Hi there,

I try to model daily stock market trading volume with a bunch of independent variables like stock returns, stock price volatility, etc. in an time-series regression model.
In the first place, I estimate a rolling AR(1) model of trading volume over a ten-day window in order to use the residuals of this model as "shocks" to volume as dependent variable in my main regression setup.

Is there any way to compute the effect of an independent variable on volume given that choice of the dependent variable?
I mean, can you make a meaningful statement like "Given the estimated coefficient of dependent variable X: If one increases X by one unit (one percent), c.p. the effect on stock market volume amounts to ..."?

Thanks!