Investigation of the Trade Durations with Autoregressive Conditional Duration Model : Evidence from Borsa Istanbul

This paper investigates duration dynamics in Borsa ̇Istanbul (BIST) stock exchange via Autoregressive Conditional Duration (ACD) model applied on time observed between consecutive trades. Investigation of a suitable error term specification reveals that Weibull ACD (WACD) model is the most appropriate choice of distribution. Applying WACD model on ten most widely traded stocks in the market documents cross sectional differences in the trade duration dynamics, where the level of duration clustering differs even among the most liquid stocks in the market. The duration dynamics in Borsa ̇Istanbul are shown to be closer to developing market examples in market microstructure and intraday liquidity rather than the developed market ones