First Consultative Paper on the Implementation of the New Basel Capital Accord

ANNEXURE A
CONSULTATIVE PAPER NO. 1/2004
IMPLEMENTATION OF THE NEW BASEL CAPITAL ACCORD 20 February 2004

ii
Table of Contents

1. Executive Summary ................................................................................................... 1
2. Introduction.................................................................................................................. 6
3. Progress to Date ......................................................................................................... 9
   Participation in international initiatives.............................................................. 9
   Initial strategy......................................................................................................10
   Accord Implementation Forum .........................................................................12
   Quantitative Impact Study.................................................................................13
   Credit-risk model validation..............................................................................16
4. Scope of Application of Basel II in South Africa...................................................18
   Principles upon which scope of application decisions are based...............18
   Scope of application...........................................................................................22
5. Minimum Level of Capital Adequacy .....................................................................26
   The BSD’s response ..........................................................................................27
6. Credit Risk – Standardized and Simplified Standardized Approach .................29
   External Credit Assessment Institution (“ECAI”)...........................................29
   Proposed treatment of certain national discretion items ..............................30
   Claims on sovereigns ............................................................................30
   Claims on banks .....................................................................................32
   Claims on corporates .............................................................................33
   Claims included in the regulatory retail portfolio ...............................35
   Higher risk categories............................................................................36
   Gold bullion .............................................................................................37
   Unsolicited ratings ..................................................................................38
   Credit-risk mitigation ..............................................................................38
   Sovereign guarantees ...........................................................................39
7. Credit Risk – Internal Ratings Based (“IRB”) approach.......................................41
   General comments on the IRB approach.......................................................41
   Consultation ........................................................................................................44
   Partial use of the IRB approach.......................................................................45

iii
Transitional arrangements ................................................................................48
IRB ratings systems ...........................................................................................49
IRB model approval and validation..................................................................50
8. Operational Risk.........................................................................................................55
Alternate Standardised Approach....................................................................55
Partial use............................................................................................................56
Data requirements ..............................................................................................57
9. Future Plans ..............................................................................................................59
Resource planning .............................................................................................59
Accord Implementation Forum .........................................................................59
Bank-readiness reviews ....................................................................................60
Fourth quantitative impact study......................................................................61
Further consultation and guidelines ................................................................62
IRB approval and validation..............................................................................62
Development of a regulatory framework.........................................................63
Appendix A: Structure of the Accord Implementation Forum..................................65
Appendix B: Schedule of AIF meetings for the year.................................................66

1. Executive Summary
   1.1. In view of the global debate on and uncertainties about the
   implementation of the proposed new Basel Capital Accord (“Basel II”), local banks were becoming anxious for the South African Reserve Bank (“SARB”) to confirm its position thereon. Banks Act Circular 14/2003 communicated the initial thoughts of the Bank Supervision Department (“BSD”) of the SARB, on the subject. The BSD also undertook to provide further details on implementation within a South African context by means of a consultation paper. As an input to the consultation paper, the BSD requested information pertaining to the scope of application of Basel II that banks were considering adopting on inception of the new regulatory framework. Input from the banks has guided the BSD’s implementation strategy. This Consultative Paper (“CP”) is the first in a series of consultation papers intended to communicate the BSD’s thinking on aspects of the proposed approach to implementation of Basel II. The consultative paper further aims to get banks views and suggestions on the implementation process, and in order to achieve this objective, we have formulated various questions for the attention of banks.
   1.2. The BSD is encouraged by the level of awareness of Basel II within
   South African banks and is therefore continuing its efforts to contribute to the development of Basel II, thereby ensuring that the interests of non-Basel member countries are taken into account. The initial implementation strategy focused on readying banks for Basel II implementation. Although, it became clear that implementation of the more advanced approaches might take a little longer than initially envisaged, the BSD believes that it is more prudent to implement Basel II appropriately than to meet strict timelines.

1.3. A new capital adequacy framework, which is based on Basel II
principles will replace the current capital adequacy framework (which is based on the principles of the 1988 Capital Accord) as from 31 December 2006 (hereafter referred to as the common implementation date) or any later implementation date which may be set by the Basel Committee on Banking Supervision (“Basel Committee”). Unlike some other jurisdictions, the BSD has decided that the current capital adequacy framework will not be available to banks as from the common implementation date. The BSD is of the opinion that it will be possible for every registered bank in South Africa to migrate to the simplest approaches of Basel II without any major disruption. Basel II offers a menu of approaches to both credit and operational risk. The BSD will make all approaches available to enable banks to select there from but the more advanced approaches (Advanced Internal Ratings Based Approach (“AIRB”) for corporate credit risk and the Advanced Measurement Approach (“AMA”) for operational risk) will only be considered on an exceptional case-by-case basis, until such time as the industry matures to effectively adopt these more advanced approaches and BSD has had an opportunity to develop appropriate processes to supervise the same.
1.4. More sophisticated banks, however are expected to adopt the more
advanced approaches over time in an effort to improve the level of sophistication of risk management practices in South African banks. Banks intending to adopt the Internal Ratings Based (“IRB”) approach for credit risk or the AMA approach for operational risk will be required to calculate their capital requirements under both the current capital adequacy framework and the Basel II framework as from 1 January 2006 until common implementation date. This period will be referred to as the “parallel run” and is intended to allow for the study of the impact of new framework and to incorporate appropriate changes, where necessary.

