HMDA Threshold Report Blog
Introduction
The CFPB released a report today analyzing differences in lending patterns for lenders below and above the 100-loan closed-end threshold set by the 2020 Home Mortgage Disclosure Act rule. While this analysis is necessarily limited and preliminary, the report does find some differences in lending patterns for lenders above and below the threshold. Lenders below the 100-loan threshold appear to make more investment purpose loans to higher income borrowers and non-natural person borrowers (i.e., trusts, partnerships, and corporations), as well as more loans secured by properties in low-to-moderate income census tracts. These findings are consistent with a possible explanation that lenders below the 2020 rule’s 100-loan closed-end threshold are making more loans to investors buying up property in low-to-moderate income census tracts for rental or resale.
Background
The Home Mortgage Disclosure Act (HMDA) is a data collection, reporting, and disclosure statute. The 2015 HMDA Rule required that financial institutions that originated no fewer than 25 closed-end mortgage loans in each of the two preceding calendar years and meet other reporting criteria such as asset and location tests report their closed-end mortgage activities. In May 2020, the CFPB raised this closed-end reporting threshold from 25 loan originations per year, where it had remained for HMDA reporting activity for 2018 and 2019, to 100 loan originations per year, effective July 1, 2020.
The annual volume of HMDA reporters varies substantially. For instance, in 2019 HMDA data, 5,473 reporters reported at least one closed-end origination. Among those reporters, 4,532 reported fewer than 1,000 closed-end mortgage loans. Focusing on these 4,532 reporters with closed-end origination volume below 1,000, on June 14th, 2021, the Bureau released a report that provides a brief overview of the general lending patterns of HMDA reporters with a closed-end origination volume between one and 1,000.
In this blog, we explain the data source and analytical approach, as well as key findings from that report. Even though the information presented in the report covers all 2019 HMDA reporters with closed-end origination volume below 1,000, the report highlights the contrasts between the reporters whose origination volume was below 100 and those whose origination volume was over 100, as 100 corresponds to the closed-end reporting threshold set by the 2020 HMDA Rule.
Data Source and Analytical Approach
This analysis uses publicly available HMDA data from 2019, collapsed to the reporter level. The information presented in the report is limited to originated loans that are not flagged as open-end lines of credit.
For each reporter in our data, we obtained the total number of closed-end originations, share of borrowers in major demographic groups, share of loans secured by properties in low-to-moderate income (LMI) census tracts, share of loans secured by properties in rural areas, share of loans secured by manufactured homes, share of loans that are conventional, share of loans that are secured by a first lien, share of loans for principal residence, share of loans for investment properties, median loan amount, and median income. We also identified whether each reporter is a depository institution (DI), and for DIs, identified their assets.
We rely on a series of binned scatter plots to demonstrate the general patterns which may or may not vary substantially with origination volume, with loan origination volume on the horizontal axis and loan or borrower or reporter characteristics of interest on the vertical axis on each of the binned scatter plots.
Findings
In our analysis, we find some differences in terms of the loan and borrower characteristics and lending patterns between the institutions whose origination volume falls below the new 100-loan threshold and those whose origination volume is above that threshold. In general, those lenders newly exempted under the 2020 HMDA Rule (i.e., with annual origination volumes that exceed the 25-loan threshold but fall below the 100-loan threshold) do not appear to be more likely to lend to Black and non-White Hispanic borrowers than larger volume lenders. There is some evidence that these lenders might be more likely to lend to non-natural persons, i.e. trusts, corporations or partnerships. Our analysis also suggests that a higher percentage of their loans are secured by properties in low-to-moderate income (LMI) census tracts, properties in rural areas, second liens, and investment properties. Their borrowers also appear to have higher incomes than larger lenders’ borrowers as well.
We note that this analysis in the report is necessarily limited and preliminary and is not an assessment by the CFPB as to the effectiveness of the thresholds change in meeting HMDA’s objectives. Additional analysis is needed to better understand these findings and to explore the impact of the threshold changes on data available for specific markets.