Read me file for the paper: Real exchange rate and international co-movement: News-shocks and non-tradable goods with complete markets
=========================================================

This readme file provides guidelines on how to run the codes and how to use the information provided in the subsequent folders. 

Computer and software requirements: Results were produced with Dynare 4.5.3. but any other version should work. For the calibrated files it takes only a few seconds to run. The estimated one would take more than 15 hours. 

------------------- Data -------------------------
The data used for the paper are in the folder "Data". They are split in two parts, "TFP Data" - used to estimate the exogenous TFP process and "Data statistics" which basically allow the user to compute the Data column in the main table of the paper. The statistics can be obtained by running the two Matlab files in the folder: "GenerateData.m" and "GenerateSectoralData.m". Another readme file in both folders provides more details - so do Comments in the Matlab codes.

Note: The codes for generating the data calls the "hpfilter.m" file to filter the data and "dyntable.m" to produce tables. Its best to keep those codes in the same folder that runs the code.  


----------------------- Codes/ Model(s) ----------------------- 

The folder "Codes" includes the Dynare and other codes that allow to re-produce the results of the paper.

Benchmark results: which are referred to as the results under "Benchmark" in Table 3 of the paper are produced by running the Dynare code "BSur.mod" in the folder "Benchmark". Variations can be done by changing the parameters in that file. For example, for the case of homogeneous capital across sectors one could simply set parameter "epsilon" to a very large number. To change between surprises/ news economies is the parameter "q", set in line 20 of the code. Setting it q=0 is an economy driven by unexpected TFP shocks and q=1 shocks are anticipated one period in advance.

Results with investment-specific technology shocks: can be reproduced by running the Dynare code "BSurist.mod" in the folder "IST Calibrated".

Resutls of Section 6: are in the folder "Estimated". Here are two folders. In "Estimation of shocks" is the actual Dynare code that would estimate the shocks as described in the paper. That takes more than 15 hours to run. To re-produce the results of Section 6 go to folder "Calibrated using estimated shocks size" and run the Dynare code "BSur_calibForEstn.mod". This code reads the mat-file BSur_mean which includes the mean of the estimated posterior (Table 4 of the paper). 

Note: To compute the steady state, Dynare codes use a function called "withunitroot.m" that employs Matlab's fsolve. Best to keep this file in the same folder that you run the code.