Footnote:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments March 15, 2023 erstellt
Description:
We suggest a procedure to predict individual stock liquidity and study the relation between stock liquidity forecasts and average stock returns. Our forecast model reduces the root-mean-squared error by 12% for the Amihud (2002) liquidity measure compared to realized stock liquidity in the previous month. Our liquidity forecasts capture economically large changes in liquidity and improve in accuracy over time. Whereas liquidity measures provide mixed empirical results in asset pricing tests, we find a strong relation between expected changes in liquidity and average stock returns. Sorting portfolios according to the expected change in illiquidity in the next period leads to a monthly excess return of the long-short portfolio of 1% for equally-weighted and 0.8% for value-weighted portfolios. Our results are robust to controlling for various predictors of stock returns in Fama and MacBeth (1973) cross-sectional regressions and to using the effective spread as an alternative liquidity measure. The large return premium of expected illiquidity changes can be explained with portfolio re-allocations of investors with heterogeneous investment horizons. Consistent with the clientele effect of Amihud and Mendelson (1986), mutual funds with short investment horizons sell (buy) stocks for which liquidity is expected to deteriorate (improve). Funds with longer horizons, the natural counterpart, do not react on expectations but base their portfolio decisions on realized illiquidity, leading to a temporal mismatch between supply and demand