0
$\begingroup$

I did a study with structural vector autoregression (SVAR model) corresponding to the IS-LM model (a macroeconomic model). I have four variables that are I(1). I have fitted the SVAR model to the first differences of the original variables. I analysed the impulse response function (IRF). The interpretation of the IRF was contradictory to the economic reality.

Questions:

  1. What could be the causes of such result?
  2. Could it be due to cointegration between variables?
$\endgroup$
1
  • $\begingroup$ I was going through my old answers and noticed this one was not accepted. Do you perhaps need further clarification? $\endgroup$ – Richard Hardy Mar 16 '20 at 15:20
1
$\begingroup$
  1. Either misspecified model (omitted variables, "wrong" lag order, neglected structural changes, "wrong" functional form, conditional heteroskedasticity -- you name it) or perhaps the theory is "wrong" such that the empirical results actually hold even if they seem counterintuitive.
  2. Perhaps it could be. If the series are cointegrated, you have omitted the error correction terms by fitting a VAR on first differences in place of a vector error correction model (VECM). Omitted variable bias could be causing the trouble.
$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.