Wednesday, May 21, 2008

That thing's operational! (More on the GSE's)

Yesterday's post on the GSE's garnered a lot of great comments and I wanted to answer some of them in a new post, as I know the majority of Accrued Interest's readers come to us via RSS or e-mail. I'm not going to hit them all, but I'd encourage you to read all the comments for yourself.

First, I made some overly simplistic, and partially just plain wrong, statements about Freddie Mac's debt/asset situation. Read about that in a note at the end of yesterday's post.

Another persistent question is whether the GSEs would be solvent right now if not for the implicit support from the Treasury. Its an interesting question to ponder, but impossible to answer completely. That's because we don't know what the GSE's would look like if they didn't have government support. In other words, Fannie Mae and Freddie Mac would look very different today had they been founded as private companies and/or had the Treasury cut ties with them many years ago. They'd likely be a lot smaller, and probably would have focused in higher margin products. You know, like sub-prime.

But for the hell of it, let's consider the GSEs' solvency with the only variable being the market perception of government support. So we're holding their portfolio composition, loss situation, funding strategy, overall size, etc. all constant. What would that look like?

I'd say the closest parallel are mortgage insurers, like PMI and MGIC, in terms of access to capital. MGIC recently did a preferred equity offering, but generally speaking those firms would have a hard time doing straight debt offerings right now. Fannie or Freddie would have a couple things going for them, when compared with the mortgage insurers. The mortgage insurers' primary exposures are where the borrower didn't put 20% down. A portfolio like that is clearly riskier than Fannie Mae or Freddie Mac's. Freddie Mac's CLTV is 67%. Now there might be some silent seconds that they aren't counting, but still, its assuredly better than a straight mortgage insurer.

We know that the GSEs have recent done preferred offerings to raise capital. Would they have been able to complete those transactions without government support. Maybe, but surely not at the levels they got. See the terms of MGIC's convertible offering from April 1.

So I'd guess that they would be solvent, but it would be close. Damn close.

Would a government bailout be just too expensive, even for the Treasury? I don't think so. I think the Treasury could just directly guarantee their debt, (see Chrysler) which would have limited direct costs to the Treasury. Or they could simply have FHA buy a chunk of bad loans from the GSEs. There are several ways to work a bailout of the GSEs. I believe very strongly that tax payers are married to the GSEs, for better or worse.

Did Freddie Mac move their ABS portfolio into Level 3 because they didn't like the bid indications they were using for valuation? The company says no. Here is the quote from their earnings conference call:

Q: (From Paul Miller of FBR) ... There is a headline out there, talking about Freddie Mac Level 3 assets of $157 billion and I don't see that in any of your releases. I was just wondering is that true... and is that related at all to the markups of the trading securities...?

A: (Buddy Piszel, CFO) No, it is not Paul. We made a determination in the first quarter, that given how widely the pricing we were getting on the ABS portfolio, that it no longer made sense to leave that in Level 2.... We were still using the mean price that we were getting from the pricing services and the dealers. So we are not using a model price.... It has nothing to do with the trading portfolio.

So if you believe what he's saying, that means that the actual valuation would be the same either way. By moving to Level 3, they are saying they no longer believe the valuations represent "observable inputs."

What about mortgage insurers? Freddie has as presentation on this from March. I think the way to think about GSEs and MI is similar to a municipal bond portfolio and the monoline insurers. Losses will be a function of both the MI going down and the actual borrower going down. Plus the GSE would have a claim on the MI in run-off. On the other hand, MI exposure is substantial and should any of them go bankrupt (a strong possibility) it absolutely could result in considerably higher losses at the GSEs.

Finally, what's the "end game" here? Here is what I think we know. We know that new business written by the GSEs can be profitable, if loss rates on 2008 vintage loans are even close to historic norms. We also know that the real cash losses on the GSE portfolios are just beginning. Freddie Mac predicts their credit loss rate will double by 2009. The good news is that they've reserved for that. The bad news is that there are a hell of a lot of variables to those loss numbers.

