================================================================================ README FILE FOR REPLICATION MATERIALS ================================================================================ Paper Title: Returns to Experience and the Elasticity of Labor Supply Manuscript Number: RED-D-24-00183 Authors: Tess M. Stafford The University of New South Wales School of Economics UNSW NSW 2052, Australia t.stafford@unsw.edu.au Scott French The University of New South Wales School of Economics UNSW NSW 2052, Australia scott.french@unsw.edu.au Editor-in-Charge: Professor Loukas Karabarbounis Date: January 2026 ================================================================================ OVERVIEW ================================================================================ This replication package contains the data and computer code used to generate the results in "Returns to Experience and the Elasticity of Labor Supply" submitted to the Review of Economic Dynamics. IMPORTANT NOTE ON PROPRIETARY DATA: The empirical analysis relies on proprietary data from the Florida Fish and Wildlife Conservation Commission (FWC) that cannot be shared. Replicators must obtain access to this data through official channels. See the Data Availability section below for detailed information on data restrictions and access procedures. ================================================================================ SOFTWARE REQUIREMENTS ================================================================================ Stata: - Version: Stata 12 (or compatible) MATLAB: - Version: R2023b - Statistics and Machine Learning Toolbox - Parallel Computing Toolbox - Optimization Toolbox Operating System: - Developed on: macOS - Code should be portable to Windows/Linux with path adjustments ================================================================================ DIRECTORY STRUCTURE ================================================================================ RED_Replication_RED-D-24-00183/ ├── readme.txt [This file - includes complete directory structure] ├── Code/ │ ├── Stata/ │ │ ├── Phase0_Price_Data/ [Price data preparation] │ │ ├── Phase1_Data_Preparation/ [Data preparation] │ │ ├── Phase2_Sample_Selection/ [Sample selection] │ │ ├── Phase3_Regression_Preparation/ [Regression preparation & bootstrap] │ │ ├── Phase4_Regressions/ [Main regressions] │ │ ├── Phase5_Results_Aggregation/ [Bootstrap aggregation] │ │ ├── Phase6_Tables/ [LaTeX table generation] │ │ ├── Phase7_Figures/ [Figure generation] │ │ ├── Phase8_Summary_Statistics/ [Summary statistics] │ │ └── Phase9_CPS_Statistics/ [CPS analysis for Appendix G] │ └── MATLAB/ │ └── Monte Carlo/ [Monte Carlo experiment for Appendix G] ├── Data/ │ ├── CPI/ [CPI data] │ ├── Calendar/ [Lunar cycle, weekday, timeseries] │ ├── Dealers/ [PROPRIETARY - dealer information] │ ├── Hurricanes/ [Hurricane data] │ ├── License/ [PROPRIETARY - permit codes] │ ├── Prices/ [PROPRIETARY - price CSV files] │ ├── Spp/ [PROPRIETARY - species CSV files] │ ├── Tkts/ [PROPRIETARY - ticket CSV files] │ ├── Weather/ [Weather data] │ ├── Zip_Codes/ [Zip code geocoding data] │ └── IPUMS_CPS/ [IPUMS CPS raw data - NOT INCLUDED] ├── Derived Data/ [GENERATED - proprietary derived data] │ ├── Phase0/ [Price data outputs] │ ├── Phase1/ [Data preparation outputs] │ ├── Phase2/ [Sample selection outputs] │ ├── Phase3/ [Regression preparation outputs] │ ├── Phase4/ [Regression outputs] │ ├── Phase5/ [Aggregation outputs] │ ├── Phase6/ [Table generation outputs] │ ├── Phase7/ [Figure generation outputs] │ ├── Phase8/ [Summary statistics outputs] │ └── Phase9/ [CPS analysis outputs] └── Results/ ├── Figures/ [Final figures (.png)] └── Tables/ [Final LaTeX tables (.tex)] Note: [REPLICATION_ROOT] placeholder is used throughout code files. Replicators must replace [REPLICATION_ROOT] with the absolute path to this replication