README DOCUMENTATION FILE

Data are in the following .txt files:
Shiller.txt: contains asset data from Shiller's website.
FRED.txt: contains interest rate data from the Federal Reserve Bank of St.Louis.
term_premium.txt: contains term premium data from the Federal Reserve Bank of St.Louis.
Livingston1990-2003.txt, Livingston2004-today and LivingstonCPI.txt: contain Livingston survey data.

Stylized facts:
The file dataJMCB.xmls contains data and calculation of empirical stylized facts. The sheets "Results" and "Results-timevarying" in particular report
the statistics in the column "data" of Table 2 in the paper.

Model results:
Those baseline model statistics (column "Baseline" of Table 2) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m.
Results of the representative agent model (column "psi=0" of Table 2) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m, changing line 39 of the steady state file from rot=0.6 to rot=0.00001.
Results of the model without habits in ratio (column "Habits in Ratio, chi=0" of Table 4) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m, changing line 30 of the steady state file from chicC=0.82 to chicC=0.
Results of the model with habits in difference (column "Habits in Difference, psi=0.2,chi=0.5" of Table 4) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m, changing line 30 of the steady state file from chicC=0.82 to chicC=0, line 39 from rot=0.6 to
rot=0.2 and line 31 from chicc=0 to chicc=0.5.
Results of the model with habits in difference (column "Habits in Difference, psi=0,chi=0.82" of Table 4) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m, changing line 30 of the steady state file from chicC=0.82 to chicC=0, line 39 from rot=0.6 to
rot=0 and line 31 from chicc=0 to chicc=0.82.
Results of the model (column "psi=0.2,X=77" of Table 5) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m, changing line 39 of the steady state file from rot=0.6 to rot=0.2.
Results of the model (column "psi=0.2,X=5" of Table 5) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m, changing line 39 of the steady state file from rot=0.6 to rot=0.2 and line 35 from chiW=77 to chiW=5.
Results of the model (column "thetapi=3" of Table 7) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m, changing line 27 of the steady state file from thetpi=1.5 to thetpi=3.
Results of the model (column "thetaR=0" of Table 7) that can be obtained with a second order approximation are computed by the dynare file baseline.mod 
with the corresponding steady state file baseline_steadystate.m, changing line 26 of the steady state file from thetr=0.75 to thetr=0.

Results of the Uhlig model in Table 6 can be obtained using the dynare file uhlig.mod and the corresponding steady state file uhlig_steadystate.m. 

Results that require third order approximations in these tables can be obtained in each case removing line 106 from baseline.mod and activating line 105.

To obtain figure 1 in the paper run the file irfs.m (which in turn uses the files final.mod and final_steadystate.m) following the instructions therein.
Running final.mod one can also obtain results in table 3 in the paper.

All codes use dynare version 4.4.3.