So whether or not the GSEs can remain in business (without help) comes down to credit losses and access to financial markets. If the GSEs can keep raising new capital, there won't need to be any bailout orchestrated by the Treasury. I think that's the best tax payers can hope for.


Anonymous said...

AI, thanks so much for the follow-up on the GSE's.

Q: With 30 yr rate at just below 6%, at what level inflation rate has to be before the vintage 2008 start to lose money.

I just think with so much uncertainty and so much risks in the economy, inflation, and housing market, 6% for 30 yr. is just insane. Are we gonna have another S&L bailout because risk models are wrong, again.


Scott Finch said...

"That thing's operational!" - GREAT TITLE period.

david pearson said...


Thanks for the correction on the loss as a % of assets calculation. Shows you are a class act.

Anonymous said...

Hulu: Why would mortgage and inflation rates contribute to loss? Presumably the GSEs are selling fixed and floating rate bonds in a way that makes sense with the fixed and floating loans they are recieving. If mortgage or market rates change in the future it doesn't change the terms those bonds and loans were placed at in the past. The fixed rates will still be the same and the floating should move more or less in lock-step. Mostly it mucks up the value of the portfolio in reserve.

Loan Default is where most of the risk lives. Prepay during rate declines can be a big issue on the non-pass-through side of things, as your high-fixed-rate mortgages turn into low-fixed-rate mortgages while the bonds they sold are still high-fixed-rate. But they do try hedge against that.

Anonymous said...

If the GSEs do need bailing out at some point, the government doesn't need to pump in a ton of money until the balance sheet balances out, it just have to cover the spread between money recieved and payments due. The loans and G-Fees will still be contributing something, so I wouldn't expect it to be outragiously taxing on the federal budget. It's not like the Fed will suddenly get a bill for several trillion dollars.

Anonymous said...

"Hulu: Why would mortgage and inflation rates contribute to loss?"

Didn't that contributed to the collapse of S&L? If your borrowing cost rate is higher than your lending rate, are you not upside down on those loans. I am pretty new to the bond market, that's why I asked.


Accrued Interest said...


I think Anon has answered your question correctly. They borrow at a rate lower than they lend (thinking of it simplistically) so it doesn't matter what happens to rates so long as they've matched their assets and liabilities well.

However in the past, they've had mixed success with hedging and ALM. At times they've made more money on their hedges than they should have, which if you think about it is a serious problem.

Typically prepayment speeds are their biggest problem. For their sake, I hope they've assumed very slow prepayment speeds, because that's what I think is happening in the real world.

gramps said...

AI - this whole discussion is getting bogged down in accounting minutia and a lot of speculation of what could'a should'a might'a happened...

Lets focus on the underlying economics. You have two entities that are, in aggregate, responsible for owning and/or guaranteeing about 40% of all mortgages in the U.S.

There is something like USD $12 trillion worth of outstanding / securitized mortgage debt. Based on long term trends in house prices (linear interpolated), homes are about 15 percent over valued. 15% of $12 trillion means housing is (in aggregate) about $1.8 over valued. One could make an argument that we spent a long time "over trend", and now we have to spend time "below trend" to even things out, but lets keep it simple and say prices only revert to trend and do not over correct (as they often do).

If you do the same analysis using home prices as a multiple of earnings, houses are overpriced by a bit more than 15%. Personally, I think price/income multiples is a better measure of housing values, because ultimately a house can't be worth more than someone is able to pay. This is the problem we are facing today in the U.S. -- people cannot afford a $1 million dollar "starter" home, no matter what the Fed Funds rate might be.

Anyway, lets stick with the simple linearly interpolated house price model -- even though I think it is less accurate. I am picking it solely because it produces the least bad outcome.

In aggregate, we expect $1.8 trillion dollars in lost home value. This part you cannot really argue with (except maybe that the number is probably bigger). Note that this is 15% of mortgage debt outstanding; it conveniently does not count lost value for persons who own their homes outright / no mortgage (ie the elderly or rich).

