There are two kinds of market timing measures, the return-based measures and the holdings-based measures. The holdings-based measures use observations from portfolio holdings, which make themselves more informative than the return-based measures. To overcome the problem of insufficient observations resulted from the short sample period of Taiwan mutual fund market, this paper employs a bootstrapping procedure to make statistical inference. We then conduct cross-sectional tests on the market riming coefficients obtained from both the return-based and holding-based measures. Our empirical evidence shows that mutual fund managers in Taiwan do not have the market timing ability.