The factory behind your home loan
Patrick reads from his 2022 Bits About Money essay on mortgages, making the case that a mortgage is best understood as a manufactured product, not a simple loan between a bank and a customer. He walks through the assembly line behind every home loan, the loan officer and back-office staff who build the 700-page document. Then he traces the supply chain it gets sold into, where GSEs insure against non-payment risk, servicers buy the right to collect monthly checks, and pension funds and other private capital end up holding the economic exposure, because they want it more than banks do.
Presenting Sponsors: Mercury & Granola
Complex Systems is presented by Mercury—radically better banking for founders. Mercury offers the best wire experience anywhere: fast, reliable, and free for domestic U.S. wires, so you can stay focused on growing your business. Apply online in minutes at mercury.com.
If meetings consistently leave you with hazy action items and lost context, Granola handles the transcription so you can actually participate and gives you searchable notes afterward. Try it free at granola.ai/complexsystems with code COMPLEXSYSTEMS
Timestamps:
(00:00) Intro
(02:26) Mortgages are a manufactured product
(04:19) Who manufactures mortgages?
(07:08) Who buys mortgages?
(07:42) The risk of non-payment
(10:08) Sponsor: Mercury | Granola
(14:35) The risk of failing to service a mortgage correctly
(17:51) Every other risk you could imagine, of which there are many
(24:04) Scratching the tip of the iceberg
(25:10) More about flow meters
(26:24) Wrap
Transcript
Welcome to Complex Systems, where we discuss the technical, organizational, and human factors underpinning why the world works the way it does.
Hi-de-ho, everybody. My name is Patrick McKenzie, better known as patio11 on the internet. I'm here alone in the studio at Lighthaven today, recording an audio version of the podcast. We've just gotten done with several video version s. But something that came up during the conferences held this week, at Manifest and LessOnline, was the topic of mortgages, and mortgages are often misunderstood by people, I feel.
So I thought I would read a classic essay, "Mortgages are a manufactured product," from Bits about Money, back on January 14th, 2022, that helps explain these for an audience that hasn't done mortgage origination professionally. As always for this format of episode, I might, or might not, throw in some ad lib commentary as we go through the essay.
So let's kick it off.
Why is it so hard for entrepreneurs to get mortgages? Why did the address I send my payment to change? Why don’t most banks own mortgages anymore? Why is there a thriving economy of small specialist lenders locally who only do mortgages? Why is a 30-year fixed rate mortgage available at all?
Welcome to the wild and wooly world for mortgages in the United States: the world’s most important manufactured product that virtually no users understand. For starters, virtually no one outside of the value chain considers a mortgage a manufactured product at all. Most people who own homes think they were the primary customer of the mortgage, but that’s not true.
(Apologies in advance for this being a very specifically U.S. flavored issue of this publication. While practice in some nations—Japan comes to mind—is trending in the direction of U.S. practice, the world’s largest and most sophisticated mortgage market is also very quirky, as a result of a hundred years of policy decisions with societal-level impact. Trying to address multiple nations’ takes on this instrument would make this issue the length of a small encyclopedia. Also, for simplicity, we’ll largely be looking at residential mortgages rather than commercial.)
I'll say as an aside, the best single essay you can read about mortgages is "The 30-Year Mortgage Is an Intrinsically Toxic Product" by Byrne Hobart. I'll link that in the notes, but it's more for people who have a financial orientation around life than this essay does.
Mortgages are a manufactured product
Let’s get the most important bit out of the way first: civilians think that a mortgage is a loan between a bank and a customer, and think of it primarily as services work (to the extent they think of it at all). This is a materially incorrect worldview, which drives much confusion about how the industry (and mortgages) function, both among people who can’t be expected to know better (customers) and people who probably could (regulators and journalists, among many others).
A mortgage is a product, which is built by specialist workers using an immense and costly capital edifice, to be sold into a supply chain for consumers of that product. It incidentally happens to involve a house and a loan, but those two facts do not drive most behaviors of the mortgage industry. The structure of the manufacturing process, and the consuming supply chain, do.
