Margot Williams is ICIJ’s research editor. She will be sharing new online search sites and research tips for investigative journalists here weekly.
There’s good news from the U.S. on a research tool for digging into non-profits and private foundations.
CitizenAudit.org, a database of the U.S. Internal Revenue Service (IRS) financial documents from nonprofits called “990s”, is now up and running. Luke Rosiak, reporter and database editor at the Washington Examiner, is devoting his off-duty hours to transforming electronic images of annual filings from the IRS to a full text searchable resource. He calls it “Doing what the IRS won’t: the finances of the tax-free….freed.”
Why does this data need to be freed?
First of all, for all of their good works, many nonprofits and private foundations deserve closer scrutiny. Recent watchdog journalism has uncovered charities that divert donations to pay-for-profit fundraisers, use donations for questionable political spending, or have been used by corrupt officials to loot taxpayer funds.
The US tax returns for charitable organizations are required to be publicly disclosed, with certain exemptions, but the IRS does not post the documents on the agency web site. Although the authoritative list of tax exempt organizations is searchable at the IRS site, to get the documents you have to go elsewhere — either to the venerable Guidestar service (which provides free access to three years of reports, but requires a paid subscription for previous years), to ProPublica’s Nonprofit Explorer, by visiting the charitable organization in person, or by requesting copies by mail from the IRS service center.
Rosiak’s new database provides free access to more than 10 years of annual 990 reports – more than 7 million documents made available through the decades-long efforts of Internet pioneer Carl Malamud and hosted at Malamud’s site bulk.resource.org. The full-text indexing of this massive document set is a work in progress, with 5000 documents a day in process and 60,000 new documents arriving each month. Rosiak is prioritizing current documents and working his way back through the years.
Right now filings from 2009 up to 2013 are indexed and searchable. But just the availability of all the filings back as far as 2000 make the CitizenAudit.org unique. And the value for looking up names and other word searches will grow as more years are indexed. Caveat: with most optical character recognition processing, text information quality is better than numbers, which are often garbled. On the search results page, however, it is easy to open the PDF file and check with the original document.
What can you find?
IRS 990 filings disclose the compensation of officers, directors, trustees, key employees, highest compensated officers and independent contractors. For example, when I was researching the Marine Stewardship Council (UK) for the NPR series on sustainable seafood, the IRS 990 for the organization showed the salaries, while the UK Charity Commission filing did not.
The filings show grants and donations from the charitable organization. But without a search tool, I was not able to search all the filings to see where the organization name might show up as the recipient of grant money. I had to download reports and read through them to see which recipients were noted each year. With CitizenAudit.org, I can easily check that the organization received funding from the Walton Family Foundation, the David and Lucile Packard Foundation and other funders.
The filings often show the property holdings, investments and expenses of non-profits. The filings also show recipients of grants and — for private foundations only – may show donors to private foundations.
You might search for the name “Pahlavi” and retrieve the tax return for the private Foundation for Iranian Studies, in Bethesda, Maryland. If you open the document and look on page 16, you’ll find the sole donor is Princess Ashraf Pahlavi, with her address in New York City.
Another feature, free text search, allows you to hunt through the tax returns for county, name, or company entities. But when you display the list of results, it’s a bit tricky to locate the exact page for your search term.
When you get your results, click on the title of the filing under “Matches found inside text of PDF” and then hit Control/F and find your search term in the text box. Then you have to open the PDF to find the original location of the term. The PDF itself is not searchable, which would make the task easier. But it’s a great start for a resource that should have been provided by the IRS in a searchable format as part of public disclosure.
Support for Rosiak’s volunteer project comes from the Sunlight Foundation and Public.Resource.org. In August, the Sunlight Foundation announced an award of $5,000 “for digitizing and hosting about 5 million IRS Form 990s filed by nonprofit organizations, making them searchable, partially extracting structured data from the digitized text, and making bulk data freely available to others for advanced analysis..”
The bulk data is also provided by Rosiak on his page for researchers & developers. Rosiak invites programmers and reporters with programming expertise to help with his project and free the non-profit tax return information for public use.