CAN SPECULATIVE TRADING EXPLAIN THE VOLUME-VOLATILITY RELATION

We derive a speculative trading model with endogenous informed trading that yields a conditionally heteroscedastic time series for trading v olume and the squared price changes. We use half-hourly price-change a nd volume data for IBM during 1988 to test the model and estimate the structural parameters using the simulated method-of-moments estimation procedure. Although the model seems to do a reasonable job fitting th e unconditional moments of the volume and the squared price change pro cesses, it fares less well in fitting the relation between current tra ding volume and lags of trading volume and squared volume's (and its l ag's) relation to squared price changes.