In Depth Analysis: CalculatedRisk Newsletter on Real Estate (Ad Free) Read it here.

Sunday, November 25, 2007

What Is "Subprime"?

by Tanta on 11/25/2007 10:52:00 AM

It’s time to revisit this months-old question. (Some days it feels like “age-old,” but we do well to remind ourselves that “subprime” became a household word only quite recently.)

Mortgage underwriting isn’t really that difficult, which is of course why some of us have been so shocked at the extent to which lenders have been screwing it up in the last few years. I do not mean to insult the professionalism or skills of mortgage underwriters, mind you. Assessing credit quality, like a lot of other things, has become increasingly complex over the years as its constituent parts—the myriad ways people earn money, spend money, accumulate and put a value on assets, plus the changes in quality, quantity, and popularity of the many types of mortgageable real property, not to mention the legal issues that accompany those changes (condominiums, PUDs, cooperatives, leaseholds, grandfathered zoning, flood hazards and insurance availability, etc.)—have all become increasingly complex. You have your days when encountering a borrower who earns a simple salary and keeps his money in a simple bank account and is buying a simple single-family detached home with no deed covenants in fee simple is kind of a refreshing change of pace. Lending is never less complex than the economy that produces the money to pay it back or the extent of control we wish to exercise on land use.

That said, what it’s about is just working through the complexity of the variations on three things that have been the core of mortgage underwriting since roughly the dawn of time: the three Cs, or Credit, Capacity, and Collateral. Does the borrower’s history establish creditworthiness, or the willingness to repay debt? Does the borrower’s current income and expense situation (and likely future prospects) establish the capacity or ability to repay the debt? Does the house itself, the collateral for the loan, have sufficient value and marketability to protect the lender in the event that the debt is not repaid?

There is no New Paradigm, there was no New Paradigm, there is not going to be a New Paradigm. The Cs are the Cs. What we “innovated” was our willingness to believe that we had established the Cs with indirect or superficial measures (that are, not coincidentally, cheap and fast compared to direct measures). We looked at FICOs—scores produced by computers—instead of full credit reports and other documents to supplement them. We looked at the borrower’s statement of income or assets, not the documents; when we got docs, we looked at the last paystub or the current balance of an account, not the documentation of a long enough period to establish stability of income or source of account balances. We looked at AVMs instead of full field appraisals. We read the Cliff’s Notes.

These practices have not worked out so well, of course, but my point is that they were simply “innovative” ways of answering the three C questions, not new questions. They’re not a repeal of the laws of physics or the laws of the Cs. They’re just wrong ways to answer the right questions.

This is important to the question of what is subprime because people do wonder, quite understandably, what exactly the whole point of subprime lending ever was. I’m guessing a variant of this question has appeared in the comment section of this blog about a thousand times: Why do people make loans to people who can’t pay them back? What did they think was going to happen? Is that like stupid, or what?

What we call “prime” lending was based on the idea that all three C-questions had to get at least a minimally correct answer before proceeding. You had to be sufficiently creditworthy and sufficiently capable and have sufficient collateral before we made the loan. If you had, say, two out of three, you might qualify for a near-prime (like FHA) or subprime loan, depending on which two and by how far you missed the third. There has only ever been one kind of (mortgage) lending that let you get one out of three—we usually call it “hard money” or “collateral-dependent” lending, as it is based solely on the value of the collateral—and regulated residential mortgage lenders haven’t been allowed to engage in it since approximately living memory goeth no further. It’s right up there with pawnshops, payday lenders, and vulture capital: a blight on society that we will always have with us as long as we have vulnerable people and unregulated “entrepreneurs” to pocket what little money they have.

“Traditional” subprime lending was about loans to people who had capacity but not creditworthiness or who had creditworthiness and capacity but not great collateral. The first instance is the borrower who has defaulted on debts in the recent past, but who has sufficient income to carry a new debt (for example, a borrower whose bankruptcy has been discharged, and whose income is now more in line with expenses). The second instance, a large portion of traditional subprime lending, was loans for manufactured housing (mobile homes), rural properties, urban mixed-use residential/commercial properties, and, of course, substandard housing in general. These are properties that either cannot be properly valued as residential, or that are unlikely to maintain value over the life of a realistic loan term, or that, occasionally, simply appall a responsible mortgage lender who just doesn’t want to participate in putting human beings into unsafe, unsanitary housing. Yes, there were sometimes lines we wouldn’t cross.

Over most of its history, subprime lending was overwhelmingly refinance business. That’s why to this day a lot of people refer to it as “home equity lending” or HEL. It began with the “finance companies” (who were not depository lenders) who usually offered second-lien mortgages to cash-strapped homeowners, including those who were having a hard time paying the mortgage they already had, which was most likely to have been originated as prime or near-prime, for a much higher interest rate than is typical for secured lending. In the 90s, the go-go deal was the 125% loan, which ended very badly.

