Editor's note: Following is the eighth 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; the fifth part; the sixth part and the seventh part
) -- In Part 7, I did an "idiot check" on my credit loss numbers. They appear pretty robust. This post does an idiot check on the pretax, preprovision profit estimate. Here I am less confident.
The massive rise in government-sponsored enterprises pretax, preprovision profits is one driving factor behind my assertion that the GSEs can recapitalize. In the
10-Q from the first quarter was this (often quoted) and profoundly bearish line:
"Our annual dividend obligation, based on that liquidation preference, will be in excess of our reported annual net income in nine of the ten prior fiscal years. If continued to be paid in cash, this substantial dividend obligation, combined with potentially substantial commitment fees payable to Treasury starting in 2010 (the amounts of which have not yet been determined), will have an adverse impact on our future financial position and net worth, and will contribute to increasingly negative cash flows in future periods."
This line, or variations of it, appears multiple times in the recent 10Q.
This is a blunt statement that Freddie could never repay the government because it owed the government $5.2 billion per annum and that is more than the earnings of almost every prior year.
There is a little that is disingenuous about this statement, and this is possibly deliberate. The statement compares the obligations to the Treasury to the
post-tax, postprovision income for the past decade
. In most years the pretax, preprovision income of Freddie was in excess of $5.2 billion (which would have allowed some repayment). But far more to the point, the current
pretax, preprovision income
is in excess of $15 billion.
After writebacks, they dealt with more than $8 billion of the $50-odd billion outstanding in one quarter in the second quarter, but they are not permitted to make the actual repayment (more on that in a later post).
Here is a cut-down version of the profit and loss account from the last quarter:
Revenues -- net of mark-to-market swings -- are up from $2.3 billion to $4.4 billion. Administrative expenses are down slightly. Pretax, preprovision profits are probably running at more than $4 billion
But that of course nails down the problem. The situation is so rosy for the preferred (and survivable for even the common stock) precisely because the pretax, preprovision income is so high. If the high pretax, preprovision earnings go away, so does the taxpayer's chance of getting repaid the
and Freddie bailout money. For that matter, so do the chances for holders of the preferred securities (which we at Bronte Capital have so carefully, and cheaply, accumulated).
What the Margin Is for
Fannie and Freddie make their margins in two ways:
- By charging guarantee fees for mortgages that they guarantee.
- By holding mortgages and earning a spread.
The guarantee fee margins were a fifth of a percent of outstanding balances, or less, for as long as I remember. I always thought that those margins were insanely low -- and indeed the very low margins for insuring credit risk is (in my opinion) the main reason why Fannie and Freddie were in long-term-trouble. Banking systems without enough profitability cannot survive bad times. I have blogged about that extensively -- see
for a controversial example. Those guarantee fees are going up but are by no means enough. It would not be unreasonable to charge 0.4% percent. However, under conservatorship and even with an absence of competition, fees have not risen to that level. In the absence of competition Fannie and Freddie
be able to raise guarantee fees sharply. They should, too -- otherwise the fees are not reflective of risk. However the fees have not risen by quite that much, and the only explanation I have is political interference. (Again you will need to wait for another post.)
However, with a guarantee book of less than $3 trillion dollars, guarantee fees -- although important -- are not the way in which this company recapitalizes. Guarantee income was about $700 million last quarter. Not small change to anyone but Fannie and Freddie -- but not enough to produce the profit stream necessary to cover forthcoming defaults and to repay the government. My guess is that the guarantee fees rise over time, but only if the regulator allows them to rise.
The driver of high pretax, preprovision profitability is high interest rate spreads. They are high because of a lack of competition. Interest margins are rising pretty well everywhere in banking, but not as intensely as at the GSEs. There is roughly $900 billion on the book. Making 1.2 % on that (which does not seem unreasonable but is much higher than the traditional Fannie or Freddie margin) will get you to solvency, although solvency for the GSEs emerges after, say, seven to eight years under this normalized income scenario. The current spreads are way higher than normal -- unsustainably high. Those unsustainably high spreads might lead to very rapid recapitalization.
