How Precision of the Sharpe Ratio Improves With Monthly Data

Abstract

Practitioners often estimate the Sharpe ratio using annualized monthly data. This paper demonstrates how the bias and precision of the Sharpe improves with monthly versus annual data. I provide small-sample and large-sample formulae for the distribution, highlighting the distinction between the annual and annualized monthly estimators. With more than two years of monthly data the large-sample distributions generally provide a good approximation, simplifying the calculation of confidence intervals; this applies for both normal and non-normal returns. Although these results apply to iid returns they are of practical use, since independence for monthly returns is a good description for many financial assets.

Keywords: Sharpe, Sharpe Ratio, sampling distribution, non-central Student-t, small-sample distribution, large-sample distribution, asymptotic distribution, confidence interval

JEL Classification: G10, G11