The subprime lenders who originated first liens were still, on the whole, doing first-lien refinances, usually for cash out of some sort. Often these were “foreclosure bailouts” or “take-outs,” where the cash taken out was just enough to pay off the past-due interest and fees owed to the original foreclosing lender. Typically, these were fairly low-LTV loans: the borrower was too cash-constrained to make payments, but did not want to sell the home to avoid foreclosure. A borrower with no or negative equity at all most frequently just mailed in the keys, then as now.

There wasn’t then and isn’t now any inherent reason that these take-out loans were irrational from the borrower’s perspective or the lender’s. Long-time homeowners do experience financial disruptions: job loss, illness, divorce, ninety days in the county clink for some regrettable excess of high spirits that one can think is unlikely to recur. If the current lender (the original “prime” lender) won’t forbear or modify or otherwise play ball, the subprime lenders were there to structure a refi deal that would keep the borrower in the home. For a hefty rate.

As long as it was plausibly a temporary or non-recurring problem, it was and is reasonable for this kind of lending to go on. Some of us, in fact, have always believed that this particular kind of subprime lending would be a very good thing for regulated depositories and the GSEs to be into: the idea was that the loan terms would be standardized, the application practices supervised, the risk analyzed more carefully, the pricing appropriate but not usurious, the servicing careful and responsible, and hence the likelihood of total recovery—of the borrower being eligible down the road for a refinance back into a cheaper prime loan—would be increased. Leaving “bailout” or “rescue” loans to the sharks and cowboys and fast-talkers has really not worked out so hot, especially once they learned that they could, with predatory lending tactics, create the crises that need the bailouts. Of course, now that we’ve thoroughly demonized “subprime,” the odds of getting the responsible regulated parties to take over the business are quite slim, as everyone is clamoring for them to stay away from it. This is a problem.

Whatever else it was, subprime lending just wasn’t much of a purchase-money business. When it was, you had things like down-sizing (a borrower who got in trouble owning the big house, sold it, and is buying the new affordable house but now has a dreadful credit history), amateur rehabbing (intentionally buying substandard properties), or generally bizarre transactions (non-arm’s-length deals, buyouts of contracts for deed, seller carry-backs, blanket mortgages, what have you). Plus the mobile homes. In other words, not just marginal deals, but deals involving a tiny margin of the real estate market. Until a few years ago, the idea that subprime lending could have real, measurable impact on the broad existing or new home sale market was, well, laughable.

The main impact of subprime lending on the overall mortgage business was the take-out function. As subprime lending grew, you saw better “performance” of prime or near-prime mortgage portfolios. This was not because subprime lending did away with the traditional default drivers of job loss, illness, divorce, or disorderly conduct; it was because loans in that kind of trouble had a place to go besides foreclosure. Prime lenders could and did congratulate themselves on their low foreclosure rates as if it were a matter of their superior underwriting skills, but that involves a high degree of naiveté. It’s really important to understand this issue, because it gets to the heart of the “contagion” thing. It is not that subprime delinquencies are “spreading” to prime loans as if some infectious agent were in play. It’s that the drain got backed up: when subprime lenders go out of business, or investors won’t buy subprime loans, there is no place for the inevitable prime delinquencies to go except foreclosure. Prime delinquencies become “visible” because they don’t move out of the prime portfolio via refinance into the subprime portfolio, where we “expect” to see them.

This dynamic is obscured because subprime has recently become a significant part of the purchase-money business, and observers who come to it recently, lacking the historical view, take that as normal. Because they don’t understand the refinance or “take out” function of subprime, they fail to grasp its relationship to prime lending. How, precisely, subprime became a purchase-money business is itself something that really needs explaining: in any informed view of the mortgage business, that’s an unusual and troubling development.

You therefore have this giant conceptual gulf between industry analysts and the media, the latter being, on the whole, those who never really spotted the problem with the idea that homeownership is always and everywhere a good thing for everybody because it’s always an “investment.” If you believe that, you don’t tend to see anything odd about lending practices that offer purchase-money (not refi money) to people who appear to have no particular qualifications for homeownership. In essence, the old “hard money” or “collateral dependent” loan went mainstream, except that it went from the margins of the housing stock—manufactured homes, dilapidated row houses, the old farmstead—to the front and center—new homes, flashy condos, high-quality existing homes whose previous owners were heading for the McMansion. Given assumptions about the collateral—like, its value always goes up and its value always goes up—you could more or less forget about problems with the other two Cs. When the RE markets were hot enough, in fact, there weren’t “problems” with the other two Cs. Sure, borrowers with loads of consumer debts and insufficient incomes failed to make mortgage payments just like they always did, but it was always possible to sell out from under foreclosure or get another cash-out. A humming RE market keeps those cash-out appraisals plausible.

