Luca Guerrieri, Dale Henderson and Jinill Kim,
"Interpreting Shocks to the Relative Price of Investment with a Two-Sector Model,"
Journal of Applied Econometrics, forthcoming.

------------------
This .zip file includes codes to replicate all the figures in the paper.

Before running any of the codes, the program setpath_windows.m needs to be 
updated to reflect the path of the local installation of Dynare. If needed,
Dynare can be downloaded from https://www.dynare.org/

The replication codes in this directory do not include the .mat files that 
contain the estimation and Monte Carlo results. The .mat files need to be 
produced anew by running the programs described in the sections 
"VAR results" and "Monte Carlo experiments" below.

Depending on the platform used, running the Monte Carlo experiments can be 
time-consuming. The .mat files with the results were omitted from this 
replication package to reduce the storage requirements, but are also 
available from an extended replication packet that can be downloaded from 
http://www.lguerrieri.com/rep_codes_1018.zip

------------------
List of figures and programs:

Figure 1 and 5, run repfisher.m
Figure 2, run run_plot_densities_var.m 
Figure 3, run call_smallmod4.m
Figure 4, run run_plot_densities.m
Figure 6 and 8, run_monte_smallmod4.m
Figure 7 and 9, run_monte_smallmod4_ist.m
Figure 10, run_monte_smallmod4_largesample.m
Figure 11, run_monte_smallmod4_ist_largesample.m


NOTA BENE: 





------------------
VAR results

repfisher.m   extends Fisher's empirical work to encompass the response of 
investment to both TFP and sectoral MFP shocks


The VAR includes 1) the change in the relative price of investment
                 2) labor productivity growth 
                 3) hours per capita
                 4) consumption growth per capita
                 5) investment growth per capita


Apart from computing IRFs, repfisher finds the correlation between 
consumption and investment at business cycle frequencies implied by the VAR 
when only shocks to the relative price of investment are turned on
or when only neutral shocks to labor productivity are turned on.

repfisher.m produces Figure 1 and Figure 5 (in the appendix). 

Results for the correlations are saved in .mat files whose names are 
selected based on the switches set at the top of the program.
These files are also needed to assemble Figure 2 and Figure 4.

There are switches at the top of the program to select the three cases 
shown in Figure 2. After running repfisher with each of the 3 alternative
switches. Run run_plot_densities_var to reassemble Figure 2.

------------------
Monte Carlo experiments

The programs runestim_smallmod4.m and runestim_smallmod4_ist.m estimate the
parameters for the shock processes used in the two alternative models -- 
make sure to line up the standard deviation of the MFP, IST, and neutral 
shock so that the long-run responses of the relative investment price and 
of labor productivity in the model match the VAR.

call_smallmod4 produces Figure 3 and saves some population moments needed 
in Figure 4. call_smallmod4_ist saves some population moments needed in 
Figure 4.

run_monte_smallmod4 runs the Monte Carlo experiment using the MFP model as 
DGP and saves data needed for Figure 4. It produces Figures 6 and 8 
(in the appendix) and part of the data for Figure 4.

run_monte_smallmod4_ist runs the Monte Carlo experiment using the IST model
as DGP and saves data needed for Figure 4. It produces Figures 7 and 9 
(in the appendix) and part of the data for Figure 4

Figure 4 is assembled invoking
run_plot_density.m, which compares the densities for the correlation 
between c and i from the VAR, the MFP model and the IST model.

------------------

Questions on this package can be directed to: luca.guerrieri@frb.gov