1.5. Although specific approval will not be required for the Standardised
(and Simplified Standardised ) approach, for credit risk as well as the Basic Indicator approach for operational risk, the BSD expect banks to consult with it on the suggested approach prior to the date of implementation. For any of the other approaches banks will be required to make a formal application in the prescribed manner.
1.6. Foreign bank branches and subsidiaries wishing to avail themselves
of the advanced approaches will be allowed to do so, however, reliance will be placed on the home country supervisor. The BSD will adhere to the principles on cross -border implementation of the Basel Committee on Banking Supervision (“Basel Committee”) when dealing with foreign-controlled branches and subsidiaries.
1.7. This paper does not seek to consult on all aspects related to the
various approaches. The items addressed in the paper relate only to the areas in which the BSD has formulated an opinion. The BSD will consult on the other areas at a later date.
1.8. During the course of 2004, the BSD will consult with the External
Credit Assessment Institutions (“ECAIs”) to gain some understanding of their ratings assignment processes and to discuss, in more detail, the eligibility criteria provided in Basel II. The BSD will produce a list of recognised ECAIs once it has had the opportunity to assess them.
1.9. Several banks have argued that under a more risk sensitive capital
adequacy framework, a 10 per cent (as opposed to an 8 per cent) capital adequacy requirement is not justified. The BSD highlights its view on this issue and request banks for further information to substantiate their argument.
1.10. This paper also addresses some issues of national discretion for the
standardised approaches to credit risk relating to:

• ECAIs.
• Claims on sovereigns, banks and corporates.
• Claims included in the regulatory retail portfolio.
• Higher risk categories.
• Gold bullion.
• Unsolicited ratings.
• Credit-risk mitigation.
  1.11. Certain banks have made some progress in the design and use of
  rating systems, but are far from meeting the minimum standards required by Basel II. Full compliance may well extend beyond the common implementation date. In view of the evolving nature of the IRB process, the BSD cannot consult fully on these approaches at this stage but will continue discussions with banks and other stakeholders.
  1.12. In this paper, the BSD provides its initial thoughts only on the
  following areas pertaining to the IRB approach:
• Adoption of the IRB approach across asset classes.
• Guidance on transitional arrangements.
• The IRB rating systems, in particular, the use of models.
  1.13. Since the quantitative approach to the supervision of operational risk
  is a new concept the BSD will continue to work with banks in order to devise a supervisory review process. The BSD will also develop further guidance on the supervisory expectation pertaining to operational risk. The BSD will allow a bank to adopt the Alternative Standardised Approach (“ASA”) for corporate and retail business lines, provided the particular bank can satisfy the regulator that this alternate approach provides an improved basis for avoiding the double counting of risk.

1.14. Many banks expressed their desire to adopt the partial-use option for
operational risk. Under the partial-use provisions, a bank could calculate the operational risk charge for some business lines using one approach and using another for the remaining business lines. The BSD is amenable to banks adopting a partial use approach to the implementation of operational risk, subject to minimum requirements being met. Banks will however have to discuss their plans with the BSD.
1.15. The last chapter of this paper deals with the BSD’s future plans. A
schedule of meetings of the Accord Implementation Forum for the
year 2004 is included as an annexure to this paper.

2. Introduction
   2.1 This Consultative Paper (“CP”) is a first in a series of consultation
   papers intended to communicate the thinking of the Bank Supervision Department (“BSD”) of the South African Reserve Bank (“SARB”) on aspects of its proposed approach to the implementation of the proposed new Basel Capital Accord (“Basel II”). Each CP will introduce aspects that the BSD has considered, and will request responses and constructive inputs from stakeholders. The consultative process is an important step in informing the BSD’s implementation initiatives. Several areas of the text pose questions to the industry to elicit their views or request further information.
   2.2 Banks Act Circular 14/2003 communicated the BSD’s initial thoughts
   on its position on the implementation of Basel II within the South African banking sector. The circular was timely and necessary given the global uncertainties following pronouncements by certain countries on their scope of application of Basel II. An announcement by the Basel Committee on Banking Supervision (“Basel Committee”) of a potential delay in the release of the final Basel II document led to further uncertainties.
   2.3 This paper provides details on the BSD’s progress to date and its
   views on the scope of application of Basel II within the South African banking sector. Some technical aspects on the Standardised and Internal Ratings Based (“IRB”) Approach to the treatment of credit risk and specific aspects pertaining to the treatment of operational risk are discussed. Finally, proposed next steps for the implementation process are communicated.
   2.4 The BSD wishes to reiterate that it welcomes the efforts of the Basel
   Committee to revise the 1988 Accord, in order to align capital

adequacy to the risks inherent in banking organisations. The BSD is particularly encouraged by the objective of Basel II to incentivise banks to improve their risk-management efforts and to align economic capital more closely to regulatory capital. The flexibility and menu of approaches embedded in Basel II make the Accord applicable to banks of varying degrees of sophistication. Although initially designed only for internationally active banks, the Basel Committee has formulated an Accord that is equally applicable to banks in lesser￾developed countries. Wider adoption of Basel II worldwide will go a long way towards ensuring a safer financial system.
2.5 This CP is based on the Basel Committee’s third Consultative Paper
(“CP 3”), released in April 2003, and available at www.bis.org. The Basel Committee received more than 200 comments on CP 3, a great deal more than expected. The comments were substantial in nature and the Basel Committee is addressing them through recalibration of or different treatment of several areas of Basel II. The final Basel II document is expected to be released in June 2004.
2.6 The BSD is of the opinion that the overall framework of Basel II is
unlikely to change, since all stakeholders are in agreement with its broad direction. The BSD, however, anticipates that there will be several modifications to the final Basel II document, as compared to the proposals of CP 3. If there are significant differences, the BSD will revise its position appropriately and may undertake further consultation at a later date. The BSD’s consultative process is further subject to international implementation initiatives, by which the BSD is guided. The Basel Committee is establishing various structures to assist regulators around the world in their implementation initiatives. Such initiatives may lead to a change in the BSD’s implementation approach.

2.7 Regardless of the potential delay in the finalisation of the Basel II
document, stakeholders are encouraged to start the implementation initiatives, given the tight deadlines set. This consultative process is intended to be a catalyst in banks gaining momentum with their implementation initiatives.

