The new HMDA data was just released. Stay tuned for insights from @PolygonRE.

Did FHFA Director @Pulte pull the plug on SPCP this morning? No. "Special purpose credit" as defined in Reg B, the regulation implementing the Equal Credit Opportunity Act of 1974, is between a creditor and an applicant. Of the three vectors for eligibility, the one that has seen a lot of activity since late-2020 (when the CFPB issued an advisory opinion that succeeded in its goal of resolving regulatory uncertainty) is for-profit organizations - i.e. mortgage lenders. The lenders who launch an SPCP (or participate in someone else's) have to create a written plan in which they establish the need for the program and identify the "class of persons" the program is for,  monitor and measure the results, and, implicitly, at some point end the "special" program. The mechanisms used by lenders to extend special purpose credit boil down to either reducing the overall debt burden of the loan by subsidizing fees or contributing to a down payment, or to change their underwriting standards. The latter might limit whether there will be a buyer for the loan, forcing the lender to keep the loan in portfolio - there is no implicit or explicit counterparty in an SPCP who is obligated to buy that loan. What Fannie Mae and Freddie Mac respectively did with their HomeReady® First and BorrowSmart programs was a mixture of launching their own SPCP initiatives and buying loans that came through lender SPCPs - possibly including, in the latter case, loans otherwise outside of their purchase criteria. But they are not for-profit organizations right now, and perhaps this is what Director Pulte had in mind when he wrote "the current level of support for SPCP's is inappropriate for regulated entities in conservatorship". As we've done elsewhere online, I commend the efforts of lenders who seek to correct "special social needs" (more language from Reg B) in our society through an effective SPCP. I also commend the GSEs for wanting to do the same. But if there is to be an SPCP round 2 from the GSE's at some point, it needs to start with clear logic for how it ties to Reg B, include a written plan, and produce clear metrics.

My company sells a SaaS solution for HMDA analysis. That's by way of disclaimer, not sales pitch, as I make the following observations: With @russvought and @DOGE putting a hard stop to @CFPB enforecement, an irony most people aren't aware of is that the CFPB, after inheriting the HMDA data function several years ago, has made it much better. The irony lies in the fact that while the CFPB might (in the past or future) use HMDA to come after you, HMDA is also what you'd use to defend yourself - e.g. showing progress over time, peer performance, alternative metrics, etc. This - with HMDA's treasure trove of competitive intel on "every" lender in every market - makes HMDA more pro-business than many people tend to think.

Where there are a lot of broker-lender relationships:

And, finally, interesting patterns in the wholesale channel:

Even more so when we drill down to the zip code level (a call out to home town Chicago):

A some interest rate variation by geography:

A great deal of YTD concentration, and a tight race for the top spot:

FHA rates ticking down:

Some FHA loan observations via HUD endorsement data via @PolygonRE FHA Pivot. Looking at data through October shows loans ticking up:

The new annual #HMDA data is out, and I was surprised to see a new state at the top of the leader board by new loan originations (all types):

@CNN recently published an article looking at a major mortgage lender's approval rates by race. Their data source was great - the Home Mortgage Disclosure Act (#HMDA) - a data set we know a whole lot about at Polygon Research. But with great data comes great responsibility. One thing CNN did well was to disclose the subset of loans they were including in their analysis (their filters). But what they didn't say was that the loans they chose to analyze represented just 19.3% of the lender's 2022 approved loans. HMDA is an extremely rich data set; there are many ways to slice it. Looking at the full pie (all 2022 approved loans), there were other large lenders with a larger gap than the one CNN highlights in their article. But...not wanting to make the same mistake: "large lenders" is subjective; the top 2022 mortgage lender approved 478,162 loans while the # 20 lender approved 69,916 (based on HMDA data published by @CFPB 12/3/2023). So...the responsibility we encourage is to avoid a narrow approach to your business questions. Great data - if you have a great tool to analyze it - always yields follow-up questions as you explore it. #HMDAVision #mortgage #housing #marketintel polygonresearch.com cnn.com/2023/12/14/business/…

By the March 1, 2023 deadline, 4,438 lenders reported their #HMDA mortgage transactions for 2022. This is more lenders - in a down year - than reported for 2021 (4,316 this time last year, growing to 4,365 via late submissions). #hmdavision #2022hmda