Forward: this was printed about four hours before Berkshire Hathaway took a $5 billion stake in Bank of America under sweetheart terms. Firstly Buffett got better terms than me - it helps being known as the world's greatest investor. Second, the rapid appreciation in my position is dumb luck.
I am long Bank of America on my own behalf and on the part of my clients. It has not been a fun experience. (We have had great returns at Bronte but) Bank of America is one of two stocks on which Bronte has lost more than 3 percent of the portfolio.* So you can see this note as written from the perspective of a Bank of America loser.
Still I am bullish on BofA at these prices. Very bullish. I think the politically driven finance bloggers (Yves Smith at Naked Capitalism) should be seen for their (I think justified) anti-bank agenda. Most of the rest are fitting their analysis around the stock price.** There is an awful lot of stock-price doing the analysis here.
You might ask if I am putting my money where my mouth is – and I am a little. We purchased some more BofA below $7 but not much. Why not much? Risk control. Nothing else. Like most “value investors” we believe the right response to a stock we like going down is to buy more of it. As people who are (sometimes painfully) aware of our fallibility, we know repeatedly doubling up on a stock with tail risk is a way of getting pole-axed. So we added but did not pile in.
That said, I should explain how I think about Bank of America right now.
The problem everyone is talking about is liability for fraudulently originated mortgages – mostly mortgages originated at Countrywide. Bears argue (correctly I think) that BofA has under-reserved for liability problems associated with past mortgage origination. Some bears suggest large numbers ($50 billion is often quoted) for the real liability. They argue that BofA will be forced either to insolvency, a further bail-out or to raise a lot of capital under highly unfavorable terms (thus crunching the existing equity holders).
There are other bear cases – and I will get to them – but I want to deal the the main bear case up front.
First lets get the capital argument sorted out. Bank capital can't be accurately measured. There are just too many estimates. Steve Randy Waldman (at the indespensible Interfluidity blog) says it better than me when he points out just how many estimates go into measuring bank capital and how large those estimates are relative to the stated equity.
Over a decade or more you know whether you got it more or less right or not. Canadian and Australian banks have simply spat out money – great gobs of it – for twenty years. I do not know whether the capital or profits are stated right (and nor does anyone else) but I can be certain that these banks have been very profitable. Japanese banks by contrast don't seem profitable on a decade-long view.
But on a day-to-day basis capital can't be measured and it is meaningless to say that the bank has $50 billion too much or too little capital because you can't measure that either.
We can find out ex-post (that is after a liquidation) whether the bank was egregiously under-capitalized or not but it is very hard to tell on a day-to-day basis. I thought Lehman was insolvent in 2006. Ex-post it probably was insolvent in 2007 – it operated for quite a while after that. I never thought WaMu was insolvent but the regulator disagreed with me. Ex-post in that case I think I was right on the regulatory capital but was not right on stock.
That doesn't mean banks don't get themselves into trouble by having too little capital. They blow up with alarming frequency. But it is not actually too little capital that is the problem. Banks can operate with negative capital for years (as per Japan). It is one of two related problems:
(Type A Problem) the bank suddenly has losses so large and so unavoidable and generally so fast that any regulatory capital amounts are just smashed and the bank goes regulatory-insolvent or is forced to raise capital under disadvantageous terms or to find a “bridegroom or a suitor” or the regulator takes them over.
(Type B Problem) the market loses faith in the bank and the funding dries up which causes the bank to have liquidity troubles.
Type A problems are rare, Type B problems are much more common.
The archetypical sudden-death of a bank problem is Barings. Nick Leeson lost US$1.4 billion in the Singapore Futures Market and the loss was revealed in a single day. This was twice the bank's stated capital and Barings was forced to find a suitor. They sold themselves for a dollar.
This sort of loss is rare for large institutions (though more common for smaller institutions with government guarantees). Large institutions reveal their losses over time. That is important because operating income (in the context of a bank income before provisions and tax) can be used to shield the loss. If a bank takes its losses slowly enough it can shield very big losses this way (albeit at the cost of appearing zombie-like for years). The champions in slow-loss revelation are Japanese banks who spread losses over more than fifteen years and never breached stated regulatory guidelines on their accounts.
