The Economic Value of Conditioning on Overnight Information

This paper provides a comprehensive economic and statistical evaluation of the predictive ability of information accumulated during nontrading hours for a set of European and US indexes. We introduce a stochastic volatility model, which conditions on lagged overnight information and distinguishes between the nontrading periods of weeknights, weekends, holidays and long weekends. Furthermore, Bayesian methods are implemented for estimation and ranking of the empirical models. More importantly, we assess the economic value of conditioning on overnight information, and find two key results: (i) a risk averse investor will pay a high performance fee to switch from a dynamic portfolio strategy which does not account for the impact of overnight information on daytime conditional expected returns and volatility to a strategy which does; and (ii) the economic value of conditioning on overnight information is high in both European and US markets, and is particularly high for stock exchanges where the opening call auction is conducted by a specialist with a price continuity obligation. These two results are robust to reasonably high transaction costs