Predictive ability of low-frequency volatility measures: Evidence from the Hong Kong stock markets

Abstract

We employ low-frequency data to estimate historical volatility measures for Hong Kong stocks and examine the relationship between these measures and the one-month ahead stock return over thirty-five years. First, we employ a stock's past three-year weekly return to compute idiosyncratic volatility. Second, we use a stock's past three-year maximum weekly return to create a MAX measure. We find that both IVOL and MAX are significant and negatively related to the one month ahead stock return. Both effects co-exist in the Hong Kong stock markets and are robust after controlling for the financial crisis, January effect, and tiny stocks.