Sunday, December 23, 2007

Balloons for Ubernerds

(Calculated Risk) A question arose the other day about balloon mortgages. Balloons are one of those things that have been around, in one incarnation or another, for at least a hundred years in this country, but they were not, by and large, a popular form of (first) mortgage structuring in the big boom of the 2002-2006 period. So we’ve really never talked much about them, our attention having been taken up by the various forms of ARMs. However, you don’t get promoted to Mortgage UberNerd Eagle Scout (Marbles and Piracies Division) until you know what a balloon is. Talking about balloons is also a good excuse to think about a few underlying issues with how mortgages are structured and, crucially, how they get restructured. That last thing is sort of a big deal these days, so it seems apropos.

First, a technical definition: what is a “balloon”? It is a loan that matures before it amortizes. In case that didn’t clear it up for you, let’s start with amortization.

Amortization, in the context of loans, is a mathematical formula that generates a substantially level payment sufficient to reduce the principal balance of the loan to zero with the last installment. Strictly speaking, we call this “full amortization.” The important thing about amortization is that it is more than just some structure that involves periodic repayment of principal.

You could design a loan that requires the borrower to pay the full interest due on each installment, plus some arbitrary amount of principal, say $100. Given a big enough loan, it could take centuries to pay that off, and most people want to retire before that. Certainly most lenders don’t want a loan on the books that long; there aren’t that many houses that have an “economic life” long enough to provide security for such loans throughout their terms. If you simply declared an arbitrary “maturity date” (legally, for mortgages, this is the date that all remaining principal outstanding under the note is due and payable), say 30 years from the first payment date, you could have a situation in which a large chunk of principal is still outstanding on the maturity date. A $200,000 loan, for instance, with a $100 per month principal payment for 30 years would have a balance due at maturity of $164,000. We would call that a “balloon” payment, the metaphor being that the final installment on this loan swells up like a big balloon in comparison to the prior installments.

However, such a loan, with an arbitrary principal payment, is not much like the modern balloon loan, although in the very distant past such loans were made. In the very distant past (just before and during the Great Depression), most mortgage loans were balloons of this type or were interest-only loans with very short terms (a year was common). Why were IOs and balloons the only loans available back in the day? You have to remember that this was long before the invention of the ARM, or a mortgage loan that can “reprice” its interest rate to market without having to actually pay off an old loan and start a new one. Banks and thrifts that funded mortgage loans with deposits had the same problems back in the old days with interest rate risk that they have now: if you make a very long-term loan at a fixed rate of interest, you might find at some point before that loan pays off that prevailing interest rates paid on deposits have increased on you. You therefore have a loan out there earning you, say, 5%, and depositors expecting you to pay them 6%. That is called a “negative interest margin” or “the road to bank failure,” depending slightly on context. Even if you are not experiencing pressure on deposit rates, back in the thrift days of loans funded exclusively with deposits, you might find (It’s a Wonderful Life) that you just have all your assets tied up in slowly-repaying loans when your depositors want cash from you. That’s the classic “liquidity” problem. You will want to bear it in mind in any discussion of whether “the end of securitization” might come, or might be a bad or a good thing.

Offering only short-term loans that had to be renegotiated (refinanced) every year was a way to manage the asset-liability problem and keep things sufficiently liquid. It was, therefore, like the modern ARM, a form of risk-shifting: the interest rate and liquidity risk was taken off the banker and placed onto the borrower. It (more or less) kept the banks solvent and liquid, but it rather famously created a nasty trail of foreclosures when suddenly borrowers just couldn’t find a lender to renegotiate those maturing loans, and didn’t have the cash to pay them off. (Sound familiar? Yep.)

Without straying too far into the history of the mortgage in the U.S., suffice it to say that the big innovation coming out of the New Deal legislation that formed FHA and, subsequently, the old Federal National Mortgage Association (which turned into Fannie Mae, the GSE, rather later) was the fully-amortizing long term loan. At first the new thing was the 10-year loan, then the 20-year loan, then the 30-year loan. (Several decades later, you notice, we’re on the 40-year and even 50-year loan.) So this thing that Americans think of as the “traditional” 30-year fixed rate loan isn’t, in the big picture, all that old.

