Standard deviations of shocks

[theta23]

Dear all,

I have a question concerning the size of estimated standard deviations of shock processes. In the model I have four shocks: monetary policy shock, technology, cost-push and housing preference shock.
Estimated standard deviations are:
technology - 0.2767
cost-push - 0.0381
monetary policy - 0.0016
housing demand - 2.3635

So housing demand shock is extremely large, does this mean that something is going wrong? If I have a good explanation that in my sample period house prices have changed dramatically, isn't it still an implausibly large standard deviation in comarison to other shocks?

Thank you!

[jpfeifer]

That depends on the scaling. If the shock is in percent and thus implies 200%, then clearly something is wrong.

[theta23]

Variables are written as log deviations, but the scaling of shocks is 0.01%, so this would be about 2%. But still I suppose it is suspicious when one shock is 10 or 100 times larger than the other one..

[jpfeifer]

It is a bit suspicious, but not unheard of. For example, estimated preference shocks are often an order of magnitude larger than other shocks. See e.g Born/Peter/Pfeifer (2013): Fiscal News and Macroeconomic Volatility, http://www.sciencedirect.com/science/article/pii/S0165188913001437 for an example.