Crazy Plan to Privatize Fannie Is ... Pretty Good?

The weird thing about this new Fairholme Capital Management proposal to privatize Fannie Mae and Freddie Mac is that it's pretty good! A little tricky, but not too bad.

The weird thing about this new Fairholme Capital Management proposal to privatize Fannie Mae and Freddie Mac is that it's pretty good! (The Wall Street Journal has the proposal slides here and the accompanying letter to Ed DeMarco, the Frannies' regulator, here.) My natural inclination is to look for tricks in things like this, and there is a pretty big trick here and we should talk about it, but if you were just an honest disinterested person trying to figure out the right next steps for the U.S. mortgage market, you'd probably want to do at least some of what Fairholme suggests.

Their proposal is to split Fannie and Freddie into two chunks. One chunk is just the legacy business -- every mortgage guarantee Frannie wrote, every bond they bought, etc. etc., from the dawn of time until some cut-off date, say June 30, 2014. That would be managed in runoff, with the government getting 80 percent of the value realized in runoff and the common shareholders -- who exist, and are aggrieved -- getting the other 20 percent. 1

The other chunk is like the computers and people and procedures used in writing mortgage guarantees, which the government would sell to two new, state-chartered, not-named-Fannie-or-Freddie insurance companies owned by Fairholme and its fellow holders of Fannie and Freddie preferred stock. After the cut-off date, these newcos would write mortgage guarantees similar to what Fannie and Freddie do now -- i.e. on mostly prime, conforming, etc. mortgages -- but without an implicit government guarantee.

The proposal, at least, involves them being capitalized to a AAA sort of level 2 with private money, and underwriting based on rational private-sector business judgment rather than whatever conflicting goals (affordable housing, government budget requirements, mindless bureaucracy) the public sector might use to underwrite risk and charge premiums. Anyone else would be free to compete with them in this business, which Fairholme et al. think will be profitable. 3 And the newcos would adapt to any level of government subsidy that Congress ultimately wants: They'll be the first-loss guarantee if the government wants to reinsure mortgages, or they'll take 100 percent of the risk if the government wants to get out entirely.

I don't know, it is not an awful idea in the abstract. You can worry a lot about the details, and because this is the mortgage market the details matter a whole ton. How much will the newcos charge for their guarantees, and how much more yield will investors demand for newco-guaranteed instead of Frannie-guaranteed mortgages? (Presumably newco's guarantees will be worth less than the government's, but will they also be more expensive? Maybe not that much!) What will this do to mortgage liquidity and the to-be-announced market? Will the newcos in fact be adequately capitalized, and will their state regulators be any good at finding out? 4

So a lot to think about, but you've got to think about the same things in any scenario where private capital takes over the credit risk in the U.S. mortgage market. Now of course, you might think about those things and say ehh, never mind, let's just leave that credit risk with the government, but Fairholme very cleverly begins its presentation with two pages of quotes from various luminaries, starting with Barack Obama and ending with Ed DeMarco, all saying "we need private capital in the mortgage market." I suppose the theory is that DeMarco might forget that he said, that but the presentation will remind him.

But of course there is a trick. The big trick is that the plan calls for the new companies to be capitalized with $52 billion consisting of "$17.3 billion in new cash raised in a rights offering" from Fairholme & friends plus, what is this, "$34.6 billion of restricted capital from conversion of existing preferred stock." That existing preferred stock is the stock that Fairholme & friends currently own, and it trades at, I don't know, not that much, call it 30 to 40 cents on the dollar. 5

That means that Fairholme & friends will stump up $17.3 billion in cash and $10 or $12 or $14 billion worth of preferred stock, and will be handed in exchange a brand new, pristine, no-debt company with $52 billion in assets, 6 for an immediate first-day gain of $20 billion or so. Which is pretty good!

Actually, this understates the issue a bit; that preferred stock might be trading for 30 to 40 cents on the dollar, but if you ask a lot of people they'll tell you it's worth zero cents on the dollar, and some of them are in the sorts of positions in the U.S. government that can make that a reality. If you believe that, then this plan is more like a $34.6 billion one-day windfall to those preferred holders.

That windfall -- look, I mean, you can think your thoughts about it. There seem to be only two schools of thought. One is, "who are these dirtbags, Fannie and Freddie exist -- and are currently making a profit -- only because of their essentially explicit government guarantee, so of course all the profits should go to the government." The other is, "these poor shareholders, the government stole their shares and never compensated them, the least it can do is give them credit for their pref at 100 cents on the dollar in this convoluted scheme." I mean either of those is probably a defensible view; obviously, Fairholme takes the latter view, and almost as obviously the government takes the former. I have my own view but why would you care about it, honestly?

