Aggregated micro consumption and macro consumption do not match in a shock

[feberbo]

I was using the one-asset Hank notebook as a base for checking on consumption on a micro level after a shock.As a sanity check, I tried to see if I was doing the aggregation of micro right.

So I changed the line :
td_nonlin = hank.solve_impulse_nonlinear(ss, unknowns, targets, rstar_shock_path)
to:
td_nonlin = hank.solve_impulse_nonlinear(ss, unknowns, targets, rstar_shock_path, internals=['hh'])
kept the rest of the notebook untouched and added a block below. However they will not match. I have included the whole notebook in a zip file as well.

Am I doing something wrong on this aggregation?

hankAggCons2.zip

[feberbo]

I have also made an attempt using Dbeg instead of D, but the plot stays mostly the same.

[thom364e]

Hi,

Did you by any chance ever find the solution to this problem? I'm running into the same problem, so I'm interested in what's going wrong in the aggregation.

Thanks,
Thomas

[feberbo]

No I haven't

[thom364e]

I've played around with it a little, and this turns out to be the solution (shown for my own project, but should apply universally I guess)

i.e. when aggregating, you'll have to also use the steady state policy functions, because in the IRF dict, everything is given in terms of deviations from steady state. You can then subtract the steady state level of the variable you want after doing to aggregation to get back to the deviation from steady state.

Hope it helps!