When I ask for a plot of Bayesian impulse responses almost always, I get these impulse responses are scaled down to 10^-4 or 10^-3. On the other hand, an impulse response plotted by stoch_simul command (with relative frequency option) does not have this scaling down problem. The shapes of the impulse responses in both cases are the same but for Bayesian impulse responses, they are scaled down to near zero. How can I fix this? I like ot get the Bayesian impulse response in proportion to he steady state. Can anybody suggest what command I should use for this.
The IRFs are always one standard deviation. Typical shock have standard deviations of 0.5 to 1 percent, which is 0.005 to 0.01. This is exactly the size you get. The scale invariance of stoch_simul is lost because the scaling of shock sizes needs to correspond to the data. See “Remark 10 (Scaling With a Factor 100)” in Pfeifer (2013): “A Guide to Specifying Observation Equations for the
Estimation of DSGE Models” at sites.google.com/site/pfeiferecon/Pfeifer_2013_Observation_Equations.pdf