Duration Dynamics in the After-Hours Electronic Futures Market during the Global Financial Crisis

This paper examines the dynamics of trade durations during financial crisis. Using break tests, we establish that an Autoregressive Conditional Duration model for electronic futures contracts on the S&P500 index displays significant parameter changes during the period 2006-2008. Some of the identified break points can be aligned with documented crisis events. Subperiod parameter estimates show that, at critical points in the crisis, observed duration becomes less clustered and more dependent on expected duration. During the crisis period, the impact of trade volume on duration weakens, and the error distribution moves closer to an exponential, consistent with duration homogeneity.