Who won the Bank of America / Fannie feud? I want the answer to be "all of us," but I guess it isn't. Unlike Bank of America's multi-front battle of deviousness with MBIA, which has spawned some genuine entertainment, BofA's battle with Fannie has been conducted almost entirely in the boring trenches of actually flinging mortgages at each other. The story so far: back in the bad old days of "January 1, 2000 to December 31, 2008," BofA/Countrywide sold approximately $1.4 trillion of mortgages to Fannie; in recent years Fannie has been figuring out that lots of those mortgages were badly underwritten, came with false reps and warranties, etc., and so it can demand that BofA repurchase them at par. It's tried to do so on around $11bn of loans, and BofA's reaction has been along the lines of (1) no and (2) "we're so mad we're going to stop selling you more mortgages," which somewhat surprisingly to the casual observer actually seems to have been interpreted by both sides as a threat rather than a reward.
But today Fannie and BofA announced a settlement that would resolve all of those claims, as well as almost all future claims on the $297 billion of unpaid principal that remains on those $1.4 trillion of '00-to-'08 mortgages.
The math here is a little funny. Let's try to parse it:
- Fannie had somewhere between $10.8bn and $11.2bn1 of BofA loans that it wanted BofA to repurchase.
- BofA is giving Fannie $10.3bn of cash in this settlement.2
- So Fannie got 92 to 95 cents on the dollar, depending on who's counting.
- But not really because only $6.75bn of that cash was actually paid to repurchase ($6.51bn principal amount of) mortgages. The other $3.55bn of it was just free cash, and Fannie doesn't have to give up the underlying (crappy!) mortgages to get it.
- The loans that Fannie is selling back to BofA are probably worth something; from Fannie's 8-K you can see that a majority of them are either (1) current (albeit usually after loan modifications) or (2) <80% LTV, which I feel like for a Countrywide-originated mortgage either of those things should get a gold star. Let's say that on average those loans are worth 28 cents on the dollar of unpaid principal - based on guesswork (but conservative) math here.3
- The loans that Fannie is not selling back to BofA are also worth something. My guess is that they're worth more than the loans that Fannie is putting back - since presumably the settlement would make BofA take the worst loans off Fannie's hands - but the public statements don't say anything one way or the other. Even if they're worth less - even if on average they're worth only half of what the actually-put-back loans are worth - that's still about $600 million worth of loans that Fannie is (1) getting paid for but (2) keeping.
- If you add up Fannie's sources of value here - cash for putbacks, free cash, and loans that it didn't have to surrender to get the free cash - then Fannie is getting at least $10.9bn of value to resolve $10.8bn of putback claims. (Feel free to fiddle with my math but, again, I think it's pretty conservative.4)
So that's a pretty big win for Fannie, no? Normally when you bring $100 of claims and your adversary says "naaah the right number is zero" you end up settling for, y'know, $20 or $50 or $70 or whatever. Some number between zero and 100. Not, as here, $101 to $111 or so. And, indeed, this seems to have been a surprisingly big win for Fannie; despite having a nebulous but probably gazillion-dollar-ish amount reserved against these putback claims already, today's "actions are expected to be covered by existing reserves and an additional $2.5 billion (pretax) in representations and warranties provision recorded in the fourth quarter of 2012."5 So this is at least 25% bigger than BofA was "prepared" for, in the loose accounting-y sense of "prepared" implied by rep & warranty reserves.
So what happened? One tempting explanation is that the threat of not being able to guarantee Bank of America mortgages was not all that terrifying to Fannie - whereas the reduced ability to sell mortgages into the lucrative GSE pipeline turned out to be not that great for BofA. Fannie had something BofA wants, while BofA doesn't have much that Fannie wants - other than, y'know, this money, which it's now given to Fannie.
But the main answer has to be that BofA is at a place in its life where it really needs to pay for certainty."6 This deal doesn't just resolve $11bn in putback claims for a very full-valued $11bn; it also resolves another $286bn of potential putback claims on the mortgages that BofA/Countrywide sold to Fannie during the Bad Old Times and that Fannie hasn't yet found a problem with. Now note that these mortgages are now somewhere between four and twelve years old; if Fannie hasn't yet found a problem with them, it's a good bet that they won't give rise in the future to lots of meritorious rep & warranty claims - that the expected value of putback claims there hovers relatively close to zero.7 That's a bet that you or I might be willing to take, on pure expected value grounds. But BofA isn't acting on expected value grounds these days: its priorities seem to be, as Moynihan put it in the press release, "resolving our remaining legacy mortgage issues, further streamlining and simplifying the company and reducing expenses over time." Fannie had a trade to offer BofA that could tick all of those boxes - so it's no surprise that BofA overpaid for that trade.
Bank of America Reaches $10 Billion Settlement With Fannie Mae [WSJ]
Bank of America Announces Settlement with Fannie Mae to Resolve Agency Mortgage Repurchase Claims on Loans Originated and Sold Directly to Fannie Mae Through December 31, 2008 [BAC]
Fannie Mae 8-K [EDGAR]
2. Plus other stuff! BofA is selling servicing rights to some loans to third parties, and is making an unspecified "cash payment to Fannie Mae to settle substantially all of Fannie Mae's outstanding and future claims for compensatory fees arising out of past foreclosure delays." I assume that these things are conceptually distinct though of course in negotiation it may have worked out that BofA is overpaying for foreclosure delays in exchange for underpaying for reps & warranties, or vice versa, or other.
3. More in the spreadsheet; I assume a mortgage is worth zero, plus 25% of principal for being current, plus 25% for being <100% LTV, plus an extra 25% for being <80% LTV. That's pretty dumb but whatever, you can play with it.
4. For instance: assume each mortgage is worth 20% of principal + 20% for being current + 20% for <100%LTV + another 10% for <80%LTV, for an average of 40 cents on the dollar among the loans being repurchased. And assume the mortgages not being put back are worth about the same as the mortgages being put back, as a percentage of principal. Then you get value to Fannie of just over $12 billion, or 111% of the principal amount of the disputed loans.
5. KEEN OBSERVERS will note that the fourth quarter of 2012 ended approximately a week ago. KEENER ACCOUNTANTS THAN I are welcome to explain this. One assumes something like "well, a week ago, they kind of knew this settlement would be bigger than they'd already reserved for." Weird though.
6. Relatedly, the FT notes that "the size of the deal is likely to wipe out BofA’s earnings for the quarter, casting doubt over whether the bank will be permitted to return a significant amount of capital to shareholders," but if you were a regulator would you be all that thrilled about a big capital return program with tens-to-hundreds-of-billions of potential repurchase liability still hanging over BofA? Especially where BofA is potentially under-reserved? Especially where BofA's antagonist is basically a government agency? This doesn't strike me as a step backward in BofA's efforts to return capital, exactly.
7. But you never know; there'll probably always be more embarrassing emails to dig up.