A note on market timing: Interim trading and the performance of holdings-based and return-based measures

Abstract

Market timing is the ability of portfolio managers to anticipate stock market return by increasing (decreasing) portfolio sensitivity in upward (downward) markets. To assess market timing, the financial literature has proposed return-based and holdings-based measures. Our objective is to analyze the performance of these measures in relation to interim trading bias. This bias is due to managers trading between the observation dates used to measure timing. As managers' timing decisions are not observable we run the empirical analysis over a data set of simulated portfolios. This paper shows how holdings-based measures may lead to biased results if the timing estimation window does not match the managers' timing decision window. Evidence is found that holdings-based measures are unable to detect daily timing ability. Also, these measures are not unbiased to measure monthly timing if estimation and decision windows do not coincide on the same days; in consequence, there is an underestimation of the actual timing parameter and its significance.