I was just going to say, that windfall affects this proposal in weird ways. In particular the proposal is not, on its face, "give us the franchise to run mortgage insurance in America." It's "open mortgage insurance up to fair competition, and sell us some assets to start competing." But if you capitalize New Fannie and New Freddie with a combined $52 billion, they'll be pretty pretty big -- probably bigger than all their plausible competitors combined. 7 And that capital will be pretty cheap, in that every dollar that investors put in will get them as much as three dollars in newco capital, depending how you count.

The newcos' competitors will not have those advantages. (Nor will they, I guess, have the Frannies' buildings and people and regulatory relationships, which should probably count for something.) Every dollar they want to put up to start a new insurance business will cost them a dollar. 8 That's not necessarily a recipe for healthy competition. And the opposite of healthy competition is something like "duopoly of too-big-to-fail mortgage guarantors with implicit government backing, whose profits go to private investors but whose losses are socialized ad hoc in a crisis." 9 Which sounds a little too familiar.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.
  1. Really 79.9 and 20.1 percent for dumb reasons. Also I mean this is its own thing to freak out about -- if you're Treasury you're really expecting 100 percent at this point -- but let's let it go. Also I'm just taking it on faith that that legacy portfolio has a value greater than zero; that doesn't strike me as mathematically certain and you could imagine all the guaranteed loans defaulting and Treasury needing to put more money into Legacy Frannie. But let's let that go too; Fairholme says "Wind down is reliably profitable" so sure.

  2. The phrase -- from slide 26 -- is "Fort Knox balance sheet," which I guess is even more secure than a "fortress balance sheet," plus it's got gold inside.

  3. See slide 25 of their deck, which sounds perfectly plausible. Obviously there is a big risk in the business of writing very correlated insurance like this: It's profitable until it isn't, so it makes a lot of sense to just collect premiums for mispriced risky business for as long as you can, then go bankrupt and retire wealthy and only moderately ashamed. Equally obviously state insurance regulation is supposed to prevent this sort of thing with prudential and capital requirements, though its efficiency in doing so is beyond my powers to assess.

  4. In that regard the proposal is $52 billion in initial capitalization with the goal of writing $1 trillion of new business initially, for capital equal to around 5 percent of their risk in force. (Though I guess they'd raise more capital from the guarantee fees.) That seems, I dunno, pretty high compared with some of the muni guys anyway. Some data points include that (1) the worst vintage of conforming loans seems to have had about a 13.7 percent default rate (slide 19 here), though losses on default are presumably not 100 percent and those defaults didn't all occur at once, and (2) at least Barclays expects conventional mortgage losses to run about 4 basis points a year, so 5 percent capital would cover a century or so of losses.

  5. I see Fannie 8.25 percent Series S at $8.50 for $25 par, the 8.25 percent Series T at $9.75, the 7.625 percent Series R at $7.50. Fairholme owns a lot of the S at least. Freddie 8.375s seem to be at $8.79, again with Fairholme the top holder.

  6. What assets? Some know-how and stuff, but mostly you'd expect it to be the sorts of assets that insurance companies invest in, bonds and stuff. The presentation (slide 13) says only "Corresponding assets transferred to NewCo as start-up capital" in exchange for the $34.6 billion of preferred; the rights offering would raise $17.3 billion in cash, presumably to go buy bonds or whatever.

  7. Compare some other monoline-y type things: Assured Guaranty has $5 billion of equity, Radian has $736 million, MGIC has $197 million, Fidelity National has $4.7 billion. Market caps are higher, and I mean Berkshire Hathaway has a zillion dollars in equity so it depends how you count, but the point is that $52 billion is a lotta capital.

  8. Sunk costs, etc., but one feature of the Fairholme proposal is that the rolled-over capital will be locked up for five years in the newcos. So it's really basically free to the newcos; they might as well compete pretty hard.

  9. Even better is slide 18 of the proposal, which says:

    [imgviz image_id:iLqW8eJs.IKQ]

    Heeheehee well. What is a superior proposal? Like, here is my proposal:

    • Do everything that Fairholme says.
    • Except instead of giving Fairholme & friends $34.6 billion for their worthless preferred stock, give me $34.5 billion for my worthless, I mean, I have this weird Bloomberg mug on my desk, I will give you that and you will give me $34.5 billion.

    That proposal is I think pretty close to Fairholme's,9a but saves Treasury $100 million in cash, plus they get the mug. So is it superior? Well that depends on how much credit you give to those last two capitalized-check-bullet-points, or in other words, on how Treasury should value what is essentially a credit bid by Fairholme. If you count the preferred stock at par, then a superior proposal requires someone to stump up more than $52 billion in cash. If you count it at zero, then superiority requires rather less.

  10. Except hahaha I'm not going to come up with $17 billion in cash for a rights offering, they've got me there.

To contact the author on this story:
Matthew S Levine at

Before it's here, it's on the Bloomberg Terminal.