The RE market, in other words, functioned for subprime purchase-money loans the way subprime refinances always functioned for prime purchase-money loans: it “took the loan out” before anyone had to report defaults. More than a few heady souls concluded that they had excellent underwriting expertise, since the subprime purchase-money loans they were making didn’t default much, just as their prime brethren assumed that it was their own underwriting, not the subprime refinancer’s risk appetite, that kept their portfolios looking spotless.

But of course it got to the point where the RE market couldn’t keep propping up subprime purchase loans without subprime purchase loans propping up the RE market. That’s why the two started to go south at the same time, and nobody can answer the question of which went first. They went down together. My own view is that overbuilding probably tipped the scale first: we just ran out of subprime borrowers before we ran out of houses. Certainly, just before the end, we saw some real desperation in the subprime lenders’ quest for purchase borrowers. There have been more than a few media reports of subprime purchase loans given to undocumented immigrants (complete with stated income and all that), which have led more than a few people (judging from our comment threads) to conclude that the entire subprime debacle was all about them illegals. Of course it wasn’t; the subprime lenders just ran out of better choices and didn’t want to give up yet.

The association of subprime lending with the brown people is just the most overtly disgusting bit of bigotry to arise from the great mess. The belief that subprime borrowers are “poor people” has taken root so deeply that you need a jackhammer to rip it out. The capacity C of traditional underwriting was, of course, always relative to the proposed transaction. A lower-income person buying a lower-priced property was, you see, not a case of subprime lending; assuming a reasonable credit history, it was a prime loan. People with quite good incomes and stellar credit histories who tried to buy way too much house got turned down by the prime lenders. That was back in the days when you could live within your means, and you were expected to do so.

The trouble with the low-income prime loan was that it was a small prime loan. And that there were, in many market areas, more lower-income people than lower-priced properties. Both industry greed—wanting to make the biggest loans possible to make the biggest profits possible—and industry overcapacity, combined with ever less-affordable housing in the employment-rich population centers, brought us to a situation in which we might not have started with poor people, but they were certainly poor by the time we got done putting them into too much loan to buy too much house. There are subprime borrowers you find. There are those you create.

And of course low income does correlate with less education, leaving a lot of low income folks vulnerable to slimy mortgage originators. And low income folks do tend to have a much harder time coming up with a sizable down payment, meaning that loans to low income folks are usually already tipping toward near-prime or subprime from the start, even before we look at other facts. Until rather recently, though, you could have a lot of lower income applicants who could quite comfortably handle the payments; they simply weren’t getting anywhere with accumulating the down payment, since renting was, in many parts of the country, actually more expensive than owning. These folks weren’t going to get off the merry-go-round of handing over income to the landlord instead of saving it for a down payment unless someone came up with some non-traditional affordable housing loan programs, which we did.

There is a growing meme out there, particularly coming from the conservative think-tanks, that it was those affordable housing initiatives of the 90s that led us down the road to ruin. Most of it is quite remarkable nonsense, as it skips over the part about how it managed to become more expensive (in terms of monthly payment) to own than to rent, how incomes stagnated, how young people started their careers with crippling educational debt, and how affordable housing initiatives (not the loans but the actual development of inexpensive homes) gave way to NIMBYism and investment in giant suburban cathedral-ceilinged heat-wasters and all that. Of course it no longer made sense to use low down payment programs for first-time homebuyers if the monthly payment was no longer cheaper than, or at most slightly higher than, renting. But it’s rather hard for me to understand how rents were dropping relative to home prices and mortgage interest rates as a direct and uncomplicated effect of 5% down programs that the GSEs offered in the 90s.

The argument goes that it was the relatively low defaults of those 90s-era affordable mortgage programs that spawned the current mess by giving everybody the impression that you could do no-down loans all over the place and not worry about it. This assumes that the lending industry is so stupid that it cannot understand the mechanisms that kept those defaults low: first, selectivity in the programs; second, the availability of home equity lenders (the old subprimers) to take out the problems; third, cheaper real house prices. Perhaps it is the case that the industry is too stupid, on the whole, to figure this out. But how that becomes the “fault of” the original affordable housing initiatives just isn’t clear to me.

What is clear to me is how convenient this argument is for certain folks whose only other option is to admit to having been stupid and greedy. Exhibit A, our favorite Tan Man, whose transformation from “I got into this business to help poor brown people” in the 90s to “those brown people made me do it” is nothing short of nauseating. Exhibit B is everybody who decided that the best way to avoid being given fraudulent income and asset documentation and appraisals was to not ask for documentation or appraisals. Exhibit C is everybody who made “investment” loans for properties that did not and could not cash-flow, and hence had to flip to survive. Exhibit D is the “bridge loan,” or the product designed to blow up in 24 months and force either sale or refinance. There are many more Exhibits in this sorry book. The point is that the whole flimsy edifice had to fall down. That it started with the weakest parts—subprime—is no surprise. That this means that it’s all about subprime is mystification.

