May 7, 2014

National Mortgage Database: Good for Regulators, Bad for You

Hester Peirce

Former Senior Research Fellow
Summary

Two years ago, in a joint undertaking with the Bureau of Consumer Financial Protection, the Federal Housing Finance Agency launched a project to build a National Mortgage Database. The database, which includes detailed loan-level, borrower-specific information about mortgages from 1998 to the present is intended to help policymakers better understand the mortgage markets. Later this month, the database is set to expand to include even more personal information about borrowers. The database might be a gold mine for curious regulators and researchers, but is not cause for celebration for the borrowers whose personal information populates the database.

Contact us
To speak with a scholar or learn more on this topic, visit our contact page.

Two years ago, in a joint undertaking with the Bureau of Consumer Financial Protection, the Federal Housing Finance Agency launched a project to build a National Mortgage Database. The database, which includes detailed loan-level, borrower-specific information about mortgages from 1998 to the present is intended to help policymakers better understand the mortgage markets. Later this month, the database is set to expand to include even more personal information about borrowers. The database might be a gold mine for curious regulators and researchers, but is not cause for celebration for the borrowers whose personal information populates the database.

The mortgage database is an ambitious, time- and technology-intensive undertaking. Most of the information comes from credit reporting agencies, but the database also draws on other government and private information sources and includes a survey component. The result is a very comprehensive picture of the mortgage, the underlying property, and the borrower. One of the engineers of the database acknowledged that "it is easy to reverse engineer and identify the people in our database." And there are a lot of people in the database. According to a 2013 presentation, the database is designed to capture one in 20 newly originated mortgages and include quarterly data about one in 20 "individuals who have ever had an active mortgage from 1998 onward ... from one-year prior to taking out their first mortgage (or 1998) until they die."

Currently, the database includesinformation such as the type and term of the mortgage; the sales price and size of the down payment; certain features about the mortgaged property, such as whether it is new construction; and the borrower's age, ethnicity, income, creditworthiness, debt-to-income ratio, and gender. The FHFA plans to considerably expand the information collected. The database will include more granular data about the mortgage and the mortgaged property. In addition, the expanded dataset may include the borrower's name; social security number; religion; race; language; educational, military and employment records; account balances and past due amounts for credit cards and other loans; financial account numbers, "financial events in the last few years," "life events in the last few years," and wealth.

In other words, government regulators will have lots of information-some of it intensely personal-about many Americans right at their fingertips. The Federal Register notice announcing the impending expansion of the database promises that "[i]n most cases, records will not contain personal identifiers." But some personal identifiers are needed to match information culled from different databases. Even in the absence of personal identifiers, the nature of the contemplated data collections leaves little to the database researcher's imagination.

The FHFA and the CFPB are not going to keep the information they collect to themselves. The information will be provided to academic researchers subject to carefully designed parameters, but data leakage or unauthorized reverse engineering are likely to occur at some point. Moreover, the Federal Register notice identifies a long list of other outside parties that may get access to records in the database. Some of these make sense, but others such as regulated entities and "contract personnel, grantees, volunteers, interns, and others performing or working on a contract, service, grant, cooperative agreement, or project for FHFA," raise concerns about just how many eyes are going to be perusing the personal data of millions of Americans and for what purpose.

Regulators argue that the data collection and research it spawns "will provide sufficient warning to allow [them] to take steps to avoid or mitigate major mortgage market crises in the future." Certainly, one can hope that this objective will be met, although history suggests that regulators-who have the same limitations and failings as the rest of us-don't always know what to do with the data they have. And the fact that regulators are collecting so much data can lull private market participants into believing that they can leave risk monitoring to the government. At the very least, the potential unintended consequences of amassing in one place so much information about the lives of individual Americans need to be taken into account as the national mortgage database is expanded and its contents are made more widely available.