Credit derivatives users argue that the proposed rules misstate risk in a variety of ways.

By Nina Mehta

How much capital should a bank hold against possible credit losses? That multibillion-dollar question is one of the most controversial topics in banking—and one that has a direct impact on the bottom line of the world's most powerful financial institutions.

The group that has the most say in determining the answer to that question is the Basle Committee on Banking Supervision, which tries to harmonize the banking regulations of the 12 G-10 countries to ensure that the same regulations are applied to the same countries whenever possible. In 1988, the Committee issued a Capital Accord, which laid out a capital adequacy framework for the credit risk on banks' books.

The sniping about the inadequacies of the Accord began shortly after it was released, and formed a continuous drumbeat of protest that colored regulatory discussions throughout the 1990s. Over the years, banks lobbied furiously for a more risk-sensitive capital allocation process that was more reflective of the actual risks faced by banks—and by the capital markets more generally.

The birth of the explosive credit derivatives market and the increasing complexity of credit risk management tools that slice and dice risks in previously unimagined ways have made these demands all the more pressing for regulators, especially as banks became more adept at regulatory arbitrage.

Last June 3, after years of quiet deliberation, the Basle Committee finally unveiled its overhaul proposal for the capital treatment of credit risk. But in the months since, this has produced a new round of aggressive sniping and complaints.

The granular way

The Basle Committee has accepted the industry notion that improved granularity in risk weightings will prompt increased risk mitigation efforts on the part of banks. To this end, it has proposed the use of external ratings from established credit rating agencies to determine risk weightings for capital charges—in what it refers to as a standardized approach to capital charges for credit risk. Sovereign as well as corporate debt will thus be risk-weighted according to ratings published by Standard & Poor's, Moody's Investors Service and other agencies.

At the same time, the Basle Committee announced it was considering allowing banks to use their own internal risk ratings to determine capital adequacy. A bit further down the road is the possibility that credit risk portfolio models might be used to determine capital adequacy levels as well. That, naturally, would provide spectacular granularity in assessing risks across portfolios—and, for banks, managing capital more productively.

BANKS are likely to rush to lend to aa- and above credits. at the same time, they will do credits in the junk category above ccc+.” —Jeremy Henderson

But that's a task easier said than done. And for now, critics see a slew of problems with the current proposal.

The core problem, they say, is that the chalk lines between weightings are at best indiscriminate and are unlikely to result in uniformly better risk management practices. The thresholds are considered nearly arbitrary. Given the 8 percent baseline capital charge on debt, corporates rated AA- and above will be weighted at 20 percent; those rated A+ to B- will be weighted at 100 percent; and those rated below B- will be weighted at 150 percent. Not everyone, however, sees a big difference between an AA- and a A+ credit. "That creates pressure not only on the rating agency process but also on a bank's management of its portfolio,” says Jeremy Henderson, managing director and head of marketing for derivatives and financial products at Société Générale in New York. "Banks, for instance, are likely to rush to lend to AA- and above credits, of which there are not enough to feed all the global bank balance sheets. At the same time, they will do credits in the junk category above CCC+.”

Dennis Oakley, managing director of global credit and portfolio management at Chase, also sees a moral hazard issue around the bend. "You don't want to incent banks to get rid of the good credit risk and keep the bad credit risk. You don't want us to deal with OECD banks that are poorly rated as opposed to highly rated insurance companies for credit derivatives. There's a lot of moral hazard in the current Accord, and I don't see a big improvement in the proposal to get rid of that.”

Another anomaly crouching in the consultative document's risk categories is the fact that unrated credits are weighted at 100 percent. This gives them an advantage over low-rated debt. Kim Petra Olson, an assistant vice president at the New York Federal Reserve Bank, who helped work on the Basle Committee's June 3 document, concedes that unrateds can run the gamut from high quality to low quality. In the end, though, because of the lack of a credit differentiation technique for the standardized approach, she says, it was impossible for the Basle Committee to finesse the issue any further, particularly since differences in ratings penetration across countries could result in competitive disadvantage if unrateds were weighted at more than 100 percent. The result: this, too, could put pressure on credit departments to look more winsomely at high-yield junk credits that are unrated—and therefore only 100 percent weighted.