folder. ================================================================================ DATA AVAILABILITY AND RESTRICTIONS ================================================================================ A. PROPRIETARY DATA (Not Included - Must Be Obtained Separately): ------------------------------------------------------------------ The following data CANNOT be shared and is NOT included in this replication package due to FWC data use agreements: 1. Florida Lobster Fisheries Data (FWC) - PROPRIETARY - Source: Florida Fish and Wildlife Conservation Commission - Data files: * Lobster trip tickets (species data): lob_spp_YYYY.csv (1986-2008) * Trip ticket data: lob_tkts_YYYY.csv (1986-2008) * License data: license_stds_fy.csv, license_stds_fy_lastfy.csv * Dealer information: dealers.csv * Fish prices: fish_prc_YYYY.csv (1986-2008) Access: Researchers must apply to FWC for data access through official channels. - Contact: Florida Fish and Wildlife Conservation Commission (FWC) - Website: https://myfwc.com/ - Application Process: Submit a formal data request including research purpose and data requirements. FWC will review and provide access if approved. Data use agreements may be required. 2. All intermediate processed data files derived from FWC data - All .dta files in "Derived Data/" folders (Phase0 through Phase8) Reason: These files contain proprietary FWC data, even if processed/aggregated. Note: The methodology for creating these files is documented in the do files themselves. 3. IPUMS CPS Raw Data Extract - Source: IPUMS (Integrated Public Use Microdata Series) - Data file: cps_00004.dat (2001-2006 CPS extract) - Location: Data/IPUMS_CPS/ (folder exists but file NOT included) Reason: IPUMS redistribution restrictions prohibit sharing raw extracts. Access: Researchers must download their own IPUMS CPS extract from https://ipums.org/ consistent with the code specifications. Note: A derived Stata .dta file (IPUMS_CPS_Basic_2001-2006.dta) IS included in Derived Data/Phase9/ and can be used to reproduce CPS results without the raw IPUMS file. B. PUBLIC DATA (Included in this package): ------------------------------------------ The following data IS included in this replication package: 1. Public auxiliary data: - NOAA Weather Data: Weather buoy data, precipitation, temperature (Data/Weather/) - Hurricane Data: NOAA HURDAT2 hurricane database (Data/Hurricanes/) - Zip Code Geocoding Data: Latitude/longitude, distances to buoys (Data/Zip_Codes/) - Calendar Data: Lunar cycle, weekday indicators, timeseries (dates only) (Data/Calendar/) - CPI Data: Consumer Price Index data (Data/CPI/) 2. Derived data (non-proprietary): - IPUMS CPS derived file: Derived Data/Phase9/IPUMS_CPS_Basic_2001-2006.dta (This file can be used to reproduce CPS results without raw IPUMS data) 3. Final results: - Regression results tables (LaTeX format in Results/Tables/) - Summary statistics tables (LaTeX format in Results/Tables/) - Figures (Results/Figures/) - Monte Carlo precomputed outputs: Code/MATLAB/Monte Carlo/Output/ 4. All code files: - All Stata .do files (Code/Stata/) - All MATLAB .m files (Code/MATLAB/, for Monte Carlo analysis) - Supporting code and helper functions C. REPLICATION WITHOUT PROPRIETARY DATA: ---------------------------------------- It is NOT possible to fully replicate this paper without access to the proprietary FWC data because: - All intermediate data files are derived from FWC data and cannot be shared - Final regression results depend on these intermediate files - Bootstrap procedures require the full dataset However, this replication package provides: 1. Complete code for all analyses (Stata and MATLAB) 2. Comprehensive documentation in this readme file 3. Final results for verification purposes 4. Public auxiliary data that is used in the analysis With access to FWC data, replicators can: 1. Follow the code and documentation