Momentum in Taiwan: seasonality matters!

Abstract

Previous studies suggest that momentum exists in international stock markets with the exception of Asia. Using a large data set of Taiwanese stocks, we show that momentum does exist, but it is restricted to the months following the deadline for annual statements. During the remaining months, a reverse momentum, or contrarian, strategy produces significant returns. These contrarian returns are particularly high during the national holidays linked to the Lunar New Year and the Lunar Moon Festival.

Acknowledgements

The author, Fu, acknowledges Providence University grant PU96-11100-B01.

Notes

¹The exception is Hong Kong. It is worth noting that the authors do find significant momentum profits for a pan-Asian strategy that excludes Japan, with an average monthly momentum return of 1.03% (t-stat. 2.53) for the period 1976 to 2000.

²This assumption is commonly made in the noise trader literature. Using brokerage data from China, Chen et al. (2007) demonstrate that individual investors are more prone to behavioural biases than institutional investors.

³Data is from the website of the TAIEX.

⁴It is interesting to note that although previous studies find a January momentum effect in both developed and emerging markets, there is less evidence in support of a general January effect in emerging markets (Cheung and Coutts, 1999; Fountas and Segredakis, 2002) while Zhang et al. (2008) shows how active governments can be an additional source of seasonality that may not be restricted to January.

⁵Comparable studies of momentum in Taiwan include Rouwenhorst (1999) with 119 stocks, Chui et al. (2000) with between 53 and 325 stocks, Hameed and Kusnadi (2002) with 92 stocks, Chen (2000) with 100 stocks and Hong et al. (2003) with 228 stocks.

⁶The TAIEX generally trades 6 days per week, with the exception of the month of February when the market is closed for a substantial period. To illustrate, during the years 1999–2005, there was an average of less than 16 trading days during the month of February.

⁷The exception to this was the brief introduction of CGT during 1989. The tax proved to be extremely unpopular and was immediately withdrawn.

⁸It is interesting to note the pattern of returns for the individual winner and loser portfolios broadly correspond with the ‘sell in May and go away’ anomaly identified by Bouman and Jacobsen (2002).

⁹Although listed firms are required to produce quarterly earnings statements, only the annual and semi-annual statements are audited by independent certified public accountants. In contrast, the two quarterly statements are reviewed but not audited by certified public accountants. Moreover, comparing the annual and semi-annual statements, the annual statement is more comprehensive. Hence, the annual statement has substantially more market impact than the other three and represents the most important source of firm-specific news.