fediskhakov / eqb

Code repository for Equilibrium Trade in Automobiles paper (JPE, 2022)

Replication code for "Equilibrium Trade in Automobiles"

by Gillingham, Iskhakov, Munk-Nielsen, Rust & Schjerning

Journal of Political Economy, 2022

https://doi.org/10.1086/720463

Research team:

• Kenneth Gillingham (Yale University)
• Anders Munk-Nielsen (University of Copenhagen)
• John Rust (Georgetown University)
• Fedor Iskhakov (Australian National University)
• Bertel Schjerning (University of Copenhagen)

MATLAB scripts used for replication

The MATLAB script run_all.m reads in the data and replicates all the main results in the paper by calling the following MATLAB scripts in sequence

1. run_illustrations.m
   • MATLAB script that produces numerical illustrations of equilibrium in a stylized economy with 2 cars and 2 consumers
   • Replicates Figure 2 and 3 in the theory section
   • Results stored in the folder results/illustration
2. run_estimation.m:
   • MATLAB script that reads in the tabulated data and runs DNFXP to obtain parameter estimates used below.
   • Results stored in the matfile results/estimation/mle_step2.mat
   • Before running the scripts below, make sure converged parameter estimates are stored as results/estimation/mle_converged.mat. In run_all.m this is done automatically by copying mle_step2.mat to mle_converged.mat.
3. run_param_tables.m
   • MATLAB script that produces tex tables with estimation output.
   • Replicates tables with parameter estimates and standard errors in the online appendix.
   • Results stored in the folder results/tables
   • Note: This code also produces the summary statistics for all car types found in the oline appendix. To obtain the table with summary statistics for each of the 8 demographic groups, you must also execute the Jupyter Notebook /data/sumstats.ipynb.
4. run_model_fit.m
   • MATLAB script that produces graphs with model fit
   • Replicates Figure 4-6 in the empirical section of the paper
   • Results stored in the folder results/model_fit
5. run_laffer_3d.m
   • MATLAB script that produces contour plots and 3d-graphs with Tax-revenue (Laffer curve), CO2-emissions, and social surplus when varying the fuel and registration tax rates
   • Replicates Figure 8 (contour plots) in empirical section of the paper and Figure 1 (3d-curves) in the online appendix.
   • Results stored in the folder results/laffer
6. run_iruc_CF.m
   • MATLAB script that produces counterfactual policy simulation results. This code run simulations assuming i) perfect pass-through (as in the paper) and ii) imperfect pass-through (not published)
   • Replicates Figure 7 (prices) and Table 1 (summary of policy simulation) in the empirical section of the paper.
   • Results (with perfect pass-through) stored in the folder results/iruc_CF.