Survivorship bias and alternative explanations of momentum effect

This paper provides the first detailed examination of momentum effect in Australian equity market. In contrast to previous research, we find that momentum effect has not been present in Australian market since late 1970s. We argue that previous research found strong momentum effect because they assumed perfect foresight of future delisting or acquisitions in the sampling process. In addition, we find that Fama and French three-factor model cannot explain the mean momentum returns although it can fully rationalize the returns on winners and losers portfolios. Our findings raise awareness in the literature that momentum effect is not robust to different sampling methods. We contribute an alternative explanation to momentum returns documented in existing literature. Momentum effect could be a product of look-ahead bias incurring from the sampling techniques. More importantly, we provide supports to the efficient market hypothesis at weak form.