2 However, some of the issues relevant in the context of the emerging markets and developing countries are yet to be fully addressed. In its attempt to strive for more accurate measure of risks in banks, the simplicity of the present Capital Accord is proposed to be replaced, with a highly complex methodology which needs the support of highly sophisticated MIS/ data processing capabilities. The complexity and sophistication essential for banks for implementing the New Capital Accord restricts its universal application in the emerging markets. Banks in these emerging markets form a significant segment in financial intermediation and are likely to find implementation of the New Capital Accord a major challenge in the medium term. Besides banks, supervisors would be required to invest considerable resources in upgrading technology systems, and human resources to meet the minimum standards. Banks in emerging markets would, therefore, face serious implementation challenges due to lack of adequate technical skills, under development of financial markets, structural rigidities and less robust legal system.

3 The QIS 3 results for the Standardised Approach show an increase in capital requirements for all country groupings in respect of both Group 1 and Group 2 banks. The QIS 3 results from the participating non-G-10 countries show that overall increase in risk weighted assets under the Standardised Approach was 19%, reflecting the impact of new operational risk charge (+ 15%) plus a credit risk contribution (+ 4%). These average contributions fall to + 11% and + 2% respectively (or an overall increase of around 13%) after some recalibration to the risk weights attached to claims on retail portfolios, residential property and past due loans.

4 The Reserve Bank of India is fully committed to implement the best international practices. However, the level of preparedness of the banking system and the supervisors would vary from country to country. In view of this, it will be desirable to assign greater flexibility to national supervisors to calibrate risk weights on different types of exposures under the Standardised Approach. For example, the CP 3 has recalibrated the risk weights on claims on retail portfolio to 75% and residential property to 35%. The CP 3 has also indicated reduction in risk weights on past due loans from 150% to 100% or 50%, depending on the level of provisions held against such loans and to encourage banks to make higher provisions for past due loans by providing capital relief. RBI welcomes such adaptability in the approach shown in CP 3. RBI also notes that the national supervisors can consider a higher risk weight on unrated claims on corporates if warranted in their jurisdictions. However, RBI feels that there are many other areas in which national supervisors can be allowed greater flexibility in assigning a lower risk weight if the country-specific situation so warrants than following a 'one-size-fits-all' approach based on the external ratings under the Standardised Approach. RBI has examined the various aspects of the proposals contained in the CP 3 and specific comments thereon are detailed hereunder:

4.1 The Committee has proposed that the New Accord will be applied to internationally active banks. However, it has been indicated in the Overview of the New Basel Capital Accord that the New Accord may be extended to include other significant banks as national supervisors deem appropriate. RBI reiterates that the focus of the New Accord should be primarily on the internationally active banks. As the main objective of the New Accord is to ensure competitive equality and providing a reasonable degree of consistency in application, it is necessary that all supervisors, across the world should have a common definition of internationally active banks. Basel Committee may, therefore, define what constitute internationally active banks.

4.2 RBI, while appreciating the Committee’s proposal that reciprocal cross-holdings of bank capital artificially designed to inflate capital position of banks should be deducted, feels that cross-holdings of equity and other regulatory investments may be allowed in principle, but may also need to be moderated to preserve the integrity of the financial system and minimise the adverse effect of systemic risk and contagion.

4.3 The Committee’s proposal that the Export Credit Agencies (ECAs) qualify for recognition only if they publish their country risk scores and subscribe to the OECD agreed methodology is appreciated. However, the OECD methodology and ECAs’ country risk classifications are still confidential.

4.4 The flexibility to provide uniform risk weight i.e. one category less favourable than that assigned to claims on sovereign to all the banks (under first option) (Paragraph 35) militates the basic philosophy of aligning capital adequacy assessment more closely with the key elements of risk. The mere location may not necessarily be a good indicator of a bank’s creditworthiness. This proposal provides competitive advantage to banks with weak financials by virtue of their having been incorporated in better-rated countries.

RBI, therefore, reiterates its earlier view that the risk weighting of banks should be de-linked from the credit rating of sovereigns in which they are incorporated. Instead, preferential risk weights should be assigned on the basis of their underlying strength and creditworthiness.

