Basel II: Need for Improvement
Basel II aims to achieve a process of continuous improvement in the risk management space. It is built upon three mutually reinforced pillars: Pillars I, II and III.
Pillar I focuses on risk components and risk capital calculation, thus making the capital allocation process risk-sensitive. Pillar-II advocates that setting aside capital is not a substitute for the risk and tries to ensure the presence of a strong risk management framework from a supervisory perspective to oversee and review the risks addressed by Pillar I. Pillar III tries to establish transparency and improve market discipline by requiring firms to publish certain details of their risks, capital and risk management processes.
Basel II entails a number of significant new, functional requirements. First of all, Basel II substantially extends the management of credit risk. Adding to this, it also presents a completely new framework for operational risk and the treatment of securitization. With the IRB approach, Basel II allows the banks to develop their own systems for calculating expected default in payments. In contrast to this, previous methods in accordance with Basel I directives were based on a ‘one-size-fits-all’ solution. It was only possible to distinguish between the asset categories to a certain degree. Subsequently, the possibility for optimizing the use of capital was also restricted. Under Basel II, the use of ratings from external credit rating agencies is permitted for the first time for calculation of risk-weighted assets.
The effort to tie the capital requirements more closely to risk and promote a disciplined approach to risk management, though largely successful, leaves a lot of room for improvement. Some issues need to be addressed either though a more sophisticated approach, elimination or substitution.
The issues can be described in three broad categories
1. Existing blocks – These are the existing building blocks in Basel II. However, the same can be improved further to cover the broader scope of risk management in banks:
2. Near-misses– These refer to areas that perhaps were considered while Basel II was being formulated but were not present in the final Basel II regulatory framework.
3. Overall impact – This section refers to the impacts that Basel II would have on the global banking and financial industry.
The accord proposes that the banks must calculate expected losses based on the following risk components:
All four credit-risk drivers are used to measure the expected loss. In this process, Basel II has assumed that PD, LGD and EAD are independent variables. In contrast, there is considerable empirical research showing a reasonably strong positive correlation between these variables. BIS working paper 113 (‘The link between default and recovery rates: Effects on the procyclicality of regulatory capital ratios’) says, ‘several regressions between the two fundamental variables (PD and LGD) and we find that one can explain about 45 per cent of the variation in the annual recovery rate with the level of default rates (this is the linear model) and as much as 60 per cent or more with the quadratic and power relationships.’
A study by Mortgage Insurance Companies of America (MICA) also says there is a high correlation between LGD and PD noted in geographically concentrated portfolios of loans. These correlations decide the volatility of losses or the unexpected loss component of credit risk. But the accord doesn’t explicitly mention the correlations between the four risk components.
With improved computing technology, the committee might well feel the need to identify some more drivers of credit risk apart from these four and try to granularize them further to improve accuracy and materiality in the absolute risk measurements.
In Basel II, the formulae for calculation of capital charge have been enumerated. But while arriving at these formulae, credit losses are assumed to have a normal distribution, whereas there is a widespread consensus that historical credit losses display a much greater frequency of extreme outcomes than would be predicted by a normal distribution.
Future work for the Basel Committee is to accommodate some alternative methods of calculating capital requirements taking different distributions into account.
The most important consideration is probably that the total capital required for all exposures is the sum of the capital for the individual exposures. This approach, by design, ignores one common tenet of risk management – diversification. In reality, the largest losses for different positions and risks do not typically occur at the same time. The risk of a portfolio of positions is therefore usually less than the sum of the risks of the positions themselves. The degree of diversification is determined both by the extent to which different market variables tend to move together and by the extent to which different businesses have similar positions. The greater this correlation among defaults, the higher the Basel capital requirement. There appears to be a consensus that Basel II is relatively conservative in its correlation assumption.
Simply, standalone total economic capital (EC) equals (=) EC (market risk) + EC (credit risk) + EC (operations risk).
If we analyze the reported economic capital in the latest balance sheets, we get the following:
| Banks | Economic risk capital | Diversification Effect | Percentage | |||
|---|---|---|---|---|---|---|
| 2004 | 2003 | 2004 | 2003 | 2004 | 2003 | |
| Credit Suisse | 2830 | 2931 | 791 | 786 | 27.95 | 26.82 |
| Credit Suisse FirstBoston | 8687 | 8562 | 2155 | 2110 | 24.81 | 24.64 |
| Deutsche Bank | 11447 | 13275 | 870 | 1152 | 7.60 | 8.68 |
Through their superior analytics, the banks in Table 1 have been able to calculate diversification benefits accruing from diversification between the credit and the market risks.
One interesting thing to note is that the extent of diversification benefits as reported varies from 7.5-28 per cent. The dispersion seems a bit larger. Although the extent of benefit definitely depends on some bank-specific factors, there is a clear need for a means to control through supervisory and regulatory guidelines.
It is very difficult to segregate the interest rate risk and credit risk in the case of the fixed income assets. And interestingly, the losses due to interest rate variation do not necessarily happen simultaneously with losses due to changes in the credit quality of the counterparties. Therefore, it is advisable to measure and manage these sources of risk in a common framework in order to take the dependencies between them into account.
The future lies in a paradigm shift in the way the risk capital is calculated. An integrated approach to market risk and credit risk management and measurement may emerge, starting from fixed income assets.
In terms of recognizing credit risk mitigation instruments, although Basel II is revolutionary as compared to Basel I, the tone is one of scepticism and conservatism. The accord discusses the broad range of risk-reducing features of some transactions, including: (1) borrower-posted collaterals, (2) arrangements to net multiple exposures with a single counterparty, and (3) third-party credit enhancements (such as credit derivatives). Among the credit derivatives, only credit derivative swaps (CDS) and total return swaps (TRS) are accepted. In the case of synthetic securitization, special purpose vehicles (SPVs) are generally not recognized as eligible protection providers, even in cases where they are fully collateralized.