If it helps you to visualize this, think of a widget going into some hyperspecialized and boring bit of capitalism. Electronic flow meters, for example. [0] Somewhere, in fact in many somewheres, there exists a factory of people who do nothing but assemble electronic flow meters. These do exactly what they say on the tin; they measure flows of liquids or gasses using electric means. There are a lot of them in parts of the economy which push things through physical pipes, whether that is New York’s sanitation department or a steel smelter. It would be a very bad thing for the electronic flow meter factory to sell flow meters that didn’t meter flow accurately. Their consumers would hate that. Cakes would taste bad. Buildings would explode.
If you replaced your current mental model for mortgages with “it’s like a paper electronic flow meter for money, possibly with less paper these days”, it would improve your ability to understand the mortgage industry. The analogy is less about providing visibility into the contents of pipes (though mortgages must do that) and more “highly specialized manufactured widget that the entire world sits downstream of.”
Who manufactures mortgages?
Mortgages are written by originators. The first and most thorough misconception the public has about mortgages is that they’re written by lenders. Originators can be lenders, and historically generally were, but these days they generally are not. The people who make electronic flow meters do not also run the cake-baking factory as a fun hobby. Those are entirely different skillsets. They sell the thing they make, through a fairly complex value chain, to the cake-baking factory.
Of the top 15 mortgage originators in the U.S., only four are banks (Bank of America, Chase, U.S. Bank, and Wells Fargo), and most of those banks’ mortgage volume originate-for-sale rather than originate-to-lend. The remainder are all pure-play mortgage firms, some of which you’ve probably heard of due to direct-to-consumer marketing (RocketMortgage, Guaranteed Rate, etc) and some of which are more obscure (despite writing tens of billions of dollars of mortgages). This is due to a difference in operational models for them; we’ll get to that in a later issue.
As a side note, Rate used to be called Guaranteed Rate, and I wonder whether a wrist slap from regulators caused them to rebrand the entire company. There has been something of a kerfuffle between the CFPB and the previous administration, and various firms in the credit industries, over whether you can guarantee a rate to someone before they've put in an application.
So who actually has their hands on the tools in the mortgage factory? Every mortgage is a collaboration between:
The front office: A sales professional, called a loan officer by convention (and usually referred to in the industry as “a producer”). Their job is to interface with the home buyer, educate them on the most complicated and high-stakes financial decision they’ll have to make in their lives, and project manage the financial side of a real estate close. This will include a backbreaking amount of passing documents to…
The back office: Approximately 1.5 professionals per loan officer, who are responsible for making sure electronic flow meters are properly calibrated. Oops, wrong side of the analogy. No, they are responsible for making sure that the physical instantiation of the mortgage, 700 pages of documents or so, will pass the strict acceptance tests of the value chain that sells mortgages to end-users of mortgages.
The industry operates, as many people who have interacted with it can attest, in a disturbingly artisanal manner for a trillion dollar manufacturing sector which props up a huge portion of the economy. As the person buying a house, you care tremendously about the professional skill levels of your loan officer in a way which you never really care about the professional skill levels of the person who assembled your phone.
Who buys mortgages?
In a typical manufacturing supply chain, the box with the thing in it passes through different hands, or the thing gets incorporated into increasingly complex assemblies. In the mortgage supply chain, the thing gets split apart and (sometimes) repackaged into other things.
The manufactured product which is a mortgage is a collection of risks. The finance industry and government in the United States, to be more efficient and accomplish a long list of social goals, has successfully disaggregated many of those risks. Different risks are bought by different entities.
The risk of non-payment
As a policy decision, the United States has effectively socialized most of the market for the risk of non-payment of mortgages, in the interests of making home ownership more predictably available. (Whether this makes it cheaper or more expensive on net is a complicated question to answer.)
