Friday 17 October 2008

What's all the fuss about marking to market?

On Sunday European heads of government discussed accounting standards. That's pretty unusual. On Monday the International Accounting Standards Board (IASB) had an emergency meeting. On Tuesday Trustees of IASB had a conference call to suspend the IASB's due process. On Wednesday the Accounting Regulatory Committee of the EU met to consider the IASB's proposal. Such haste is unusual. What was all the fuss about, and why is accounting intimately involved in the solvency of banks?

It all comes down the technicalities of an IASB accounting standard IAS 39 on Financial Instruments. Financial instruments have been a problem for financial reporting because a lot of people think that historical cost accounting doesn't do a good job of reporting them. For one thing you can acquire financial instruments without laying out any immediate cash, so the cost is zero, and for another the value of the things can fluctuate wildly and may not bear any relationship to the cost, if any. IAS 39 tries to solve this problem by classifying financial instruments into several boxes - I won't go into the details.

The main importance of which box instruments get classified into is that some boxes allow the instruments to be valued at 'amortised cost' on the assumption that the company will hang onto them until the end of their lives. Other boxes require the instruments to be valued at 'fair value' - which is mostly market value. This is what is called 'marking to market'.

Now one of the main effects of the credit crunch has been that the market price of many instruments, especially collatoralised debt obligations (CDOs - rights to the cash flows from sub-prime mortgages and similar things), have collapsed because nobody wants to buy them. If the holders of CDOs are required to value them at market value then that creates a huge hole in the balance sheet. And if the holder is a bank, that can mean that the bank either has to raise a lot of new capital, or reduce its lending drastically in order to maintain its required level of reserves.

What the banks want to do is to revalue the CDOs and similar on an amortised cost basis, which means they look at the likely payments that they are going to receive over the life of the CDO and value that stream of cash flows, using a net present value method. Since everyone seems to agree that, although currently unsaleable, most of the CDOs will eventually produce a respectable stream of cash flows, this will result in a much smaller balance sheet hole and make life a lot easier for the banks.

Sounds like a good idea? Well, yes, it probably is. But if it works in one direction it works in the other. That is to say, the reason that banks got into the current pickle is that they marked up the value of financial instruments when the market was valuing them above a realistic value, and the banks didn't complain then. In fact they declared record profits and paid their senior executives and employees mind-blowing bonuses.

My personal take on this is that marking to market always was a bad idea and encouraged market instability. Accounting should be providing a reality check on markets, not reflecting their errors back at them. The head of IASB has often defended market value by saying, in effect, that it is the only way of portraying reality. But that's a view that is rooted in a touching faith that markets can't be wrong. Users of financial reports want to know what the long term cash flows that they'll receive from various financial decisions will be.

Market prices reflect the value of those long term cash flows, but with an enormous amount of error. If the market gets the value of one bank's CDOs wrong, it gets the value wrong for all banks, so there is a huge 'systemic' effect. Moreover, if an upwards error in market prices is reflected in profits and bonuses, then banks buy more CDOs pushing up the prices still further and creating incentives to make more CDOs. When the error reverses, the herd instinct among investors, ensures that they all try to sell at the same time and prices collapse.

'Okay, clever clogs', I hear you say, 'what would you do?' Well, I'd use historical cost principles. Either, I'd treat the assets as 'held to maturity' and value them on an amortised cost basis. Or I'd treat them as trading items and value them at the lower of cost or net realisable value, as is the practice for inventory. For these purposes 'cost' could be zero for things like interest rate swaps which don't require up front payments. I'd also require extensive 'footnote' disclosures showing market values, but I wouldn't incorporate these in the balance sheet and income statement which form the basis on which contracts are written and solvency calculated.

For more on why marking to market is not such a good idea see John Kay's FT column of 14th October.

No comments: