Reforming finance: Bank capital

Target practice Jul 9th 2009 From The Economist print edition How much capital is enough? The second in a series on financial reform ASKING how tight gun-control laws would need to have been to have prevented any assassinations in Sarajevo in 1914 is not a rewarding exercise. For banks, however, looking at “what if” scenarios is more worthwhile. Most regulators now think that gradually building up capital buffers as the economy recovers is the best way to make banks safer. Britain’s Treasury was the latest to tread this line in a white paper unveiled on July 8th. For a long time the amount of capital required to withstand a genuine meltdown has been a near-complete unknown. The minimum capital levels set by the Basel 1 accord in the late 1980s in effect endorsed the overall amount of capital in the system at that point. The crisis changes all that. Estimates of ultimate losses remain highly uncertain and the state of the economy is unclear. But regulators now have a real-life example of a systemic collapse with which to calibrate their new rules. So how much capital would the system have needed to have survived the crisis unscathed? It depends how you define “unscathed”. Assume, in an ideal world, that no bank should be in the position of having to raise equity once a crisis has begun. And assume, too, that at no point in the cycle should its level of core capital (basically, its common equity) drop below the minimum now demanded by the American and British governments of 4% of risk-weighted assets. Based on the estimates used in the American government’s stress tests, the 19 firms involved collectively need to hold minimum core capital of $313 billion. Between the start of the crisis in mid-2007 and the end of 2010, these firms are expected to eat up about $317 billion of capital (with gross losses of almost a trillion dollars offset by core earnings). On this basis, they would need to have entered the crisis with $630 billion to have avoided breaching the floor. That is equivalent to 8.1% of their current risk-weighted assets. The American and European systems actually had a core-capital ratio of nearer 6% at the end of 2007, according to the Bank of England. The Basel rules allowed core-capital ratios to go as low as 2%. If anything this calculation is generous. Using Bloomberg’s higher estimates of the losses recorded so far, a peak core-capital ratio of nearer 10% would have been needed for the big American banks. And of course losses are not evenly distributed across the system. Citigroup would have needed a starting core-capital ratio in the mid-teens. Using a similar methodology, Switzerland’s UBS would have needed a pre-crisis ratio of about 12% to have avoided hitting the floor of 4%. This is not far off the level it actually had. The fact it has still raised so much equity reflects a tough regulator and the perception that to command the confidence of its private-banking customers, its minimum core-capital must always be far above the 4% level. That points to one problem with the exercise: a 4% minimum at the bottom of the cycle may be too low. There are others. Estimates of losses will be wrong. Bulls point to the recovery in the prices of risky assets such as leveraged loans. Bears worry that unemployment will rise above the level assumed in the stress tests. There is also a risk of double-counting. The adjustments used to calculate risk-weighted assets are becoming more conservative, as regulators seek to penalise certain types of activity, especially trading. That will also have the effect of building up capital, even if the minimum ratio remains unchanged. Still, for all these caveats, the basic message seems pretty clear. Going into another crisis of this scale, absolute capital levels would need to be at least double the existing minimum level to avoid breaching that floor at the trough of the cycle. If regulators think the worst banks should set the benchmark for all in the future, the buffer could be even higher. For banks, that is bad news. Their best chance is to persuade regulators that a repeat is unlikely, but such pleading may fall on deaf ears. Right now, governments are interested in stitching taxpayers a bulletproof vest.