The mechanism for this are the GSEs (government-sponsored entities), like Fannie Mae, Ginnie Mae, Freddy Mac, and the Federal Home Loans Bank. These are all privately owned entities who have CEOs, shareholders, etc etc, but they’re also policy arms of the U.S. federal government and everyone knows it. (If there was any ambiguity about that, and there was very little, the financial crisis dispelled it.)
The business model of the GSEs rounds to this: they publish the specifications for the mortgage industry. This lets you know that the electronic flow meter will work as advertised and not blow up the factory you install it in. A mortgage which fits their specifications is called “conforming.” Buyers of a conforming mortgage need inquire no further about the 700 pages of documents. They treat mortgages as black boxes with some observable metadata to them.
The GSEs monetize these specifications by writing insurance against all conforming mortgages. The insurance is against, specifically, non-payment risk. Originators of mortgages who desire to sell them into value chain overwhelmingly choose to get conforming mortgages guaranteed. The cost is below 60 basis points and ultimately borne by the home purchaser, in the form of points or interest. Points, if you haven't bought a mortgage recently, are prepaid interest on a mortgage.
The GSEs, operating as galactically sized insurance companies, profit or lose in a similar manner to insurance companies: they can use capital markets to invest the float (though, due to risks and their quixotic relationship with the Treasury, they cannot be as aggressive as e.g. Berkshire Hathaway) and hope to earn an underwriting profit. That is, they hope the specifications they write and rigorously enforce correctly predict and price default risk. This was resoundingly not true in the run-up to the global financial crisis, which resulted in large government bailouts and a sort of quasi-nationalization of them.
(There has been some limited experimentation with privatizing the risk of non-payment again, via mechanisms like Freddie Mac’s Credit Risk Transfer product, but the GSEs will happily tell you that most is currently borne by the taxpayer.)
I will say as an aside, this is currently in the course of changing. There's been a concerted push to re-privatize the GSEs and float shares of Fannie Mae and Freddie Mac again. This would effectively involve a transfer of tens of billions of dollars between the government and those who currently, and perhaps in the future, own those shares.
This has been covered by Matt Levine, among many others. It doesn't look very likely that there will be action on this until after the midterms, for various political reasons. Currently, smart money is divided on whether this is likely to happen. It's heavily in some people's interest, and not very much in other people's interest. When investment banks survey professionals involved in the mortgage market, about 50% think the GSEs are likely to be refloated in the near future, and about 25% think it's much less likely. I suppose the remaining 25% or so are on the sidelines.
The risk of failing to service a mortgage correctly
A mortgage has a quirky little subcomponent called a Mortgage Servicing Right (MSR). Every month, it needs to collect money from the borrower and send that money… somewhere. This implies, minimally, a mailbox where you can send checks, someone to open the mail, and a phone number with a CS representative who can answer questions like “What is my current balance?” and “Did you get the last check I sent you?”
The holder of an MSR has an obligation to the owner(s) of the mortgage to not screw this up, enforced both by contract and by quite a bit of regulatory supervision. In return for taking on the boring business of opening letters and answering phone calls, they’re paid a small cut of the incoming payments. This generally works out to about 25 basis points of outstanding principal.
Their obligations are actually more complicated than this quick sketch, particularly in unhappy cases like e.g. when someone defaults on their mortgage. It’s an operationally intensive business to be in.
Importantly, MSRs trade some amount of work you could build a giant human-powered machine to do for a stream of ongoing payments. The finance industry loves that sort of stuff, and has built a large system to buy and sell MSRs. This results in the servicing of mortgages migrating to large players who have huge economies of scale to open letters and answer phone calls. It also confuses many people who walked into e.g. a Citibank to get their mortgage and think that this means Citibank will own the mortgage (probably not) or take checks from them for it (probably not).
