Dynare forums

Hi,

I have added a volatility shock to the simple RBC model (attached), and I wanted to see the impulse response of a shock to volatility. Could someone please let me know why the responses look so wiggly?

Thank you,
Ghazika1

There are several issues here. First of all, you need to enable pruning. Second, volatility shocks require order=3. Third, this type of generalized IRF requires many replications (replic option in the thousands).

Thank you very much jpfeifer. I used third order and pruning, and also used large replic numbers (.mod file is attached), and attached the results in a PDF for replic options: none, 2,000, 5,000, 20,000. The responses to k and c become smoother, but the response to z (technology) still remains fairly wiggly (also the initial response to c is a bit counterintuitive). I had worked with the codes that Andreasen had put in his website (from Andreasen et.al ReStud paper with pruning and generalized IRFs) and I remember that responses to volatility shocks were fairly smooth. Can you tell me please if there is anything else that I should take into account?

Thank you,
Ghazika1

Unfortunately, this is a common issue. At the current stage, you can either use the Andreasen et al. codes or use IRFs as deviations from the ergodic mean in the absence of shocks (EMAS) as in Fernandez-Villaverde et al (2011), Born/Pfeifer (2014).
See the discussion in http://www.dynare.org/phpBB3/viewtopic.php?f=1&t=5348

Thanks a lot jpfeifer, I appreciate it.

Best,
Ghazika1