The simple analysis would suggest that the GSEs are on the line for 40% of that loss, or $720 billion. Such a number would easily overwhelm the GSEs.

Of course, the GSEs to a large extent "cherry picked" loans by predominantly focusing on conforming loans. So $720B probably overstates their exposure. On the other hand, there was a lot of "creative accounting" by many borrowers, so there are a lot of loans labeled as conforming that in reality are not.

But lets look at the $720B potential exposure and ask ourselves how "lucky" do the GSEs have to be to remain solvent?

If their exposure is 50% of the simple number, or $360 billion -- they are finished. The GSEs only have about $80 billion in shareholder equity (assuming we believe their reported numbers, which is another debate).

Lets say their exposure is 25% of the simple number, or $180 billion. or 2.25 times share holder equity. They are toast.

If their exposure is 11% of the admittedly simplistic number, it wipes out ALL their shareholder equity. 5% would pretty much destroy them.

There is no doubt that the $720 billion overstates their exposure. To suggest that the actual number is under 10% is, to be blunt, overly optimistic.

The GSEs are insolvent. We can waste vast amounts of time arguing accounting details and whether a type III asset is really a type II asset... but the problem isn't accounting classifications nor willingness of FASB to bend the rules.

Most mortgage lenders made a lot of really stupid loans that don't make economic sense. That is the problem. And the mother of all lenders were the GSEs.

All the big lenders who got themselves leveraged 30-1 are bankrupt. Yes, I know Citi and Morgan Stanley did not formally file Chapter 7 papers, but they raised so much new capital that they essentially have new ownership. And they probably aren't done raising even more capital.

Now the GSEs are levered by at least 72-1, assuming we believe the numbers published by the GSEs. The GSE's accounting is so bad that even Congress wasn't buying it, leading to FNMA kicking out their CEO. But even with the 72-1 leverage, it strains all credibility to suggest the GSEs are economically solvent.

Will Congress and the Treasury allow all sorts of accounting rules to be bent and others to be simply ignored? Obviously the answer is yes.

Fifteen years or so ago, the Bank of Japan through out the accounting rules in order to pretend that Japanese banks were not insolvent. No matter how much the BoJ protested, the rest of the world just wasn't buying it.

Guess what? The world isn't buying the U.S.'s bullshit either.

Bernanke isn't helping anything by setting Fed funds at absurdly negative real rates. This keeps the zombies "alive" a while longer, but it hurts all the economically sound lenders and businesses that would otherwise be available to power NEW growth.

Let the insolvent institutions fail gracefully. Stop pretending they are solvent. Don't even suggest for a moment that taxpayers should bail out the stupid at the expense of the prudent -- the U.S. is supposed to be a meritocracy; we all know what happens to standards of living when governments go around propping up bad, but politically favored, businesses.

If you truly believe a GSE failure would destroy the markets (any worse then what is happening anyway)-- then freeze their assets and wind them down gradually. Same thing with the big banks.

Yes, I know AI is just trying to be "practical" in saying that Congress is too stupid to actually make real fixes. Fair enough. But they have always been too stupid. The difference this time is that the markets are ALSO being stupid. We have a big problem, and the first step is DON'T MAKE IT WORSE. Don't keep propping up bad businesses. How screwed up is Wall Street that this suggestion is not considered obvious?

Accrued Interest said...


In order to get to your $720 billion number you have to assume...

1) GSE borrowers are "average" borrowers. They aren't, as you admit. In fact, the worst borrowers, the ones most likely to default, aren't in conforming mortgages.

2) That ALL of the negative HPA turns into real dollar losses. That doesn't make any sense at all. In order for that to be the case, you'd have to see 100% defaults with all recoveries being equal to the "fair value" of housing.

I've seen this kind of analysis all over the web. But to say that paper losses of $X Billion (or Trillion) will certainly turn into cash losses of the same doesn't make any more sense than to say the opposite during a bull period.

In other words, take today's Case-Shiller number, which showed a -14.1% YOY HPA. But the 5-year figure is still +22%. Should I write a post trying to devine where all that 22% in profits lies? How ridiculous would that sound? One could easily counter the 22% figure by saying that the 5-year period isn't the holding period for any particular person.

But neither is 1 year!

Now look, I think its clear from my writings that I think there are plenty of problems with the GSE's. But if you are going to start from the assumption that they are going under, and then work the numbers until they support your conclusion... well... I'm sure you'll find what you are looking for.

gramps said...

Now look AI, you really need to read my comment in full.

I said the $720 billion was an overestimate. I said it in so many words. Why would you claim otherwise?

We don't know how much of this unrealized loss would get realized... all I said was that if a mere 11% gets realized, that would completely wipe out ALL shareholder equity at the GSEs.

I also addressed the GSEs being "above average" lenders in assuming only a 15% loss. Markets like California, Florida, Arizona, etc are seeing losses more like 25-30%.

If you don't agree with me, that's one thing. But don't re-write what I said.

Accrued Interest said...


My point is that the net negative HPA from top to bottom doesn't have anything to do with the GSE's loss potential. Their losses will be based on foreclosures and recoveries. Now obviously the fact that many foreclosed properties have more than 100% CLTV will have a big impact on recovery. That's not really debatable. But the way to estimate GSE losses is to estimate foreclosures, see where those foreclosures will take place, and estimate CLTV's in those areas. You also have to consider the impact of mortgage insurance and modifications.

We also know that GSE loans were approved under a certain set of standards that did not include Option ARMS and 100+ OLTV in most cases. They also did not do loans larger than 417,000, which meant that their coastal exposure is less than average. In effect, their bubble exposure is less than average. This bears out by comparing OFHEO's HPA survey vs. Case Shiller's.

I'm not arguing that everything is fine and we can all go about our business. I'm just saying let's analyze the situation for what it is, not what its not.

gramps said...

AI, I am not trying to reach a pre-ordained conclusion. I am just looking at the numbers published by Freddie and Fannie, and knowing that they are exposed to at least 40% of the mortgage market, and saying the problem is far bigger than either company is letting on.

I am not alone in this. Warren Buffet called the two companies "the biggest accounting disasters of the last 10 years". Buffet was a major shareholder at Freddie and arguably had access to a lot of information the rest of us do not... and he sold his shares.

And even though I would agree that the majority of the GSE portfolios represent "above average" mortgages (or at least conforming anyway) -- I would draw your attention to the fact that the two GSEs have reported over $160 billion in Level III assets, which is roughly twice the shareholder equity of the two companies. Agency pools don't get such a classification-- so according to GSE management, they have a lot of crap on their books in addition to the good stuff. Roughly twice as much crap as equity.

The GSEs are leveraged far more than any private company could ever get away with (at least 72-1). So even though they may have above average selection of mortgages, they have basically zero room for error.

The fact that they are operating on such absurd levels of leverage tells you that speculating on a federal bailout is pointless: they are already being treated like they are part of the government. If they were private, they would already have been shut down. Heck, as private entities they would never have survived the earlier accounting scandals that led to the CEOs of both companies being fired. How bad do things have to be when CONGRESS is firing a CEO for accounting issues?

So please, instead of some vague hand waving gestures and empty assurances that Fannie and Freddie are fine, please tell us what you see that Buffet does not.

Explain why the GSE management has classified $160 billion of assets as level III, when the companies are supposedly only holding "superior" conforming loans.

Other than new management, what changes did the GSEs make since their accounting was so bad that not even the U.S. Congress could tolerate it? A new PR face on the shareholder reports does not mean any of the business procedures, risk management techniques or accounting practices have improved.

And explain how any responsible portfolio manager would tolerate 72-1 leverage, never mind the even higher levels that Barney Frank is forcing on them for political reasons.

No one disputes that the politically connected can massage accounting rules for quite some time, but eventually the underlying mess is going to come out. Its just a question of when.