The quasi-operational risks have been included under Pillar II. For example, banks’ use of collaterals, guarantees or credit derivatives to offset credit or counterparty risk brings its own set of challenges, including legal risk, documentation risk and liquidity risk. Pillar II says supervisors should require banks to have written credit risk mitigation policies and procedures to control those types of residual risks.
Future work from the Basel Committee is expected to try to explain these risks more specifically, both in terms of their occurrence and their impact on capital. Other operational-type risks left out of the definition, including reputation and strategic risks, will no doubt be included in any Basel III project.
‘Credit Risk Mitigation (CRM) in the form of guarantees and credit derivatives must not reflect the effect of double default.’ – BCBS (June 2004, paragraph 301).
For both single-name and portfolio credit derivatives, Basel II capital charges are not determined by the substitution approach. Under the substitution approach, a bank can substitute the PD and LGD of the guarantor for those of the obligor if this would result in a lower risk weight. This substitution approach does not adequately recognize the lower risk of joint default or the benefit of double recovery associated with guarantees. It therefore brings ambiguity to the very objective of the new accord promoting better risk management practices and aligning regulatory capital with risk.
Again, the benefit of a guarantee is limited to a reduction in the effective PD. But real life experience shows that in many hedged transactions, the true LGD is also reduced either through the benefits of double-recovery or because the LGD associated with the guaranteed loan is clearly lower than that for the underlying facility. But better banks take into consideration other aspects of a guarantee or credit derivative – joint default probability, joint recovery, legal risks, etc. The best part is that the committee has recognized the issue of double default and has already started working on a prudentially sound solution.
Basel II has a distinct corporate credit focus with less emphasis on retail credit despite the large retail exposure in the economy. Of the $16.3 trillion in outstanding debt, including securitized debt, in the US by the end of 2003, approximately 54 per cent ($8.8 trillion) was composed of consumer outstanding debt including revolving and non-revolving debt and home mortgage loans. Corporate outstanding debt was just more than half of the all retail debt. Non-corporate business accounted for another $2.4 trillion or about 15 per cent of the outstanding. Out of 604 paragraphs under Pillar I, only around 80 paragraphs deal with retail credit risk. The approach makes no distinction between a foundation and advanced IRB approach for the retail asset class. The suggested approach limits the number of loss distributions to three product categories. But in reality not all products (the home equity lines of credit) fit exactly into the three prescribed categories, and again within each category, certain specific products with the same PD, LGD and EAD may exhibit inherently different risk. For example, automobile lending may include direct (bank-originated), and indirect (dealer-originated) credit, with the two sub-portfolios exhibiting very different risk attributes. Given the advanced analytical modeling prevalent in the retail credit management (models for application, behavior, collection, attrition, and pre-payment scores by individual product lines) flexibility (advanced approach) within IRB would address many issues and make retail credit capital consistent with other portfolios.
The Basel II Accord is described in over 500 pages of text, but the word ‘agriculture’ does not appear once. The exposures are broken into six groups: corporate, retail, bank, sovereign, equity and project. Again, it’s not clear where agriculture fits. With the cap on individual loan size of $1m for retail exposure limit, it implies that large agricultural loans would be treated as corporate loans and small agricultural loans as retail loans. Basel III would need to take into account the particular characteristics of farm loans when setting capital charges for organizations involved in agricultural lending. Because characteristics like cyclical performance, seasonal production patterns, high capital intensity, leasing of farmland, participation in government programs, and annual payments of real estate loans make the farm business very unique.
It seems that the scope for ‘lending arbitrage’ is in-built in the IRB approach. The options provided by the accord put non-G10 corporates and those in developing countries at a distinct disadvantage under the IRB approach to credit risk. Given the costs of compliance, the non-Basel banks will not be able to lend to high-rated corporations. Possibly, a caste system in the banking industry would develop with two types of banks.
The objective of the accord was to establish a stable financial system. But the capital tied to the riskiness of the assets will fluctuate during the times of economic boom and bust. The creditworthiness of the clients will be high during the good times, thus leading to less capital requirements and will be low during the recessionary periods, demanding more capital and thereby squeezing the economy. So the very objective of the accord gets defeated. This phenomenon can be fixed by simply setting capital charges not on current risk ratings, but on an average going back over the years.
Many of the corporations in emerging countries do not have a debt rating. Even if they have one, the efficacy of the same can be far from the international standard. For materializing the benefits of the accord, more companies will need to be rated and (probably) more rating agencies will need to be born. In this scenario, it becomes the prerogative of the national supervisors to verify the robustness of the rating agencies. Meanwhile, the rating agencies will come under more and more pressure to deliver good ratings to win business. Expressing concerns on the over-reliance on credit rating agencies, their quality and the extent of coverage, some of the countries have suggested different ways of setting risk weights under the standardized approach for credit risk management.
A lot of effort has gone into realizing Basel II. Though the issues raised are not radical in nature, they challenge some of the assumptions of the accord and can be resolved. Suggestions can be sought from industry bodies in the areas of insurance, securities, academia, and banking. In fact, the committee has already started working on some of the issues. Although it might seem a little too early, a sequel – Basel III – is required with prospective solutions to the listed issues. Hopefully, the accord would prescribe formulae and guidelines for models for credit portfolio management, taking care of the correlations between the risk components PD, LGD, EAD, and so on. This should address the issue of diversification benefits acknowledging the correlations between credit risk and market risk, as also with operations risk and business risk.