Because banks have quite a lot of discretion about how they book their losses and most losses can be spread over-time just running out of stated capital is not the common way for a bank to get into trouble. [The exception is small banks with guarantees as per the US and the only way they ever seem to leave is by rolling losses until they are comically enormous compared to stated capital.]
You see banks deferring losses every cycle. When the crisis hits everyone screams the bank is under-reserving and guess what – everyone is right. But the bank gets to take its losses over time and that suits the bank because the bank tends to have high pre-tax pre-provision earnings in a crisis and time cures things. When you take losses is subjective most of the time. In bad times banks lie about losses because they can and it is their interest to lie. This is – as Buffett has noted – a self-assessed exam where the penalty for failure is death.
Of course if the losses are too fast and too sudden the bank can't spread them. When someone is not paying their mortgage extend and pretend is an option. When someone actually throws back the keys and walks out you have to take the losses. Enough of them at once and you get the Nick Leeson situation, big losses which you have to book now. But it is easy to extend and pretend so banks do.
The more common way banks get into trouble is when people don't trust them any more and they can't fund themselves. This happened a surprising amount in the crisis. Sometimes we discovered the “bank” was grotesquely insolvent (Lehman Brothers). Sometimes it was only marginally problematic (Bear Stearns). I still believe Washington Mutual was solvent and the run was a panic.
You can't measure bank capital accurately but there is a way in which banks can have too little capital. They have too little capital when they can't convince regulators or creditors that they are themselves a good credit.
Lets look at Bank of America in this light
The credit default swap (one year, illiquid) says that BofA is having some trouble financing itself. People are willing to pay 4 percent for a one year BofA default bet.
But absent that (rather hairy) data point I lean on the fact that the “too big to fail” rules of the game are well understood at the moment whether Yves Smith or Paul Krugman likes them or not. No big bank in America is going to be let fail. Ultimately the credit of Bank of America is synonymous with the credit of the United States of America and last I looked at US bond pricing that credit was good.
In other words BofA has enough capital to raise money in the bond market because its real capital is not something on its book. Its real capital is faith and credit of the United States. And because of that the bank won't fail through a wholesale run. Besides Bank of America has a lot of short-term liquidity. Not enough to save them from a mega-catastrophic run of course but they can deal with most things and a mega-run relies on the too-big-to-fail consensus breaking down.
The only capital risk to BofA then is one that regulators find them poorly capitalized and force them to raise capital or the like. That is definitely possible but in my view unlikely.
It would happen if BofA were to book a sudden 50 billion in provisions for mortgage-fraud settlements. But that is the legendary self-assessed exam where the penalty for failure is death.
You see these are litigation losses not credit losses and the one thing that everyone agrees on about litigation is that it is slow.
For Bank of America slow is good. Very good. You see BofA has more than 10 billion dollars in pre-tax pre-provision earnings every quarter. This number is falling but it still very large.
If Bank of America really has 50 billion – no – lets get really bearish – 70 billion in additional losses to take but the litigation lasts seven years it will eat only a quarter of the pre-tax, pre-provision earnings over that period. It will dampen earnings but can't cause BofA to run short of regulatory capital.
And I am pretty sure they could stretch the litigation five years if not seven. I have seen court cases where discovery lasts that long.
So in summary Bank of America won't fail because the market does not want to fund it. It is “too big to fail” and its credit is really the credit of the US Government. And it can't fail because it needs to take too many losses too fast and runs out of regulatory capital. These are litigation losses and they offer plenty of time for deferral against future income.
In other words this Bank of America panic is just a panic.
And at the risk of sounding like Jim Cramer: Buy.
Where I can be wrong
Above I am talking my book. I own Bank of America shares. My clients own Bank of America shares. I want to explore the ways I can be wrong. After all it could cost our clients a further 5 percent if Bank of America fails.
The first way I could be wrong is if there is another big credit cycle (not old stuff, new stuff) caused by a deep, nasty double-dip recession. In that case the pre-tax, pre-provision income of the bank may be committed to paying off another round of credit losses and not be available for litigation losses.
