Welfare calculation contradicts IRF graphs

Stochastic simulations
1

Dear all,
I am trying to perform and compare welfare evaluation between two cases of flexible ER and fixed ER, and how macroprudential policy affects welfare. This is the code I use for welfare calculation (omega_e):

util_e = log(ce) - (he^(1+psi))/(1+psi);
omega_e = util_e + beta*omega_e(+1);

What puzzles me is why the results of welfare calculation seems contradict with my calculation for volatilities and the IRF graphs. One of my confusion is: the welfare for a fixed ER is higher (better welfare) compared to flexible ER (no macroprudential policies) when the graphs clearly show that a fixed ER creates more volatilities in key variables.

Below are the codes. I am afraid my welfare calculation is not correct. (I use different mod files to (a) compute volatilities and plot IRF graphs; (b) calculate welfare; but they are actually identical). As I am only allowed to send 2 files at a time, I will send the other two files in separate message.

I would really appreciate if anyone can help me solve my confusion.
Thank you very much.
Ratih

Welfare_17Aug_exchrate.m (4.9 KB)

Fed_paper_17Aug_wel.mod (11.2 KB)

2

Here are the other two files:

Run_plot_17Aug_exchrate.m (10.2 KB)

Fed_paper_17Aug.mod (11.3 KB)

3
  1. Using pruning and simulated moments is not advised.
  2. Your logic is flawed. People do not care only about volatilities, but also about levels. If you are considering unconditional welfare, the mean can be very different. Take an extremely volatile world where people do precautionary savings like crazy. As a result, the capital stock will be really high and so will be consumption. That is why you may want to consider conditional welfare and check whether your macroprudential policy alters the mean variables that enter the utility function.
1 Like