3 Implementation Progress to Date
3.1 This section sets out progress with some of the initiatives towards the
implementation of Basel II in South Africa over the last two years. The BSD is encouraged by the level of awareness of Basel II within banks, and their willingness to adopt Basel II principles, not only to comply with regulations, but also as a tool to improve business efficiencies. Participation in international initiatives
3.2 The BSD is continuing to contribute to the development of Basel II,
thereby ensuring that the interests of non-Basel member countries are taken into account. This contribution is made through South Africa’s representation on the Core Principles Liaison Group (“CPLG”) and its sub-group, the CPLG Working Group on Capital (“CPWG”). The Secretary General of the Basel Committee chairs the CPLG, which reports directly to the Basel Committee. The CPLG has representatives from the Basel Committee, the World Bank, the International Monetary Fund and developing countries. The CPLG was formed to deliberate mainly on broader supervisory issues arising in non-Basel member countries, whereas the CPWG was formed specifically to deal with capital-adequacy issues.
3.3 Initially, the Basel Committee sought to apply Basel II to
internationally active banks only. Over 100 countries have, however, adopted the 1988 Accord, and most of these countries have expressed their intention to adopt the revised framework. The driving factors behind the proposed adoption of Basel II in these countries are, inter alia, the adoption of international standards, the objective of improving the general level of risk management in the banking sector,

financial stability, and competitive equality. The Basel Committee therefore, had to take into consideration the interests of these non￾Basel member countries. The Basel Committee did in fact take their interests into account, as is evident by the introduction of the Simplified Standardised Approach to the calculation of the capital requirements of credit risk. The CPWG was instrumental in lobbying for these changes.
3.4 The CPLG and CPWG meets four times a year, and each meeting is
attended by a representative of the BSD, who actively participates in discussion, as well as regularly give feedback on the implementation initiatives in South Africa. This helps to raise the awareness of South Africa’s implementation initiatives and could in future, benefit foreign banks considering South Africa as a potential investment destination. Their decision will partly be based on the strength of the banking sector in this country.
3.5 The BSD is also represented on the global Quantitative Impact Study
(“QIS”) group, through South Africa’s participation in the third Quantitative Impact Study (“QIS 3”), which is discussed in detail below. The BSD successfully hosted one of the global QIS 3 meetings in Cape Town in 2002. Initial strategy
3.6 The BSD commenced implementation initiatives in the last quarter of
2001. The initial strategy focused on readying banks - especially the
large internationally active banks - for Basel II implementation. The BSD’s thinking was that the banks’ level of readiness would dictate the appropriate strategy that the supervisors should follow in the supervision of Basel II-compliant banks. The BSD carried this strategy out through, inter alia:

3.6.1 Creating awareness of Basel II and its implications among
banks, through various interactions.
3.6.2 Giving regular feedback on international developments with
Basel II implementation.
3.6.3 Engaging banks in discussion on complex implementation
issues, not only with the BSD but also with other banks, through the Accord Implementation Forum (discussed in more detail below).
3.6.4 Bilateral meetings with some banks to discuss
implementation.
3.6.5 Undertaking QIS 3, which was the most comprehensive
consultation process thus far.
3.6.6 Ensuring that the larger banks developed and communicated
their implementation plans – some banks were required to present Basel II implementation plans in their organisations as
part of the “flavour-of-the-year” topic at the annual trilateral
meeting.
3.7 The strategy paid dividends, as evidenced by the progress made to
date by some of the larger banks. Examples of progress made includes, the following:
3.7.1 The majority of the banks embraced the Basel II proposals,
and recognise that there are business benefits to be gained from the effective implantation of Basel II principles.
3.7.2 The larger banks have devised Basel II implementation
strategies, which have been approved by the board in some cases.
3.7.3 Banks have board-approved budgets for Basel II
implementation projects; some banks have boosted their risk￾management budgets, since they are incorporating Basel II principles into their risk-management framework.
3.7.4 Banks have established Basel II implementation committee
structures, driven typically by a steering committee, which comprises board members and senior managers.

3.7.5 Data-identification and -collection initiatives are underway in
some banks
3.7.6 Basel II principles are helping banks to develop their
economic-capital models.
3.7.7 A simulation through the QIS 3 exercise enabled the banks
involved to conduct a gap analysis.
3.8 Based on the size of assets, most of the banking sector is therefore
well placed for implementation of Basel II on the common implementation date. This of course does not imply that banks will be compliant with the most advanced approaches of the proposed capital-adequacy framework. Implementation of these more complicated approaches will be evolutionary in nature, and may well extend beyond the common implementation date. The BSD however, consider it more prudent to implement Basel II appropriately, even if it takes a little longer, than to meet a strict timeline and carry out substandard implementation. Accord Implementation Forum
3.9 Banks Act Circular 5/2002 of 4 April 2002 gave rise to an Accord
Implementation Forum (“AIF”). The mission of the AIF is “To promote the implementation of the standards prescribed by the New Basel Capital Accord (“Basel 2”), in co-operation with all relevant stakeholders, in a manner that promotes a sound banking and financial system”. The AIF evolved into an elaborate structure of committees established to deliberate on various specialised areas of Basel II. The structure of the AIF is appended to this CP as Appendix A.
3.10 The various committees very enthusiastically commenced meetings to
deliberate on technical issues pertaining to implementation of Basel II. Much progress was made with discussing, among other things,

requirements for capital adequacy for the different risk types, the data requirements of Basel II, and Pillar III issues. The AIF even proposed a regular data-collection exercise, which proved to be impractical at that stage owing to the evolving Basel II framework, as the Basel Committee sought to finalise the new Accord. Accordingly, some of the initiatives may have been premature, and it was decided to focus on AIF discussions only after banks had had an opportunity to deliberate on these issues internally, and the Basel Committee and/or other supervisors around the world had developed more implementation guidelines.
3.11 The committee discussions were comprehensively recorded and are
publicly available on the SARB’s website: www.reservebank.co.za. The BSD will be transparent in its interaction with banks and auditors, through the AIF, and will make documents publically available as far as possible.
3.12 One of the milestones of the AIF was the commitment of the chief
executive officers (“CEOs”) of banks represented on the Steering Committee to attend at least one Steering Committee meeting for the year. The CEOs’ participation will ensure that some of the more significant strategic issues facing the industry as a whole are debated, discussed, and amicably resolved. The CEOs’ participation is also serving to ensure that the Basel II implementation teams within banks are given the necessary executive support. Quantitative Impact Study
3.13 The Basel Committee undertook a most comprehensive field study of
Basel II in the form of QIS 3, which commenced around June 2002. QIS 3 included 365 banks from 43 countries. There were 188 banks from 13 group-of-ten (“G10”) countries and 177 banks from 30 other countries. Of the 177 banks from other countries, 111 banks were

from 18 non-European Union (“EU”) and non-G10 countries. Eight South African banks participated in this comprehensive impact study. The eight banks represented approximately 80 per cent of the total South African banking sector (as measured by the size of assets).
3.14 The QIS 3 process involved intensive work with the banks, including
the following steps:

• Banks were given comprehensive guidelines to complete the
  QIS 3 questionnaire which was in the form of a very detailed spreadsheet with embedded formulae.
• Banks were guided through the spreadsheets by way of a
  workshop, and were given an opportunity to pose questions at any time that they required assistance.
• This was followed by several months of intensive data collection,
  during which the banks deployed teams to retrieve – in most cases –information manually. The spreadsheets were then populated with the data to calculate capital requirements automatically.
• The completed spreadsheets were submitted to BSD, which
  analysed the spreadsheets in detail.
• The BSD populated a template, which had been designed to
  consolidate summarised individual bank data and perform analysis of the data.
• The data were validated through extensive discussions with the
  banks individually. Each validation session lasted a full day.
• The final validated data were then consolidated and analysed,
  and a report of the findings was drafted and submitted to the Basel Committee, for analysis.
• Finally, the banks were invited to a presentation at which the
  findings were revealed and discussed.

3.15 There was an overall increase in risk-weighted assets for South
African banks, regardless of the approach that was adopted. The average increase in risk-weighted-assets is broken down as follows for the different approaches:

• For Group 1 banks (assets over Euro 3 billion) using the
  alternate standardized approach for operational risk:
• 8 per cent increase for the standardized approach.
• 15 per cent increase for the foundation IRB approach
• 27 per cent increase for the advanced IRB approach
• For Group 2 Banks (assets less than Euro 3 billion) using the
  alternate standardized approach for operational risk:
• 6 per cent increase for standardized approach.
• 2 per cent increase for foundation IRB approach.
• 4 per cent increase for advanced IRB approach
  3.16 This finding was consistent with the experience of non-EU, non-G10
  countries, where the capital requirements for all approaches increased, ranging from 9 per cent to 19 per cent on average. In the opinion of the BSD, the increase can be attributed to several factors, inter alia:
• QIS 3 was done on a best-efforts basis, since banks were not
  expected to be Basel II compliant.
• The higher capital requirement is reflective – as argued by
  certain critics – of South Africa’s emerging market status. The risk-quantification methodology may have factored a higher risk weighting for emerging markets.
• There was a general lack of credit-risk mitigation data. The
  implications, therefore, are that exposures did not receive a preferential risk weight if credit-risk mitigation mechanisms were mitigating some of the risks inherent embedded in the

exposures. A higher risk weighting will therefore result in an additional capital requirement for such exposures.

• Generally, banks were very conservative in their estimates.

• There were significant increases in the risk-weighted asset
  requirement for the sovereign, bank and equity portfolios, owing to the current regulatory treatment thereof, which is not sensitive to risk.

• The operational-risk charge (a charge additional to the current
  regulatory requirements), made the highest contribution to the overall increase in risk-weighted assets.
  3.17 QIS 3 was an excellent starting point. Since it, however, was done on
  a best-effort basis, conclusive reliance cannot be placed on the results. The BSD therefore, plans a further QIS during the latter half of 2004 or early in 2005. All banks will have an opportunity to participate in the proposed QIS and will reap the same benefits as the initial eight banks that participated in QIS 3. Further details of the additional QIS are discussed in the last section of the CP. Credit risk model validation
  3.18 During the past six months, we have carried out some work on the
  validation of credit risk models for corporate and retail (excluding residential mortgage retail exposures) exposures.
  3.19 The objectives of the work carried out on model validation were:

• To carry out an in-depth documented study of existing credit risk
  models for corporate and retail portfolios, including vendor supplied models used in the rating systems of some banks.

• To undertake a documented comparative study of the
  researched credit risk models.

• To document Basel II minimum requirements for the use of
  credit risk models, and to understand these requirements in line with the findings of the research into the credit risk models.

• To document current thinking on the various credit risk models
  validation techniques emerging globally.

• To document the corporate and retail models used by certain
  pilot banks chosen for the study. Banks were asked to supply comprehensive details about their models, as well as to discuss the assumptions made when applying the models, and the manner in which the models are used within their organizations.

• A benchmarking technique of validating internal models was
  performed on several pilot banks.
  3.20 The findings of the above-mentioned exercise will be discussed
  confidentially with each individual pilot bank to understand its model - validation process better, and to give it some guidance on its development, relative to the development of the rest of the industry.
  3.21 The work performed will also assist the BSD with developing a data
  specification for regularly collecting data from banks during implementation period. This aspect is further discussed later.

4 Scope of Application of Basel II In South Africa Principles upon which scope of application decisions are based
4.1 The 1988 Accord was a move in the right direction, since it focused
on capital requirements, which are a vital element in reducing the risk of bank insolvency and the potential costs of a bank’s failure to depositors. Rapid developments, in the banking sector however, have led to the increased sophistication and complexity of the business of banking. The principles of the 1988 Capital Accord have thus become outdated. One of the unintended outcomes of this disjuncture between regulations and modern developments in the banking sector is the concept of regulatory arbitrage. Regulatory arbitrage results as banks devise strategies to improve their return on regulatory capital, by keeping low-quality high-margin exposures on their books, while securitising the high-quality low-margin exposures in order to generate additional funding. This is possible because the market, in some cases, places a lower economic capital requirement on certain exposures than required by the regulations. Today, even the smallest of banks in South Africa have outgrown the 1988 Accord, which, inter alia, fails to align regulatory capital to the inherent risks taken by a bank.
4.2 The BSD subscribes to (and is encouraged by) the Basel Committee’s
belief that an improved capital-adequacy framework is intended to foster a strong emphasis on risk management and to encourage ongoing improvements in banks’ risk-assessment capabilities. This can be achieved by closely aligning banks’ capital requirements with prevailing modern risk-management practices, and by ensuring that this emphasis on risk makes its way into supervisory practices and into market discipline. The BSD therefore, considers Basel II an