I'll say as an aside that in the continued machinations of the mortgage markets, there's been some spinning out and recapture of mortgage servicers from the large mortgage companies. One of them said on a call with investors, "We're not a mortgage company, we're a homeownership company." So implicitly, they want to bundle the servicer in with them.
Why rebundle the servicer? Well, among other things, this essay was written at a point of historic lows in the mortgage market. We've since had an increase in interest rates, so mortgages are more expensive these days. You can buy less house than you used to be able to for a given monthly payment. As a result, people in mortgages struck in the 2020 or 2021 timeframe are quite tied to those mortgages. So instead of doing the more typical American thing and moving out of that house in seven or eight years, they might stick around for ten years or more to enjoy the fact that their mortgage has an interest rate around 2%, rather than the 8% I'm currently paying on my own mortgage. I should probably refinance at some point. That's neither here nor there.
Anyway, if you stretch out the length of time someone is actually paying a mortgage, the associated mortgage servicing right is mechanically more valuable, because you're servicing the mortgage for longer, so the servicer gets a longer stream of payments. Mortgage servicers as businesses have become more valuable as interest rates have increased, because the servicing right gets spread out over more payments. It's interest rates all the way down. Back to the essay.
Every other risk you could imagine, of which there are many
Private capital buys all the other risks. The How of that will have to wait for a future issue.
We can get into part of the What, though, right now.
A mortgage, like any manufactured product, has a value. You can go to your friendly neighborhood seller of mortgages and buy mortgages in any size or shape. If your factory needs a flow meter, someone has a flow meter for you and can tell you what it costs; they probably didn’t build the flow meter with their own two hands.
The value of the mortgage is not the outstanding principal, any more than the value of a flow meter is the maximum rate of water flowing through it. That is an important technical feature of the artifact at issue, but is not the price tag.
The value of flow meters goes up when there are more companies wanting to buy them in a time period than factories can manufacture them in the same time period, and down when the reverse is true. That happens to mortgages, too; supply and demand is a thing.
One driver of supply and demand is the interest rate environment, which is not entirely external to the mortgage market but which you can assume is for simplicity. All existing loans, including mortgages, become less valuable when interest rates go up and more valuable when interest rates go down. (The easy way to remember this: pretend you bought a widget for $100 which spit out $5 a year. If the new model of widgets spits out $6 a year and also costs $100 new, will anyone buy your perfectly functioning used widget for $100? No, that’s silly; they’d buy the new one. You have to offer them a discount until the yield matches that of the new model.)
This is one of the reasons why society has arranged it so that banks don’t own most mortgages. For a very, very long time, banks were primarily engaged in the business of maturity transformation; they would borrow money short term from depositors at low interest rates and lend it long-term, mostly to mortgage borrowers, at a higher interest rate.
We have learned, through long and painful experience, that this is an extremely repeatable recipe for banking crises. Interest rates go up, which increases banks funding costs and decreases the value of all the mortgages they own. (It also decreases the number of re-finances they sell, because they are now mechanically more expensive, but that’s a relatively minor factor.) If the bank is not sufficiently capitalized, to maintain their reserves (and have adequate liquidity to meet responsibilities to customers and regulators) they’ll have to sell some of those now less-valuable mortgages.
All banks experience the same pressure effectively simultaneously; supply and demand hits the value for mortgages and their values start to plummet; the book value of mortgages at other banks now means they’re under-reserved and so have to sell mortgages to raise liquidity; this creates a vicious circle in the banking sector.
A widespread misconception about mortgage securitization is that it was created to make Wall Street rich. This is a popular narrative, especially due to Michael Lewis’ best book (and one of the best finance movies ever made). This narrative is… incomplete.
Mortgage securitization, and secondary sales of loans, and other mechanisms cause mortgages to migrate from the banking sector to pools of capital which are more structurally insulated against the interest rate cycle.
A pension fund turns out to be a great counterparty to a mortgage, for example. They have known needs for cash in the short term and, if interest rates rise, that is generally good news to them. (It decreases the net present value of their funding shortfall for future commitments.) Their pile of mortgages would be worth less if they sold them but they do not structurally need to be sellers.
