Data Frequency and Estimation Risk

Abstract

We discuss the widespread belief that using a higher data frequency, i.e. calculating with more observations, will lower estimation risk. For example, it is often believed that calculating annualized volatility from daily returns will be more “accurate” than calculating annualized volatility from annual returns over the same period of time. We show that in the best case, estimation risk will not be affected by the data frequency. We also show that in more realistic cases, estimation risk may even increase.

Keywords: Data Frequency, Estimation Risk, IID, GARCH, Volatility, Expected Return, Estimation