Hello everyone,

I have time series data with over 1000 daily observations regarding financial assets. My dependent variable is an asset return (y) and I have 3 variables that represent risk aversion (m, n and o) and I was looking for a way of sort of "stripping" the dependent variable out of the effect of risk aversion in order to use it in other regressions later on.

I have looked at the Fama-MacBeth two step regression because it essentially regresses y on m, n and o to determine how exposed it is to each one in the first step, while in the second step y is regressed on the found factor exposures derived from m, n and o, so that y nowoffers the premium of exposure to m, n and o.

Since I have time series data, attempting to use the - asreg - or - xtfmb - commands gives me an error message of either the need to define xtset or of no observations.

Does anyone know if there's a way of using this for time series data? Or if not, an alternative to strip a variable of others to obtain the "pure effect", so to speak?

Best,

Danielle