This question is about the economic and not the econometric interpretation of the shock decomposition, which can be performed by dynare. Lately I’ve presented my work and people have asked questions why certain shocks in my historical decomposition are higher in magnitude than others. Somebody also asked why some shocks didn’t show up at the beginning of the sample and only towards the end. So, my question is the following: to what extend can we give the shock decomposition an economic interpretation? Because the shocks in use absorb everything we can’t write down with our model and that can be a lot of things. And in papers I’ve never seen discussions why certain shocks are so dominant than others.
For example, lets say we take the observable “US Households Loans” and we perform a shock decomposition on this variable. The decomposition shows that loan-to-value shocks have contributed massively to the pre-crisis expansion of household debt. Would it be wright or wrong to say that the lower lending standards by banks during this time could be behind those loan-to-value shocks?
Thanks for all your help.