Modelling the Time Between Trades in the After-Hours Electronic Equity Futures Market

This paper models the time between trades of the after-hours electronically traded equity futures market, a market which is previously unstudied in this regard. Using a relatively long 2 year data set, trades in the NASDAQ and S&P500 equity futures are shown to require different forms of autoregressive conditional duration models, including longer lag lengths than previous spot data applications. Volume provides an informative mark in both cases. The S&P500 necessitates a threshold model where the majority of trades display the typical low autocorrelation and strong clustering evident in other assets, but with large durations more autocorrelated with low clustering.