The Term Structure of Stock Return Predictability

Abstract

Periodic spikes and waves in daily stock return predictability appear across quarterly and other frequencies and dissipate at more distant lags in the term structure of predictability. A 'long ripple' across the term structure spans more than one year of lags. The term structure's level and slope are forecasted by market conditions such as proxies for liquidity and panic, volatility and several standard factors. When returns are aggregated over longer intervals, these patterns manifest as familiar reversal, momentum and seasonality patterns. An out-of-sample monthly investment strategy which conditions on disaggregated daily returns has 1.4% monthly alpha. These empirical results support theoretical models in which short term reversal and momentum are related phenomena and models in which predictability reflects mispricing caused by periodic clientele flows across factors.

Keywords: momentum, reversal, seasonalities, predictability term structure, stock returns

JEL Classification: G11, G12, G17