When sticker prices fall: posted-tuition reductions, price signaling, and net-price pass-through in U.S. higher education

# Tuition resets final replication package This archive reproduces the quantitative results, manuscript tables, and figures reported in the final manuscript. The package is anonymous. It contains replication inputs, code, reference outputs, and documentation needed to reproduce the results. It does not contain manuscript drafts, local-machine paths, cache files, virtual environments, obsolete output duplicates, or materials outside the replication workflow. ## Included inputs - `data/processed/panel_analytic.parquet` - `data/processed/panel_with_resets.parquet` - `data/external/public_communication_coding.csv` - `data/external/public_communication_event_validity.csv` The processed panel files are derived from public IPEDS data and contain the variables needed for the analyses. Raw IPEDS Access Database files are not included because they are public files distributed by NCES. The package includes a precise IPEDS download guide at `docs/ipeds_download_guide.md`. ## Software setup Create a Python environment and install dependencies: ```bash python -m venv .venv source .venv/bin/activate pip install -r environment/requirements.txt ``` ## Reproduce all results from bundled processed inputs Run the full replication pipeline: ```bash python code/run_final_replication.py --config config.yaml ``` The command regenerates model outputs, manuscript tables, figures, and validation reports. Generated files are written to: - `outputs/generated/results/` - `outputs/generated/manuscript/tables/` - `outputs/generated/manuscript/figures/` Reference files included in the archive are stored in: - `outputs/reference/results/` - `outputs/reference/manuscript/tables/` - `outputs/reference/manuscript/figures/` Successful completion writes: - `outputs/generated/results/validation_final.json` - `outputs/generated/reference_validation.json` Both validation files should report no issues. ## Rebuild from raw IPEDS downloads To rebuild the processed panels from raw public IPEDS Access Database files, follow `docs/ipeds_download_guide.md`, place the required ZIP archives in `data/raw/downloads/`, and run: ```bash python code/run_full_from_ipeds.py --config config.yaml ``` This raw-data path stages and extracts the IPEDS Access databases, rebuilds the analytic panel, identifies posted-price reductions, and then runs the final replication pipeline. ## Main manuscript artifacts Main tables: - `table_01_validity_and_public_communication.csv` - `table_02_event_year_estimates_by_validity_sample.csv` - `table_03_pass_through_by_income_band.csv` - `table_04_institutional_aid_offset_ratios.csv` - `table_05_discounting_architecture.csv` Appendix tables: - `appendix_table_A1_specification_checklist.csv` - `appendix_table_A2_public_communication_coding_rubric.csv` - `appendix_table_A3_validity_sample_pretrends.csv` - `appendix_table_A4_cohort_pandemic_sensitivity.csv` - `appendix_table_A5_heterogeneity_by_baseline_discounting.csv` - `appendix_table_A6_income_band_difference_tests.csv` - `appendix_table_A7_strict_public_campaign_descriptive_sensitivity.csv` - `appendix_table_A8_exploratory_ftft_enrollment_checks.csv` Figures: - `figure_01_pass_through_by_income_band.png` - `appendix_figure_A1_clean_sample_dynamic_estimates.png`