Evidence of herding and positive feedback trading for mutual funds in emerging Asian countries

Abstract

When analysing the behavior of investors, the emphasis is usually on positive feedback and herding behavior, and the existing literature abounds with studies on the domestic strategy of mutual funds or on their impact. Due to the advantages in terms of the data, many studies investigate US data. However, with the increased flows of capital into emerging markets, studying the behavior of international mutual funds in emerging markets has become more and more important. Nevertheless, studies involving emerging markets are relatively rare. This study examines whether the positive feedback effect and herding behavior exist in Asian markets based on mutual fund data covering the period from 1996 to 2004. The long period enables us to test the sensitivities under the following four conditions, namely the capital volatility (volatile vs. stable), the degree of suffering during the Asian crisis (more suffering vs. less suffering), and the timing of the Asian crisis (pre-, during, and post-crisis), using the exchange rate regime. It was found in this study that mutual fund inflows into the Asian market were attracted by positive stock returns and currency appreciation. Furthermore, it was found that the positive feedback effect and herding behavior did exist in the Asian markets. However, the extent of the above behavior is not the same under different conditions.

Keywords:

• Herding
• Financial crisis
• Mutual fund
• Positive feedback effect
• Behavioral effect

Notes

⊥ Regarding the origin of the Asian financial crisis, financial and corporate sector weaknesses that were combined with macroeconomic vulnerabilities sparked the crisis. Formal and informal currency pegs, which discouraged lenders and borrowers from hedging, also contributed to the outbreak. Capital inflows helped fuel rapid credit expansion, which lowered the quality of credit and led to asset price inflation (Lindgren et al. 1999).

† Studies using US data, include Grinblatt et al. (1995), Warther (1995), Fant (1999), Nofsinger and Sias (1999), Wermers (1999), Edelen and Warner (2001), Badrinath and Wahal (2002), Sias (2004), Dass et al. (2008). Other studies include Wylie (2005) for the UK market, Kim and Nofsinger (2005) for the Japanese market, and Kim and Wei (2002a, b) for the Korean market, while Borensztein and Gelos (2003) focused on developing countries and Richard (2005) on the Asian market.

† The correlated coefficient and VIF test have been conducted to test the problem of co-linearity. For example, the mean VIF is very low, 1.26, showing there is no co-linearity problem among variables. Also, the Granger causality test was conducted. The F test failed to reject the null hypothesis that the independent variable (KI) does not Granger cause the dependent variables. Thus, a model with lagged variables may not be necessary.

† Davis and Lleo (2008) conclude that a risk-averse manager will tend to track the benchmark very closely, giving up appreciation potential in order to reduce the chance of falling behind the index.

† Proykova et al. (2003) find that emerging markets, such as Bulgaria, show higher fluctuations than American stock markets.

‡ Previously (shown in appendix A) it was found that the positive feedback effect and the herding behavior applied during the financial crisis and after the financial crisis, but did not apply before the crisis. However, in this revision, positive feedback and herding behavior during the crisis are not always supported. For example, the coefficients of StockR_{t −2} before, during and after the financial crisis are 1.918, −4.796, and 5.664, respectively, indicating that positive feedback is more prevalent after the crisis. On the right-hand side of table 7, testing the herding behavior, the coefficient of KI_{t −1}_BC is significantly negative, but this does not apply to KI_{t −1}_AC, which does not attain any significance. Thus, reverse herding behavior does exist before the crisis. Appendix A is provided for reference.

† We appreciate a referee's suggestion for dealing with the fixed exchange rate problem by excluding China and Malaysia from the sample. As a result of the decrease in the number of observations from 1034 to 860, the empirical results are not satisfied. For example, the coefficients of the exchange rate return remain negative, but become insignificant in many cases. Compared with the original empirical results which include China and Malaysia in the sample, the new results become insignificant. In addition, the most interesting finding in the previous work, i.e. that the positive feedback effect and the herding behavior only apply during the financial crisis and after the financial crisis but do not apply before the crisis, no longer applies when China and Malaysia are excluded. The reasons behind this unsatisfactory result probably include the following. (1) China and Malaysia are classified by the IMF as countries adopting fixed-peg arrangements (according to the annual report of the IMF in 2002 and 2004). However, the truth is that the Malaysian Ringgit (against the US dollar) was still devalued from 2.53 to 3.80 during 1996–2004; the devaluation rate thus reached a level of 33.42%, which is far from zero. On the contrary, China's RMB appreciated by 0.24%. Thus, the exchange rates in these two countries are not fixed in practice. (2) About one-fifth of mutual funds flow into China and Malaysia in terms of flows in Asian markets (refer to table 2, the maximum column). (3) China is also one of the BRIC countries, and was recommended by Goldman Sachs in 2003, during our sample period. The report stated that the BRIC countries, namely Brazil, Russia, India and China, will become leading economic powers in the world by 2050. In this revision, the referee suggests including China and Malaysia in the sample accompanied by a caution when referring to the coefficient of the exchange rate.