HMDA Analysis for More than Just Compliance...
Are you missing an opportunity when you review your HMDA LAR to look beyond technical reporting errors? Does the LAR data have more to tell you?
When I first met John, he asked me: Is it possible for our clients to not only survive these regulatory changes (Dodd-Frank), but actually benefit from them? Well, classic case of John asking a question that seems innocent enough at first, but that actually guides you towards a much larger breakthrough. This is a question, like many others like it, that I wound up thinking about for a long time- notes scribbled down, things discussed ... probably could have written a research report on the topic. Ultimately it left me with a bunch of ideas and several conclusions.
One conclusion that I came to is a distaste for doing anything purely for compliance sake. What a waste! I mean, if Massachusetts forces me to renew my driver's license this year at the DMV, I don't have a choice - I have to go wait in line for a couple of hours. But doing nothing other than sulking is just letting that requirement hurt me more than it should. Maybe I take the opportunity to meet a friend for lunch who lives near the DMV. Maybe I use it to file some paperwork that will save me time next year. You get the idea - just get over the fact that it's required and make the best of it. Take control back.
So that train of thought leads me to our topic:
HMDA Now - Going Through the Motions
Every institution does some form of analysis of its HMDA LAR. I think often it's done to try and catch any HMDA or CRA errors. Often software is used to accomplish this.
I think this is a missed opportunity.
Vision - Making the Most of It
The HMDA LAR can speak volumes about an institution. And not just about whether HMDA reporting errors exist. Obviously there's CRA compliance. But more importantly, you can analyze the LAR to identify concerns with ECOA notification requirements, with fair lending issues, and with numerous non-compliance items related to efficiency, lost opportunities, and peer comparison.
If you're using software, there's normally much more you can get from the statistical reports than you currently are. It can be very helpful in assessing the health of your organization and how it compares to other institutions, nationally and locally.
What You'll Need
Just looking at the HMDA LAR can be revealing. But you'll want the ability to compare yourself to peers. Well one way to gather peer data is through the publicly reported HMDA data from the prior year.
For finding out information about other organizations, for free, I like to use the CFPB's HMDA reporting tool - available here. As an example, it took me about 2 seconds to look up that there were 17,037 originated loans made to male borrowers in Delaware last year, as compared to 8,238 loans to women. Now, I admit I don't exactly see how that piece of information is helpful to me. But then again, I'm not an executive for a mortgage lender trying to develop a strategy for that region.
For additional comparison purposes, you may need to buy data from The Warren Group or search for articles like this (an article that shows this isn't a completely original idea of ours today).
Apply Some Real-Life Examples
Without further delay, let's go through some real examples of this type of analysis. (Again, imagine that you're at your desk and already have to go through fair lending reports - let's make the best of it!)
Number of Denials & Origination Rates
Relying on national statistics, it's common to see this at about 16% of your reportable transactions. Look for large deviations from this, from competitors, and from past years.
In a similar vein, what's your origination percentage overall? See the article above to compare yourself to peers (although you could get more specific data for local peers): For example, small lenders and community banks are at about 70% origination rate. How about an average 67% origination rate for conventional home purchases, 65% to middle/upper income census tracts, 60% by lenders with $10 billion or more in assets, NCUA credit unions at 58%, etc. If you have a middle/upper income census tract and a 40% origination rate - maybe it's worth asking if something is wrong? Same thing if you're a small community lender with a 55% origination rate.
Time for Denials
How many loans are you denying beyond 100 days? Whether or not you see a fair lending risk here, this is a sign of a department that may have an opportunity to improve customer service ... imagine trying to get a loan but ultimately getting denied after working on it for 100 days? Sounds like an upset customer and a clogged up pipeline.
Number of Loans Coded Application Approved/Not Accepted
Expect this to be in the ballpark of 4% of your LAR. Much more than that is unusual.
Late Withdrawals
How many Withdrawals do you have over 30 days? Over 50 days? While there are always special cases, it'd be unusual to have a large number of Withdrawn loans this late in the process. Why? Reg B requires action within 30 days and we only code a loan as Withdrawn if the borrower withdraws before we have enough information to make a credit decision. Most of the time we should know enough to make a credit decision 30 days in right? What about 50 days in? This could be a) a mistake we're making in HMDA reporting, b) a sign of issues with complying with ECOA, or c) a sign that your underwriting department is struggling, whether they're to blame or not.
Of course you can think of dozens more. The point is that this should all be available to you to monitor fair lending performance. Anything better than nothing, right?
In Other News
New CFPB tool launched to show deficiency statistics. Worth a minute of your time to check out this Mortgage Performance Trends tool.
Helpful article in Bloomberg from the reliable firm Reed Smith on the so-called "Black Hole" issue that continues with TRID.
And for a really fantastic read- check out the great newsletter (click here) from my colleague Galina where she provides great insight and helpful tips on running effective underwriting meetings.
Just noticed that the information provided as part of MIT Professor Albert Saiz's presentation for the annual Warren Group event earlier this year have been posted online. One of the most interesting I've seen lately!
On My Mind ...
Professor Saiz's presentation, referenced above, was intriguing because it offered a new take on Boston's housing issues. First of all, he framed housing supply as the single most important thing distinguishing Boston from several other key cities competing for the likes of Amazon, etc.
But more to the point I found interesting, MIT research shows that, of all the things that affect housing supply, the so-called "regulatory index" - a measurement of how difficult it is off the charts in Massachusetts. It is the third most difficult state by this index, behind Hawaii and Rhode Island, and followed by New Hampshire. Few other U.S. states even come close - with regulatory burden measuring at 1/3 of that in Massachusetts in California, Connecticut, and even further in other areas.
So it's not really the T. It's not really the commuter system. It's not unions or simply the lack of land to build on. Boston's housing issue really comes down to how difficult it is to get permits, construction licenses, pass inspections, and get through other regulatory limitations? Wow, pretty eye-opening.
“It takes discipline to focus only on high-value targets instead of giving in to the temptation of the low-hanging fruit life serves up daily.”
- Mark Divine
Thanks so much for reading our weekly newsletters. We're not always going to be perfect, but because we always do our best and try not to overpromise, we hope that we're always going to be trustworthy. Your calls and e-mails are very helpful - please keep contributing.