Dollars and Jens
Tuesday, October 21, 2008
bank solvency, mark-to-market, and capital requirements
A lot of people have complained that mark-to-market accounting has contributed to bank solvency problems, forcing them to write down assets. Most of the time "mark-to-market" is going to be a much more reliable gauge to the value of an asset than is "mark-to-whatever-the-market-was-at-when-I-acquired-this-asset", or "mark-to-wishful-thinking" for that matter. Sometimes markets are inefficient, and it may well be that right now some assets for which markets exist are putting unduly low prices on assets that, with a more sober assessment of likely future outcomes, would make banks solvent that are currently, based on market prices, insolvent. As a general principle, though, I don't like the idea of encouraging banks to sit on those assets that are overvalued on their books when all evidence suggests that they have lost value. I think many people who come from a trading quant background feel the same way — mark-to-market is just obviously the right thing to do, and anything else is an attempt to ignore reality.
It's worth noting, though, that the fact that a company's assets are lower than its liabilities does not mean the company, as a going concern, is worthless; when Pepsico spun off Tricon (now "YUM brands", viz. Pizza Hut, Taco Bell, and KFC), it gave the new company more debt than assets, and the shares had a negative book value for a year or two. Operating profits were healthy, and the company's debt never had wide spreads, because bondholders knew they had nothing to fear; profits were high enough to make interest payments as they came due and to pay down some of the debt, and after a while the company had a positive net worth. It could well have paid a small dividend to shareholders, even with a negative book value, without ever imperiling the bondholders.
The term "solvency" can mean different things in different contexts; the uniform fraudulent transfer act renders certain transfers "constructively fraudulent" based on a definition of solvency as having assets worth more than liabilities. (Perhaps this explains why Tricon didn't pay its shareholders a dividend.) This is the usual definition, when a clear definition is needed; it is also the definition used by Chapter 9 of the bankruptcy code, which requires that a municipality be insolvent in order to file for bankruptcy. No other chapter makes that requirement, though; a company looking to effect an orderly liquidation can file for bankruptcy if it simply makes it easier (perhaps more transparent) to wind things down that way, or a company with a large contingent liability that would, if realized, make the company insolvent might well find it preferable to seek bankruptcy protection even if the accountants
don't deem it insolvent yet. On the other hand, a company that is "solvent" on paper can be pushed into bankruptcy if it can't make a debt payment, perhaps because its assets are illiquid (e.g. factories). When we distinguish between a company having a liquidity problem versus a solvency problem, we typically intend "solvent" to mean that a company would be able to pay off all of its debts if it were able to obtain credit at some reasonable rate for some finite period of time. In this sense, Tricon was viewed as solvent all along.
I've argued before — I think I've said this here, but maybe not — that I think the big financial institutions that haven't failed yet are all solvent in that last sense of the word, and I think any investment the government makes that returns a higher rate than the relevant treasury note is likely to result in a profit to the government if it leads the markets to believe that the company is both solvent and liquid. Any financial company without access to credit on reasonable terms is doomed, regardless of its actual solvency, at which point it's left in bankruptcy marking its assets to market, possibly straining those markets at the same time. This can lead to self-fulfilling "runs" — you don't want to be the last person to lend money to a bank, even if it would be viable, if it isn't viable — and conceivably even to speculative attacks, though I expect those are far less common than is popularly imagined. Of course, if you give more money to any institution that is not, in the long-run, solvent, you're throwing good money after bad, and (perhaps not disinterested) bankruptcy lawyers believe that, in general, most bankruptcy filers file much later than they should rather than sooner.
Washington Mutual, though, to hear some reports, was making $8 billion a year on its retail operations, and, if it had to mark its assets to market right away, would have been, making the absolute worst assumptions, $40 billion in the hole. It was in part hoping that those worst assumptions would not come to pass, but was also trying to delay write-downs as long as it could, in the hope that it could stretch out realizing its losses over a couple of years, and thereby never technically be insolvent. The company was by no means as solid as Tricon, but it may have been in a qualitatively similar situation, in which by a market reckoning it had negative equity, but would not have been unable to dig itself out without asking anyone to make an unprofitable investment.
The problem for Washington Mutual, taking this story as correct, is that it had regulatory (and even covenant) obligations that required that it remain solvent in the UFTA/Chapter 9 sense. Whether the assets were maturing soon or not — and thus, under any alternative to mark-to-market, would have had to be recognized — doesn't seem terribly salient to this fundamental element of the story. It may well be that regulators want to err in the direction of crude measures that are sometimes overly cautious and will prevent financial firms in holes from digging themselves, and the financial system and even taxpayer, in deeper; a company that wrote more than $40 billion in bad assets is not on the top of the list of companies you might expect to produce salutary results in the future, and there's something to be said for capping the size of the potential problem. Still, I wonder, if the possibility of being seized by regulators hadn't been in play, whether a WaMu might have had other options available to it.
One of the triggers for my thinking about all of this, over the last month or two, has been Arnold Kling's call for reduced capital requirements, at one point suggesting that one of the reasons for having significant capital requirements during good times is so that banks have funds that they would be able to lend — if capital requirements were lowered — when things look bad. This would be a bad idea from a cut-our-losses standpoint, but I agree with his proposal; solvent banks right now are unlikely to seek risky new loans, and a small change in capital requirements would do more to give room to healthy banks to make new loans than it would to increase the size of potential losses from zombie banks. On the other hand, I'm not sure "capital requirements", in terms of a straight-up look at a bank's balance sheet, is the best proxy for "how likely is this bank to cause how big a problem?" in the first place.
Some speakers at the Fed's conference at Jackson Hole in August noted that the Basel II international model banking regulations are too focused on making banks keep down the day-to-day volatility of their assets, rather than on tail-risk, systemic-type concerns, and I'm sure I agree; idiosyncratic risk, and especially day-to-day volatility, is the sort of thing that should be left to conventional market discipline. While we're looking at rethinking the way we do bank regulation, though, I think looking at balance sheets, whether with or without mark-to-market, without any attention to sources of future cash flows that aren't discounted and included as assets, is myopic, and might be amenable to improvement.