appropriate framework for South African banks and will certainly adopt Basel II standards within the South African capital adequacy framework.
4.3 The current South African capital adequacy framework which is based
on the principles of the 1988 Capital Accord and comprises the Banks Act, 1990 (Act No. 94 of 1990), and the Regulations Relating to Banks, will be replaced by a new regulatory framework based on Basel II principles. The new framework will come into effect on the common implementation date (of 31 December 2006 or any later date that the Basel Committee may set).
4.4 The BSD’s decision is based on the premise that a Basel II-compliant
regulatory framework has the potential to enhance financial stability, mainly because capital adequacy will be more risk sensitive, and because the provisions of Pillar II and Pillar III of the new framework will ensure that the appropriate qualitative framework is in place to ensure that banks manage risks well.
4.5 Unlike some other jurisdictions, the BSD has decided that the current
capital adequacy framework will not be available to banks as from the common implementation date. The BSD’s decision is based on the fact that it will be possible for every registered bank in South Africa to migrate to the simplest approaches of Basel II without any major disruption or high cost.
4.6 The new Accord offers a menu of approaches to deriving the risk￾based capital charge. For credit risk, the approaches available are
the Standardised Approach (and the Simplified Standardised Approach), the Foundation Internal Ratings Based (“IRB”) approach and the Advanced IRB approach. For operational risk, the approaches available are the Basic Indicator approach, the Standardised (and Alternated Standardised) approach; and the Advanced Measurement Approach (“AMA”). The simplest of banks

can use the simplest of approaches, to derive their credit, and operational-risk capital charge, whereas the more sophisticated banks can make use of the most advanced approaches. A reduced capital charge is one explicit incentive provided to banks to improve their risk-management capability. The BSD encourages the approach adopted by the Basel Committee and believes that the regulatory framework has to cater to the needs of a variety of banks, which are at different stages in their efforts to develop their risk -management framework, and has to provide incentives to improve.
4.7 The BSD believes that a bank’s risk-management capability has a
direct impact on its competitiveness in the market. Their desire to improve competitiveness will encourage banks to improve their risk -management capabilities. The availability of the different approaches of Basel II will therefore foster an environment in which market forces will drive banks to improve their risks-management capabilities. This objective is further strengthened by the Pillar III disclosure requirements.
4.8 The BSD will therefore make all approaches (to deriving a risk -
sensitive capital-adequacy requirement for credit and operational risk) available to banks on common implementation date, except for the Advanced IRB approach for corporate credit risk and the AMA approach for operational risk. The latter approaches will, however, be available only on an exceptional case-by-case basis.
4.9 Each individual bank will therefore have to elect the most appropriate
approach to the calculation of its capital-adequacy requirement. The BSD however, will expect the more sophisticated banks to adopt the more sophisticated approaches to calculating their capital adequacy requirements, albeit at a later date. Failure of such banks to do so may lead to higher capital requirements, in terms of the supervisory review process of Pillar II.

4.10 Banks will be required to be Basel II compliant on the common
implementation date. Banks intending to adopt the foundation IRB approach (or Advanced IRB and AMA approaches on an exceptional case-by-case basis) will be required to calculate their capital requirements under both the current capital adequacy framework and Basel II requirements, as from 1 January 2006. This one-year period prior to the common implementation date is referred to as a “parallel run” period and is designed comprehensively to assess the impact of Basel II on a specific bank and the banking sector as a whole. In effect, the “parallel run” period can be regarded as a quantitative impact study, which will assist the BSD with resolving any implementation issues arising prior to the common implementation date. The statutory capital requirement that banks have to maintain, however, will be based on the current capital adequacy framework until the common implementation date.
4.10 Given the complexity of the Basel II implementation, the BSD will
adopt a flexible approach when applying Basel II rules to banks. The BSD will take into consideration the evolutionary nature of the implementation process, bearing in mind that disruption to the banks’ day-to-day business has to be minimised as far as possible. The BSD will be transparent in its thinking and will provide adequate and specific guidance on its supervisory expectations from banks. The BSD will, furthermore, create the appropriate environment to encourage banks to communicate with and learn from each other. The Accord Implementation Forum is the ideal platform to achieve this objective.

Scope of application
4.12 The base (or minimum) requirements for all registered banks in South
Africa will be compliance with the Simplified Standardised Approach to credit risk and the Basic Indicator Approach to operational risk, as well as compliance with Pillars II and III of Basel II. Although banks will not require specific approval to comply with the base requirements, the BSD will expect banks to consult with it on the appropriateness of the approach chosen, prior to the common implementation date. After a due consultative process, the BSD reserves the right to request a bank to adopt any alternate approach that the BSD may consider to be more appropriate for that specific bank, taking into consideration, inter alia, its level of sophistication, its size and its systemic importance to the banking system as a whole. Alternatively, the BSD may impose an additional capital charge – under Pillar II – on those banks which are considered to be adopting the simpler approaches solely to minimises their capital requirements.
4.13 Because there is no formal approval process for compliance with the
base requirements, banks will be deemed to be compliant if the BSD will review compliance with the qualitative requirements through its regular supervisory processes, involving, inter alia, analysis of regulatory returns, on-site reviews, prudential meetings and reviews through audit work done by internal and external auditors. The same process will apply to banks wanting to adopt the Standardised approach to the calculation of the credit-risk capital-adequacy requirements.
4.14 Banks wanting to adopt any other approach will be required to make a
formal application to the BSD, requesting the BSD’s express approval, which will be subject to compliance with minimum standards. Banks will be required to make such an application, in the prescribed manner, one year prior to the common implementation

date at the latest (i.e. 31 December 2005). Owing to the anticipated complexity of this process, banks are encouraged to commence this application process much earlier.
4.15 To inform the BSD’s implementation plans, Banks Act Circular
14/2003 requested all registered banks to complete a survey, to indicating the approaches that they were likely to adopt on the common implementation date and their plans to adopt the more advanced approaches after implementation. The results of the survey were a direct input into the BSD’s determination of the scope of application of Basel II in South Africa.
4.16 Table 1 details the results of the survey:
Table 1: Result of survey of Banks Act Circular 14/2003
Risk Approach
Percentage of total banking sector
(asset size)
Percentage of total no. of banks
Simplified Standardised
Approach
0,66% 40,5%
Standardised Approach 1,93% 10,8%
Foundation IRB Approach 90,01% 27,0% Credit risk Advanced IRB Approach 7,40% 21,6% Basic Indicator Approach 1,41% 43,2% Standardised /Alternate Standardised Approach 72,07% 32,4% Operational risk Advanced Measurement Approach 26,52% 24,3%

