.
Upon estimation of the likelihood of a certain thing, investors often made inaccurate assessment because they assume themselves smarter or consider the information they receive as more valuable than other investors on the market. The reasons for this psychological state are the tendency in which investors find complementary information to the existing and the desire to become more professional and proficient than others.