I ran my VAR model with inflation, real gdp, a proxy for fiscal policy and a policy indicator. I used the function externalinstrument in R and followed this tutorial to derive IRFs.

After transforming all the variables to induce stationarity (first or second differencing), I ran the VAR and got the following IRFs (1 standard deviation shock):

enter image description here

As you can see the IRs don't behave as nicely and smoothly as in usual IRF applications. Especially, real GDP follows a very weird path (it increases when it should be decreasing and it's not mean reverting but it rather follows a random walk). The policy indicator as well decreases initially while it should be increasing. Any idea why this is happening and what I can do to correct it?

  • 1
    $\begingroup$ Are the responses for transformed variables (differenced once or twice) or original variables? Also, I think second-order differencing is quite uncommon and difficult to justify for macroeconomic variables; are you sure you need it? $\endgroup$ Mar 14, 2020 at 13:46
  • $\begingroup$ the responses are for transformed variables. Am I wrong? Shall I do it for orginal variables? Only HICP is twice differenced (log). I can try with one diff only but is it going to change that much? $\endgroup$
    – Rollo99
    Mar 14, 2020 at 13:51


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.