The city of Richmond, CA, is trying to use eminent domain to seize and refinance underwater mortgages, and yesterday the trustees of some mortgage bonds holding those mortgages sued to stop them. On its surface the lawsuit is about constitutional law but really it's about option valuation. To stylize it a bit, the plan is for Richmond to:
- use eminent domain to "seize" a performing mortgage with a balance of, say, $300,000, but on a house worth $200,000;
- pay the owner of the mortgage $200,000 in compensation;
- issue a new mortgage to the homeowner for $200,000; and
- sell the new mortgage to an investor for $200,000, funding the costs to pay off the old mortgage lender.
The main stylization there is that actually the compensation will be more like $160,000, not $200,000, to account for the expected costs of foreclosure, and to provide a profit margin to the new investors (and fees to Richmond's financial advisor).1 That is important for the lawsuit but not that important for our purposes so let's ignore it.
The question is then: is a $300,000 performing mortgage on a $200,000 house worth $300,000, or $200,000, or something else? If the answer is $300,000, give or take, then the Richmond plan can't work: it's unconstitutional to take a mortgage without "just compensation," and if you have to pay the mortgage owners the full principal amount in compensation then you can't use this plan to write down any mortgages. If the answer is $200,000, give or take, then the plan is fine: it achieves its goal of solving a collective action problem, writing down the mortgages to their actual value, and reducing homeowners' debt without actually costing mortgage owners any money.2
So what's the answer? It seems like an empirical question susceptible to an empirical answer, insofar as mortgages are sort of tradeable. But only sort of, so in fact there are dueling claims about the empirical answer, with Richmond basically saying "it's less than $200,000" and the mortgage trustees saying "it's much closer to $300,000ish."3
But it's also a theoretical valuation question insofar as a mortgage is, again stylizing slightly, a (1) non-recourse4 (2) secured loan, which is to say a contract that gives its owner the right to receive:
- a stream of principal and interest payments worth the principal amount of the loan, or
- the value of the house securing the loan,
- whichever is less.
In other words the mortgage contains an embedded option in which the homeowner can put the mortgage back to the lender at the price of the house. So on fairly simple arithmetic the mortgage's value should be capped at the value of the house: if the mortgage were worth more than the house, the homeowner would get rid of both of them and make a risk-free profit.
Now this is obviously wrong as you can tell from the dispute over the empirical answer; if the value of the mortgage were really capped at $200,000 then why would it trade for more? Or you can look at other embedded options in mortgages and see how they're valued; the answer is not "they're valued as options." A 6% BBB-rated non-callable corporate bond with five years to maturity should be worth about 115 today; that same bond if callable at par should be worth par. And so: it will normally trade at around par.5 Freely prepayable mortgage-backed securities with above-market coupons trade above par all the time6; mortgage traders talk about "prepayment speeds" - basically the percentage of people who will rationally exercise their options - rather than just saying "well, they can prepay at par, so why would I pay above par?" The percentage of people who do not rationally exercise their options is large.
Or you can tell even more simply by the fact that these are performing loans, meaning that their borrowers keep making their monthly payments. If the homeowners were going to put back the mortgage to the bank because its principal amount was higher than the value of the house, they'd have already done so. You can see why Richmond is pioneering this plan and not, say, Hyde Park.
But what does it mean that it's wrong? In a Black-Scholes-y market where that simplistic option valuation was right, you wouldn't need Richmond's plan; homeowners could do it themselves:
- default on your $300k mortgage
- bank forecloses on your $200k house
- borrow $200k
- buy your house from the bank out of foreclosure for $200k
- Voilà, write-down!
I hope I do not need to explain the ways in which our world differs from that Black-Scholes world.7 But "the terms of these underwater mortgages" are not among them: no feature in those existing mortgage contracts would prevent you from following exactly that plan. Features of the world are what's holding you back.
