Abstract: The objective and contribution of this study is to analyse market timing over non-simultaneous periods. This approach considers that decisions on portfolio risk could affect the fund return in subsequent periods and not only the simultaneous period. Robust estimates of changes in beta are computed by Kalman filtering. Initial results for a sample of Spanish mutual funds do not evidence market timing ability in general, although a higher number of funds, particularly larger funds, present negative timing. The study shows how the evidence of negative timing is more robust and persistent for a longer term window. For shorter terms the evidence is driven by an omitted benchmark bias from negative timing of small cap stocks. A comparison of these results with those achieved by a set of passive benchmarks following a buy and hold strategy demonstrates that the long term evidence of negative timing in mutual funds is the result of management contrary to a buy-and-hold strategy.