For years now, Carl Malamud has spent thousands of dollars to buy DVDs from the IRS containing images of nonprofits’ 990 tax forms — public data that is an essential tool for watchdogging charities, political groups and other organizations in the massive nonprofit sector of the economy. He laboriously rips those DVDs and puts the data online, free for anyone to use. This archive has become an important resource: Pro Publica’s Nonprofit Explorer is built, in part, on Malamud’s service, as is Citizen Audit. Sunlight and countless others have benefited from Malamud’s work and generosity.
The IRS has not been helpful to these efforts. Releasing data by DVD is plainly ridiculous and unnecessarily expensive. The IRS also refuses to release the data in the e-file format in which it is originally submitted, which is structured and easily searchable and analyzable. Instead, the agency continues to insist on distributing 990s as image files, which inhibits analysis and can introduce errors.
Worst of all, the dataset contains the improperly unredacted social security numbers of thousands of Americans. Malamud has done his best to detect and report these problems to the IRS, White House and Congress. But his efforts have attracted little attention from government.
Yesterday, Malamud made good on his warnings and took the 990 archive offline. You can read his message about the service termination here. It’s impossible to begrudge Malamud regarding this decision, but it’s clear that the internet has just lost an important resource. We join him in calling on the White House, Congress and IRS to engage with the problems he has identified: data distribution, redaction of sensitive information, and the release of e-file data. Our government should do better.