Basel II framework: a work in progress

On June 26, the Basel Committee on Banking Supervision published its long-awaited revisions to the 1988 Capital Accord, known as Basel II, and formally titled it International Convergence of Capital Measure and Capital Standards: A Reviewed Framework. The Framework will serve as a blueprint for the national regulatory authorities to issue implementation guidelines, and supervisors will have a certain amount of discretion to craft standards to best suit their own national banking systems. In the U.S., for instance, only the most "Advanced" approach will be applied, although the Framework provides three methods for computing regulatory capital requirements (Standardized, Foundation, and Advanced). Also, in the U.S., the Advanced approach will be used only for the largest, most internally active institutions (banks with more than $250 billion in assets or more than $10 billion in cross-border exposures), although other institutions may voluntarily adopt it, provided they can meet its rigorous validation standards.

As the formal title of the Framework implies, Basel II is designed to align required regulatory capital with risk measurement. When the process to revise the Basel Capital Accord began more than six years ago, most believed that the old Capital Accord was not risk sensitive, as it assigned a flat 8% capital charge to all exposures, regardless of their relative risk. Advances within the industry had moved far beyond a simple one-size-fits-all capital requirement. In 1995 the Basel Committee allowed the industry to use an internal models approach to assign capital for market risk, and the industry argued that it should be allowed to use its own internal risk-estimation procedures for credit risk as well.

Basel II does not yet allow an internal models-based approach for credit risk, but it does represent a big step forward in that it allows the industry to use its own internal estimations of probability of default (PD), loss given default (LGD), and exposure at default (EAD). However, it leaves much work still to be done since it does not allow the industry to use its own internal models for capital assignment purposes, but instead requires that a Basel II regulatory model be used.

Implementation of the Advanced approach is set to begin in 2008, although a "parallel run" period begins in 2007, in which institutions will be required to compute capital under Basel II as well as the existing Accord. A two-year transition period will also be in effect beginning in 2008 in which capital floors will remain in place, allowing regulators to assess the overall impact of the new Accord. U.S. regulators have stated repeatedly that they may recalibrate the Accord if they are not satisfied with the results. Indeed, the Basel Committee itself stated that it will review the calibration of the revised Framework prior to its implementation and could apply a "single scaling factor," possibly 1.06%, based on the results of the parallel run.

Many details of the Accord remain a work in progress and could very well remain so for quite some time, particularly given the number of issues that remain to be resolved by the Basel Committee and the overall level of national discretion contained in the Framework. Nonetheless, U.S. regulators intend to get a jump start on analyzing the overall capital impact of the New Accord this fall when they will conduct a Fourth Qualitative Impact Survey. Results of the survey will be released in the first half of 2005.

As stated in the Framework itself, "the Committee has designed the revised framework to be a more forward-looking approach to capital adequacy supervision, one that has the capacity to evolve with time." Moreover, the Committee stated that the Advanced approach ...

   represents a point on the continuum between purely
   regulatory measures of credit risk and an approach
   that builds more fully on internal credit risk models. In
   principle, further improvements along that continuum
   are foreseeable subject to an ability to adequately
   address concerns about reliability, comparability, validation,
   and competitive equity. In the meantime, the
   Committee believes that additional attention to the
   results of the internal credit risk models in the supervisory
   review process and in banks' disclosures will be
   highly beneficial for the accumulation of information
   on the relevant issues.

While the Committee's intent to have the Framework evolve is encouraging, many are finding it difficult to begin the implementation process when the Framework remains a work in progress. Supervisory guidance necessary to implement the Framework is also far from complete. In the U.S., draft guidance for Advanced Internal Rating Systems for Corporates and the Advanced Management Approach for Operational Risk were released last summer. Draft supervisory guidance for retail exposures is expected by the close of 2004. Additional draft guidance for residual corporate, commercial real estate, securitization, and other wholesale has yet to be released. U.S. regulators plan to release final guidance for all exposures for industry comment by mid-2005 and plan to publish the Final Rule in mid-2006.

With the parallel run set to begin in 2007, institutions that are required to adopt the Advanced approach have little choice but to begin implementation before the final guidelines are issued. And many are finding this prospect troubling, since a number of the systems changes necessary to capture and warehouse the multiple data requirements of the Advanced approach will be quite costly. Moreover, many within senior management believe that they should not be forced to adjust their internal risk management systems just to satisfy Basel II requirements, particularly if they are not yet finalized.

To help resolve this problem, U.S. regulators have suggested that institutions adopt formal, written Implementation Plans, based on the draft guidance issued to date, and that institutions share those plans with their supervisors. Most institutions that are required to adopt the Advanced approach have formal policies in place outlining their internal risk measurement and management procedures. U.S. regulators would like for the Implementation Plan to discuss how internal policies can be mapped to requirements specified by Basel II to date. Moreover, they have stated that the qualification process will not be a one-time event, but will be "iterative," evolving over time much as the new Accord is designed to evolve.

This will no doubt be, as the saying goes, "easier said than done." Most of the Basel II requirements reflect best practice within the industry. However, in a number of places, Basel does not reflect best practice. Over the past six years, the industry and regulators have worked together to improve the Framework, moving it closer to best practice. This process will continue, but it is a timely and costly process, and many already suffer from "Basel fatigue." The fatigue is all the more tiring when such basic concepts as the definition of capital and a reexamination of leverage ratios are not even being considered at the moment. Alas, while much has been accomplished, much more remains to be done.

Pamela Martin can be reached by e-mail at pmartin@rmahq.org.

[c] 2004 by RMA. Pamela Martin is executive editor of The RMA Journal and director of Regulatory Relations and Communications at RMA--The Risk Management Association.

Basel Time Frame

End-2004--Draft Supervisory Guidance for Retail

1Q 2005--Results of Qualitative Impact Survey

Mid-2005--Publication of Guidance for All Portfolios

Mid 2006--Final Rule of Guidance

2007 Parallel Run (data maintenance requirements begin--three-year mininum)

2008--Implementation (with floors in place)

2010--Floors removed (five years of PD and seven years of LGD data required)

COPYRIGHT 2004 The Risk Management Association
COPYRIGHT 2005 Gale Group