VAR and impulse response

[Jamie]

Hi, I'm just learning how to do a VAR analysis and plot the impulse response. I have a couple of beginner's questions which I hope someone can help to clarify.

First of all, when plotting the impulse response, what is the difference between choosing residuals versus Cholesky as the decomposition method? The variables in my model are ordered so does that mean I should be using the Cholesky decomposition?

Second, I use first differences of the variables in my VAR model (to ensure stationarity). When examining the impulse response functions, how do I standardise a 1% shock to one of the variables (say, exchange rate)? Consequently, how do I transform the variables into levels so that the cumulative level responses are shown?

Would appreciate any advice and many thanks in advance.

[arthur]

Hi gentlemen

Could you please help me with a question?
when choosing a variable as an exogenous one, why VAR doesn't allow us to give them shocks?

Thanks in advance

[grijpmak]

@ arthur

I see that you haven't receive an answer to your question and I was wondering if u were able to find a solution yourself. Especially for the second part of your question. Half of my data consist of logged variables and this makes an impulse response function hard to interpret. Therefore I want to compute level-IRF's for my VAR model with log-transformed variables. Maybe you are able to help me?

Thanks in advance!!!
Karin