Dufour and Engle (2000) have shown that the duration between subsequent trade events carries informational content with respect to the evolution of the fundamental asset value. Their analysis supports the notion that no trade means no information derived from Easley and O'Hara's (1992) microstructure model. This paper revisits the role of time in measuring the price impact of trades using a structural model and provides challenging new evidence. For that purpose we extend Madhavan et al.'s (1997) model to account for time varying trading intensities. Our results confirm predictions from strategic trading models put forth by Parlour (1998) and Foucault (1999) in which short durations between trades are not related to the processing of private information. Instead, they are caused by strategic trading of impatient non-informed agents who use market orders more intensively when order book liquidity is high.