Overnight News and Daily Equity Trading Risk Limits

Abstract:     
Value-at-Risk (VaR) is the most prevalent risk measure used in banking. This paper investigates the role of overnight information for setting daily equity trading VaR limits. We propose a novel bivariate modeling approach that deals separately with daytime and overnight returns, and allows for a non-negligible ex ante covariance between them. This modeling strategy is motivated by inefficiencies in the opening price discovery process due to price staleness and news spillover effects. We provide evidence on the forecasting outperformance of the bivariate model against widespread overnight-adjustment approaches in the literature. For a small cap portfolio, the bivariate VaR forecasts are also competitive against VaR forecasts generated instead at the market open, when the overnight return is realized. For a large cap portfolio, this result is reversed. The contrast reveals that price discovery at the market open is less efficient for small capitalization, thinly traded stocks.

Number of Pages in PDF File: 31

Keywords: Overnight, Price discovery, Realized volatility, Risk management, Value-at-Risk

JEL Classification: C52, C53, G15

working papers series