It required the invention of FHA and FNMA because somebody had to take that long-term fixed interest rate risk and tied-up capital off the banks and thrifts, which were still (the thrifts, at least) mainly deposit-funded. (Note to those who keep insisting that the New Deal-era invention of the government mortgage agencies started the whole problem: this was an attempt to put risk where it could be withstood. We had just gone through a Great Depression which had more or less conclusively proven that individual middle-class citizens have a hard time taking on risks that even well-capitalized financial institutions cannot absorb. The real “road to hell,” in my view, crystallized in 2004 when Alan Greenspan told consumers to go ahead and take ARMs—take interest rate risk—at a time when fixed rates were much more to their benefit. We put the risk exactly on the weakest party, the small-time homeowner, which some of us consider voodoo economics. Ahem.)

This fancy new loan product that FHA and FNMA started buying up had a cool new feature: amortization. That is, full amortization. Not only was the interest rate fixed for the life of the loan, and not only was the life of the loan much longer than in mom and dad’s generation, and not only was the monthly payment stable, but the monthly payment was enough to bring the balance of the loan to zero at the end of the term (whereat one could have a “mortgage-burning party”). We take amortization of loans so much for granted these days that it’s hard to grasp what an “innovation” this was in residential mortgage lending.

It was not, however, a free lunch. The old balloon and IO mortgages actually had, adjusting for home prices and interest rates, a lower monthly payment than the newfangled amortized loans, on the whole. That is, exactly, what pushed the “standard” term from 10 years to 20 years to 30 years, and lately to even longer. The math of amortization is pretty simple: the more scheduled monthly installments, the smaller the installment payment. From the lender’s perspective, those longer terms, while making it possible for more and more people to handle mortgage payments (especially those returning GIs after WWII who were busy creating the Boomers), also made the interest rate risk of those loans harder and harder to bear.

One solution to that problem was to drag out the old balloon concept, and apply it to an amortizing loan. This is the “modern balloon” that we know these days. You calculate a fully-amortizing payment for a loan over a 30-year term—so the borrower gets the benefit of the lower payment based on the long loan term—but you arbitrarily make the maturity date of the loan occur before—sometimes very much before—the point at which the loan amortizes down to zero. For a while, in the 80s and 90s, the popular balloon product was the seven-year: the loan payment reflected a 30-year amortization, but the maturity date was seven years after the first payment date. Because the rate of principal payment in this method is the same as for a plain old 30-year fixed loan, the borrowers who got a seven-year balloon were not racking up equity any more slowly than their neighbors whose loans didn’t balloon. In fact, they probably racked up equity slightly faster, since the balloons were offered with a slightly discounted interest rate (to encourage people to take them). All other things being equal, the lower the interest rate on a loan, the “faster” it pays down principal, and so the lower the total interest charges over the term of the loan.

Detour for beginners: I know some people get confused over the concept of amortization “speed.” A 30-year FRM at 6.00% and a 30-year FRM at 7.00% will both pay off completely in 30 years. “Speed” here is not a matter of time to maturity. It is a matter of the slope of the balance curve. A $100,000 loan will have a payment of $599.55 at 6.00% ($500 in accrued interest for the first month and $99.55 in principal), and a payment of $665.30 at 7.00% ($583.33 in accrued interest and only $81.97 in principal). The amount of the payment does not change over time, but since the balance is declining each time some principal is paid, the accrued interest is less each month and therefore the portion of the payment devoted to principal paydown is greater each month. After seven years of payments, the 6.00% loan has a balance left of $89.639.39 and the 7.00% loan has a balance of $91, 147.41. The formula for amortization always assures that the payment is sufficient to cover the full interest due (plus some principal); so when there is more interest due (because the rate is higher), more of the monthly payment is devoted to interest in the early years of the loan. We therefore say that the higher-rate loan amortizes more “slowly” in its early years.

The actual rate discount on seven-year balloons was less than 1.00% in most cases. Of course the difficulty in calculating the appropriate discount was (and is) complicated by the availability and attractiveness of refinances generally. We all know that a plain old 30-fixed rate loan doesn’t have to stay on the books for the full 30 years, as long as there is someone willing to make a refinance loan. In periods in which prevailing mortgage rates are declining, the average life of a 30-year mortgage loan easily can converge on seven years (or actually any balloon term you can think of). In such an environment there is no point (to the lender) in discounting a rate in exchange for the borrower accepting a balloon term: the lender concludes that any loan it writes is likely to pay in full in less than seven years or so, balloon term or no balloon term. (Right when the lender might like those loans to stick around, since they yield more than a new loan would.)

