Jaeho Kim and Wang Le, "Hidden Group Patterns in Democracy Developments: 
Bayesian Inference for Grouped Heterogeneity", Journal of Applied 
Econometrics, Vol. 34, No. 6, 2019, pp. 1016-1028.

The data used in the article are from the personal homepage 
of Professor Stéphane Bonhomme, who is the author of the paper 
that our study is replicating:

  Professor Stéphane Bonhomme
  Department of Economics
  University of Chicago

The link to the data is:

https://sites.google.com/site/stephanebonhommeresearch/

All files are zipped in the file kl-files.zip. All text files are in
DOS format. However, the zip file also contains binary files, so users
of non-Windows operating systems may need to be cautious.

Our econometric analysis is based on the three variables stored in the
file 'democracy_data'. The dependent variable is the democracy measure
(the Freedom House indicator). The two independent variables are the
lagged democracy and the lagged income (log-GDP per capita). The data
cover 90 countries for the period 1970-2000. 

The matlab codes used in our study are in five directories under
'matlab_codes'.  To obtain the reported empirical results, please run
the m-file 'main_result'. The m-file 'main_likelihood' is used to
compute WAIC values for model comparison.  The same programs can be
used to replicate all other results in the supplementary appendix 
with slight modifications.