For Ginnie Mae, data analytics lends disaster relief

When disaster strikes, data can help organizations keep a pulse on the fallout. But without an effective analytics strategy and platform, all the data in the world won’t help facilitate difficult decisions when they matter most.

In 2017, Hurricanes Harvey, Irma, and Maria spread destruction through Florida, Puerto Rico, Texas and the U.S. Virgin Islands. As the individuals affected tried to pick up the pieces in the wake of the storms, the Government National Mortgage Association (Ginnie Mae) scrambled to understand the magnitude of the impact on its mortgage issuers. Ginnie Mae’s answer was a new analytics tool: The Disaster Response and Relief Dashboard.

Ginnie Mae is a wholly owned government corporation within the U.S. Department of Housing and Urban Development (HUD) with the mission of expanding affordable housing. To help lenders get a better price for their loans in capital markets, Ginnie Mae bundles mortgages into mortgage-backed securities (MBS) and sells them to investors. This $2 trillion portfolio, in turn, enables Ginnie Mae lenders to make new mortgage loans available to first-time home buyers, low-income borrowers, and veterans.

“With these hurricanes bringing widespread destruction to Texas, Florida, Puerto Rico, and the U.S. Virgin Islands, Ginnie Mae was in need of a tool that would enable the business to proactively manage the portfolios associated with the lenders’ affected areas,” says Tamara Togans, vice president of enterprise data management at Ginnie Mae.

The tool would have to leverage internal and external data to assess the potential loss exposure in Ginnie Mae’s portfolio, understand which Ginnie Mae issuers were affected, and by how much. It would also have to help the company understand how much of loss exposure came from various loan-insuring agencies such as the Federal Housing Administration (FHA), Veterans Affairs (VA), and Rural Housing Services (RHS). The company also needed to project potential delinquencies of the affected loans using a variety of scenarios, and then combine those projections with issuers’ financial data to assess potential financial stress on issuers, monitor the risk of an issuer default, and make key decisions on disaster relief actions.

Defining the disaster relief effort

Ginnie Mae first assembled a task force of executives and key staff from its securities operations, policy, enterprise risk, capital markets, and data/IT organizations. Here, coordination was key to ensuring a holistic approach to solving the challenges that resulted from the disaster, Togans explains.

“Reporting and analytics were a critical requirement for the task force, as availability and timeliness of accurate data would be key in making program and policy decisions,” she says. “Early on, 40-plus reports would be used to support the task force’s decision-making. Eventually, the Director of Single Family Account Management, Harlan Jones, began evangelizing the idea of creating a technical solution to optimize reporting.”

Jones solicited the help of Ginnie Mae’s data and IT resources to build and deploy a dashboard that would give the organization timely information on the disaster impact. The resulting Disaster Response and Relief Dashboard earned Ginnie Mae a 2019 Digital Edge 50 Award for digital innovation, but executive leadership at the company wasn’t immediately sold on the idea.

“The project initially drew some concerns from those who legitimately challenged the idea of investing in a tool that would be used for less-than-frequent, season decision-making,” Togans says.

Jones worked with Togans and her team to broaden the utility of the tool to support predictive analytics that would allow the dashboard to provide disaster impact forecasts in addition to post-disaster analysis.

“Users would be able to define the impact area, simulate the severity level of the impact, and understand the immediate as well as potential long-term impact based on historical disaster data,” Togans says. “This was a key feature that better positioned Ginnie Mae to plan for future hurricane seasons.”

Quick response, rich visualization

Togans’ team identified requirements for the dashboard by focusing on the standard and common ad hoc reports that Ginnie Mae’s leadership and staff frequently requested for daily decision-making. She notes that functionality that would promote ease-of-use was a primary focus.

“The result was a tool with intuitive functionality that delivered information in an easy-to-navigate, centralized dashboard with rich visualization,” she says.

By presenting data to executives in a way that made it easy to digest results and socialize the benefits, Togans says the tool was able to turn those executives into ambassadors for the dashboard’s functionality and use.

“The project team learned early on that such a project that can positively integrate mission, vision, and core values with technology solutioning is one that resonates most with executives and staff,” Togans says. “The dashboard was very well received across Ginnie Mae, largely because it managed to speak to all user profiles — providing high-level presentation of data for executives as well as more parameter-based capabilities for managers and staff.”

Ginnie Mae built the dashboard using an iterative process. It leverages Tableau for data visualization and draws on monthly issuer-reported loan level data (there are currently about 12 million loans in the portfolio), along with HUD UPS Zip Code Crosswalk data, Census Bureau data, and FEMA data to determine whether a loan is located within an impacted area. It also has the capability to ingest third-party data, including property damage and repair cost estimates. The dashboard visualizes the geographical data, delinquency trends, issuer’s financial exposure, and highlights the key data points.

The need for quick action in the wake of the hurricanes necessitated an informal effort to gather requirements. The team developed wireframes to validate the project’s requirements and stood up an initial version of the tool within three weeks. The team then added functionality and features over the course of eight to nine months. Ginnie Mae turned to Deloitte as a technology integrator for the project. Deloitte helped the company integrate the tool with Ginnie Mae’s existing information assets as well as data from the Federal Emergency Management Agency (FEMA), the U.S. Census Bureau, and the U.S. Postal Service (USPS).

“The tool brought more enriched intelligence to the fingertips, literally, of Harlan and his team of account executives,” Togans says. “It generated a genuine intellectual curiosity amongst staff who now are empowered to make more nimble decisions in real time.”