Here’s my simplistic, probably naive take on mark market accounting. In recent years some of these banks loaned $500,000 to buy California houses that in today’s market are only worth $250,000. Mark to market accounting would require them to value the asset (the mortgage) at $250,000 (the market price) rather than the nominal price the asset is carrying on their books (the original $500,000, minus any payments on the principal of the loan).
Banks don’t want that to happen, because it makes their books look bad. Banks have capital requirement ratios they have to maintain and that becomes difficult when the underlying value of their assets is shrinking. Hence the calls from banks to Washington to suspend or modify mark to market accounting.
Here’s an example that demonstrates why I don’t think we should suspend mark to market. Let’s say the bank has two $500,000 mortgages exactly like the example above on two identical houses next door to each other in the same subdivision. One homeowner is still making the mortgage payment. The other one stops making payments and the bank forecloses, eventually selling the house for $250,000.
The bank wants to pretend that the first house is still worth $500,000 even though the second, identical house only brought $250,000 on the open market. That’s what a suspension of mark to market accounting would accomplish.
If you really want to see how absurd that is, let’s swap positions with the bank. I’ll go to California and tell the bank I want a loan to buy that foreclosed house for $250,000. But I don’t want a $250,000 loan. I want a $500,000 loan. After all, that’s the amount of the original mortgage from a few years ago. As a comp I’ll use the first, identical house that the bank is valuing at $500,000 on its books.
Naturally, no bank in their right mind would let me claim a house is worth $500,000 in loan collateral when the market says it’s worth half that. Yet if regulators suspend mark to market rules on mortgages the banks will be doing exactly the same thing because it benefits them.
One reason this example is somewhat simplistic is that if the second homeowner is making payments the asset is generating revenue for the bank. One argument in favor of ignoring mark to market accounting, or not giving it full weight, is that some assets, like mortgages, have both an asset value and a revenue-generating value that needs to be taken into consideration.
Comments and criticism appreciated.
Some other peoples’ thoughts on mark to market here and here. You might also read the relevant part of the Wikipedia entry.
RSS Posts Feed
Les,
The main problem with Mark to Market is its volatility, and the fact that it is just as poor a method for accounting assets as the previous method was. In a down economy, it can be worse.
The problem with Mark to Market is that volatility in the market will lead to volatility to the balance sheet, obscuring the actual health of the company. If the value of the assets changes rapidly, it may be indicative of something OTHER than the actual value of the asset — a temporary situation.
The problem with Mark to Market is that bad news sticks around and good news is fleeting; if a bank or other company has to mark down an asset 30, 40, or 60%, people are going to remember that the company had asset problems and market confidence in the company will weaken, all over a temporary short term dip that actually indicated no problems in the company.
I think the best solution would be for some sort of tiered system where Assets are graded both by a market and assayed REAL Value.
But, that would be for a real economist to come up with.
You start seeing the full picture right at the end, almost. With a mortgage, the two values you mention can really be viewed as three. There is the revenue stream from monthly payments, which is what troubled firms were packaging and trading. The asset value has two faces. To the bank, it is only collateral, but for the mortgage holder it is property. Unless the mortgage holder defaults, the bank has no ability to liquidate the asset. If the market value of the home rises, only the mortgage holder has the option to sell and take the gain; the bank just gets the principal and can clear the capital for other uses.
This is why I think mark-to-market was more dangerous on the upside of the real estate market than the downside. On the downside, the revenue value remains unchanged but for the increased risk of default. Adjusting the asset value on the bank’s ledger just makes the loan appear undercollateralized. On the upside, mark-to-market makes loans look more secure, overcollateralized.
This was the fundamental problem: poor accounting of risk in MBS. Think of the hassle of constantly having to adjust the books to reflect market conditions over the life of a mortgage. Why would a bank want to do that? They wanted to do it on the upside because it distorted risk in their favor. They also distorted risk by bundling mortgages with default swaps and with complex probabilistic formulas. That was the big game that set up the markets for collapse.
I’d say mark-to-market is a practice that should be ended looking forward. Ending it retroactively is a more complex decision.
Volatility is part of it. But here’s the real problem: Last month the house was worth $500k, this month it’s $250k, next it’s $350k, next it’s $200k, next it’s $300k, next it’s $350k again before dropping to $175k.
So tell me, what is the asset “truly” worth? If volitility were the only problem you could try a 180 day rolling average, but people don’t buy on the average of the last 6 months. They buy *today*.
So, the real answer is “I don’t know”.
And since nobody else does either, no one really wants to buy the asset right now. What happens to prices when there are no buyers? This means, for all intents and purposes “I don’t know” = $0.
So, you have a $100mm dollar loan pool that may truly yeild between $75mm to $85mm in receivables. But, no one really knows for sure what that number is. Two years ago someone might pay $65mm for it figuring that even if they’re wrong they would still be ahead. Today, however, you compound uncertainty about the likelihood of repayment with fears about the general economy, inflation, unemployment and suddenly there are no buyers.
Now, this asset which has a non-trivial non-zero value (because barring total gov’t collapse in which case none of it will matter anyway there will still be many debtors that will pay back their loan) will, under MtM, be reported as worth essentially $0.
So, to return to your housing example, my choices are to either report that this $250k house is worth twice it’s real value or that it’s worth nothing at all.
Take your pick as to which is worse.
Good comments here so far. I quoted an article a few months ago on my blog, Les. (http://blog.craigdthomas.com/2008/09/alternate-solution-to-financial-hiccup.php).
Yu-Ain Gonnano points out the biggest problem. The uncertainty and the few bad assets in the package cause the perceived value to be driven to $0 because there are no buyers. However, in actuality, if you liquidated the houses in the package, they would bring much more than $0. They may only be worth 40%-50% of the original value, but it is more than 0% that they had to be marked down to.
Persimmon: that’s an interesting point about how MTM worked to the banks advantage on the way up and to their disadvantage on the way down. It’s the bookkeeping equivalent of being highly leveraged.
The frightening thing about this crisis is the number of leverage points. It is really hard to tell how far out of balance the trading value of debt bonds deviated from their worth.
We know they simplified risk equations by disallowing real estate values to drop in value, and there is evidence that buyers lost track of the inherent limits on debt assets versus the hypothetically unlimited nature of real estate assets.