www.graincon.com
The new capital-adequacy rules have proved tricky to draw up
BANKS might imagine themselves to be the financial equivalents of Wal-Mart. Their regulators, however, are unlikely to be persuaded that they are watching over nothing more complicated than big shops. In most industries, companies are, or should be, allowed to go under if they become insolvent. Up to a point, the same goes for banks. Yet bank supervisors go to special lengths, and draw up special rules, to ensure that their charges have sufficient capital to keep them out of difficulty. These rules are due to change at the end of 2006. The changes are detailed, possibly far-reaching and certainly controversial. "It's the elephant in the room," says Andrew Kuritzkes of Mercer Oliver Wyman.
Left to take its course, the failure of a big bank can wreak economic havoc. Its depositors lose their money. The bank cannot meet obligations to other banks, which may fail in turn, ruining their depositors. The depositors of failed banks are likely to have creditors of their own, who will also suffer financially. In fact, the mere fear that a bank may be in danger of failing could be enough to spark a run, which would then spread to other banks. Most governments underwrite bank deposits, up to a limit, so that depositors have no need to panic should banks look wobbly. In the United States, the Federal Deposit Insurance Corporation (FDIC) guarantees that savers at insured institutions will get at least their first $100,000 back.
The trouble is that this creates a moral hazard for banks. Knowing that the state will bail out depositors, they might take more risks than they would if deposits were not guaranteed. Some banks, moreover, are thought to be "too big to fail". All this helps to explain why banks are subject to special regulations. They may be banned from doing things that regulators consider too risky, such as owning shares. Bail-outs may be conditional on the managers being sacked. The most important rules, however, are those setting minimum capital requirements.
The current rules, known as the Basel accord or Basel 1, were drawn up in 1988 and implemented in 1992. They require a bank's capital to be at least 8% of its risk-weighted assets. The weight for each class of asset ranges from zero (for assets thought to be very safe, such as the government debt of developed countries) to 100% (for unsecured loans to consumers and companies). In practice, the rules vary slightly across countries: in Japan, for example, shares in other firms can be counted as capital, and the minimum capital ratio for banks that are not internationally active is only 4%.
The plan is to replace this accord with a new one, Basel 2. During the 1990s, big banks' risk-management systems became far more sophisticated than required under Basel 1. Moreover, Basel 1 had an unintended consequence: its weights did not match the market assessment of the risks that banks faced. So banks indulged in "regulatory arbitrage": they disposed of risks for which Basel 1 required more capital than the market did, such as credit-card debt or residential mortgages; and they retained assets for which the market demanded more capital than the regulators did. The rise in banks' capital-adequacy ratios (see chart 1, next page) thus reflects regulatory arbitrage as well as better risk management. The ratios therefore mean less than supervisors would like.
Basel 2 is intended to bring regulatory capital requirements more into line with actual risk and to reflect improvements in the best banks' practices. The broad outline has been clear for some time, and almost all the details have now also been sorted out. The new set-up will have three "pillars". Pillar 1 requires banks' capital to be at least 8% of risk-weighted assets, the same as now: the weighting system, however, will change. Pillar 2 says that banks must plump up their capital cushions if national supervisors consider them too thin, even if they are above the minimum. Pillar 3 stresses the importance of market discipline, and says that banks should become more open about risks to their capital positions and profitability. This has already had an effect. In 1999, most big American banks kept mum about such risks in their annual reports; by 2001, the proportion was down to one quarter.
Three pillars of wisdom Pillar 1 breaks up regulatory capital into three parts, to match credit risk, market risk and operational risk. The market-risk element, dealing with trading losses, is unchanged from Basel 1 (which was amended for this purpose in 1997). The operational-risk part is new: it says that banks' capital should reflect the risk of mistakes and wrongdoing. An example might be a fine levied on a bank for overcharging credit-card customers.
For measuring credit risk, banks will have three options, each of which will assign more detailed weights to their assets than Basel 1 did. Under the simplest "standardised" approach banks will use external data, from credit-rating agencies and others. There are also two methods based on banks' own internal ratings. The more sophisticated of these is known as A-IRB.
