Though the paper is mainly based on the numerical results, a simple regression is also included. 

(1) Empirical Part
a. Software and operating system used: STATA MP 14 on Windows 10 
b. Expected computation time:  a second
c. The details are as follows:
If the zip-file named 'empirical_part.zip' is extracted, you can see five files including the 'regression_result.log' file showing the results. The files named 'bc_indicator.dta' and 'bds_f_szst_release_modified.dta' are the data files. (As stated in the paper, Table 1 is based on the data provided by Professor Teresa Fort, http://faculty.tuck.dartmouth.edu/teresa-fort/data, and these two files are parts of it.) 'net_growth_diff_sm_lr.do' is the file to clean up and merge these two data files. 'conduct_regression.do' is the file to generate the results of the regressions that correspond to Table 1. To replicate Table 1, run 'net_growth_diff_sm_lr.do' and 'conduct_regression.do' in order. Then, you can find a new file named 'regression_result.log' where the results are stored. 

(2) Numerical Part
a. Software and operating system used: PGI Visual Fortran (ver. 14.6) on Windows 10 
b. Expected computation time
   - The fortran code is accelerated with OpenACC, a way of utilizing GPGPU (General-purpose computing on graphics processing units)
   - The system on which this code was run has a GeForce GTX TITAN X Graphics card and it took about 2 days until the results were obtained
   - If this code is run on the computer without the recently developed graphics card, it may take 2-3 weeks for the same results to come out  
c. The details are as follows:
If the zip file “fortran_code.zip” is successfully extracted, you should be able to see three main folders: (i) basic_solve, (ii) “transition, and (iii) results. The first one is a collection of all files, which are grouped into some sub-folders under the parent-folder, for obtaining the results of the steady-state economy in ‘4.1 Steady State’, the second one is a collection of all files for deriving the transition paths of the model economy in ‘4.2 Transition Path: the Housing Market Boom and Bust in the 2000s’, and the last one contains all the results
In the folder “basic_solve”, there are three sub-folders: (i) “basic_solve”, (ii) “input”, and (iii) “output”. The first sub-folder “basic_solve” contains all the fortran source files (*.f90) required for calculating the steady state model economy, the second sub-folder “input” contains a file named “parameters.txt’, which lists all the values of parameters (and variables used throughout the program), and the last sub-folder “output” contains four empty folders named “benchmark”, “experiment_1”, “experiment_2”, and “experiment_3”, in which the key statics are stored when all relevant fortran files (*.f90) are compiled and executed. When you actually run the code, a window will pop up to ask you to put 1 for benchmark and 2 for experiment. If you want to get the program to calculate the steady state, put 1; or if you are interested in performing an experiment in which the parameter governing the disutility of being a renter, which is denoted as “rent_P” in the code, is changed, put 2, and then you will see, “Put 1 for version 1 or 2 for version 2 or 3 for version 3: ”. The version itself doesn’t mean much other then it allows the different sub-folders in the folder “output” to be assigned for the different values of “rent_P” for which you are asked to put immediately after you put the number for the version.
In the folder “transition”, there are three sub-folders: (i) “transition”, (ii) “input”, and (iii) “output”. The first sub-folder “transition” contains all the fortran source files required for obtaining the transition path, the second sub-folder “input” contains a file named “parameters.txt’, which is the same as the one used for the steady state case, and the last sub-folder “output” is just a place spared for storing the results from the code. Importantly, this code (for the transition path of the model) must be run after the code obtaining the steady state economy.