During this week's Tech Sprint, the Polygon Platform team distilled an antidote to three things that increase uncertainty, cost, and bias in the HMDA submission process:
First, we'll stop data quality problems before they happen by capturing loan transactions at the source of truth. This means we’ll ingest individual loan document files in any format, as well as multi-loan batches from any loan system. This will solve the game of telephone currently degrading data quality each time an application goes through data entry or system-to-system transfers. Since bad data obscures the presence of bias, removing it is a crucial first step in understanding and eliminating bias, keeping lenders compliant in the process.
Second, 4,939 mortgage lenders each originated less than $2B in mortgages in 2019 (we’ll have the 2020 numbers in a few days). This group includes first-time HMDA filers and other lenders who don't have dedicated HDMA compliance personnel. These lenders will gain continuous compliance by using myHMDA as an Agile Mortgage Workbench throughout the year. Under the hood, the Polygon Research Machine Learning Platform will stage loans in the proper LAR format, flag improbable and inconsistent data, and provide a feedback loop to lenders to understand the source of problems and correct them.
Finally, we bring HDMA data to life through our visual analytics tool HMDAVision®, which will be bundled with myHMDA for insights and use cases ranging from sales to compliance to M&A. In addition, lenders will have a private edition – myHMDAVision – which will give them insights into their LAR throughout the year leading up to the annual submission deadline. Our tools represent a democratization of analytics for small lenders, providing best-in-class solutions at a transparent price point.
Please keep an eye on www.myhmda.com as well as this blog to learn more, including our announcement of a free tier for lenders serving distressed and underserved areas.