4.5 The proposal to assign preferential risk weight to short-term claims (Paragraph 38) may lead to arbitrage of regulatory capital through roll-overs, concentration of short-term borrowings and serious asset-liability mismatches, which could trigger systemic crisis and contagion in the domestic inter-bank market. It would also be very difficult to monitor and control the rollovers of short-term claims, given the high volume of transactions in the inter-bank market.

RBI, therefore, reiterates that preferential risk weights should not be linked to the maturity of the claims.

4.6 Banks are strongly regulated and supervised entities. Risks inherent in inter-bank exposures are not comparable to that of the corporates. There is, therefore, a need for a modified treatment for claims on banks. The Basel Committee has provided discretion to national supervisors in paragraph 28 to assign a lower risk weight to the exposures to the sovereign of incorporation, denominated in domestic currency and funded in that currency. A similar flexibility should be provided in respect of claims on banks as well under option 2.

4.7 The Committee has indicated that if banks are allowed to use unsolicited ratings in the same way as solicited ratings there may be the potential for ECAIs to use unsolicited ratings to put pressure on entities to obtain solicited ratings. Therefore the Committee has proposed that such behaviour, when identified, should cause supervisors to consider whether to continue recognising such ECAIs as eligible for capital adequacy purpose.

4.8 RBI appreciates the Committee’s efforts in evolving a range of risk-sensitive options for assessing capital for credit risk. However, the reliance on external credit assessment institutions (ECAIs) under the Standardised Approach for assigning preferential risk weights may not be a better option. First, the credibility of the rating agencies is at stake and there is no system of accountability for sharp deterioration in the credit quality of rated entities immediately after assigning a rating. Secondly, their access to information, especially in the absence of transparency and good corporate governance principles is severely restricted; whereas, banks are privy to customer information and are less exposed to customer-related informational asymmetry. Thirdly, the population of rated entities, even in the advanced countries, and especially in the emerging markets, which have exposure to the banking system, is very few in number. Fourthly, the use of external credit rating agencies in the regulatory process may act as a disincentive for the banks to improve their credit risk rating systems.

4.9 It is appreciated that the expanded role envisioned for IRB Approach provides positive incentives to banks in improving their credit risk management techniques. However, the adoption of the IRB Approach, even under the foundation approach, requires considerable investments in IT / human resources and rigorous supervisory oversights. Thus, most of the banks may not be able to adopt, even in advanced markets, the IRB foundation approach and would initially adopt Standardised Approach.

4.10 With a view to encouraging the banks using Standardised Approach, to move over to the IRB Approach at the earliest and also to equip them during the interregnum to adopt robust internal rating systems, they may be allowed to use the internal ratings for assigning preferential risk weights, on certain types of exposures, subject to compliance with the minimum standards prescribed by the Basel Committee for internal ratings under the IRB Approach.

4.11 This could be gradually extended to a larger portion of the banks’ asset portfolio. This will encourage banks to refine their credit risk assessment and monitoring process, which would facilitate better management of their asset portfolio. This will also avoid the use of ECAIs in the regulatory process and reduce the burden of additional cost on this count. Besides, the scarce supervisory resources will be optimally utilised for validating the banks’ internal rating systems rather than for approving ECAIs. This would also avoid conflict of jurisdiction over rating agencies.

Internal rating based approach

4.12 RBI appreciates the Basel Committee’s proposal to offer a range of options of increasing sophistication for providing explicit capital charge for credit risk. RBI recognises the inherent attractiveness of the IRB Approaches, which will result in better internal credit risk management. However, the minimum requirements stipulated even under the IRB foundation approach are difficult to be implemented, especially in the emerging markets. Most of the banks do not have robust rating systems and historical data on Probability of Default (PD), nor do the supervisory authorities maintain time series data for estimating Loss Given Default (LGD).

4.13 It is well recognised that the proposal to assign banking book exposures into six broad classes of exposures with different underlying credit risk characteristics - corporates, sovereigns, banks, retail, project finance and equity under IRB Approach would better discriminate the likely pattern of portfolio losses. However, a common framework for definition of these segments, without recognising the institutional framework, value of accounts or geographical spread, may pose severe implementation problems to banks in emerging markets.

RBI, therefore, re-iterates that national supervisors may have discretion and flexibility in defining the exposure classes, such as corporate, retail, sovereign and project finance.