to recreate all intermediate files 2. Generate their own regression results 3. Compare their results to the provided results to verify replication ================================================================================ REPLICATION INSTRUCTIONS ================================================================================ IMPORTANT: Replicators must first obtain access to the proprietary FWC data through official channels. See Data Availability section above. STEP 1: OBTAIN PROPRIETARY DATA -------------------------------- 1. Contact Florida Fish and Wildlife Conservation Commission (FWC) for data access - Website: https://myfwc.com/ - Submit a formal data request with research purpose and data requirements 2. Obtain required FWC data files (see Data Availability section above for details) At minimum, you will need: - Trip ticket files: lob_tkts_YYYY.csv (1986-2008) - Species files: lob_spp_YYYY.csv (1986-2008) - License files: license_stds_fy.csv, license_stds_fy_lastfy.csv - Dealer information: dealers.csv - Price files: fish_prc_YYYY.csv (1986-2008) 3. Place FWC data files in appropriate Data/ subfolders according to paths referenced in do files STEP 2: SET UP REPLICATION ENVIRONMENT -------------------------------------- 1. Extract this replication package to your desired location 2. Replace [REPLICATION_ROOT] in all do files with the absolute path to this replication folder, OR: - Set [REPLICATION_ROOT] as an environment variable - Or update paths directly in do files to match your local structure 3. Verify that all do files start with: cd "[REPLICATION_ROOT]" 4. Ensure all public auxiliary data files are present in Data/ subfolders STEP 3: RUN STATA CODE (In Sequential Order) --------------------------------------------- All do files should be run from [REPLICATION_ROOT]. PHASE 0: PRICE DATA PREPARATION Run all files in Code/Stata/Phase0_Price_Data/ in numerical order: 1. (1.0) lobster trip tickets, population.do 2. (1.1) merge csv price files.do 3. (1.2) cpi.do 4. (1.3) prices - all.do 5. (1.4) prices - stone crab.do 6. (1.5) prices - black grouper.do 7. (1.6) prices - dolphin.do 8. (1.7) prices - grunts.do 9. (1.8) prices - hogsnapper.do 10. (1.9) prices - kingfish.do 11. (1.10) prices - mangrove.do 12. (1.11) prices - mutton.do 13. (1.12) prices - red grouper.do 14. (1.13) prices - yellowtail.do 15. (1.14) prices - lobster.do PHASE 1: DATA PREPARATION Run all files in Code/Stata/Phase1_Data_Preparation/ in numerical order: 1. (2.2.1) prep - set of possible lobster fishermen.do 2. (2.2.2) prep - set of possible lobster trips.do 3. (2.2.4) prep - trip values.do 4. (2.3) prep - mode area fished.do 5. (2.5) prep - birth year.do 6. (2.6) prep - zipcode.do 7. (3.5) license - experience.do 8. (4.1) full panel - fill in non-participation days.do 9. (4.2) full panel - merge in explanatory variables.do PHASE 2: SAMPLE SELECTION Run all files in Code/Stata/Phase2_Sample_Selection/ in numerical order: 1. (5.1a) sample selection - remove outside fishing option.do 2. (5.2x) sample selection - define lobster trip.do 3. (5.3) sample selection - define trappers and divers.do 4. (5.4n) sample selection - define lobster fisherman.do 5. (5.5.2.3) sample selection - hours record.do PHASE 3: REGRESSION PREPARATION Run all files in Code/Stata/Phase3_Regression_Preparation/ in numerical order: 1. (6.0) regression prep - prepare sample for regressions.do 2. Run bootstrap sample generation for all three clustering methods: a. (6.1.1) regression prep - draw bootstrap samples, ID.do b. (6.1.1) regression prep - draw bootstrap samples, YMD.do c. (6.1.1) regression prep - draw bootstrap samples, YMDID.do PHASE 4: REGRESSIONS Run all files in Code/Stata/Phase4_Regressions/ in numerical order: 1. (6.2.3) regressions - run, bootstrap loop.do (This file calls the following regression files:) - (6.2.2) regressions - pooled wage, bootstrap