Of course, the Basle document is based on the views and needs of a dozen countries, with—quite often—discrepant opinions about the function of capital allocation, given their different national histories and accounting regimes. So agreement on the need for a more protean, risk-responsive capital adequacy framework, and the acceptance of both external ratings for risk weightings and a proposal to rely eventually on internal risk ratings, is itself an accomplishment.

Mitigating circumstances

Olson notes that the consultative document doesn't offer a fleshed-out proposal for how to recognize credit risk mitigation techniques in lowering the capital charge for regulatory capital. But, she adds, the Committee is actively seeking industry comments and advice on how to strike a balance between simplicity and complexity in determining minimum capital charges, rewarding good credit risk management practices, and minimizing capital arbitrage—while also ensuring that banks hold enough capital against unexpected losses.

Indeed, she points out, the current overhaul of the 1988 document was set in motion in large part because regulators realized that regulatory capital arbitrage didn't always reduce risk. The retained residual risk in securitizations and credit enhancements, for instance, was frequently higher than the 8 percent capital charge—and internal economic capital within banks was not always being diligently applied against that.

While the June 3 consultative document isn't detail-studded on the credit risk mitigation front, it nonetheless proposes recognizing a fuller range of such efforts. On collateral, for instance, the document proposes expanding eligible collateral beyond cash, OECD government securities and the securities of multilateral development banks to include all financial assets with readily determinable value. Other recommendations concern guarantees, greater recognition of credit derivative products and, potentially, on-balance-sheet netting.

The Committee's desire to doff its cap at a wider range of credit risk mitigants, however, has thrown two central, quantitative concerns into high relief. The first is how much risk-reducing benefit should be recognized by regulators, and the second is how to address the residual risks that arise from, say, basis risk as a result of changes in market prices, which could cause the value of the underlying instrument and the value of the hedge to change differently over time.

IF you're a bank, you want other banks to be held to a sufficiently high standard in terms of their internal risk rating process, so they can't simply game the system and get lower capital charges.” —Kim Petra Olson

Other significant residual risks that regulators are trying to wrastle to the ground include maturity mismatches, in which the hedge doesn't cover the full maturity of the underlying; asset mismatches, in which the asset to which the underlying is referenced for payout is different from the underlying; and double defaults, or two-name paper, in which the risk of loss to the bank is based on the probability that both the underlying and the hedge will default.

True to form, there's a disconnect between regulators and credit portfolio managers on some of these issues.

Certain mismatches make sense economically, but not for regulatory purposes—and therefore don't help a bank manage its capital effectively. "There's clearly some benefit to doing a one-year default swap against a seven-year underlying,” opines SocGen's Henderson, "and whether the capital reduction should be significant or minor, it should clearly be more than zero. It's extremely frustrating to pay for something that gives you absolutely no benefit. And if it gives you internal benefit but not external benefit, it's just that much less useful.” The same can be said for monoline guarantees and other transactions that skim off certain risks but aren't yet recognized by regulators as risk-reducing.

It's unlikely, of course, there will ever be a perfect overlap between regulatory-capital requirements and economic capital needs, since regulatory-capital requirements are not likely to provide the most up-to-date, precise indicators of risk. Still, regulators are seeking to come up with risk-reflective capital requirements that can be applied to banks with different internal risk management schemes.

In evaluating whether to recognize double-defaults in a standardized approach, for instance, the Committee is grappling with how to address the slippery nature of correlation. The current Basle Accord substitutes the collateral's risk weight for that of the underlying risk weight, assuming a correlation of 1, but the issue now is how to take shrewder correlation metrics into account when assigning capital charges for credit risk positions. This is difficult to do in the context of a standardized approach without relying on a model—and when relying on a model, it's important for banks to ensure that their data are robust and that their estimates of correlation will hold up during times of economic downturn. This boils down to what's becoming an increasingly standard worry about models for banks: Just because a model walks like a duck and talks like a duck, does that mean it truly is a duck under all market conditions?

