Discussion: (0 comments)
There are no comments available.
Statement No. 323
View related content: Financial Services
One of the many requirements of the Dodd-Frank Act is that all federal agencies must remove references to and reliance on credit ratings from their regulations and replace them with alternative methods for evaluating creditworthiness. These requirements have potentially sweeping implications, including their impact on the Basel bank capital standards, which have relied since Basel I (and now Basel III) on credit ratings to assign risk weights to assets for purposes of computing required minimum capital requirement for most banks.
On December 21, 2011, the federal banking agencies proposed guidelines for U.S. banks to use to measure the risk weights on assets when implementing the risk-based capital rules. For the most part, these guidelines are based on models of varying degrees of complexity, while giving a nod to, but ultimately rejecting, alternative market-based indicators, such as bond yields or CDS spreads.
Over the years, the Shadow Financial Regulatory Committee has argued (beginning with Statement No. 44, September 18, 1989) against the wisdom and usefulness of determining minimum capital standards based on arbitrary risk weights for bank assets. Instead, the Committee believes that bank regulation would be better served by replacing the risk-weighted standards with a leverage ratio standard.
The bank regulatory agencies’ proposed methods for replacing credit ratings, while well-intentioned, strengthen the case for abandoning risk weights. As summarized in the Appendix, which simplifies the detailed proposals, the proposed methods for replacing credit ratings represent an additional set of arbitrary tests and models that further complicate the risk weighting procedures already contained in the Basel II and III framework. These new methods would not be necessary if the minimum capital standards were defined in terms of a simple leverage ratio.
If, however, the agencies retain the risk weighting framework, the Committee encourages the agencies to adopt the hierarchy already established by U.S. and international accounting standards for valuing bank assets. The hierarchy recognizes market prices as the appropriate measures for assets that are traded in orderly markets, but also recognizes that such prices are not available for all assets. Where market prices for identical assets do not exist, the second-best prices according to the standards are those for similar assets, with adjustments for differences in assets being valued versus the comparables (Level 2 valuations). If there is no identical or sufficiently similar asset, then banks use “Level 3” valuations, which are model-based and which maximize the use of available market-based inputs and minimize the use of unobservable inputs. The accounting profession worked to define a hierarchy for using market data in asset valuation before the crisis and revised and improved their approach as the crisis unfolded to take account of the implementation issues that arose. In short, if the agencies continue to use risk weights in lieu of a leverage ratio to determine a bank’s capital requirement, there is no reason for them to proceed down a new path when a well-established approach to determine market data reliability is already in place.
The Proposed Alternative Risk Measures
The agencies’ proposal offers various different measures of risk, depending on the type of credit. We review only the most important credit categories outlined under the proposal in what follows.
The most controversial aspect of the agencies’ proposal relates deals with its proposed methods for risk weighting securitization provisions. Under the current Basel system, securitization risk weights are determined by credit ratings, and the highest-risk securities must be backed entirely by Tier 1 capital.
The proposal sets out a series of complicated new models, which in essence, would require a bank that held every tranche of a securitization to hold more capital than if the bank held the underlying assets in the portfolio. The proposal specifically explains that the purpose of this result is to reduce the ability of banks to engage in “regulatory arbitrage” through the use of securitization.
The agencies considered using or allowing third party vendors to assign ratings, much as the National Association of Insurance Commissioners does now for insurance companies. The proposal rejected this idea, however, primarily because it entails the same basic problems with using credit ratings now, including the limited number of vendors and the potential for conflicts of interest when a vendor retained by the agencies also does business with issuers.
For sovereign debt, the European Union assigns a 0 risk weight to all debt issued by governments belonging to the OECD, including Greece, and a 100 percent risk weight to all other sovereign debt (implying that such debt has to be backed by 8% of “Tier 1” capital, or essentially equity).1 Under the proposal, banks would calculate their risk weights according to OECD Country Risk Classifications (CRCs), which are used to reflect the risk of non-payment of export credits. The CRCs have seven risk categories are regularly updated for over 15 countries, and most importantly are recognized by the Basel Committee on Bank Supervision (BCBS) as an alternative to credit ratings.