So who are “subprime” borrowers? Right now, they are simply those borrowers whose bridges all burnt: can’t refinance, can’t sell, can’t cut the budget any further to stay in place. A large chunk of them are currently in loan programs that were designated “subprime.” Bunches more are in “Alt-A” and “prime.” We don’t have more subprime borrowers than we used to have; we have fewer subprime lenders. Perhaps it’s clearer if we just talk about subprime customers, rather than “borrowers,” since you aren’t a borrower until someone makes a loan. There are lots and lots of subprime customers right now. There just aren’t very many subprime lenders left to fill the orders. To be surprised that a lot of these subprime customers are right now hanging out (by their fingernails) in “prime” pools and portfolios is to be very naïve.

Another way to look at this is that what a credit crunch (or contraction) does is, precisely, move the goalposts: the universe of “subprime” borrowers can go from, say, 10% of the customer pool to 20% with the stroke of a pen on some underwriting guidelines. An RE bust does the same thing: a few unfortunate comps, and somebody who was a nice middle class homeowner with equity last week is “subprime” this week. As the dynamic feeds on itself, it begins to affect employment (and hence medical insurance) and, undoubtedly, the bliss of a lot of people’s domestic arrangements. I’m pretty sure it causes at least a few bouts of disorderly conduct. In that sense, the crunch/bust creates the “normal” or “historical” causes of mortgage delinquency, or at the least contributes to them mightily. It therefore produces “subprime.” It does not help to keep claiming it was produced by “subprime” defined as them.

My new motto—“we are all subprime now”—is an attempt to keep reminding the attention-impaired (that would include but isn’t limited to reporters and politicians) that demonizing “subprime people” isn’t going to prevent you from going there yourself. The fact is that huge numbers of people who have “prime” mortgage loans couldn’t refi or sell right now to—literally, for some of the uninsured—save their lives. They may well still be making payments on the mortgage, but they’re rapidly approaching upside-down if they’re not there yet, they’ve spent the proceeds of the previous cash-outs that kept up the lifestyle or just kept life together, and if the truth were known about credit card balances, their current FICOs probably aren’t the envy of the neighborhood either. They are, in short, subprime. They just don’t recognize themselves in the stereotype of the deadbeat serial bankruptcy filer or the undocumented immigrant or the waitress in the McMansion or whatever extreme case you can dredge up and label “typical” for subprime. They are, increasingly, “us.”

The invisible subprimers will do okay if this blows over and they don’t lose their jobs: if and when the RE market recovers, anyone who hung onto the mortgage he has will come out “prime” again. That’s the good news. The bad news is that attributing this solely to your own prudence and good judgment and inherent worth as “not one of those subprime people” is a form of delusion. Subprime is like the weather: when it rains, everybody gets wet. The political pressure for “bailout” measures has a little bit, in my view, to do with bleeding-heart sympathy for the poor and a lot, a whole lot, to do with the dawning recognition of too many middle-class homeowners that, well, we are all subprime now, but nobody’s filling our orders.

I for one fail to understand how the American economy can get by without subprime lenders. Sending everyone with a problem directly to bankruptcy just won't work: that's a recipe for a permanent credit crunch. Where I part company with those who fret that over-regulation of the subprime industry will "hurt the poor" is precisely on the matter of the subprime purchase money market. If people cannot afford to own homes, then encouraging them to buy them anyway with impossible loan terms is not solving that problem. And contracting the refinance function of the subprime industry instead of its purchase function simply produces the kind of situation we have now, where you can be taken in but you can't be taken out.

What I think we should be most worried about is that reaction to the "Home ATM" problem--overconsumption financed by a home equity bubble--will create a long-term backlash against distressed loan take-out financing, because it's cash-out as much as because it's "subprime." Certainly a large number of the over-consumers will have to end up in BK and/or foreclosure; too many people just won't live long enough lifetimes to earn the income they've already spent. But subprime refinancing can have a real utility for the rest of us in the job loss/illness/divorce/bender category, especially when a large RE price correction makes us all upside down.

But then I am generally convinced that the only way out of a credit crunch is selective loosening. "Selective" here meaning that one selects the market segment in which risk-taking has the biggest payoff, economically, and then one takes the risk there. So instead of working ourselves up over making sure that cheap prime credit jumbo purchase loans are available, we should be worrying about whether we are making enough mid-market subprime refinances, not too few. This will continue to sound like some odd variation on homeopathy until we stop thinking about "subprime" as the cause, not the effect, of a much larger problem.