This repository contains Python code that transforms bulk PAC data from the Center for Responsive politics into GVKey-quarter level data ready to be used in further analyses.

In the last step I use a Stata do file for convenience, but this could of course be done in Python too.

The structure of the code is as follows:

• code/1_prepare_data.py (a) imports all firm names and GVKeys from WRDS Compustat and (b) imports and cleans the raw PAC data from the CRP.
• code/2_match_companynames.py matches the company names of the two data sets. I use FuzzyWuzzy to string match the company names.
• code/3_create_final_dataset.do merges the files into a final data set.

To run this code, you will need to first obtain the following input files:

• the company and g_company file from WRDS
• the CampaignFinXX.zip files from the CRP.

I provided a bit more information about these input files in the readme files of the respective input folders.

Then

• adjust the homedir in the first few lines of each code file;
• create an output/ folder; and
• run each file separately and in the order indicated in the file name.

Contents of requirements.txt

fuzzywuzzy==0.17.0
joblib==0.14.0
pandas==1.1.5

Use at your own responsibility.