Which means I guess that I think this valuation dispute is a genuinely interesting issue. The mortgage investors here say "our mortgages are worth $300,000, because that's the number on the sticker and because that's where they trade." Richmond says "no, your mortgages are worth $200,000, because the homeowner has an option to put them to you for $200,000 of value and should, rationally, exercise that option." The investors say "yeah but that option is bullshit everyone knows that." Richmond says "we have found a way to make it not bullshit, does that change your valuation?"
Does it? What do the mortgage investors deserve to be compensated for: their rights under the mortgage contract, which are basically "the house or the mortgage, whichever is less," or their expectations based on the world as it currently exists, where performing underwater mortgages trade above the collateral value because everyone knows that those borrowers aren't just going to walk away from their mortgage? That strikes me as an intriguing question of constitutional law,8 and a delightful question of option valuation. If you valued the non-recourse walk-away option in these mortgages as a genuine Black-Scholes option, you were a fool: it was never that. But if you valued it based on historical likelihood of homeowners actually walking away from these mortgages, Richmond may yet make you look like a fool. I wouldn't count on it, but just the fact that they raised the question should make people nervous.9
Wells Fargo Bank, NA v. City of Richmond [Ropes & Gray]
Pimco, BlackRock Seek to Bar California Mortgage Seizures [Bloomberg]
Battle of Richmond: Is Eminent Domain Plan a Win-Win or Lose-Win? [WSJ Law Blog]
Lawsuit Challenges Richmond, CA’s Plan to Use Eminent Domain to Condemn Mortgages [VC]
1.The complaint may be biased but here's how it describes the mechanics, based on the proponents' own materials:
44. Under the Program, once a loan is selected, Richmond will attempt to seize the loan through eminent domain powers for approximately 80% or less of the underlying home value. Because these are underwater loans (i.e., those with home values below the outstanding principal balance of the mortgage), a payment of 80% of the home value is far lower than the unpaid principal balance of the loan. ...
47. On information and belief, once Defendants secure the loan at a steeply discounted price, they then intend to replace it with a new loan to be sold into a FHA securitized pool in an amount equal to approximately 95% of the underlying home value. Richmond, MRP, and their investors and partners thus would instantly profit by sharing in the spread between the 80% seizure price and the 95% refinancing price.
48. In an example provided by MRP, an underwater loan on a home worth $200,000 would be seized by eminent domain for $160,000 (or 80% of the home’s value), and then refinanced into a new FHA loan for $190,000 (or 95% of the home’s value). The $30,000 spread between the seizure price and the refinancing price would be divided (after expenses) among Richmond, MRP, and MRP’s investors.
MRP is Mortgage Resolution Partners, the private company pushing this plan.
2.Because, that is, their loan was only worth $200k anyway so the write-down is just an accounting matter, not an economic one. Robert Hockett, the Cornell law professor who more or less invented this, has a paper about it called "Paying Paul and Robbing No One: An Eminent Domain Solution for Underwater Mortgage Debt that Can Benefit Literally Everyone," which should give you a sense of the win-win rhetoric. A sample of the logic:
Indeed for much underwater mortgage debt, principal write-downs actually maximize value. We find evidence for this in the rates at which portfolio mortgage loan holders, as distinguished crucially in ways we’ll soon see from securitization trusts, write down debt. ... Unfortunately, a host of classic collective action problems stand in the way of the win-win solution in this case [of securitized mortgages]. ...
Most decisive among the additional challenges facing securitized loans is the fact that so many of the pooling and servicing agreements (PSAs) pursuant to which most loans are securitized, drafted as they were during the bubble years when few appear to have appreciated the prospect of economy-wide housing price crash and all rushed to push or to purchase newfangled product, require unanimity or supermajority voting among mortgage-backed securities (MBS) holders before loans can be modified within or sold out of trusts. And today’s fragmented owners of MBS cannot even find one another, much less negotiate with borrowers or would-be buyers and then reach agreement on what’s best for all. The same agreements, crucially, likewise prohibit or otherwise prevent even trustees and loan servicers, who are duty-bound to act on behalf of the MBS holders and hence could in theory address their collective action problems, from modifying or selling bad loans in sufficient number. ...