Now obviously some of the excess spread is due to the steep yield curve. That will go away. But if the company were solely playing the yield curve the spread would be
than it currently is. Last I looked the spread between floating-rate Fannie obligations and
was several hundred basis points.
The income is also inflated at the moment because charges that were taken as the companies went into conservatorship are being reversed through the net interest income line. I wish I knew how to quantify this. (I described this issue in Part 2.)
Unfortunately, at some point the trend in income will be down. When income goes down so does the ability to repay. Close observation of the margin between GSE Treasuries, GSE debt and wholesale guaranteed mortgages indicates that the margin peaked a couple of weeks ago. Two weeks of data are not convincing, but my guess is that pretax, preprovision operating income will be about flat (maybe slightly down) in the third quarter and will trend down (perhaps slowly) from there.
Risk of Being Forced to Shrink
The first and obvious risk is that the GSEs will simply -- by government fiat -- not be allowed to earn the spread. When the GSEs were put into conservatorship they were obliged to shrink their balance sheets fairly rapidly after the first two years. If Fannie and Freddie shrink their balance sheets, they will shrink its spread income. If this is done rapidly enough they will never repay government. The requirement to shrink the balance sheet has been reduced dramatically, and is unlikely to be enforced in the absence of a robust private sector mortgage market. Obviously the reality (that these companies are by far the dominant mortgage providers at the moment) has sunk in. Shrinking them now would blow up a good part of the recovery. But I suspect that some politicians will want them to shrink. (Others will have different feelings -- again the subject of a later post.)
When the Republicans (for example,
) want to force the issue on Fannie and Freddie right now, that is what they are suggesting. If you allow them to shrink, they inevitably die -- and they cost government when they do so. Indeed, it appears that the Republican agenda was always to destroy these companies.
I will discuss the politics in a later post. The politics is interesting -- as in the Chinese curse. We live in interesting times...
The second risk to the GSE income is that somehow competition comes back into the mortgage market. I suspect that is a few years away. We only need to last a few years for the securities to be visibly money-good. Nonetheless I can't imagine the spreads remaining this wide indefinitely.
The third risk is that Fannie or Freddie massively stuff up their interest rate hedging and fail to adequately hedge the mortgage refinance risk or the short-term interest rate risk on their books. Fannie had a (relatively) minor hedging problem I think in 2002 in which it was short duration and interest rates moved against it by about 10 basis points in one day. My count at the time was that it lost $8 billion. The company could do that again. I have no way of estimating the chance of that, but I am relatively comfortable with the interest rate risk in the book at the moment. (Losing $8 billion in a day is relatively minor only when compared to the losses that Fannie has had on the credit cycle. Interest rate risk is part of these businesses.)
The Commitment Fee
At the end of this year the government has the right to charge Fannie and Freddie a commitment fee (mentioned in the quote above). The size of this fee has not been determined. This fee does not change the end loss to taxpayers but it may change the value of Fannie and Freddie's preferred and common stock. An excessive fee could lead to a Fifth Amendment complaint by preference shareholders. However it is a real risk to this thesis.
I am a preferred shareholder and -- as a shareholder in anything -- nearly always worried about risk. But if I had to tell you what keeps me awake at night it is essentially
political decisions crimping Fannie and Freddie's ability to earn revenue
. In particular, they may not -- by government fiat -- be allowed to charge adequate guarantee fees. They may -- by government fiat -- have to shrink their books very radically, thereby reducing spread income. They may -- also by government fiat -- be kept perpetually insolvent by way of the forthcoming commitment fee.
A note: I was a little more sloppy about the costs at Fannie and Freddie in Part 6 than I should have been. Some accurate criticisms have been received as to how I broke up cost items. However, I note that costs are seldom more than 12%-15% of revenue at the GSEs. The GSEs are large wholesale institutions -- buying bulk mortgages and doing finance in bulk. Costs do not matter much. What matters is revenue (this post) and credit losses (last post). When it comes to the 10Qs I have always read the cost section relatively fast as it is relatively unimportant.
The real risks to my thesis are on the revenue line and in the credit cost estimates.
-- Written by John Hempton in Bondi Junction, Australia
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.