4.17 Ninety per cent of the banking sector (in terms of assets) have
expressed a desire to adopt the Foundation IRB approach to credit risk and the Standardised/Alternate Standardised approach to operational risk on the common implementation date. Many of the IRB banks have expressed a desire to migrate to the more advanced approaches at a later date, within one to three years of the common implementation date.
4.18 Most of the BSD’s preparatory work will thus be directed towards
gearing up for the supervision of all approaches up to the Foundation IRB approach to corporate credit risk (the Foundation IRB approach is not available for retail credit portfolios, and therefore the BSD will devise processes to supervise such portfolios under the advanced framework) and the Standardised/Alternate Standardised approach to operational risk. This does not imply that the BSD will be unable to supervise the advanced IRB approach and the AMA approach. The BSD will prepare its supervisory process for these advanced approach once it has the opportunity to work more closely with banks using these approaches. The BSD may accordingly be in a position to supervise such banks only after the common implementation date. Alternatively, the BSD will work closely, and for an extended period of time, with such advanced banks and approve them on a case-by-case basis only. Foreign bank branches and subsidiaries wishing to avail themselves of the most advanced approaches will be granted approval to adopt these approaches, but the BSD will place reliance on the home-supervisors and work closely with them in approving the use of such advanced approaches. The BSD will apply the Basel Committee’s principles on cross-border implementation1 when working with such foreign banks. Nevertheless, the BSD reserves the right to impose additional capital requirements, under Pillar II, on a foreign bank, should this be considered to be prudent within the context of the South African banking system. This discretion is

consistent with the principles of Pillar II, and Principle 3 of the “High￾level principles for the cross-border implementation of the New Accord” document of the Basel Committee.
4.19 The BSD will therefore utilise its resources most appropriately by
initially preparing for the less complex approaches. The BSD resources will be directed to the more advanced approaches only when banks indicate that they are ready to migrate to the more sophisticated approaches of the new Accord. Questions:

• Do you agree with our decision on the scope of application in South
  Africa? If not, please propose alternate approaches to determining the scope of application for South African banks, taking into consideration, inter alia, financial stability, competitive equality and various constraints.
• Do you have any level playing-field concerns? If so, please point out the
  relevant provisions of the new Accord giving rise to your concerns, as well as the potential implications if the specific issues are not addressed.

1 Basel Committee on Banking Supervision (April 2003) “High-level principles for cross￾border implementation of the New Accord” Bank of International Settlements

5 Minimum Level of Capital Adequacy
5.1 The minimum capital adequacy ratio for South African banks is 10 per
cent as opposed to the minimum 8 per cent ratio established by the Basel Committee. The BSD’s decision to impose a higher capital adequacy requirement was partly based on South Africa’s emerging market status and the accelerating pace of change and financial innovation in the banking sector, driven by the steady march of technological advancement. International markets required South African banks to hold additional capital (a premium), providing them with added comfort when engaging South African banks in business. This is evident by the average capital adequacy ratio of South African banks being at 12.6 per cent (as opposed to the 10 per cent minimum requirement) as at the end of November 2003.
5.2 Certain banks however argue that the Basel II IRB models used by
them explicitly takes into account the riskiness of the South African industry and the exposures within the industry. Therefore, they suggest that there should not be a need by the BSD to subjectively impose a higher capital requirement on South African banks, as the higher risk will in deed be factored into the final capital adequacy ratio. They therefore suggest that consideration be given to revert to a minimum capital requirement of 8 per cent in order to avoid double counting of risk.
5.3 Furthermore, some banks argue that a capital adequacy ratio formula
calibrated on 10 per cent will increase the capital adequacy ratio technically (and artificially) due to the smaller multiplier (10 instead of 12.5) in the denominator.

The BSD’s response
5.4 The BSD takes note of the banks’ argument pertaining to the 10 per
cent minimum capital adequacy requirement for South African banks. The BSD is of the view however that there is insufficient information and evidence, at this stage, demonstrating that the models used to determine the parameters for the IRB approach to credit risk accurately and consistently take into account country specific systemic risk.
5.5 In addition, the BSD would like to emphasise that the capital
requirement curves for the various portfolios in the IRB approach to credit were calibrated by the Basel Committee based on portfolios of internationally active first world banks, where the level of systemic risk is considered to be lower than that of most emerging markets. The arising implications in the view of the BSD, for South African banks using the IRB approach is that a more favourable treatment is likely if the capital adequacy calculations were based on the 8 per cent calibration.
5.6 In light of the above discussion, the BSD will at this stage retain the
10 per cent capital adequacy requirement for South African banks. The BSD however encourages banks to supply it with information in support of their arguments to lower the capital adequacy requirements. In addition, the BSD will explore this issue quantitatively during the proposed fourth quantitative impact study.
5.7 With reference to banks’ technical argument suggesting an increased
capital adequacy ratio due to the lower multiplier in the capital adequacy ratio formula calibrated on 10 per cent, the BSD acknowledges the potential for such a technical anomaly. A possible solution is to calibrate the multiplier to the 8 per cent standard (i.e. a multiplier of 12.5 instead of 10), while maintaining a minimum capital

adequacy requirement of 10 per cent for South African banks. This approach may enhance comparability with banks of other countries. The BSD will also explore this issue further during the proposed fourth quantitative impact study.