The charge that eminent domain proceedings must always represent a ‘zero sum game’ is simply false, reflective of misunderstanding the nature of market failure at best, or of outright mendacity at worst. Literally everyone can win under the eminent domain solution, if only the actual stakeholders think for themselves, examine the numbers, and don’t squander resources – as the securitization industry has (bluffingly) threatened to do since the summer – on pointless litigation, lobbying, or strong-arming efforts.
Key to understanding why the eminent domain solution can benefit literally everyone is the notion, already appealed to above, of a needlessly wasteful collective action problem. The solutions to such market failures, by definition, recoup needless losses – recoupments which then can be distributed Solomonically over all stakeholders.
That is, if these stupid bondholders would just stop suing, these write-downs would be great for them.
3.From the complaint:
50. MRP has claimed that the loans the Program seeks to target do actually trade, and one can pull the trading histories and determine that, for example, a performing $300,000 loan on a $200,000 house is worth about 80% of the value of the house. But this is inaccurate. There is no trading market for performing underwater loans held by private-label RMBS trusts. Indeed, the Trusts are structured under federal tax laws as “real estate mortgage investment conduits,” or “REMICs,” and, as such, are prohibited from selling performing loans. Regardless, the value of such performing loans to the RMBS Trusts is clearly not the current value of the underlying home.
51. Additionally, the entire Program is premised on undercompensating the owners of the loans. It could not function in any other way, because the Program is profitable for its participants only because the loans are seized for heavily discounted prices and are then refinanced with a new loan purportedly worth more than the amount for which the homeowner’s existing loan was seized. The new loan can be sold to a new securitized pool, creating a profit spread. So compensation for the seized loans under the Program must, ipso facto, be at an artificially deflated value – and hardly the “just” compensation that is constitutionally required.
52. In fact, not only is the 80% price not a fair value for a performing underwater loan with a low risk of default owned by an RMBS Trust, it would not even be a fair price for Richmond loans not part of RMBS Trusts that are in default or foreclosure. On information and belief, defaulted residential mortgage loans available for sale have recently traded at far above 80% of the underlying home value.
4.Casual Googling suggests that California purchase-money mortgages are non-recourse, though I suppose refi mortgages might be part of the deal here and I guess those are (can be?) recourse.
6.For giggles, here is Fannie Mae's 6%er:
If you have a 6% mortgage, you should probably refinance it!
7.Just in case, though: it's hard to get a mortgage when you just defaulted on another mortgage. For starters. Credit scores and whatnot. Also foreclosure takes a while and you need to, like, live somewhere and stuff.
8.It's not the only question in the case of course, though it ties to the other big ones. The plaintiffs, and a lot of other people, make a lot of public-policy arguments to the effect of "actually doing this would make mortgages more expensive." Which I'm sure is correct! It amounts to making the non-recourse option found in mortgages actually a real option, which means making it more valuable, which means that banks, not being charities, would charge you for it. There are various policy implications to that; the plaintiffs' view is basically "that option should be valueless because the people who'd get value out of it are deadbeats while the people paying for it ex ante are everyone." I am not unsympathetic to that but it's just like a choice man.
9.My gut is that the Richmond plan is not going to work - that a court or the federal government or someone will squash it. That leaves you wondering: is there a way to do the self-help thing that I outlined? Like, Mortgage Resolution Partners could just finance that directly - sign a contract with a homeowner that is like "Homeowner will default on his mortgage, hang out until the bank forecloses, and then buy the home back out of foreclosure, while MRP will agree to lend Homeowner the purchase price out of foreclosure and take a new mortgage on the house." There's a lot of administrative messiness there but my guess is that the main obstacle is what this would do to the homeowner's credit score. Are there others? If not, wouldn't this be fun to try?