I'm tempted to bastardize Churchill in saying that markets are the worst way to price these assets except for all the other ways to price these assets.

I'm a big fan of mark-to-market valuation. I think it's the best--and really only legitimate--way to place a financial value on anything. That said, I don't actually think that the market can ever give us the absolutely correct value of an asset. In fact, I'd argue that finding the absolutely correct value of an asset is impossible unless you have perfect knowledge of the future. If you knew exactly what was to happen in the future, you could calculate the future cash flow of any asset you're looking at (and the cash flows of all the alternative assets available to you) and pick a perfect price. But, of course, we never have that perfect knowledge. And therefore, a market (even a vibrant, active, and fair market) only produces a price that is the consensus value of the asset based on highly imperfect knowledge of the future.

In ordinary times, the imperfections in market pricing aren't important. People are expecting the market to get some things wrong and some things right, but they understand the variability and risk and adjust accordingly. In unusual times (like now), however, people no longer know what to expect and begin to distrust their assumptions and therefore discount assets heavily to account for higher levels of uncertainty about the future. This discount is valid--uncertainty about the future should reduce the price of assets. But it also is problematic, because it tends to create self-fulfilling prophecies. Everyone is worried the prices of assets will fall, so everyone discounts the assets they are considering purchasing, and therefore the price actually does fall, increasing the level of uncertainty and leading to further discounts. This is how recessions start. Usually a period of over-optimistic assumptions in which the market inflates values is followed by a correction in which the market adopts overly pessimistic assumptions and underestimates value. Both on the upswing and the downswing, the market is valuing assets "incorrectly" (which we'd know if we had a crystal ball), but of course those "incorrect" values are the only real and legitimate values, because they are the values that determine the actual price at which assets are bought and sold.

Alternative (non-market) forms of valuation work by adopting assumptions about what will happen in the future and doing the math. The alternative valuation approach has two problems:

1. There's no way to know the assumptions in the non-market valuation equation are any better than the consensus assumptions of the buyers and sellers in the market.

2. In reality, at any point in time any asset is actually only worth what someone will pay for it at that time, so non-market valuations are meaningless if you actually have to sell (or buy) an asset.

These serious and obvious flaws in non-market valuation methodologies make them generally dubious for accounting purposes--but they don't make them useless for all purposes. In fact, someone hoping to make a profitable investment will use non-market models to try to identify areas where the market is possibly wrongly assessing the value of an asset. Locating such opportunities can lead to big investment gains. One could argue as well that in extraordinary periods like the one we're in now, where uncertainty about the future is extreme, markets must (legitimately) discount assets to such a degree that many assets will turn out eventually to have been priced far too low today. In these periods of extreme uncertainty and extreme discounts, non-market methods of valuation may be valuable to help make guesses about which assets might eventually prove to have been over-discounted today. This doesn't mean the non-market models will be right. In many cases, they'll be completely wrong. But simply accepting the market's current valuation when we know the market is affected by unusual degrees of uncertainty is maybe not a good idea either. It may lead to us think our assets are worth far less than they will turn out to be worth, and therefore may cause us to miss real opportunities or to make poor decisions about what to do with our current assets.

What does all this mean for our policy toward the banks? Personally, I think it means we should not be too ideological about valuation methodolgies. I think we should look at both market and alternative valuations of bank assets. Market valuations should always have precedence, but for some purposes, some non-market valuation may actually be prudent and useful. Our present age is marked by extreme, almost fanatical, faith in markets. But markets aren't infallible and we shouldn't allow ourselves to be slavishly ruled by them. Mark-to-market accounting, like the free market, is a good thing most of the time. But there are exceptions and we may be in a period when exceptions should be the rule.