Nonetheless, the agencies recognize that there are drawbacks to the CRCs: (1) they automatically apply their best classifications to high-income countries, which may not reflect a particular high income country’s relative risk of default; (2) while CRC uses qualitative factors to assess risk on a monthly basis, the quantitative factors it uses are available only on annual basis; and (3) in some cases, the CRC may not fully reflect the risks of sovereign debt restructurings. For all these reasons, the agencies propose supplementing the CRC classifications with a requirement that 150% risk weight be applied to any sovereign debt where the government has defaulted on any exposure in the past five years. In addition, the proposal would continue to assign a 0 risk weight to debt issued by the US government (but a modestly higher risk weight for GSE debt, up to 20%), while sovereign entities having no CRC would be assigned a 100 percent risk weight.
The agencies considered market-based alternatives for assigning sovereign risk weights, including CDS spreads or bond interest rate spreads. The agencies acknowledge that both of these indicators could be more “forward looking” than the measures it proposes. Accordingly, the proposal seems to suggest that the agencies might in the future refine their sovereign risk weighting methodology to permit banks to use as a risk weight the higher of the weight indicated by the CRC or a five-year CDS premium (the five year CDS being the most liquid of these kinds of contracts). To limit volatility in the CDS measure, the proposal suggests using a one-year rolling average of the CDS premium.
The agencies also recognize the potential usefulness of bond interest rate spreads, but note that most sovereign debt is denominated in a home country’s currency, which would make it more difficult to compare spreads of different countries’ debt. This problem could be solved for those countries that issue dollar or Euro denominated debt, but this would cover a much smaller group of countries than the CRC classifications, and often these dollar or Euro bonds are issued infrequently and in small amounts.
For corporate debt, the current Basel rules generally assign a 100 percent risk weight to all corporate debt, but only a 20% risk weight to bank claims on or guaranteed by securities firms. The proposal claims it would permit a bank to use a methodology that uses both market-based and historical information to assign risk weights to corporate, non-financial debt issued by publicly held companies, and to assign a 100% risk weight to all other corporate debt excluding exposures to depository institutions. Broadly speaking, the risk weights vary from 0 to 100% for publicly held debt, except for a 150% risk weight assigned to “high risk” debt that is more than two categories below investment grade or its equivalent based on a bank’s internal ratings. In addition, the weights are also determined by a combination of “indicators,” including measures of leverage, cash flow and monthly stock return volatility. While the proposal says the agencies considered the use of market-based alternatives (such as CDS and bond interest rate spreads) as well, it says only that each of these alternatives had “significant drawbacks”, and that they would study these alternatives further (the proposal did specifically identify several problems with bond spreads, however: that they allegedly frequently misprice risk, reflect factors other than issuer-specific credit risk, and that the spreads can vary a great deal over short time periods). In addition, the proposal summarily concludes without providing specific reasons that all debt issued by financial institutions will be assigned a 100% risk weight.
The proposal says that the agencies are still “considering” whether to use “investment grade” determinations for higher-risk debt issues, in particular, pointing to the OCC’s investment securities regulations which require banks to determine whether or not a security is investment grade. In this respect, the proposal seems to elide without explanation the requirement in Dodd-Frank motivating this entire exercise that agencies abandon credit ratings in their rules. Nonetheless, the proposal does not permit sole use of credit ratings for this purpose since it also requires that banks supplement the ratings with “due diligence processes and analyses that are appropriate for the bank’s risk profile and for the amount and complexity of the debt instrument.”
There are no comments available.
1150 17th Street, N.W. Washington, D.C. 20036
© 2015 American Enterprise Institute for Public Policy Research