I have some questions about estimating parameters of a DSGE model. I am quite new to this whole topic. I shall be very thankful if anyone could answer these.
I have a DSGE model and since I wanted to log linearize I entered all the variables as exp(W) for wage, exp© for consumption, exp§ for price etc. The original data (when taken in logs) has unit root of order one.
How should I enter the data for Dynare? Should I de-trend it and de-mean it? Or is it ok to just de-trend it? Does it make sense to de-trend using log-difference?
The model has price (exp§) and inflation (exp(Pi)). If I am using log-difference to de-trend, then should I use it for inflation or for original price? To be consistent, if I am using log deviation for W, then the log deviation of price should be the data for P. But then log deviation is actually the inflation. I am very confused in this part.
How sensitive is the estimation to the de-trending method?