this code solves the continuous time version of the two asset models with adjustment cost presented in kaplan and violante (2014) and kaplan, moll and violante (2018).
it is mostly a replication of ben moll's matlab code.
once cloned on your computer, run
p, hh = TwoAssetsContinuousTime.hjb() # solves for the policy functions
TwoAssetsContinuousTime.kde!(p, hh) # solves for the stationary distribution
to change the calibration, modify the src/params.json
file.
dependencies
this code uses pardiso.jl, which requires you to have the mkl library installed. while the installation can be a bit cumbersome, this allows us to use parallel computing to solve the hjb and kfe.
the stationary distribution over illiquid and liquid wealth for the high income households looks like
the remaining figures can be found in the fig/
folder.