17 November 2005

FT: Chance to correct the defects of the Basel II framework

Paul Sarbanes, the senior Democrat on the US Senate’s banking committee, warned last week that Congress could still stop the Basel II accord on banking regulation from being implemented in the US. He was questioning the necessity of the accord in the US following problems encountered with the testing of the rules on US banks and a consequent decision to delay the accord’s implementation for one year. The International Institute for Finance, the primary lobbying group for the world’s largest banks, expressed its displeasure at the delay.

The test, the fourth quantitative impact study, shows that bank capital of the largest US banks would fall by about 20 per cent on average following implementation of Basel II, but with a wide difference in impact on individual banks. The unexpected variation in amount of capital between banks reflects different assumptions in the measurement of risk. Since low bank capital is a competitive advantage, the implications of the study are likely to be a race to the bottom in the quality of risk assessment. Basel II gives banks considerable leeway in how they measure risk.

The central idea of Basel II is that banks estimate their risk by means of internal models and use these risk estimates to calculate minimum capital, an approach known as “internal rating based”. Regulators want overall bank capital to remain relatively unchanged following the switch to Basel II and use quantitative impact studies to measure the likely outcome of particular regulatory mechanisms. This requires the rules guiding how banks calculate their risk to be sufficiently unambiguous that different banks measure the same risk in roughly the same way.

However, the results from the US impact study should not have come as a surprise since the lack of reliability of financial risk models and their sensitivity to specific assumptions have been widely documented. Herein lies the dilemma facing the architects of Basel II: they can either become more prescriptive, and lay down specific standards for risk models, or use discretionary capital to top-up capital to the level felt appropriate by the supervisors. Neither approach is appealing.

The first option implies that bank risk methodologies become more harmonised, leading banks to react to crises in a similar way. Consequently, business cycles are exacerbated and the financial system becomes more unstable, as is shown in several academic studies. Moreover, the adverse effects of banking supervisors dictating what banks can and cannot do are now well documented. It would be unfortunate if regulators resumed such practices through the back door by means of Basel II and risk-management systems.

The alternative is for banks to develop their own risk assessment methodologies, but for the supervisors to “correct” undesirable outcomes by requiring on a bank-by-bank basis a supplementary amount of capital. For this to be effective, banking supervisors would need to run sophisticated risk models to determine what they deem is the optimal level of bank capital. It is unlikely that they will have the resources or ability to do so with any degree of accuracy.

The spirit of Basel II is to design regulations to ensure financial stability while not overly burdening banks. This objective is much better served by a strong emphasis on market discipline, minimum standards for risk management and contingency planning if a banking crisis occurs.

Instead of requiring bank capital to be risk sensitive, banking regulation should simply require the use of high quality risk models in banks, without using their output to determine capital. Minimum bank capital is better calculated as a simple fraction of bank activity in broad categories.

Europe is still determined to push ahead with Basel II, but that may change when the results from the European quantitative impact studies become known in early 2006. In contrast to US politicians, the Europeans have shown a curious lack of concern about the impact of Basel II on European banks.

The Basel II accord has come under considerable criticism for its complexity and reliance on statistical risk models. In spite of this, the Basel committee has been determined to push ahead. The results from the quantitative impact study in the US provide banking supervisors with an opportunity to revise their stand before it is too late.

© Financial Times