Another worry concerning correlation in the standardized approach is that banks will simply go out and buy default protection from unworthy sources, such as banks with low credit ratings. That could lead to a daisy chain of cross guarantees, which, in a downturn, could undermine financial stability in the capital markets.

Regulators are, however, trying to march across the frozen tundra of complex credit derivatives even as these structures make further inroads in the market. Take synthetic collateralized-loan-obligation structures, for example. On November 17 last year, the Federal Reserve Board of Governors and the Office of the Comptroller of the Currency issued interagency guidance regarding the preferential capital treatment for synthetic CLOs and other instruments. According to that document, if a sponsoring bank hedges a portion of a reference portfolio through a series of credit default swaps and credit-linked notes and retains a high-quality senior risk position that meets certain criteria, it could receive a risk weighting of 20 percent instead of the normal 100 percent. "This is quite an important development on the regulatory front,” says Olson, "since it seeks to align the risk weight with the high-quality nature of the underlying.”

Naturally, one of the growing concerns is that the current smorgasbord of sliced-and-diced risks could become too complex for banks to assess adequately. Regulators, for their part, consider it essential that banks appropriately assess the risk of complex transactions and hold adequate capital against those risks. Consequently, the Federal Reserve provided supervisory guidance last July when it issued a letter announcing its expectation that banks engaging in sophisticated transactions "have their own internal capital allocation processes that are sound, and have a robust way of determining the capital they need to hold, particularly for complex transactions,” notes Olson. It would then be up to regulatory supervisors to review banks' internal capital allocation processes.

While many in the markets believe the days of regulatory-capital arbitrage—or at least pure regulatory-capital arbitrage—are winding down because of improvements on the regulatory front, the use of credit derivatives and structured credit risk management tools is on the rise. And credit derivatives will clearly get a boost in the future—from tax advantages, the desire to take on previously unavailable or economically inefficient exposure to corporates, the increasing ease of transferring credit risk to the capital markets, and the irreducible fact that credit derivatives are sometimes the only way to hedge long positions. But equally obvious in this brave new world is that the regulatory process must be dynamically risk-responsive in its capital requirements—even as it sets out rigid hoops through which banks must jump.

The Lure of Self-Determination

For now, the internal risk ratings approach to determining capital adequacy for credit risk is the road not taken, although all signs are that this is the path down which regulators will head. The key elements currently being worked on are best-practices issues, the creation of sufficiently stringent regulatory parameters to prevent gaming and the need to ensure proper validation processes. "If you're a bank,” notes Kim Petra Olson, an assistant vice president at the New York Federal Reserve Bank, "you want other banks to be held to a sufficiently high standard in terms of their internal risk rating process, so that they can't simply game the system and get lower capital charges, while you, because you're very conservative or are supervised by somebody else, are held to an extremely high standard.” The more sophisticated moneycenter banks, of course, see the internal risk ratings approach to regulatory relief as the magic bullet that could help align regulatory capital with economic capital. By allowing theoretically infinite granularity, internal risk ratings methodologies could stimulate better capital management by denaturing the appeal of regulatory capital arbitrage. Dennis Oakley, managing director of global credit and portfolio management at Chase, argues that this will only happen when the Basle Committee takes a full-model approach to credit risk—a model, that is, that takes into account correlation across instruments in a bank's entire portfolio. Although Chase understands the impact of regulatory capital, its protocol is to look instead at economic capital to determine how a new transaction or securitization will affect the credit quality of the bank's entire credit portfolio. Most banks, however, aren't prepared to ignore the return on regulatory capital quite so blithely. Indeed, the largest credit derivatives transactions for many banks continue to be centered around regulatory relief.

—N.M.