All this requires a lot of data to be collected and processed, on which banks have already started. You need several years' worth of statistics to be able to assess the likelihood that the ball-bearing company to which you lent $500,000 will default next year, especially if you are allowing for the swing of the business cycle. The cost of compliance is correspondingly high. Mr Kuritzkes reckons it will be 0.05% of assets--or $100m-200m for a big bank--over the next few years.
For many activities, and thus for a fair few banks, the new regime implies big changes in capital requirements. The new operational-risk element, for instance, will hit banks specialising in areas--such as custody or asset management--that involve little lending and therefore in the past have needed little capital to keep supervisors sweet. In the credit-risk area, the weight attached to mortgages looks likely to drop by 30-50%, cutting the minimum capital requirements of banks specialising in home loans. Almost any bank with a lot of retail business--except for sub-prime lending and perhaps credit cards--can expect its minimum capital to fall.
However, working out the precise effects of Basel 2 is difficult. First, the rules are complicated and not quite complete. The main piece of outstanding business is the weight attached to lending on credit cards. In their accounts, banks include "expected losses" in operating costs. This is their best estimate, from past experience, of credit-card debts that will not be repaid. Because they have counted this as a cost, they say that minimum capital requirements should be calculated on the basis of credit-card assets net of expected losses. Originally, regulators disagreed, saying that expected losses should be counted in the asset base. Banks persuaded them to change their minds late last year. Now the formulas are being worked out again.
Second, the majority of banks, at least in America and Europe, are comfortably capitalised according to Basel 1. A change in their capital requirements might therefore make no difference: if they are not capital-constrained, their minimum could go up or down without changing their behaviour. Third, even if a bank's minimum requirements under pillar 1 go down, under pillar 2 its supervisors could still add to the minimum. Fourth, financial markets might frown on a bank that reduced its capital ratio, so even if Basel 2 relaxes its capital constraints, the market might not.
Most banks that will have to implement Basel 2 seem content with it, or at least resigned to it. Most of them, and their regulators, expect it to come into effect more or less on time--where, that is, it is being introduced. Every bank in the European Union, down to the last Sparkasse, will have to comply with Basel 2. In fact, many European banks think that new accounting standards, due to be adopted by all listed companies in the EU by next year, are just as big a worry. The proposed treatment of derivatives, in particular, could cause big swings in their accounts.
In America, only a few big banks will be required to switch to Basel 2. The Americans propose that only "big, complex, banking organisations" will have to adopt the new rules. In essence, this means the top dozen or so banks; however, other banks are free to switch if they want to, and regulators expect that a total of 20 or so will do so. America's Basel 2 banks, however, will be obliged to use A-IRB, the most advanced method of assessing credit risk. They will also have to employ the most sophisticated of the three ways of measuring operational risk.
Not everyone in the United States is happy about this. Mr Kovacevich at Wells Fargo, which will have to change over, does not see why it should: "We're large, but we're not complex. There are banks smaller than us and not in Basel 2 that are complex. We're not internationally active...I don't think that they have any right or credibility to say how we should manage our credit risk." The change will cost tens of millions of dollars, he complains, and regulators might not allow him to keep the "effective" risk-assessment system he already has. Indeed, "it might make us less effective." Wells Fargo's main competitors, he says, are small enough not to have to adopt the new system.
Small banks, and the politicians who speak for them in Congress, take the opposite view: that Basel 2 will benefit big banks at their expense. More generally, it is possible that having two sets of capital-adequacy rules will distort competition between Basel 1 and Basel 2 banks. Because of the complexity of the regulations and of America's banking markets, it is hard to tell from first principles whether it will or not. So economists at the Federal Reserve have been studying four areas of possible distortion: lending to small and medium-sized companies (SMEs); mergers; the mortgage market; and credit cards.
The first two studies were published in early March. In the SME market, the worry is that small, community banks might lose out because bigger banks that had adopted Basel 2 would have lower capital requirements and therefore would be able to lend at lower rates. This looks unlikely, finds Allen Berger, the author of the SME study. Community banks have an advantage in serving "opaque" SMEs, which publish little information but whose bankers know them intimately. Big banks' extra edge on interest rates--at most 16 basis points--would probably not overturn this. More at risk might be the bigger Basel 1 banks, which compete with larger institutions to serve bigger, more "transparent" SMEs.
