Understanding the Tradeoffs
Liquidity and credit boxes
My Uncle Jerry passed away 5 years ago, but he was a great storyteller. Among his many stories,[1] as the longtime executive of a small Milwaukee-based thrift, he told me what the mortgage loan “secondary market” was like for midwestern Savings & Loans before the government sponsored enterprises (GSEs), Fannie Mae and Freddie Mac, dominated the mortgage finance arena. Jerry would wryly reminisce about conference rooms at regional trade association meetings where bankers would bring boxes full of loan files (all on paper of course) and start haggling about the details of each loan they wanted to buy or sell. “What ya got?”, Jerry said, was how most conversations started: not a very liquid market as you can imagine. Jerry spoke of a very different sort of credit box….
Of course, what made Uncle Jerry’s old school S&L trading market so illiquid (i.e., difficult) wasn’t different loan interest rates and terms. Bankers and bean counters can do easy math[2] determining what to invest in and what to hold if the loan risk profile is uniform. The issues complicating loan liquidity for those S&L guys was that each bank had its own unique heterogeneous[3] origination nuances, requirements, and policies designed to serve local markets, quirky underwriters and Boards of Directors, and relationship-based customer models creating a lack of uniformity for loan documentation, underwriting requirements, credit standards, appraisals, servicing requirements, data quality, etc.
Fast forward to today, we no longer have paper files in boxes.[4] Instead, we have massive amounts of digitally stored information, underwriting engines, and AI to sort through, assess, and price all the file data Uncle Jerry and his peers haggled over to facilitate our highly liquid mortgage secondary market. Critically, however, through their Seller/Servicer guide requirements, Fannie and Freddie created a liquid home finance market essentially by solving the heterogeneity issues in home loans. With the market power of a government sponsored duopoly to enforce loan fungibility,[5] the GSEs established and continue to define the credit box for what qualifies as a conventional conforming loan permitted to back the mortgage-backed securities (MBS) they issue.[6]
Relating this to GSE privatization chatter
As noted in a June 5, 2025 podcast interview by Robbie Chrisman of his father, Rob Chrisman, most of the current debate about consumer impact of privatizing the GSEs focuses on the importance of the “implicit”[7] federal guaranty of the GSEs MBS in keeping interest rates lower. There is certainly merit to that point,[8] but our extremely concentrated, government-controlled home finance system[9] also poses other considerations for consumers due to the need for uniformity to facilitate liquidity. Most folks don’t consider the impact the GSEs have on who gets a loan or how they standardized the credit box.
GSE driven uniformity, however, is enforced with market advantages conferred by the government to further its housing policies.[10] As shown by the FHFA’s (thankfully aborted) 2023 DTI grid fiasco, this power in the hands of a government lacking regulatory humility can be misused to further political objectives instead of simply facilitating orderly market function and subsidizing the housing industry and consumers broadly.[11] Meanwhile, through their specially granted status, the GSEs developed (and still hold) enormous stores of valuable and sensitive (but proprietary and confidential) data about borrowing risks and consumer behavior.[12] Release of the GSEs will need to contend with the question of whether that data is sold along with the going concern value and liabilities.
Mortgage loan uniformity analyzed
Using that proprietary data (but without revealing it), the GSEs have had loan level pricing adjustments in place for a while now; slicing and dicing various credit data points into pricing adjustments. Yet, the basic credit box underwriting requirements of downpayment, credit score and income remain. But, is credit box uniformity still important to the MBS market today? Shortly after the rise of LLPAs, in 2013, a Federal Reserve Bank of New York paper highlighted the importance of uniformity by comparing mortgage-backed securities trading (specifically MBS/TBA trades[13]) to corporate bonds, noting,
“The agency MBS market is substantially more homogenous than the corporate bond market, however, and TBA trading helps homogenize the market further, at least for trading purposes.”
But is uniformity (homogeneity) needed for MBS liquid market operation? The report continues,
“There is little consensus on exactly how much actual homogeneity in the underlying mortgages and securities is necessary to support the fungibility and liquidity created by the TBA market, …. However, beyond some unknown point, fragmentation of the MBS market through greater diversity of loan and MBS features would likely reduce liquidity. In contrast, standardization of documentation, structuring, and mortgage underwriting criteria within the TBA-eligible universe is likely important to help maintain fungibility across securities, and thus promote market liquidity.”
So, the New York Fed’s eggheads[14] say uniformity is good for liquidity, but they can’t quantify how important that is. Well, given recent news about privatizing the GSEs (Tweets/Truths whatever) from Trump and his administration, we may be about to find out.
Credit box and the federal guaranty
Despite growth in the so-called “non-QM market”,[15] for the vast majority of consumers, if you want to a buy a house with financing, you need to meet the credit box underwriting requirements established by the government. Moreover, the psychological impact on potential homeowners’ common understandings arising from the GSEs’ credit box (such as the need for a large downpayment) can create a chilling effect on the market. This means the government’s credit box not only picks winners and losers of government housing support, but also may prevent some people from even thinking they are eligible to be homeowners.
Now that we have the federal Ability to Repay Rule (ATR),[16] however, why should the government care about credit box requirements beyond making sure the borrower qualifies under ATR? While uniformity of home loans is important to liquidity, is uniformity and control over the credit box needed to limit the federal guaranty? What does all of that proprietary data tell us about the risk of loss to the government from expanding the credit box to all borrowers who meet ATR? As we undertake the privatization discussion and what will become of the federal guaranty, it is important to also consider the role that the credit box plays in that risk equation.