The main defense the bank would have against that is that the litigation losses can be deferred for very long periods so the bank might deal with one fire (new credit losses) first whilst it leave the other fire (litigation losses) to simmer. It would not be nice for shareholders (no dividends or capital growth for seven years) but it is not devastating.
More worrying is my assumption that the pre-tax, pre-provision earnings are solid. They are clearly falling right now. In Japan these fell for ten straight years – and when you thought that bank margins could not get any lower they went into a bit of a decline.
I once wrote two posts on why I did not think the Japanese outcome was likely for American banks. Go back and read them (links here and here). Those posts don't look that good any more. America is looking more and more Japanese. Pre-tax, pre-provision earnings for banks is falling faster and further than I thought possible.
A small amount of inflation (surely the Fed will print money until that happens) fixes that problem – but that was my argument 24 months ago and that argument is looking less right every day.
The third way I could be wrong is if litigation is not as slow as I am guessing. There could be some kind of summary judgement along the way. But then maybe not. I genuinely do not know enough about American litigation practices to offer an informed opinion.
The fourth way I could be wrong is if the Government guarantee is called by a really skittish market. Government guarantees (implicit or explicit) are hardly what they used to be. That said America prints its own currency. And the credit markets seem to think the US government is solvent. Nonetheless I can't rule it out either.
The fifth way I could be wrong is if the regulators themselves call it. Bank of America is a beneficiary of an implicit government guarantee. That in a normal world gives the government some over-arching regulatory rights. They could determine that there really are $50 billion in provisions necessary and force Bank of America to provide rapidly. If they did so then a capital raise would be necessary because the Government forced it.
Governments have power. If they use it against existing Bank of America shareholders then I will lose. Some of the anti-Bank-of-America bloggers (and I am thinking mostly of Yves here) have that view because they believe government SHOULD use that power against banks. I suspect in that she is right – but my job is to make money for my clients – the question is not whether they should use that power, the question is whether they will. In that I think I can rely on the cravenly pro-finance Obama administration.
There is a final way I can be wrong - and it is a way that worries me more than almost any other. That is Bank of America just falls apart from a systems basis.
I have a friend who had a tax lien put on her house because Bank of America misreported something. It took her six months to get it taken off. There are stories of houses with no mortgages being foreclosed on. There are stories about people getting free houses because the bank loses the mortgage documents.
A bank only has value if it can perform the function of a bank - and top of that list is keeping the client accounts straight - knowing who owes and who owns what and having the documents to prove it.
On that Bank of America is surprisingly inept. That is what comes of doing too many bank mergers.
And of all the things that worry me about BofA their systems failures are the ones that worry me most. You hear too many stories and the stories are from credible people without an axe to grind.
If it were not for that worry I might have added (yet another) percent to my Bank of America holding. But risk control is not lost on me - so I am sitting pat.
There you have it...
My case for Bank of America right now. It is the case of someone who has marred a very good record (about 100 percent above index in just over two years) by owning a lot of Bank of America stock. It is the case of another Bank of America loser.
Make of it what you will.
*Bronte has only been running just over two years. Given more time I am sure we will find more ways to both make and lose the clients' money. We are proud of our record (even if I do focus a little on the losers...)
**Incidentally I think Yves anti-bank agenda is justified. She is correct that the banks have Washington wrapped up, that real reform is off the agenda and is necessary and that the banks have behaved terribly and that BofA/Countrywide has behaved utterly atrociously. That of course does not mean BofA is insolvent no matter how devoutly she wishes it. Indeed the Washington consensus is a major reason for thinking BofA is not insolvent.
Thursday, August 25, 2011
Posted by John Hempton at 6:54 PM
The content contained in this blog represents the opinions of Mr. Hempton. Mr. Hempton may hold either long or short positions in securities of various companies discussed in the blog based upon Mr. Hempton's recommendations. The commentary in this blog in no way constitutes a solicitation of business or investment advice. In fact, it should not be relied upon in making investment decisions, ever. It is intended solely for the entertainment of the reader, and the author. In particular this blog is not directed for investment purposes at US Persons.