State Fiscal Stress and Higher-Education Retrenchment across Public and Private Colleges - Replication Package

# State Fiscal Stress Replication Package This archive has two replication paths. ## Option A: From Cleaned Analysis Data This is the fast path. It starts from the included cleaned analysis panels in `data/analysis/` and rebuilds the manuscript tables, figures, and diagnostics. One-line setup and run: ```bash python -m venv .venv && . .venv/bin/activate && pip install -r requirements.txt && python code/run_from_cleaned.py ``` Outputs are written to `outputs/`. ## Option B: From Raw Public Data This path rebuilds the cleaned panels from raw public files, re-estimates the model-output inputs, and then runs the same table/figure build used in Option A. Prerequisites: Java, `ucanaccess-uber.jar`, the SHEF FY25 workbook, and the BEA `SASUMMARY.zip` file. IPEDS Access files, BLS LAUS data, CMS/NASBO fiscal inputs, and NASSGAP Table 9 can be downloaded by the raw pipeline where public direct-download URLs are available. After placing the manual raw files as described in `docs/raw_data_download_instructions.md`, run: ```bash python -m venv .venv && . .venv/bin/activate && pip install -r requirements.txt && python code/run_from_raw.py --download-ipeds --ucanaccess-jar /path/to/ucanaccess-uber.jar ``` Raw rebuild outputs are written under `data/raw_work/`; the cleaned inputs used by the results scripts are then refreshed in `data/analysis/` and `data/model_outputs/`. ## Contents - `code/run_from_cleaned.py`: cleaned-data replication path. - `code/run_from_raw.py`: raw-data replication path. - `code/raw_pipeline/`: raw public-data acquisition, parsing, panel construction, and model-estimation code. - `data/analysis/`: included cleaned analysis panels for Option A. - `data/model_outputs/`: included compact model-output inputs for Option A; overwritten by Option B. - `docs/data_dictionary.csv`: variable definitions. - `docs/raw_data_download_instructions.md`: source files, expected locations, and raw rebuild notes.