The pension fund cannot manufacture mortgages. That requires complex technology and a team of professionals that they do not have. They do not ever desire to write mortgages. They just want the economic exposure those mortgages represent, and they want it more than banks do.
It turns out, counterintuitively, that there are many, many buyers in the world who want to own mortgages more than banks do. They end up owning almost all of the mortgages, for fundamentally the same reason that the electronic flow meter factory ends up owning very few of the world’s current stock of flow meters. Someone has a better use for them and they exist to supply that better use, not to finance it.
I'll say as an aside, this piece was written in 2022, but if you remember early 2023, we had a mini banking crisis in the United States, and mortgages were, dun dun dun, again central to that crisis. But they weren't central for the same reason they were in 2008.
In 2008, the fundamental problem was that people were taking subprime loans and then were not in a position to pay them back. That's not what doomed, for example, SVB. SVB didn't take a hit on its mortgage-backed securities held for sale because those securities had payment risk underlying them.
What actually happened was a swift ratchet in interest rates from roughly 0% to roughly 5%, and then bond math rears its ugly head. Bond math says that if the interest rate goes up by 1%, the value of a bond goes down by about 1% for each year of duration it has. Mortgage-backed securities typically have an outstanding duration of plus or minus seven years or so. Seven years times about a 5% hit means the value of those securities goes down by about 35%.
That hit is what caused SVB to become insolvent, as originally reported by, among others, Byrne Hobart. Once a bank is insolvent, things tend not to work out very well for it. There was a great effort to save SVB and also First Republic, which had very similar dynamics. Neither of those efforts came to much. SVB had a run on it and was beyond saving. First Republic limped along for a couple of months trying to raise enough equity to cover the hole in its balance sheet, caused by the decrease in the value of the mortgages and mortgage-backed securities on its books. It ultimately wasn't survivable for First Republic. As a result, I now bank elsewhere. Chase ended up buying them out.
Scratching the tip of the iceberg
We will return to the topic of mortgage finance, and the plumbing underlying them, in future installments of Bits about Money and Complex Systems. But if you wanted to read about it right now, here are some excellent entry points:
Digitally Transforming the Mortgage Banking Industry. If you want to know how the sausage is made, this is the best account of it. I would read it critically; the author is from the sales side of the industry and is, ahem, in a position where they have to understand software to a greater degree than they actually do. Still, it’s a great look at mortgage origination where the forged bank statement hits the underwriter.
The 30-Year Mortgage is an Intrinsically Toxic Product by Byrne Hobart (who also writes a very good newsletter). This is lay-accessible but deep on both financial wonkery and housing policy.
The Dead Pledge. A scholarly history work on the pre-war years of mortgage finance, in whose shadow much of the U.S. economy (and social organization) now sits.
More about flow meters
My quixotic level of interest in electronic flow meters is on account of my best salaryman friend working at the company which makes the best electronic flow meters. I have had many long chats with him about the science of flow meters, the challenges of being a Japanese salaryman walking through customs in the Middle East with a suitcase full of sample flow meters, and the surprising difficulty of translating flow meter marketing copy into English. (For social reasons, I'll elide naming him or the company, but statistically speaking you are overwhelmingly likely to be... downstream of their work.)
And making an observation I've made many times: the world is fractal in its intrigueness and difficulty. Everywhere you look, there's a supply chain behind it, with arbitrarily deep levels of subspecialization. So there very much is someone sitting in an office in central Japan right now, preparing to translate marketing copy from Japanese into English, so that your water department has the very best electronic flow meters available. There are also mechanical flow meters out there, but trust that guy: the electronic ones are much more reliable over the timescales your water department wants between expensive capital upgrade cycles.
Thank you very much for listening to Complex Systems this week. Hope it wasn't too much of an ad for electronic flow meters, although they are awesome. We'll see you next week.
