I made a list of six intriguing analyses that can be done with the public 2018 HMDA data. Possibilities are almost endless.
- You can find where the borrowers are who can afford homes above a particular threshold (e.g., $1 million) using debt-to-income ratios, combined loan-to-value (CLTV) ratios, annual incomes, and appraised value. For example, identify borrowers who have substantial equity or can afford to take on more debt under current debt ratio limits.
HMDA Channel and Purchaser Type
- Study whether a particular lender accepts applications through a third party and/or originates loans in another institution’s name. For each channel, examine whether the institution kept the loan in portfolio and if not, who purchased the loan.
Neighborhood Equity and Downpayment Analysis
- Study the distribution of transactions by open vs closed end, lien status, and occupancy status within the refinance, home improvement, and “other” purpose lending (e.g., for education expenses). Within a drilled-down product, see how much credit was extended and average income. What is a typical appraised value for the area? How are those home values distributed in the neighborhood between debt and equity?
- Study home purchase activity as to the percent of the appraised value that the home buyer is putting down.
Take Some Noise Out of HMDA Product Distinctions
- Important new attributes allow you to flag reverse mortgages, ARMs, open-end lines of credit, “other” purpose loans, commercial purpose loans, and loans secured by 1-unit, rather than 2-4 unit, properties. The common filters in the past were loan type, lien status, loan purpose, occupancy status, and property type.
Refined Loan Pricing Comparisons
- Study the average rate spread by prohibited basis group (race/ethnicity, gender, and age), including information about origination fees, discount points, lender credits, and CLTV.
What would you like to do with the data? Let’s keep the conversation going. Please share your thoughts with us at firstname.lastname@example.org.