Some have also argued that the dual system of regulation will increase the consolidation of America's banking system. Broadly, Basel 2 banks are likely to have lower minimum capital requirements than Basel 1 banks. They will therefore have an incentive to buy Basel 1 banks in order to free the extra capital that these institutions have to put aside. The Fed's study of this question finds little convincing evidence to support this. Past changes in regulatory capital or capital standards have had little significant effect on merger activity.
Regulatory competition However, there are other unresolved doubts about Basel 2. Officials at the FDIC point out that American banks are subject not only to Basel rules but also to domestic "prompt corrective action" (PCA) requirements. These were introduced to strengthen the banking system in the early 1990s. To be called "well-capitalised", a bank must have tier-one capital of at least 5% of its unweighted assets; below 4%, it is considered undercapitalised. The FDIC estimates that Basel 2 banks can expect their capital requirements to fall sharply--maybe by 30% on average over a business cycle. This is enough, says the FDIC, to drag the Basel 2 minimum below the level needed to keep equity ratios above 4% for most of a typical economic cycle. So regulators will have to choose between weakening the PCA framework that has served America well, and ignoring the implications of Basel 2.
Swing lower Another concern is that Basel 2 might make business cycles more severe, by causing banks to tighten credit too much during recessions and loosening it too much during booms. In a recent paper, Anil Kashyap, of the University of Chicago Graduate School of Business, and Jeremy Stein, of Harvard University, argue that this criticism is well founded.
In essence, the Basel 2 rules convert the probability that a borrower will default, the size and maturity of the loan and a bank's exposure at default into a capital charge. The higher the probability of default, other things being equal, the higher the charge. Thus banks must assign higher charges to riskier borrowers. And if recession increases everyone's chances of default, then capital charges rise across the board. To reduce the riskiness of their portfolios and to hold down the probability of default, banks will tighten their lending conditions. The danger is that this will make the recession worse.
In an ideal world, say Mr Kashyap and Mr Stein, capital charges would not vary as much across the cycle as across borrowers. The charge is, in effect, the price of capital to the bank: it should therefore reflect the scarcity of capital relative to the demand for it. In recessions, because bank capital becomes more scarce relative to lending opportunities, it makes sense to accept that during recessions the probability of default will rise.
Mr Kashyap and Mr Stein have estimated what the effects on minimum capital requirements might have been had Basel 2 been in force between December 1998 and December 2002, when there were marked economic slowdowns in both America and Europe. Using default probabilities calculated for some 17,000 firms by Moody's KMV, a credit-risk analysis firm, they estimate that minimum capital charges would have risen by an average of 36% over the four years. Under Basel 1, charges would have gone up by 22% in any case, making Basel 2's contribution 14%. For investment-grade borrowers, the effect is stronger: a 111% increase in capital requirements, of which 82 points are attributable to Basel 2. The impact on junk-rated firms is much smaller, presumably because their initial capital charge and thus the implied probability of default at the outset is high. When conditions tighten, many of them default (and thus drop out of the sample); for those that survive, there is not much room for their charges to rise.
These estimates cover only minimum requirements, and most banks will have much more capital than they need, so the effect of recessions on actual capital might be much less dramatic. A lot might rest on pillar 2 of the new accord--ie, on whether supervisors allow banks to lose a bit of padding during recessions. A lot depends also on the timespan over which default probabilities are estimated: longer periods of calculation would make capital requirements less sensitive to business cycles. Mr Kashyap and Mr Stein found the effects were somewhat smaller when they used default probabilities based on Standard & Poor's credit ratings, which change less over a cycle than those of Moody's KMV.
So there is still much to argue about before Basel 2 comes into effect. Despite regulators' attempts to deal with the objections, such a complicated set of rules is almost sure to throw up unintended consequences that no one has yet spotted. "Certainly this is not going to eliminate regulatory arbitrage," predicts Mr Kuritzkes. Basel 1, he says was "more humble", as regulators tried to establish good principles. Basel 2 is a much grander, and perhaps riskier, endeavour.
All rights Reserved; The Economist
The opposition against Turkey in the EU has begun to present an ugly face 