Another way to say this is that lenders don’t need to write balloon terms onto loans when they expect fast prepayments of loans generally. That is actually one reason why if you just started paying attention to mortgage matters in the last few years, you many not have heard much if anything about balloons. They go out of fashion in low-rate fast-prepayment periods, and the one thing you can say with 100% confidence of the period beginning in 2003-2004 is that we had low rates and fast prepays. As a matter of fact, some of us old farts remember when the legal maturity date on a balloon loan was referred to as its “call date.” A balloon has an imbedded call in the sense that the lender is calling it due before it fully amortizes. In the absence of prepayment penalties (which was mostly the case in the mortgage world until recently), every mortgage has an embedded “put”: the right of the borrower to prepay in full by refinancing. The point is that lenders don’t need to protect themselves with calls in periods of time when borrowers are busy exercising their puts. The prepayment penalty (which is conceptually kind of the opposite of a balloon, in the sense that it tries to keep a loan on the books longer while the balloon term functions to force it off the books sooner) will arise in a market environment, theory predicts, exactly when balloons will fade away.

Probably the most important development in mortgage history that killed the balloon as a popular option was the ARM. (It was actually only in the 1980s that all financial institutions, including the thrifts, were finally allowed by law to offer ARMs.) The old balloon—the pre-Depression loan you had to refinance every year—existed in an environment in which there were no variable rate loans offered to consumers for mortgage loans. In essence, the balloon structure was what allowed the interest rate on the loan to “float.” The trouble, then and now, is that one can never guarantee that there will be a lender ready to make a new loan in the future when the balloon matures. If no one will make you a refinance loan and you can’t pay the thing off, you lose your house.

The modern ARM was not invented in order to shoe-horn consumers into artificially low payments. Nor was it invented to give borrowers the “opportunity” to benefit from falling rates in the future. That’s all marketing materials we developed after the fact to keep loan officers busy; sadly, there are too many loan officers and brokers who haven’t been in the biz very long or have never studied its history, who are perfectly willing and able to confuse marketing with the real straight dope. The modern ARM was invented to avoid the major problem with balloons, the risk that there is no refi money available at maturity, while maintaining the major benefit (to the lender): they make the borrower absorb the rate risk.

The original residential home mortgage ARM (things were a bit different in commercial lending) was a simple one-year or six-month deal. Whatever the interval, the rate adjusted at each interval from the very beginning. That’s a tough loan for a lot of folks, especially first-time homebuyers. (Yes, we knew that fact a long time ago. We “forgot” it in the latest craze.) The big mortgage product innovation in the early 90s was the “hybrid ARM.” That’s the one that is a hybrid of a fixed rate and an ARM: you get your rate fixed for three or five or seven or even ten years, then the loan turns into a 6-month or (originally) one-year ARM after that for the next 27 or 25 or 26 or 20 years. Like the balloon, these loans could (in certain market environments) be discounted from the competing 30-year fixed rate. (NB: it is extremely difficult to discount a 10/1 ARM; that is really only possible in occasional and fairly brief rate periods. That’s why you see them pop up every now and again and then quickly disappear. Most of the time, given typical prepayment patterns of 30-year FRMs, there is no way a lender can offer a better rate on a 10/1 (or even at times the 7/1) than on the FRM, so they just come off the rate sheets, like the balloons do.)

The important difference between the hybrid ARM and the balloon is that the ARM does not, legally, have to be refinanced at the end of the initial period, because the hybrid ARM still has a 30-year maturity date. The borrower may not want the loan after it starts adjusting, but the borrower is not forced to refinance it or else pay the entire balance due. In other words, the hybrid ARM had a built-in “fallback position” for the borrower. If you ignore, for the moment, the question of whether a borrower can afford the rate adjustment, compare the likely fortunes today of a borrower who took an “Alt-A” balloon five years ago, versus a borrower who took a 5/1 ARM. The ARM borrower may be pretty unhappy right now about how his payment has increased, and how he can’t find a lender offering any way to refinance it. The balloon borrower, however, is in foreclosure: he isn’t facing a higher monthly payment, he’s facing coughing up 70% or more of the original balance of the loan in cash or else, so it’s practically speaking else, since it’s quite likely nobody is writing the same kind of Alt-A loans now that they did then. In this sense, ARMs are “safer” for consumers than balloons.