6 Credit Risk – Standardised & Simplified Standardised Approach
6.1 This chapter sets out some aspects of the BSD’s initial thinking on the
Simplified Standardised approach and the Standardised approach to credit risk. This paper does not seek to consult on all aspects related to the Standardised and Simplified Standardised approaches. The items addressed in this consultative paper communicate only the BSD’s proposed approach to the areas in which it has formulated an opinion on how best to treat the particular issue. The BSD will consult on the other areas at a later date, as implementation work progresses. Some items may require qualitative and quantitative testing before the BSD will be in a position to make a decision. Such items may, in due course, be discussed at the AIF where the BSD will seek guidance from the industry.
6.2 The Basel II proposals remain incomplete without the national
supervisor making pronouncements on several national-discretion items available in CP 3. The following section addresses some of the national-discretions items. Items of national discretion that are not specifically addressed here will be addressed in future consultative papers. External credit-assessment institution (“ECAI”)
6.3 National supervisors are responsible for determining whether an ECAI
meets the eligibility criteria set out in paragraph 60 of CP 3. During the course of 2004, the BSD will consult with ECAIs to gain an understanding of their ratings-assignment process. The BSD will also

discuss how the ECAIs intend complying with the eligibility criteria and develop its own internal guidelines for the recognition of ECAIs. The BSD hopes to provide a list of recognized ECAIs by year-end.
6.4 The Standardised approach is based on an exposure being assigned
a risk weight mapped to an ECAI’s risk rating of that counterparty. ECAIs have developed a two-tier rating system, namely, local￾currency ratings and foreign-currency ratings. The former cater for local transactions, whereas the latter were developed to largely reflect country risk. The BSD’s approach to the issue of ratings is as follows:

• When claims are denominated and funded in the local currency,
  and when the transaction has been legally concluded in the jurisdiction of the Republic of South Africa (“RSA”), the local￾currency ratings will be applicable.
• All other claims will be subject to the foreign-currency ratings.
  Questions:
• Do you have any comments on the level of ratings penetration in this
  country, and the potential impact on capital requirements under the Standardised approach? Please provide detailed comments. Proposed treatment of certain national discretion items
  6.5 Claims on sovereigns
  6.5.1 At national discretion, a lower risk weight may be applied to
  banks’ exposure to their sovereign (or central banks) of incorporation denominated in domestic currency and funded in that currency. When this discretion is exercised, other national supervisory authorities may also permit their banks to apply the same weight to domestic currency exposures to

such sovereign (or central bank) funded in that currency (paragraph 28, CP 3).
6.5.2 The first part of the discretion conforms to the BSD’s current
treatment of sovereign and central bank treatment and, therefore, the BSD proposes exercising this part of the discretion. Claims against the central government of the RSA and claims against the central bank of the RSA will be 0 per cent risk-weighted, regardless of the external rating assigned to the sovereign or the central bank.
6.5.3 The second part will require careful consideration of the
implications if the discretion is exercised. Any decision could give rise to level playing-field concerns. The BSD therefore, will not make any proposal on the treatment of the latter part of this national-discretion item at this stage.
6.5.4 The lower risk weight may be extended to the risk weighting
of collateral and guarantees (paragraph 28, footnote 11, CP 3). The BSD proposes to exercise this discretion, since it is in line with the current regulatory treatment whereby a weight of 0 per cent risk is applied to performance-related guarantees on behalf of the RSA central government. The same treatment is proposed for collateral and guarantees by the central bank of the RSA.
6.5.5 For the purpose of risk weighting claims on sovereigns
supervisors may recognise the country-risk scores assigned by Export Credit Agencies (“ECAs”). To qualify, an ECA must publish its risk scores and subscribe to the methodology agreed by the Organisation for Economic Co-operation and Development (OECD”) (paragraph 29, CP 3). The BSD proposes to recognise the country-risk scores assigned by ECAs, subject to the minimum requirements being met.

6.6 Claims on banks
6.6.1 There are two options for claims on banks (paragraph 34 to
38, CP 3):

• Under Option 1, all banks incorporated in a given country
  will be assigned a risk weight that is one category less favourable than that assigned to claims on the sovereign of that country. For claims on banks in countries, with sovereigns rated BB+ to B- and on banks in unrated countries, however, the risk weight will be capped at 100 per cent.
• Option 2 bases the risk weighting on the external credit
  assessment of the bank itself, with claims on unrated banks being risk weighted at 50 per cent. Under this option, a preferential risk weight that is one category more favourable may be applied to claims with an original maturity of three months or less, subject to a floor of 20 per cent. This treatment will be available to both rated and unrated banks, but not to banks risk weighted at 150 per cent.
  6.6.2 Section 21(9) of the Regulations relating to Banks provides
  that banks in the RSA are risk weighted at 20 per cent. Exposures to non-OECD banks attracts a risk weight of 100 per cent. For the Standardised approach, the QIS 3 results revealed that there was an increase of 49,6 per cent in risk￾weighted assets. This can be explained by the range of risk weights (20 per cent to 150 per cent) applicable under Option 2 of the Standardised approach, as opposed to the current fixed 20 or 100 per cent, depending on whether the bank is in an OECD or non-OECD country. The contribution to an

increase in the overall risk-weighted assets was merely 0.9 per cent for the Standardised Approach.
6.6.3 The BSD considers Option 2 to be more risk sensitive and in
line with the objectives of the Basel Committee to improve of the current capital-adequacy framework. The BSD, therefore, proposes to adopt Option 2 for the treatment of banks.
6.6.4 The BSD will furthermore, adopt the preferential risk weight
that is one category more favourable for claims with an original maturity of three months or less, subject to a floor of 20 per cent. The BSD’s decision is based on the belief that the short-term interbank market is less risky and, therefore, open to a more favourable treatment. Questions:

• Do you agree with the BSD’s proposed decision for the treatment of claims
  on banks? Please substantiate your response.
• Because of the disparities between the current and the proposed
  treatment of claims on banks, the capital requirement may be increased under the proposed treatment. Please provide your comments on the likely impact of the proposed treatment of claims on banks.
  6.7 Claims on corporates
  6.7.1 Supervisory authorities should increase the standard risk
  weight for unrated claims when, it is the authorities judgment that a higher risk weight is warranted by the overall default experience in their jurisdiction. As part of the supervisory review process, supervisors may also consider whether the credit quality of corporate claims held by individual banks should warrant a standard risk weight higher than 100 per cent (paragraph 41, CP 3).

