My question might be slightly dumb, however I could not find out which choice would be better.
I am trying to construct a SVAR model including the french inflation rate, the unemployment rate and the ECB Refi rate. I am interested both in forecasting and explaining moves related to a "surprising" increase in the ECB Refi rate (which will be shown by impulse response functions).
However, should I adjust my series for seasonality ? I know that this would be better for forecasting purposes (and put back seasonality after the forecast). However, would impulse response functions still be "correct" without those seasonalities ?