Trading Intensity and Intraday Volatility on the Prague Stock Exchange: Evidence from an Autoregressive Conditional Duration Model

Using high-frequency trade and quote data from the Prague Stock Exchange, this paper investigates the empirical behavior of price durations defined as the time needed for a quote midpoint to move by a given amount. Focusing on the three most liquid securites traded on the exchange - Cesky Telecom, CEZ and Komercni Banka - we estimate the autoregressive conditional duration (ACD) models for price duration series and test several market microstructure hypotheses suggested by the infomation-based models of market microstructure. Similarly to earlier studies, we find that price durations exhibit diurnal patterns, overdispersion and substantial persistence which can be adequately captured by the ACD model. The market microstructure hypotheses, however, do not find sound empirical support in our results.