Tradeoffs
The GSEs have essentially controlled the US home loan credit box for over 50 years, and many people believe their guidelines unfairly limit homeownership possibilities for creditworthy borrowers.[17] I believe this has also stifled demand and innovation that might have otherwise expanded homeownership opportunity. On the other hand, the liquidity and interest rate subsidies provided by the GSEs have no doubt led to a well-functioning housing finance market for most people.
Privatized GSEs may or may not have an implicit (or explicit) federal guaranty[18] or may expressly limit the amount or application of the guaranty to certain securities. Eliminating the rate subsidy and government-imposed uniformity will likely increase consumer costs, but escaping the tyranny of a federal guaranty for an ATR standard could also enable the GSEs, banks and Wall Street to really open the credit box to wider homeownership opportunities for everyone.[19]
Sometimes market power can be used for good and sometimes not. Maybe that’s why former SCOTUS Justice Stephen Breyer always spoke of balancing tests in connection with antitrust law. I’m not sure where I come down on the tradeoffs with the GSEs, but, if you’ve been paying any attention at all to what I write about, I really don’t like tyranny.
Rob Hahn Interview
Totally separately, here’s a better link to my interview with Rob Hahn about RESPA and the MLS System.
[1] To know Jerry is to remember at least one good metaphor or story he told you still valuable to your own conversations. Many Jerry Levy stories have appeared in these Musings (with and without footnotes).
[2] “Bankers can do easy math”. I do not consider myself a banker in that respect, but the business of banking is borrow short (deposits) at lower rates and lend longer term. Vastly oversimplifying and ignoring leverage from the Federal Home Loan Banks and other sources, the haggling resulted when one bank had more loans than deposits and another more deposits than loans. The secondary market made that discussion moot and enabled significantly more home mortgage lending generally.
[3] This is the right word to describe these loans. It means “different in kind; unlike; incongruous. composed of parts of different kinds; having widely dissimilar elements or constituents.” Please do not confuse this word with heterogenous which is a medical term meaning “originating outside the organism”. Small spelling difference, but big meaning difference.
[4] False. Astonishingly, despite having the GSEs massive market control and all manner of digital progress and learning from COVID-19 about remote transactions, we still have too much paper in boxes. Let’s get rid of the paper folks once and for all. We can do this! (For the environment and for cost savings/efficiency).
[5] I think of it like a McDonalds. You can order different things, but it doesn’t matter what McDonald’s you go to because they all menu items taste the same regardless of location.
[6] The two GSEs still have spent the last 15 years trying to create even more uniformity to have fully common data sets and common securities issued. They are getting closer, but the goal would seem to be an asymptote.
[7] Bloomberg’s Matt Levine says, “That’s what an implicit guarantee is: You say there’s no guarantee, but there is.”
[8] Chrisman said experts estimated the value of the federal guaranty at between 50bp and 150bp.
[9] Through its control of the GSEs in conservatorship and including FHA/VA/USDA, the federal government controls well over 90% of the mortgage loans made in the US. See e.g., this Urban Institute report.
[10] Congress enacted policy decisions many years ago to enable the GSEs with this power to support the housing market. Today, after 17 years of conservatorship, it is safe to say Congress has not made any decision about the future role of the GSEs in housing policy and has abdicated its legislative role to the executive branch. Any release plans for the GSEs implemented without congressional authorization, however, could face a variety of legal challenges.
[11] As I note below and in prior Musings, however, by creating and enforcing a government-imposed credit box, the GSEs subsidize certain consumers over others, but perhaps the past subsidies didn’t pose the same political implications as the FHFA’s DEI-infused “grid fiasco”.
[12] I would argue the credit bureaus pose similar concerns.
[13] TBA trades were said to account for over 90% of all MBS per the NY Fed paper which also clarified that “in a TBA trade, similar to other forward contracts, the two parties agree upon a price for delivering a given volume of agency MBS at a specified future date, but the identity of the specific security is not known on the trade date.”
[14] This is intended only as term that economists working in the Federal Reserve system might sometimes call their colleagues. I do not mean to dismiss their work or make fun of them. Well, maybe a little when they write a complicated paragraph about something they studied only to conclude that they don’t know any more than when they started.
[15] Many loans that are sold to non-QM investors might still be eligible for QM and/or GSE financing. The ability to pay rule (ATR) is only an income test for determining whether a borrower can reasonably afford the mortgage loan. LTV and credit score are irrelevant to ATR. But the QM safe harbor was tied to the GSE credit box and remains somewhat intertwined, but clarity on that is challenging with the CFPB guidance in flux.
[16] As I noted back in 2020 in just my 4thMusing and again in my 71st Musing, the idea that a lender must underwrite a loan to have a reasonable belief the borrower will pay it back, seems so obvious that it is hard to believe we need a law making it illegal and “predatory” if lenders fail to do it. Yet, In the run up to the 2008 meltdown, mortgage lenders and borrowers chased unaffordable subprime and stated income loans right into the worst housing and financial crisis since the Great Depression of the 1930s
[17] Talk about sources of institutional racism…
[18] Lawyers debate whether the implicit guaranty would be legally enforceable. Maybe upon release they will call it a “possible” or “probable” guaranty instead of “implicit”. Trump’s a real estate guy, so for enforceability purposes I’m going to say it will be more like a verbal agreement subject to the Statute of Frauds.
[19] Targeted housing finance programs could be run out of FHA to pick winners and losers.