Operational risk

4.14 In the context of increasing globalisation, enhanced use of technology, product innovations and growing complexity in operations, RBI agrees, in principle, with the Committee’s proposal to assign explicit capital charge for operational risk. RBI also acknowledges that the range of approaches of increasing sophistication - Basic Indicator, Standardised and Advanced Measurement - would set the basic framework for estimating capital for operational risk. Given the sophistication and database required for Standardised and Advanced Measurement Approaches, most of the banks, especially those domiciled in emerging markets would be adopting the Basic Indicator Approach.

4.15 The Committee has proposed that at national discretion banks can use Alternative Standardised Approach (ASA) for calculating operational risk capital charges (footnote 91, paragraph 615). This would serve as an intermediate stage for banks which are migrating from the Basic Indicator Approach to the Standardised Approach. It is observed that under the ASA, the beta will be 15% for retail and commercial banking if they are aggregated and the banks unable to disaggregate their gross income into the other six business lines can aggregate the total gross income for these six business lines using a beta of 18%. This suggests adoption of a higher beta under the ASA as compared to the beta applicable to the Basic Indicator Approach which is 15% and may not, therefore, effectively serve the intended purpose of serving as an intermediate stage for banks migrating from the Basic Indicator Approach to the Standardised Approach.

4.16 It has been proposed that, under the Alternative Standardised Approach the exposure indicator for ‘retail banking’ and ‘commercial banking’ business lines may be the ‘volume of advances multiplied by m (which is 0.035)’ instead of ‘gross income’. It is also proposed that loans and advances for the purpose would be taken gross of provisions. Since this measure is intended to serve as an alternative to the measurement of gross income of these two business lines, it would be in order to reckon the advances ‘net of non performing loans’ under the Alternative Standardised Approach.

RBI is of the view that the proposal to alternatively consider volume of advances (instead of gross income) would imply a substantial increase in capital charge for operational risk. Hence, RBI feels that the volume of performing advances may be considered under the Alternative Standardised Approach.

International lending to developing and emerging economies.

4.17 Under the CP 3, banks have the choice to adopt any one of the following methods for measuring credit risk:

• Standardised Approach (SA)
• Foundation Internal Ratings Based Approach (FIRB)
• Advanced Internal Ratings Based Approach (AIRB)

Under the SA, the risk weight for sovereign exposures would depend upon the rating assigned to such sovereign exposures by export credit agencies. Under the IRB (Internal Ratings Based) Approaches, the risk weight would depend upon the rating by the banks’ internal ratings model and is computed as a function of the following four factors – probability of default, loss given the default, exposure at default and maturity. While the risk weight for exposures with the lowest rating (Below B-) under the SA is 150%, the same is likely to theoretically go up to 1250% under the IRB Approaches. This clearly illustrates the extent to which the IRB Approaches are more risk sensitive than the Standardised Approach.

4.18 It is unlikely that a developing economy would receive the best of the ratings. It is also largely unlikely that an entity in a developing economy would attract a rating better than the sovereign rating of that economy. In the circumstances, a bank adopting the IRB Approach is likely to be more averse to exposures to developing economies both directly (to the sovereign) and indirectly (to entities in that economy). As has been brought out convincingly in the paper ‘Basel II and Developing Countries: Diversification & Portfolio effects’ by Stephany Griffith-Jones, Miguel Angel Segoviano and Stepphan Spratt – this aversion may translate into either avoidance of risk or appropriate pricing of the risk resulting in the following scenario * :

Widespread adoption of the IRB Approach by internationally active banks would lead to a significant increase in capital requirements for loans to lower rated borrowers. To the extent that the pricing and availability of international bank loans is influenced by the capital requirements that relate to them, this would imply a sharp increase in the cost and/or reduction in the quantity of international lending to developing and emerging economies. The expressed purpose of the Basle II norms is to better align regulatory capital with actual risk. Therefore, failure of the proposals to take account of the benefits of international diversification suggests that, risk has not been measured accurately. By excluding the possibility that banks’ capital requirements should take account of portfolio and diversification effects, the proposals effectively impose an inaccurate measure of risk, at the portfolio level. The fact that the proposals under Basle II will not allow these diversification benefits to be taken into account, suggests that the regulatory capital associated with lending to developing countries will be higher than that which the banks would - and currently are - choosing to put aside on the basis of their own models.