Of course it didn’t take lenders long to start innovating the balloon. Heretofore I have been describing what we in our charming way call the “bullet” balloon: this loan is simply due and payable on its maturity date, or else it’s see you in foreclosure court. Fannie Mae and Freddie Mac, in particular, were getting some pressure from the industry in the 80s to buy balloon loans, but they, in their actually quite frequent more-pro-consumer-than-the-bankers mentality, hesitated over the “bullet” part. This was not stupid on their side, whatever hogwash the industry was insisting on at the time. So they came up with what Fannie called the “conditional right to refinance” and Freddie called the “reset option.” I will here use the term “reset,” hoping it doesn’t cause too much confusion with the concept of ARM resets. They’re not exactly the same thing, by a long shot.

The “reset option” balloon offers a possible, but not guaranteed, extension of the loan past its balloon date. Regardless of Fannie Mae’s use of the term “refinance” in this context, these loans weren’t actually refinanced at the balloon date; they were modified. But the modification happened under contractual terms and conditions. The standard deal, as spelled out in an addendum to the old bullet balloon note, said that the loan could be reset if the home was still owner-occupied, the borrower had made payments on time for the last 24 months, there were no subordinate liens, and the rate on the reset was not more than 5.00% (five points) higher than the old rate. If you met all those requirements, the lender would simply modify the loan so that 1) the maturity date was extended to a full 30 years after the original first payment date and 2) a new fully-amortizing payment was calculated on the outstanding balance at the balloon date over the remaining term at the new interest rate. In other words, it became a regular old fixed rate loan. This is very different from an ARM reset: not only does an ARM just reset to a new ARM (not to a permanent fixed rate), but the ARM reset is unconditional (as long as you are making payments, it doesn’t matter if you still occupy the property or you were once delinquent or anything else).

The balloon note addenda that lenders used generally specified a formula for determining what your rate would be if you exercised the reset option; it was typically 0.375-.625 (a spread) over the Fannie Mae or Freddie Mac 60-day required net yield for mandatory deliveries of 30-year fixed rate mortgage loans (a kind of index, or more properly a baseline “market” mortgage rate), rounded to the nearest eighth. In practice, that worked out to a rate that was generally somewhat higher than what you could get with a regular refi, but since it usually required you to pay only a modest reset fee ($100-$250 or thereabouts was typical), it was in those days a lot cheaper than refinancing. The limitation on how high the new rate could be compared to the old rate was a way of forcing a loan that would suffer severe payment shock in a reset to have to be processed as a full refinance, with complete new underwriting and a new appraisal and all that.

The restriction on subordinate financing was less a credit risk issue, actually, than it was a function of the legal issues of mucking around with an existing mortgage. Modifications are cheaper for everyone involved than refis, because they don’t involve paying off an old loan in full and processing the release of the old lien. You just file a new document in the land records that supercedes (some of) the terms of the old recorded mortgage (like changing its legal maturity date), but you keep your old lien priority because your old lien is still out there. That won’t work if someone has filed junior liens subsequent to your old first lien balloon loan: your modification could mean (at least in some jurisdictions) that your loan suddenly becomes a second or third lien instead of staying in first position. If a balloon borrower had one or more junior liens, he or she would have to do a full-blown refinance and either take enough cash out in the new loan to pay off those junior liens, or refinance the second lien as well with a different or the same lender, or convince the junior lienholders to execute a thing called a subordination agreement that “moves” those junior liens back into junior position when the new refinance loan is recorded (back in first place).

