Friday, March 13, 2009

Mark to Market

The practice of requiring banks and other businesses to mark their investments to market has drawn increasing criticism in the last several weeks from pundits and high-profile investors who would like to see it at least relaxed, if not rescinded. When I traded commodities in Texaco's international oil trading operation in London in the early 1990s I acquired some first-hand experience with "mark to market." Although I wasn't dealing with multi-billion dollar investments in exotic credit derivatives, the principles are sufficiently similar for me to offer some thoughts on the benefits and risks of this methodology, which appears to be feeding a vicious cycle of asset deflation in the financial sector.

The other night while watching our favorite TV cop show, "Life", my wife and I got a laugh out of the bumbled attempts of several of the characters to explain a derivative, and the confusion that greeted the accurate definition when it was finally given. I suspect the writers were reminding us how few people truly understand some of the financial instruments and regulations at the heart of the current crisis. Mark-to-market accounting likely falls into this category.

In the case of the futures market and physical oil market deals in which I was involved in London, the mark to market (MTM) provided a way to issue a daily report card on each of the trading positions we had taken on behalf of the company. This removed much of the element of surprise, if the value of something we had bought or sold changed significantly before the deal was ultimately completed. It entailed assigning a market-based price to each component of the deal at the end of every trading day, as if the product had been delivered or the position unwound that day, even though that might not actually happen for weeks. When the commodities in which we were dealing were ones for which there was an active, liquid market at all times, this accounting was relatively easy to perform. For some of the more unusual things we dealt in, for which there was no futures market and only occasional, sometimes unreliable reports of recent transactions, it generated uncertainty and anxiety.

The purpose of undertaking this effort, which consumed valuable time and was not exactly popular with the trading team, was to promote accountability and action. If the MTM on a particular trade showed a steady negative trend--particularly if it had moved from an expected profit to a loss--this triggered a discussion with management about why it was happening and what should be done. When handled well, this sometimes led to new insights about the market that we had failed to recognize. It normally resulted in a decision on whether to hang in there a bit longer, because we could justify our view that things would turn our way, or to modify or unravel the position--even at a loss--and regroup. Of course, that wasn't always possible; sometimes the cargo was on the water, bought and paid for, and there was nothing we could do but watch the red ink swell. That gets at the essence of my concern with the application of MTM to the big banks and institutions that the government has been forced to assist, for fear of "systemic risk"--the chance of the whole financial system crashing like the Blue Screen of Death on your PC.

Crucially, the reliability of mark to market depends on the ability to obtain an accurate reading of the value of what you are holding. That requires credible reporting of current transactions--preferably many of them--in something that, if not identical, at least looks enough like your asset to serve as a good proxy. If the only deals reported are distressed sales by desperate firms, you must write down your position to that level, even if you would never willingly sell it for so little. In the worst case a series of such write-downs causes a large enough deterioration in the balance sheet of the firm that it is compelled to sell some of these assets, driving their market value even lower and triggering a cascade of further sales by depressing the MTMs of other institutions.

Throughout the financial crisis, the practice of MTM has been defended as an unpleasant but necessary discipline to prevent an outcome such as was seen in Japan after its property bubble collapsed, with numerous "zombie banks" that were effectively insolvent but kept alive by the fictitious value of assets that were worth only a fraction of the level at which they were carried on the books. That argument still has some merit. However, it seems equally possible to destroy investor (and ultimately depositor) confidence in otherwise profitable, solvent institutions through the steady mechanical deflation of their illiquid assets, the potential buyers for which understand clearly that time is on their side.

Having run this experiment in its pure form until now, I'd like to see the administration test the opposite hypothesis for a few months: suspend MTM for bank capital purposes and restore the "uptick rule" on short-selling, while they're at it. We'd quickly find out whether these steps helped to stabilize the system. If they made things worse, they could quickly be reversed. It wouldn't be the first course correction we've seen during this crisis.

No comments: