Toward Automating Shredding Nonprofit XML Files: The Case of IRS Form 990 Data

This paper presents and describes data for nonprofit IRS filings in the United States of America. The data contains 831 attributes and 1,102,884 records for the years 2016-2021. Among other items, the data include nonprofits’ comparative financial data, governance disclosures, and hired contractors, as well as management compensation, a detailed statement of revenue, statement of functional expenses, external audit, federal audit election, and reconciliation of net assets. The data is generated using Structured Query Language (SQL) self-developed code to convert the IRS form 990 Extensible Markup Language (XML) tax filing files to a dataset in Excel. This paper is the first to convert these XML files and provide much-needed open access to nonprofit data in a long format that is useful for researchers to conduct cross-sectional analysis. The 2,174 lines of source code that we developed, and a step-by-step guide are included in this paper.