OK, *Now* We Have All the Financial Conflict Information for Judges, Right?
Our Financial Disclosures Database now includes hundreds of nomination disclosures and covers 34 years going back to 1987.
We've been collecting financial conflicts of judges for many years, but a law on the books has hindered our work. By law, disclosures are only available from the government for six years before they must be destroyed. This means that while we have been able to build a huge database of judicial conflicts from 2011-2019, getting older records has been almost impossible.
Until now.
Today we are very happy to share that we were able to add 874 additional disclosure records going back to 1987 that were previously only available in Congressional archives. We were able to accomplish this with a combination of a machine learning algorithm and volunteer data entry.
Do you know of judicial financial records we don't have? Do you know somebody that might have some tucked away in a drawer or a box? We are building the most complete collection possible and will work to scan or import conflict records. We want them all.
Why it matters
Even though disclosures older than six years are regularly destroyed, we were able to build a first-of-its-kind financial disclosure database with over a million investment transactions and roughly 250,000 pages of disclosures.
This database was used by the Wall Street Journal to engage in ground-breaking investigations into ethics violations of the judiciary. Despite our best efforts, that database did not include any disclosures prior to 2003 since older disclosures have been almost impossible to obtain.
Today we add disclosures for 874 judicial nominees and expand our database further.
We think we might finally have all of the disclosures that are out there, and if we don't we certainly are getting closer.
This expansion includes nearly every nominee to the federal bench since the end of the Reagan Administration. We hope this expansion helps provide researchers, journalists and attorneys greater insight into the federal judiciary.
How we did it
We knew that judicial nominees filed financial disclosure documents with the Senate Judiciary Committee, so earlier this year we began digging through old Senate hearing records to see what we could find.
This led us to discover that nomination disclosures are available in the Congressional Archives, and that they contain the disclosure records we were looking for.
This is great, but unfortunately the disclosure records we needed were tucked away in extremely long documents that also contained the resumes, questionnaires, litigation histories, and letters of recommendation of judicial nominees. Making matters worse, there are about 250,000 pages of relevant Congressional records to somehow analyze.
Sifting through this many pages by hand is impossible for a small organization like ours, but as experts in data processing, we were able to build a machine learning model to identify pages in the record that looked like financial disclosure documents. This method worked exceptionally well and was able to accurately identify disclosures that were handwritten, typed on typewriters, or printed from a computer.
In total, we found 874 financial disclosures, and with the help of dedicated volunteers, we were able to match these disclosures up to the judges in our database. Beginning today you can find these disclosures in our Financial Disclosures database.
You can learn more about how we built and used the machine learning model by looking at our Nomination Extactor.
Of Note
-
Our expansion includes disclosures for 30 judges in which we had no prior financial disclosure.
-
Many nomination disclosures from the 1980 and 1990s are handwritten. They have been added to our database but have not been extracted. Some are barely legible. This means that the data is not searchable in our API.
-
Financial disclosures for failed judicial nominees are not included in our database.
-
Some judges filed executive branch financial disclosures. These are included in our database.
-
Our Nomination Extactor tool is available to everyone. If you have a similar project and need to extract a needle in a haystack, please let us know. We hope our tool can help you.
A Thank You
We want to thank all of our volunteers who have helped us with this project. We are grateful for your time and effort. We could not have done it without you.
Special thanks to:
-
Danielle Barcena
-
Daniel Elliott
-
Aidan Dolik
-
Daniel Zummo
And all the other volunteers who have helped us with this project.
As always, we appreciate your feedback and suggestions and if you have a secret stash of financial disclosures you would like added to our database, get in touch.
