Dear Johannes,
I have a number of shocks, one of the shock is related to the stock market, and I think it may have a higher volatility compared with other shocks, however, I am confused about setting prior standard deviation to this shock, should I set standard deviation of the stock market shock to 2 or infinity? will the results change a lot between setting standard deviation to 2 and setting standard deviation to infinity?
Thank you very much and look forward to hearing from you.
Best regards,
Jesse
I am not sure about your question. What do you mean with “prior standard deviation”? Are you estimating the model? And how would an infinite standard deviation work? In the context of perturbation, that would be a problem as second moments would be infinite.
Dear Johannes,
Please refer to the attached PDF document, in Table 2 on page 619, the standard deviations of shocks/innovations are set to infinity, why this paper allows to set standard deviations of shocks to infinity?
Thank you very much and look forward to hearing from you.
Best regards,
Jesse
DSGEmodelofStockBubblesandBusinessCycles.pdf (522.6 KB)
You were not terminologically precise. They do not set the prior mean of the standard deviations to infinity, but rather the prior standard deviation on the standard deviation of shocks. It just governs how uncertain you are about the mean.
Regarding 2 vs Infinity, the differences are rather small as revealed by a prior plot (infinity on top, two at the bottom)
