Impulse response function
Posted: Sat Jan 30, 2016 5:16 pm
I'm running a VAR model to investigate the effect of increase and decrease of oil prices on the GDP growth and unemployment. The estimated VAR model using OLS has shown that the sign of the coefficient of the oil price increase is +0.057401 and the oil price decrease coefficient is +0.048091. However, the response of the GDP growth to shock to oil price decrease using IRF is negative!! I have checked all model adequacy tests and found the model is well specified. So, why the impulse response function is not consistent with the sign of the oil price coefficient?