I bring all that major mortgage trivia up because this problem of junior liens has become quite an issue lately with “workouts.” We’ve seen some reports of borrowers who are trying to use the FHASecure program to refinance their resetting ARM first liens, but who are of course so far underwater with their second liens included that they can’t, in fact, borrow enough in the refi to pay off the second lien (that is, to roll two loans into one new one). But FHA will not let itself get to second lien position, so these borrowers are struggling to get their second lien lenders to execute subordination agreements. (Refinancing the second lien isn’t really an option, since second liens are so dangerous right now that they’re hard to get, and when you can get one the new interest rate would be even more exhorbitant than what these borrowers are currently paying, and the whole point is to keep the monthly payment affordable.) There seem to be some people who assume that this problem is occurring because second lienholders would rather foreclose, and are—gasp!—being “pressured by the government” to work things out with a subordination agreement. Nothing could be further from the truth: these second liens are so far upside down that they wouldn’t even recover costs in a foreclosure, let alone any principal. As far as I can tell, the problem with getting a subordination agreement is an operational one: either the second lien lenders are just understaffed (or have recently hired staff who never lived through a bust and hence have never seen this old-fashioned thingy called a “subordination agreement”), or the second liens were securitized and nobody knows who has to execute the subordination or where to send it or it’s piled up at the trustee or, you know, the usual chaos created by securitization of loans outstripping operational capacity.

Of course things don’t always work out that way: historically, one of the problems with the resetting balloon was that the RE market was a little better than it is today, and so a borrower who did take out a junior lien might find that in fact, the junior lienholder has no particular interest in subordinating because it could quite possibly recover its loan fully in a foreclosure. In that case, the borrower isn’t allowed the reset option, and is stuck trying to find a refi lender who will lend enough to pay off all liens. In some cases, you had (a few years ago) borrowers who had taken a conforming dollar balloon, but who had then, as the property appreciated, taken one or more junior liens that brought the total indebtedness into jumbo territory. So when they went to refinance, they went from conforming rates to jumbo rates, and from purchase rates to cash-out rates (if the junior liens aren’t sufficiently seasoned, this kind of refi would be priced as a cash-out). If there were also market movement in there, the rate shock on the new loan could really suck. If jumbo lenders just weren’t interested in making cash-out refis at the time, it would more than suck.

I hope this provides some context in which you can fully understand the disingenuousness of the subprime lenders who claimed for all these years that the 2/28 ARM was about helping those poor folks by offering a “recuperation” loan. The idea being that the borrower could rack up a clean 24-month payment history during the fixed period (as we’ve seen, that 24-month payment history is a key part of defining eligibility for a “prime” refi or reset) and then refi into a cheaper loan; if the borrower didn’t “recuperate” his payment habits, the “worst” that would happen is that his monthly payment went up. Really, to claim that is to say that the 2/28 is a better deal than a two-year bullet balloon. Well, yes. That’s true. If you don’t remember anyone offering two-year bullet balloons all over the place before the invention of the 2/28 ARM, you’re with me. Subprime lenders used to offer straight 30-year fixed rate loans, or at worst a five or seven year balloon term on a 30-year amortization. In comparison to that, the 2/28 is hardly a super dooper deal. In the real world, outside the marketing department, the point of the “2” in the 2/28 was that that was only as long as the lender was willing to deeply discount the rate far enough that the borrower could qualify at the start payment. The part about the “2” being coincidentally equal to the usual “recuperation” period is something we liked to keep telling you to convince you that we’re just here to help you.

Finally, you may also hear mention of balloons in the context of workout options. There are lots of different ways one can work out a past-due loan. One possibility is to take the amount by which the borrower is past due, add it to the balance, reamortize the loan over the remaining term, and get the borrower started on the new payment. That works only if the borrower can stand the marginal payment increase, and many borrowers just can’t. So an alternative way of handling this is to add the past-due amount to the balance, but not recast the payment. In essence, that creates a balloon payment at the end of the loan (on its original maturity date), because the monthly payment is now not enough to bring the balance to zero with the last scheduled installment (while it is enough to cover at least interest due, so it doesn’t create negative amortization). The argument for doing this is, of course, that maturity is usually a long time away and that gives the borrower plenty of time to recover financially and either start making higher voluntary payments down the road (to bring the loan back onto its original amortization schedule) or sell the home or just have enough in reserves at maturity to pay off the balance in cash.

