Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We propose a structural model for durations between events and (a vector of) associated marks, using a multivariate Brownian motion. Successive passage times of one latent Brownian component relative to random boundaries define durations. The other, correlated, Brownian components generate the marks. Our model embeds the class of stochastic conditional (SCD) and autoregressive conditional (ACD) duration models, which impose testable restrictions on the relation between the conditional expectation and conditional volatility of durations. We strongly reject the SCD and ACD specifications for both a very liquid and less liquid NYSE-traded stock, and characterize causality relations between volatilities and durations.