%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % README for Skill-Biased Entrepreneurial Decline by Jiang and Sohail % Contact Info: faisal.sohail@unimelb.edu.au %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ########################### Code Description ########################### This repository contains the replication codes for "Skill-Biased Entrepreneurial Decline" by Jiang and Sohail (2022). The main directory includes seven subfolders which includes the code required to conduct the empirical and quantitative analysis in the paper. Each subfolder (XXX) also contains a seperate file "readme_XXX.txt" which provides more detail on file/path dependencies and data required to execute the code. All code files are also commented to aid in their use. The following six subfolders replicate the empirical analysis and are listed in the order in which they should be run: #------ 1. cps - This folder contains all the code required to replicate the analysis which uses CPS data.The file main_cps.do conducts the empirical analysis of CPS data. In order to succesfully execute this file, users should update the directories in all .do files in the folder and download the relevant CPS data from IPUMS at https://cps.ipums.org/cps/. Additional details on all the files in this subfolder are discussed in readme_cps.txt. #------ #------ 2. sipp - This folder contains all the code required to replicate the analysis which uses SIPP data.The file main_sipp.do conducts the empirical analysis of SIPP data. In order to succesfully execute this file, users should update the directories in all .do files in the folder and download raw SIPP data from the NBER SIPP data repository which are available at https://www.nber.org/research/data/survey-income-and-program-participation-sipp. Additional details on all the files in this subfolder are discussed in readme_sipp.txt. #------ #------ 3. flowdecomposition - This folder contains all the code required to replicate the flow decomposition analysis of Section 2.2 and Online Appendix Section C.1. The file main_flowdecomp.m calls files that are created after main_cps.do from the subfolder "cps" is executed. Additional details on all the files in this subfolder are discussed in readme_flowdecomp.txt. #------ #------ 4. polarization - This folder contains all the code required to replicate the results in Section 2.5. The file main_polarization.do conducts the empirical analysis regarding entrepreneurial polarization. In order to succesfully execute this file, users should update the directories in all .do files in the folder and download the relevant American Community Survey data from IPUMS at https://usa.ipums.org/usa/. Additional details on all the files in this subfolder are discussed in readme_polarization.txt. #------ #------ 5. scf - This folder contains all the code required to replicate the analysis which uses SCF data.The file main_scf.do conducts the empirical analysis of SCF data. In order to succesfully execute this file, users should update the directories in all .do files in the folder and download raw SCF data from the Federal Reserve Board data repository which are available at https://www.federalreserve.gov/econres/scfindex.htm. Additional details on all the files in this subfolder are discussed in readme_scf.txt. #------ #------ 6. gem - This folder contains all the code required to replicate the analysis which uses GEM data in Online Appendix Section E.4.The file main_scf.do conducts the empirical analysis of SCF data. In order to succesfully execute this file, users should update the directories in all .do files in the folder and download raw SCF data from the Federal Reserve Board data repository which are available at https://www.gemconsortium.org/data. Additional details on all the files in this subfolder are discussed in readme_gem.txt. #------ The below subfolder includes the replication codes for the quatitative anaylsis which should be executed after the empirical analysis. #------ 7. model - The file main_quant.m should be executed to replicate the quantitative analysis in Section 6 and Online Appendix E. This file calls other .M files in the model subfolder that simulate the transition dynamics of the model economy, conduct decomposition exercises and replicate figures from the paper. #------ ########################### System Requirements ########################### All codes were run on a HP ProDesk Desktop equipped with 3.60 GHz 4-Core Intel i7-7700 processor and 16 GB RAM operating Windows 10. All .m files were executed using MATLAB 2021a Update 1 (64-bit). The code also requires MATLAB to be called within Stata so users should ensure that the MATLAB directory is included in a Windows PATH System Variable. All .do files were executed using Stata MP 16.1 for Windows (64-bit). Succesful execution of the .do files require the installation of the Stata module "gtools". This can be installed within Stata with a web connection using the command "ssc install gtools". Information about this module is available at: https://ideas.repec.org/c/boc/bocode/s458514.html. ########################### Computation Time ########################### The expected computation time for each subfolder using the above system specifications is listed below: 1. cps: approx 90 minutes 2. sipp: approx 40 minutes 3. flowdecomposition: <1 minute 4. polarization: approx 30 minutes 5. scf: approx 7 minutes 6. gem: approx 10 minutes 7. model: approx 5 minutes ########################### Random Numbers ########################### No random numbers are used in the analysis. ########################### References ########################### Jiang, Helu and Sohail, Faisal. 2022. "Skill-Biased Entrepreneurial Decline."(conditionally accepted at Review of Economic Dynamics).