That was a very practical solution back in the old days when people started out in amortizing loans at not-so-deeply-discounted initial rates. The problem we have now is that too many people started out in “workout” loan terms. For borrowers who were making only interest payments to start with, you’re got a lot of them who can’t handle any amortized payment on any balance. (That goes extra-especially double for Option ARM borrowers who have been making even less than full interest payments.) For these borrowers, you would have to go beyond capitalizing a past-due balance, and actually create an even bigger balloon by calculating a new payment on a ridiculously long term (while not moving the maturity date out). Some of these loans started out with a payment calculated on a 40-year term with a balloon maturity at 30 years. To work them out would require calculating that payment on 50 or 60 or 70 years, which just keeps making that final balloon bigger and bigger or else it gets to the point where it does create negative amortization (if the rate is too high and so the recast payment can’t cover all interest due). So it might work out OK if in fact the borrower is committed to long-term homeownership; it will do nothing for a borrower who tries to sell in a year or two because all that principal is still due and payable when the property is sold.

Eventually these “workouts” start to resemble the old pre-Depression ticking time bomb balloon. They might not legally force the issue every 12 months, but then again the American population is used to a whole lot more mobility—the employment scene has become structured to require it—than their 1920s forebears. As far as I’m concerned, to think that “ballooning” these loans is a workable strategy for getting a sizeable number of borrowers out of the soup rests on a pretty Pollyannaish view of how long it will take for home values to recover: I think we’ll have to “balloon” more principal than eventual return to home price appreciation can “disappear,” even if you could talk investors into accepting such a very slow rate of principal return, or avoid blowing to Kingdom Come some of these securities which rely on fully-amortizing mortgage payments to keep the cash-flow to investors working out as planned.

We have observed before now that a lot of these whacky mortgage products—IO, neg am, balloons—just don’t work out in the mass market. They “work out” insofar as they imply a savy borrower with the ability to save substantial amounts of money (and therefore make periodic voluntary principal reduction payments or one lump-sum cash payoff). The economy hasn’t put most of us in that position (stable or increasing earnings in excess of the cost of living creating piles of disposable cash to prepay mortgages with) in a long time, even when it wasn’t busy inflating home prices out of reach. The current situation with “toxic” mortgages was brought to you courtesy of an ideological fixation on “free markets” and lenders not being “paternalistic” about the mortgage choices they offered or withheld from certain borrowers, as well as a covert agenda (I’m talking to you, Dr. Greenspan) to make borrowers think they were getting a deal (rate risk) that the banks wouldn’t take if you paid them. As is usually the case, that’s all just fine and dandy until it blows up, then the free marketers scurry to the government looking for help getting out from under a pile of exploding loans. That’s hardly surprising, given that anyone who willingly put borrowers into loans like this is, objectively speaking, a sociopath. If you expected them to take their licks like grownups, you don’t understand much about the essential dynamics of sociopathology.

And, predictably, we start talking obsessively about disclosures again. Of course loan disclosures can always be improved, but there gets to a point where some loan products are just simply so complicated that there isn’t any easy, clear way to explain them fully to most people. This is nature’s way of saying you shouldn’t offer those loans to most people. I have personally been known to get rather impatient with people who start nattering on about those stupid borrowers who took loans they didn’t understand (like balloons or IOs or OAs or whatever dumb thing we’re talking about). They weren’t supposed to understand them. By that I do not necessarily mean that they involve math-like concepts that are over a lot of people’s heads, although sometimes the weirder ones do. I mean that borrowers weren’t supposed to understand themselves. We were all supposed to look in the mirror and see the upper-middle-class people on TV sitcoms or financial-planning infomercials who will always be able to make those voluntary principal payments or handle those balloon payments or—snort—“invest” the principal portion of the payment in some fantabulous risk-free deal that made mortgage financing free after taxes or whatever that long song and dance is. You can write any pristinely clear mortgage loan disclosure document you want, but if the world keeps telling you it is your patriotic duty to confuse yourself with your economic betters, it won’t do you any good. To some extent, what we need are clearer “disclosures” of our paycheck math and our “necessary” expenses. As that would not likely reinforce a nation of optimistic spenders, it isn’t on the table as a priority. So we’ll spend inordinate amounts of regulatory time and energy teaching people what a balloon loan is and stuff like that. If you close your eyes you may be convinced that it's just spring and the world is puddle-wonderful.

No comments: