================================================================================= README—"THE LOOMING FISCAL RECKONING: TAX DISTORTIONS, TOP EARNERS, AND REVENUES" ================================================================================= Last Updated: May 23, 2023. This repository contains computational data and results associated with the research paper titled "The Looming Fiscal Reckoning: Tax Distortions, Top Earners, and Revenues," authored by Nezih Guner, Martin Lopez-Daneri, and Gustavo Ventura. --------------------- COMPUTING ENVIRONMENT --------------------- The computations were executed on a machine equipped with a Core i7-7820X CPU (operating at 3.6GHz with 16 cores) and 125.5 GiB memory. The operating system used is Ubuntu 18.04.6 LTS (64-bit). ---------------------- SOFTWARE AND LANGUAGES ---------------------- All computational routines are written in Fortran 90. These routines leverage the OpenMP (Open Multi-Processing) library for efficient parallel processing. They also utilize the NAG Numerical Libraries (Mark 27, NLL6I272BL) to provide high-quality random number generation and specialized mathematical functions.(1) Within this code, random number generation is crucial and used at various stages. The seed for the random number generator is specified in the main.f90 file, ensuring consistent and replicable results. If necessary, the seed value can be located and modified directly within this file. The Fortran code was compiled using Intel® Parallel Studio XE 2020 Update 2 for Linux (Copyright © 2009-2020 Intel Corporation). For advanced statistical analysis, the R statistical software package was employed. The version used is 4.0.1 (Release Date: 2020-06-06), titled "See Things Now." Please note that R is copyrighted by The R Foundation for Statistical Computing (© 2020). ----- USAGE ----- To execute these codes, you should run the provided Bash script "RUN_FILE.sh" in your terminal or command prompt. The time required to solve the model is typically around 160 minutes, with each iteration taking approximately 8 minutes. Please note that the total runtime may vary depending on your computer's resources. For Linux users, open a terminal window and navigate to the directory containing "RUN_FILE.sh." Once there, execute the script using the following command: bash ./RUN_FILE.sh For Windows users, Bash is not natively supported. However, you can set up a Bash-like environment on Windows using software such as Git Bash, Cygwin, or the Windows Subsystem for Linux (WSL). Once you've set up a Bash environment, open the Bash terminal, navigate to the directory containing "RUN_FILE.sh," and execute the script as follows: bash ./RUN_FILE.sh Alternatively, Windows users can open the Bash script in a text editor, copy the command line that compiles and executes the code, and then paste it into their command prompt or PowerShell window.(2) Each sub-directory contains its own "RUN_FILE.sh" script that executes the code specific to that folder. If you need to adjust the parameters of the model, modifications can be made in the "parameters.f90" file. This file contains a self-explanatory list of all the parameters used in the computations and their sources. Input data that the code requires can be found in the "Data" folder. ------- RESULTS ------- There are two main subfolders: SCF_Estimates_Replication-Table-2 and Model_Equilibrium_and_Results. The first one contains the Survey of Consumer Finances (SCF) 2013 data and codes necessary to generate the inequality statistics reported in Table 2. In contrast, the second one contains all the codes needed to compute the models' equilibria and all the results in the paper. For this particular subfolder, we have categorized the results into separate subfolders for easier navigation. Below is the list of subfolders, along with the Tables and Figures that can be reproduced using the contained information: 1. Benchmark Economy - Table 2 and Figure 9 2. Income Taxes - Table 3 and Figure 6: - 2.1. Gamma 1—0.053 - 2.2. Gamma 1—0.07 - 2.3. Gamma 1—0.09 - 2.4. Gamma 1—0.114 3. Linear Consumption Tax - Table 4 and Figure 7: - 3.1. No Transfer - 3.2. Transfer 3% - 3.3. Transfer 5% 4. Optimal Mixes - Table 5 and Figure 9: - 4.1. 15% Revenue Increase - 4.2. 30% Revenue Increase - 4.3. 45% Revenue Increase 5. Additional Tax Instruments - Table 6: - 5.1. Include capital income taxes - 5.2. Include changes level PIT 6. Wealth Taxes and Debt - Table 7: - 6.1. 1% Wealth Tax - 6.2. 2% Wealth Tax 7. Wealth Taxes and No Debt - Figure 8: - 7.1. Tax Rate 1% - 7.2. Tax Rate 2% - 7.3. Tax Rate 3% 8. Constrained by Mixed Tax Changes - Table 8: - 8.1. CT + Transfers - 8.2. CT + No TR+ Curvature PIT All the results are saved in the corresponding "Results" subfolder within each of the previously mentioned subfolders. We have preserved the original files by appending "- Published" to their names to prevent any accidental overwriting. Additionally, we have created Excel versions of these files to enhance the readability of the results. The filenames are self-explanatory, and they are ranked according to their importance. Therefore, the file starting with the number 1 should be the first one to examine, followed by the rest in numerical order. Lastly, we have prepared a comprehensive file, "Tables+Figures.xlsx," located in the subfolder Model_Equilibrium_and_Results. This file encompasses all Tables and Figures referenced in our paper. Importantly, it features hyperlinks directly connecting to the corresponding files within this repository, thereby facilitating the effortless reproduction of these Tables and Figures. Please note for effective utilization of these hyperlinks, it is essential to preserve the original directory structure of the downloaded repository. -------- CONTACT -------- If you encounter any issues or have questions about the contents of this repository, please feel free to reach out: * Nezih Guner: nezih.guner@cemfi.es * Martin Lopez-Daneri: martin.lopez-daneri@mlopezdaneri.com * Gustavo Ventura: gustavo.ventura@asu.edu We welcome your questions and feedback. Please ensure your queries are related to the data and code included in this repository. ----- NOTES ----- (1) The NAG library is not included in this repository and must be obtained separately. We utilized this library for the results presented in the paper. However, the code is structured to be compatible with alternative, freely available libraries. To use an alternative library, comment out the indicated sections in “main.f90” and “parameters.f90.” Please note that using different libraries may cause variations in the results. (2) While copying and pasting commands from the bash script to the Windows Command Prompt or PowerShell can work for simple scripts, it may not work if the script uses commands or syntax specific to Bash or another Unix-like shell. Therefore, using a Bash environment on Windows (like Git Bash, Cygwin, or WSL) is generally the most reliable way to run Bash scripts on a Windows system.