The purpose of this thesis is to examine whether or not liquidity timing funds outperforms in Korean equity funds. To do so, we use 3,918 equity funds, which exist more than 24 months for the sample period of January 2001 to May 2013. In this thesis, we adopt two market liquidity measures because market liquidity is not readily observable: namely, the Pastor and Stambaugh(2003) liquidity measure and the Amihud(2002) illiquidity measure. Furthermore, to examine liquidity timing ability of each fund, we use the augmented four-factor liquidity timing model by incorporating liquidity timing component into standard four-factor model, which are proposed by Carhart (1997). Using this augmented four-factor model, we divide our sample funds into two categories (skilled and unskilled funds), depending on the significance and sign of their liquidity timing coefficients. After dividing our sample funds, we construct two equally weighted portfolios using skilled funds and unskilled funds. We estimate 17 performance measures, including the Sharpe ratio, Jensen''s alpha etc, based on the equally weighted portfolio returns. In addition, we adopt the in-sample test and out-of-sample test because investors do not know which funds have liquidity timing ability in prior.
Overall, we find that skilled funds generally do not outperform compared to unskilled funds in the in-sample test using the Pastor and Stambaugh(2003) liquidity measure. As a robustness test, we also examine the liquidity timing ability using the Amihud(2002) illiquidity measure. However, the results are qualitatively different.
In addition, we divide our sample period into two sub-periods: pre-2006 and after-2006 to consider the effect of increasing market size of Korean equity funds, which are reported by Kim and Seo(2011). The estimated results are consistent with the whole period analysis.
Investors do not know which funds have timing ability when they choose the equity funds. This implies the importance of the out-of-sample test. To construct equally weighted portfolios for a one-year rebalancing interval at the end of each year, we estimate the liquidity timing coefficients and their t-values for existing funds using the previous two-year. As in the in-sample test, we use the significance and sign of the estimated coefficients to divide the skill and unskilled funds. This portfolio is held for one year. In other words, the selection procedure is repeated after one year. In the out-of sample test, we also find that skilled funds do not outperform when adopting the Pastor and Stambaugh(2003) liquidity measure. In addition, the results of the Amihud measure show skilled funds do not show higher outperformance than unskilled funds.
When we look at the result of 17 individual performance measures in the in-sample and out-of-sample tests, they show that the fund performance could be different depending on which performance measures are adopted. This implies that the fund performance is related to the choice of performance measure, as shown in Ornelas et al.(2010).