loop.do - (6.2.2) regressions - separate wage, bootstrap loop.do PHASE 5: RESULTS AGGREGATION Run all files in Code/Stata/Phase5_Results_Aggregation/ in numerical order: 1. (6.3.0) regressions - combine bootstraps - step 1.do 2. (6.3.0) regressions - combine bootstraps - step 2.do PHASE 6: TABLES Run all files in Code/Stata/Phase6_Tables/ in numerical order: 1. (6.4.1) regressions - latex, step 1.do 2. (6.4.1) regressions - latex, step 2.do 3. (6.4.2) regressions - latex, construct.do 4. (6.4.3) regressions - latex long, step 1.do 5. (6.4.3) regressions - latex long, step 2.do 6. (6.4.4) regressions - latex long, construct.do 7. (6.5.1) constructing the relevant population.do 8. (6.5.2) constructing the relevant population - run.do PHASE 7: FIGURES Run file in Code/Stata/Phase7_Figures/: 1. (6.7) regressions - wage with experience.do PHASE 8: SUMMARY STATISTICS Run all files in Code/Stata/Phase8_Summary_Statistics/ in order: 1. (7.1) summary stats - calculations.do 2. (7.2) summary stats - full latex tables.do 3. (7.3) summary stats - short latex table.do PHASE 9: CPS STATISTICS (Appendix G) Two options available: Option 1 (Recommended): Use provided derived data - Run: Code/Stata/Phase9_CPS_Statistics/CPS_basic_calc.do - This uses the pre-generated file: Derived Data/Phase9/IPUMS_CPS_Basic_2001-2006.dta Option 2: Build from raw IPUMS CPS extract (requires user download) - Download IPUMS CPS extract for years 2001-2006 from https://ipums.org/ - Place the raw data file as: Data/IPUMS_CPS/cps_00004.dat - Run: Code/Stata/Phase9_CPS_Statistics/cps_00004.do - Then run: Code/Stata/Phase9_CPS_Statistics/CPS_basic_calc.do STEP 4: MONTE CARLO EXPERIMENT (Appendix G) --------------------------------------------- The Monte Carlo experiment implements a labor supply model with two work options to quantify bias in elasticity estimates. This analysis is documented in Appendix G. Two Replication Modes: Mode 1: Quick Replication (Recommended / Default) - Uses precomputed simulation outputs in Code/MATLAB/Monte Carlo/Output/ - Run: Code/MATLAB/Monte Carlo/Output_Org/Results_gather.m (Compiles results from precomputed Output/ folder) - Run: Code/MATLAB/Monte Carlo/Output_Org/Graphs.m (Generates Monte_Carlo_4a.png figure from precomputed outputs) - Outputs: Results displayed in MATLAB command window and figure generated Mode 2: Full Rerun (Requires HPC / Cluster Computing) - Runs complete Monte Carlo simulations (computationally intensive) - New outputs are written to Output_New/ (never overwrites Output/) - HPC submission: Use PBS/SLURM scripts in Code/MATLAB/Monte Carlo/Cluster/ Example: qsub Labor_Supply_Monte_Carlo_0.pbs - After completion, run Results_gather.m and Graphs.m pointing to Output_New/ - Random seed: 598402 (set in Labor_Supply_Setup_*.m files) - Note: Full rerun requires significant computational resources and time ================================================================================ RANDOM NUMBER GENERATORS ================================================================================ MATLAB: - Seed: 598402 - RNG Type: 'twister' (Mersenne Twister) - Location: Labor_Supply_Setup_0.m (and other Setup files) - Used in: Monte Carlo experiment for Appendix G Stata: - Seed: 100 - Location: Used in bootstrap sample generation files (Phase3) and regression bootstrap loops (Phase4) - Note: Bootstrap procedures use Stata's set seed command to ensure reproducibility ================================================================================ COMPUTATION TIME ================================================================================ - Bootstrap regressions: Long - Monte Carlo simulations: Medium - Full replication (excluding data preparation): Long Note: Computation time will vary significantly depending on: - Hardware specifications - Number of bootstrap repetitions - Whether cluster/parallel computing is used