
 Replication files for Kuang and Mitra (JMCB) Potential output pessimism and austerity in the European Union.

   Important note: the results are produced using Windows 10 + Matlab version R2018a + Dynare version 4.5.7 (where applicable).
   For a few programs, there might be compatibility issues if you use more recent versions of Matlab, or Dynare, or using Mac. 


 Empirical results:
 
   Figure 1 is produced by running Figure_1.m in the folder "Figure 1". 

   Figure 2 is only a flow chart, so no computer program is needed. 

   Figure 3, Figure 4, and Figure 5 are produced in Figure 3.xlsx, Figure 4.xlsx, and Figure 5.xlsx, respectively. 


  Model results:

   Figure 6 is produced by running Figure_6.m in the folder Figure_6.

   Figure 7 is produced by running main_optimal_gain.m in the folder Figure 7. 
    
            Note: it could take some time (say half an hour to a few hours, depending on computing power) to get the result/figure due to a large number of repetitions in the simulation. 


   Table 1 is a table of parameters. No program is needed. 

  
   Table 2: 

   Findings 1(b), 1(c), 2(a) and 2(b): 

   Column Q, Table 2.xlsx in folder 'Table 2 Finding 1b 1c 2a 2b' provides the calculated cross-country correlation coefficients reported in Table 2 of the paper (Finding 1(b), 1(c), 2(a) and 2(b)).  
   Each column represents a country. The raw numbers for the models in this spreadsheet are computed using Matlab programs for each country. 

   As explained in the paper, we use country-specific calibrations. There is one folder for each country. 
   For instance, the folder 'Austria' is used to simulate the model for the country Austria. In each country folder, you run 'main_calibration_km_fiscal.m'. 
   You can see from the Command window all results/numbers for this country which are used to compute cross-country correlations. 
   For example, you see a number under 'change in potential growth from 2012Q4 to 2016Q3' in the command window. That number is then copied and pasted to Cell J11 in Table 2.xlsx. 

   Finding 1(a):

   Use the folder 'Table 2 Finding 1a'. There is a folder for each country. In each folder, run 'main_calibration_km_fiscal.m'. The model correlation is displayed at the end of the command window, under the heading 'correlation between surprises to growth rates and revisions to potential growth forecasts'.  

 
   Notes: There is a dynare file in each folder which solves the RE model and saves the RE solution coefficients. 
          The RE solution coefficients are used as input for solving the learning model. 
          If you want to experiment with different parameter values, you need to change the parameter values 
          in both the Dynare file solving the RE model and the main Matlab file solving the learning model.
          Then, you need to firstly run the Dynare file and then run the main Matlab file to get results. 
    
 
   


