Mark To Market

One of the stranger aspects of the debate over the bailout was the way both progressives and conservatives latched onto the idea of suspending mark to market accounting. (The final bailout bill (pdf) reiterated that the head of the SEC had the power to do this (sec. 132), and called for a study of mark to market (sec. 133), but did not require that it be suspended.)

If you’re not up to speed on accounting regulations, here are the basics:

Mark to market accounting requires that you value (“mark”) your assets at their market price. This is straightforward in the case of something like a share of IBM: lots of people are buying and selling those shares, so figuring out the market price is easy. It’s harder in the case of assets for which there is no active market.

A familiar example: houses. Most people hold their homes over the long term, and there are often no exactly identical homes. So when someone estimates the value of your home, they rely on the sales of houses they think are enough like yours to allow for comparisons. In normal real estate markets, you can find enough sales that are comparable enough to allow you to come up with a more or less decent valuation. But when the real estate market tanks, most people don’t want to sell if they can possibly avoid it. This has two results: first, the number of sales goes down, and so you have fewer even remotely comparable sales to look at in valuing your own house. Second, the people who do sell tend to be desperate, and will therefore sell their homes for very low prices.

This isn’t a problem for many homeowners: so long as I don’t plan to sell my home or to take out another loan on it, its value can go up or down as it pleases without affecting me. But suppose that my creditors required me to keep tabs on the value of all my assets, and to maintain a given balance between my assets and my debts. In this case, the fact that my home’s market value went down would cause the ratio between my assets and my debts to go down as well, thereby causing me to have to sell assets in order to maintain it. To do that, I’d have to sell assets.

In the worst case, I’d have to sell — you guessed it — my home. And if enough homeowners found themselves in a similar situation, it could produce a horrible downward spiral in home prices: home prices decline, causing some homeowners to have to sell their houses, causing a further decline in home prices, causing more homes to be dumped on the market: wash, rinse, and repeat. (Note: the same thing works in reverse, during asset bubbles: people get to book fantastic paper profits, and can use them to offset ever-increasing amounts of debt. Oddly, we tend not to hear banks complaining about mark to market then.)

The possibility of that kind of vicious spiral is, basically, the problem with mark to market. It’s a real problem. But I don’t see why anyone would think that suspending mark to market is a good solution to it. For one thing, what’s the alternative? Well:

“Because securities in the $1 trillion CDO market trade infrequently, it is difficult for hedge funds and other investors to mark their values to recent sale prices, called “marking to market.” Hedge funds instead use mathematical models of their own to estimate and report the value of their CDO holdings to investors — a practice known as “marking to model.””

And who, you might ask, comes up with these models? The people who own the assets. Might they have an interest in tweaking the models so that they showed inflated values for those assets? Why, yes, they might. Warren Buffett (pdf):

“Those who trade derivatives are usually paid, in whole or part, on “earnings” calculated by mark-to-market accounting. But often there is no real market, and “mark-to-model” is utilized. This substitution can bring on large-scale mischief. As a general rule, contracts involving multiple reference items and distant settlement dates increase the opportunities for counter-parties to use fanciful assumptions. The two parties to the contract might well use differing models allowing both to show substantial profits for many years. In extreme cases, mark-to-model degenerates into what I would call mark-to-myth.

I can assure you that the marking errors in the derivatives business have not been symmetrical. Almost invariably, they have favored either the trader who was eyeing a multi-million dollar bonus or the CEO who wanted to report impressive “earnings” (or both). The bonuses were paid, and the CEO profited from his options. Only much later did shareholders learn that the reported earnings
were a sham.”

As best I can tell, our main alternatives here are: (a) require firms to value assets at market value, unless it is literally impossible to do so, or (b) allow firms to construct their own models for determining how much their assets are worth. The second alternative might be OK if there were some widely accepted way of modeling the value of these assets that left no real discretion to the firms that owned them. But as far as I know, there isn’t. And that means that mark-to-model is, in my judgment, an invitation to wishful thinking and outright fraud.

As I understand it, one of the reasons why we got new rules on mark-to-market accounting was precisely that the models that had been used to value a lot of the newer financial instruments turned out to be completely wrong. Whatever the problems with marking to market during a panic, the solution cannot be to let firms disregard market data in favor of models they make up themselves.


One way to think about this is as follows. The problem with mark-to-market is, basically, not that assets are valued in a certain way, but that firms are required to do certain things as a result of those valuations. The argument against mark-to-market is that we think they should not have to do those things.

Suppose, for the sake of argument, that you agree. It doesn’t follow that mark to market should be suspended. There are two ways to allow firms not to sell their assets in response to low market prices. One is to stop asking firms to reveal the market value of their assets; the second is to change the requirement that they sell those assets whenever their asset/debt ratio gets too low. The second seems obviously preferable: it solves the problem directly, while allowing us to have as much information about the companies we invest in as possible.

But, one might say, in a lot of cases, it’s the lenders who impose that requirement, not the government, and so we can’t just change it. In that case, people who advocate changing mark to market should be straightforward enough to say: we do not trust lenders to react appropriately to information about the value of their borrowers’ assets, and that’s why we’re not going to require companies to reveal that information to them. Because that’s really what it comes down to.

I see the problems with mark to market. But why anyone thinks that the best response is to let companies go back to mark to model is a mystery to me. This is especially true now: one of the (many) reasons why we’re in trouble is that no one seems to know how solid companies’ assets are, and which unpleasant surprises might crop up in whose balance sheets. I can’t see why anyone would think that what we really need right now is less transparency.

Our ideas can save democracy... But we need your help! Donate Now!