Editor's note: Following is the sixth in a series of blog posts by John Hempton, the chief investment officer and founder of Bronte Capital. They originally appeared on Hempton's blog on the Bronte Capital Web site. Here is the first part; the second part; the third part; the fourth part and the fifth part
We now have enough to do some basic modeling of
. I will do it for Freddie Mac only -- and leave it to the more ambitious readers to do it for Fannie Mae.*
In the second post in this series, I demonstrated how the losses that have been booked to date (rather than provisioned to date) have come primarily from outside the traditional guarantee book of business. Those losses are primarily mark-to-market losses on mortgage securities (especially subprime securities), mark to market losses on the hedge book and the write-off of tax assets.
None of those loss categories are going to expand -- and indeed some will reverse.
In the fourth post, I estimated the losses in the traditional guarantee book of business. I have asserted that the model is fairly robust (and will cover that in the next three posts); however, I showed under quite reasonable assumption that there were $37.6 billion in losses to be realized at Freddie Mac at year-end 2008. Since then, $2.9 billion have been realized, so there are $34.7 billion left to come.
Of these losses, $25.2 billion have already been provided for. From now until when the problem years of business loans run off, Freddie will need to take only another $12.5 billion in provisions. It may elect to take more than $12.5 billion in provisions -- but if it does and my models are reasonable -- then in all likelihood the excess provisions will be reversed through the income statement.
Now if you go to the last Freddie Mac results, you will see they have a positive net worth of $8.2 billion. However, they owe the government $51.7 billion, as the government has injected $51.7 billion in senior preferred securities. They are thus $43.5 billion in the hole.
They will also -- over time -- take another $12.5 billion in provisions. So now, until all the problem years of business have run off, they will be $56 billion in capital short.
The government can get its money back on its "investment" in Freddie Mac, provided Freddie can earn more than $56 billion over a reasonable time period and meet the government interest charges.
This would be more certain if some of the losses described in Part 2 reversed. I am pretty sure that they will -- but let's ignore them (until a later post). Pre-tax, pre-provision operating profits of Freddie Mac are running at over $15 billion. If the government were not demanding 10% on its preference shares, the companies would be sufficiently well capitalized to repay their interest in four years.
With the drag of having to pay the government $5 billion per annum, it will take a bit over five years. Either way, the operating profits of Freddie Mac are big enough to ensure the government gets its money back.
If you do the same analysis for Fannie Mae, it's is even better. However, Fannie has less aggressively marked private-label securities to market so it has less chance of recoveries from their current marks. The consensus view that the GSEs (government-sponsored enterprises) are forever toast -- and forever a drain on the U.S. government, is very likely wrong.
I have tried modeling this half a dozen ways, and the result is fairly robust. If anything, the GSEs (especially Freddie) are solvent quicker than the model I have presented suggests. Indeed ,if the tax losses are allowed to be bought back as capital, they will reach solvency a year and a half earlier -- and will be in the position to repay substantial government money during 2012.
The losses (even after all losses are booked) come from primarily outside the traditional business of guaranteeing small well-secured and documented mortgages.
Traditional GSE business (guaranteeing lower-value mortgages with reasonable terms on full documentation and with a down payment) was very effective at raising home-ownership rates, whereas modern subprime lending, it seems, just caused a blip in home-ownership rates that corrected with much pain.
Later in this series I am going to go through the politics of this issue. For now it suffices to say that by the time Obama is up for re-election, the government will be in a position to ask for and receive considerable repayment from the GSEs. One of the festering sores from this crisis will appear healed.
One more implication for my investor readers (and this, after al,l started as an investment blog): If the GSEs can repay their debt to the government -- and I think that they can -- then the common stock in both companies has value. That is a nonconsensus view. However, the real value is in the preferred securities.
The preferred securities are currently trading between 4 and 6 cents on the dollar (and went down while I was writing this sequence, indicating my readers either do not believe me, do not have money or had no idea where I was going).
The preference shares are all noncumulative, so you are not entitled to back-coupons when they resume paying, but they will resume paying sometime in the next four to seven years. At 4 to 6 cents on the dollar, that makes them a real bargain -- offering 16 to 25 times your money over four to seven years. That is a better return than you will get in most places.
Even the lower end of the range offers a 50% annualized return. The return on the preference shares is substantially better than any possible return on the common stock. However -- and it should be noted -- the conservatorship agreement gives no time period and specifies no criteria for the government to release Fannie and Freddie for conservatorship. This means that even if this model is right -- and Fannie and Freddie do recapitalize internally -- there is still no guarantee you will get paid on the preferred. Political risk is omnipresent.
At Bronte we have thought that the pre-tax, pre-provision profits were sufficient to recapitalize the GSEs for a while. We purchased large holdings of these securities below 2 cents on the dollar. Eventually, the preferreds started rising, leading to some financial-press skepticism that they would ever be worth anything. All I can say: At Bronte, our money and our client money is where our mouth is.
There are plenty of risks to this rosy hypothesis. These fall mainly into the political risk camp (there are many people who will fight a resurrection of the GSEs). However, there is model and economic risk as well. I will examine the risks (model, economic and political) in later posts.
The next three posts are (unfortunately) a little disjointed because all I am trying to do is subject my model to different data-tests and see if it is robust. You will find that I am much more comfortable about the credit-loss estimates in the model (Part 4) than I am about the income estimate (Part 5).
What makes me most uncomfortable, though, are the political risks, and those I have very little idea how to analyze. Late in this series I will be very keen to see if I can get a robust discussion at Talking Points Memo -- because those readers know far more about politics than me or most the regular commentators on my blog. For the moment, though, what we have is Republicans (and a much smaller number of Democrats) who are extremely keen to put Fannie and Freddie into liquidation now and hence make all of this modeling entirely redundant.
*At Bronte we have done the models for both Fannie Mae and Freddie Mac. If the relevant Treasury or NEC officials wish to contact me we will provide our models more generally
John Hempton is chief investment officer and founder of Bronte Capital, an Australian based global asset management firm. He was formerly a partner at Platinum Asset Management and has served as chief analyst of tax policy for the New Zealand Treasury.