Hi Prof. Pfeifer and to all,

I have a theoretical question.

Is there a relationship between the impulse response functions and the decomposition of the variance in the unconditional variance_decomposition (e.g. the greater the impulse response the greater the proportion of dispersion in the variance decomposition)?

Thank you,

Leonid.

The variance decomposition considers the sum of the effects of the IRFs, so what matters are the size and persistence of the IRFs. Thus, ceteris paribus, the higher the effect and the more persistent, the larger the variance contribution.

Thanks for your prompt reply, jpfeifer!