Insight: True impact of mark-to-market on the credit crisis - FT

Back in April 1993 the eyes of the world were on the beseiged Balkan town of Srebrenica, which the UN declared a safe haven for Bosnian muslims, and on Northern Ireland, where secret talks between leaders from rival factions kick-started a tentative peace process. In the same month, a less-noticed development saw US accountancy regulators approve a rule that paved the way for today’s widespread use of mark-to-market accounting standards. This rule, which forced US banks to carry more securities at market value, emerged from the wreckage of the US savings and loan crisis when losses on loans had been hidden by the use of “historic cost” accounting. Only now, in the middle of a global credit crisis, is the impact of the broad introduction of mark-to-market accounting becoming clear. The critical concerns are around how much these changes helped to inflate the credit bubble and whether they will increase the speed and destructive power of its collapse. To be fair, the US banks protested at the outset that the move would change their role in the economy. So did the French banking federation before similar changes came to Europe in 2005. It warned that fair-value accounting “could even further increase the euphoria in a financial bubble or the panic in the markets in a time of crisis”. Tobias Adrian, an economist at the New York Fed, and Hyun Song Shin of Princeton University, have produced a string of work about this kind of “pro-cyclicality” in finance and the economy, culminating in a paper last September entitled Liquidity and Leverage. This paper examines the links between asset prices and the value of banks’ capital bases when mark-to-market accounting is used. It postulates that banks are driven to lend more and grow their balance sheets as the value of their capital rises. This is because they target a more or less constant leverage multiple on their balance sheets. The paper concludes it was inevitable that an industry buoyed by rising asset prices would pursue increasingly aggressive lending growth. This pushed credit upon ever more risky clients and loan structures, which then fed into asset price growth. This of course added more fuel to the fire – or created “positive feedback loops”. The most disturbing conclusion is that this system should behave in exactly the same way in reverse, creating “negative feedback loops” with a destructive impact on all kinds of asset values – from structured finance to house prices and equities. In a way this is nothing new – the old adage about a banker is that he gives you an umbrella when it is sunny and asks for it back when it starts to rain. But there are two important differences. First, fair-value accounting will speed up the process. One of Adrian and Hyun’s conclusions is that it was the speed of balance sheet expansion that caused the most blatant excesses of US mortgage lending. Second – this isn’t mentioned in the paper – there is the impact of securitisation, the practice of converting illiquid individual loans into saleable securities. This accelerated the speed at which banks could increase lending because it reduced the amount of capital needed for each new loan. It was so widely adopted because of the way it turbo-charged returns on capital. Banks’ use of pseudo off-balance-sheet vehicles to house securitised bonds further boosted this process, particularly from 2005 onwards, as can be seen in the asset-backed commercial paper market. What we will see over the remainder of this year is an ongoing painful reduction of capital values and leverage levels throughout the economy, centred around banking. Investors need to believe this re-adjustment has at least stabilised before they will return. Each new wave of price falls in leveraged loan markets or asset-backed securities markets – and each post-earnings restatement of losses from the likes of AIG, Credit Suisse or whoever is next – illustrates that the prognosis is not good either for the financial world or the wider economy. The lesson for regulators is that the solution to one problem almost always contains the seeds of another.