The volatility effects of nontrading for stock market returns

Abstract

The effect of periods of nontrading on volatility is examined. The empirical evidence suggests that volatility is higher on days which follow a period of nontrading. A nonparametric kernel regression is used to estimate a diffusion model with a volatility term dependent on the number of days of prior nontrading. The nonparametric estimates suggest that the presence of a prior period of nontrading may increase the volatility as much as 35%. A moving blocks bootstrap, taking into account the dependence in observations, is used in conjunction with the nonparametric regression to show that the differences estimated are statistically significant.