Modeling the contemporaneous duration dependence for high-frequency stock prices

This paper proposes a novel copula approach to model the contemporaneous duration dependence for high-frequency (HF) stock prices via the bivariate hazard function. This method is useful in understanding the mechanism through which the prices of financial assets jointly adjust to reflect new information. In the empirical analysis, we use the HF data on the APPLE and IBM stocks to illustrate the feasibility of our approach. In brief, the main findings are as follows: (1) there is a strong evidence of contemporaneous duration dependence between the prices of these stocks and (2) as a result the estimators of the bivariate hazard function are sensitive to the choice of copulas under our study.