Facilitating Innovations in Mortgage Data Through Public-Private Partnerships

Facilitating Innovations in Mortgage Data Through Public-Private Partnerships

Erika Poethig, Acting Assistant Secretary for Policy Development and Research Early on in this Administration, Assistant Secretary Bostic worked with PD&R staff to develop a vision for our collective work that supports HUD’s mission to “Create strong, sustainable, inclusive communities and quality affordable homes for all.” Together, PD&R staff crafted the following statement that guides our work:

Our vision is to be the preeminent housing and urban research organization, one which conducts policy analysis and creates and synthesizes data and evidence through open, challenging, creative, collaborative, diverse, and respectful exchanges of ideas to move policy and improve American communities and lives.

Over the past three years, we have worked in partnership with many different international and domestic organizations and with other HUD program offices to promote the “respectful exchange of ideas.” These conferences, convenings and roundtable discussions have enabled us to improve a full range of policies that affect many peoples’ lives. Past conferences have reviewed what existing evidence teaches us about approaches to measuring the impact of energy efficiency on households, linking housing investments to education, health and self-sufficiency outcomes, improving neighborhoods through investments like Choice neighborhoods, and building resilient cities and regions.

Last month, we co-sponsored a roundtable discussion with the Urban Institute, the Ford Foundation, NYU’s Furman Center for Real Estate and Urban Policy, and Harvard’s Joint Center for Housing Studies to examine how the housing field can improve its analytic capabilities to prevent another housing crisis. Specifically, the people gathered focused on the complexities of linking loan-level mortgage data to other sources to better understand the emerging trends in the mortgage market. Why is this important? One of the challenges of even analyzing the impact of the current crisis is the absence of a single database that tracks mortgages across institutions and markets, let alone links this data to other kinds of outcomes that can measure impact on households and neighborhoods. Using public-private partnerships, we need to find ways to better link data across markets at a national level to help policy makers monitor trends and shape housing finance policy going forward.

Throughout the day’s conversation, a sense of camaraderie developed between the roundtable participants (mostly researchers and other Federal partners) as they discussed survivor stories of seemingly impossible efforts to generate useful analyses with diffuse data in the midst of the housing market crash. During those early tumultuous days in 2008, our economists in PD&R faced numerous challenges trying to size up the foreclosure problem at a national scale, not to mention breaking the analysis out by local levels of geography or by race and ethnicity. However, since those difficult days, the housing finance research field has enhanced public and proprietary mortgage datasets on a national scale and developed innovative ways to link data to rich local sources while maintaining the integrity of the Privacy Act and protecting confidential information through strong public-private partnerships. Even as we improve our ability to measure the impact at the local level, we still need to improve our measurement of these trends at the national level and across a variety of markets.

One early example of such efforts highlighted at the roundtable discussion is the Community Advantage Program (CAP) Database from the University of North Carolina-Chapel Hill’s Center for Community Capital. Researcher Sarah F. Riley described the CAP as a partnership dating back to the late 1990s between UNC, the Ford Foundation, RTI International, Self-Help, and Fannie-Mae that created a database of about 46,000 loans made to low-to-moderate income households. What makes this database so distinctive is its combination of a longitudinal survey from a sample of the original pool of borrowers with zip-code price-estimates and small-geography census data. Such a rich set of indicators allows the research partners to untangle which aspects of the borrower experience influence performance over a loan’s lifecycle. One downside is that the sample is heavily skewed to southern states, raising some concerns of generalizability at a national scale.

Professor Vicki Been from NYU’s Furman Center for Real Estate and Urban Policy described a more recent example of such herculean data linkages that required strong public-private partnerships. In a sneak preview of her study coauthored by the Treasury’s Office of the Comptroller of the Currency (OCC) researchers Ioan Voicu and Andrew Tschirhart and fellow NYU researcher Mary Weselcouch, Professor Been described merging monthly loan performance data from the OCC with Home Mortgage Disclosure Act (HMDA) data and various NYC administrative sources such as property deeds and neighborhood crime rates. The linkages allowed the researchers to control for a host of variables that help unlock critical information on the role of foreclosure counseling and the effectiveness of loan modifications (payment, interest rate and/or balance reductions). They find that, in NYC, foreclosure counseling is associated with a greater likelihood of receiving modifications. In turn, households that receive modifications are less likely to end in re-default than households without modifications. Moreover, loans that receive modifications through the Home Affordable Modification Program (HAMP) significantly perform better than non-HAMP mods, which is evident throughout the HAMP’s history. Again, one downside is that this analysis is focused only on New York City.

As the researchers and Federal agency staff finished comparing their notes on challenges and opportunities in linking loan level data, it was apparent that these efforts could not have been accomplished without partnerships across all levels government, philanthropy, private industry and the social science community. However, these partnerships too often have materialized in limited scope, creating the need for a national platform for the standardization of methods for data matching and indicator development. I hope that PD&R can help continue to facilitate these exchanges and ignite broader partnerships to further our collective understanding of the emerging trends in the mortgage market.