6.7.2 Further work is required before the BSD can reach a
proposed decision on this item of national discretion. There may be level playing-field concerns between the Standardised and the IRB approach. The IRB approach does not impose a cap of 100 per cent or 150 per cent on corporate claims. A higher risk weight (greater than 150 per cent) may well result for low -quality corporate exposures. Therefore, the treatment of the specific low quality corporate exposure under the Srandardised approach, may be more favourable than the treatment required in terms of the IRB approach. The BSD’s work with IRB banks will inform it of low-quality corporate exposures, and the BSD will accordingly exercise national discretion by revising the risk weighting for such corporate exposures from time to time and on an ongoing basis
6.7.3 A risk weight of 100 per cent will be applicable to all corporate
exposures under the Simplified Standardised approach. This is in line with paragraph 42 of CP 3, which provides that, at national discretion, supervisory authorities may permit banks to risk weight all corporate claims at 100 per cent without regard to external ratings. Questions:

• Do you agree with the BSD’s proposed approach for corporate
  exposures? Please substantiate your response.
• Do you have any level playing-fields concerns? Please provide
  detailed comments.
• Do you agree with the BSD’s proposed treatment to address
  potential level playing-fields concerns? If not, please suggest alternatives.

6.8 Claims included in the regulatory retail portfolio
6.8.1 The retail portfolio is an area in which the Basel Committee
has introduced most significant changes compared with the current regulatory treatment of retail exposures. The risk weighting is lower, and the IRB curves are flatter, resulting in a potential overall decrease in capital requirements, which was confirmed by QIS 3. Retail exposures will be risk weighted at 75 per cent in terms of CP 3, whereas the current treatment imposes a risk weight of 100 per cent.
6.8.2 Fully secured mortgage exposures will receive a risk weight of
35 per cent - provided certain conditions are met – as opposed to 100 per cent and 50 per cent when the loan-to￾value (“LTV”) ratio is 80 per cent, under the current treatment.
6.8.3 QIS 3 demonstrated the overall decrease of 19.7 per cent for
risk-weighted assets for all retail portfolios. This contributed a reduction of 5 per cent in the overall change in risk-weighted assets as compared with the current capital requirements.
6.8.4 There may be significant changes in the nature of retail
banking in South Africa. The specific aspect that comes to mind is the high growth rate in house prices over the past number of years. The emergence of a buy-to-let market is another factor that has to be studied. These events signal the need for an in-depth analysis review of the environmental changes occurring and the likely impact of such changes on the retail portfolios of banks.
6.8.5 Application of the lower risk weights to portfolios that may be
undergoing significant changes may have unintended consequences. Supervisors in some jurisdictions are of the

opinion that the risk weight of 75 per cent for retail exposures is too low, considering the nature of their market, and are applying a risk weight of 100 per cent Supervisors are also considering the risk weight of 35 per cent applicable to residential mortgage exposures, which in some cases may be too low, whereas in other cases it may be too high, considering the low risk inherent in such exposures.
6.8.6 The BSD is therefore not yet in a position to make any
pronouncement on the treatment of retail exposures. The BSD working through the AIF, will study the impact of the CP 3 proposals on the retail portfolio of banks. Questions:

• Do you have any comments on the potential impact of the lower
  risk weight on the retail portfolio in South Africa? Please substantiate your comments.
• Do you have any comment on the nature of developments in this
  market, especially the residential retail property market, and the implications of Basel II on the development of the market?
• Please provide historical trends in retail banking over the past ten
  years. This will be used as input into a more formalised study.
• Please provide any other information pertaining to retail portfolios,
  which you may consider to be relevant.
  6.9 Higher-risk categories
  6.9.1 National supervisors may decide to apply a risk weight of 150
  per cent or higher to reflect the higher risks associated with some other assets, such as venture capital and private-equity investments (paragraph 53, CP 3).

6.9.2 Through a data-collection exercise, the BSD will learn more
about the different types of asset, and the risks inherent therein. The BSD’s work with IRB banks will inform this process further. The process may reveal that certain asset types may indeed require a higher risk weight. The results of the IRB approach will be a good guide to the inherent risks within such assets. In order to ensure, inter alia, a level playing-field between banks using the Standardised approach and these using the IRB approach, the BSD may indeed be required to exercise its discretion and to apply higher risk weights to certain high-risk exposures. The BSD cannot yet make any pronouncement on this item of national discretion, save to say that it will be a topic for ongoing review and discussion. Questions:

• Based on the experience of your bank, can you reveal assets that
  would merit a risk weight of 150 per cent or higher?
• Do you have any level playing-fields concern? If so, please
  provide further details and proposed recommendations.
  6.10 Gold bullion
  6.10.1 At national discretion, gold bullion held in own vaults or on an
  allocated basis to the extent that such bullion is backed by bullion liabilities can be treated as cash and therefore risk weighted at 0 per cent (paragraph 54, footnote 24, CP 3).
  6.10.2 The BSD will exercise this discretion, since it is consistent
  with the current regulatory treatment of gold bullion.

6.11 Unsolicited ratings
6.11.1 As a general rule, banks should use solicited ratings from
eligible ECAIs. National supervisory authorities, however, may allow banks to use unsolicited ratings in the same way as solicited ratings. There, however, may be a potential for ECAIs to use unsolicited ratings to place pressure on entities to obtain solicited ratings (paragraph 78, CP 3).
6.11.2 Under the Standardised approach the BSD recognises that
rating agencies have a significant influence on the capital￾adequacy requirements for corporate and other exposures. In the light of this influence, the BSD believes that the activities of ECAIs’ should be controlled, while still ensuring that a fair rating is applied to counterparties. In order to avoid potentially unscrupulous behaviour, the BSD will allow solicited ratings from eligible ECAIs only.
6.12 Credit-risk mitigation
6.12.1 In the case of repo-style transactions in which certain explicit
conditions are satisfied, and when the counterparty is a core market participant, supervisors may apply a zero H (haircut) (paragraph 141, CP 3).
6.12.2 Core market participants may include, at the discretion of the
national supervisor, the following:

• Sovereigns, central banks and PSEs.

• Banks and securities firms.

• Other financial companies (including insurance
  companies) eligible for a risk weight of 20 per cent.

• Regulated mutual funds that are subject to capital or
  leverage requirements.

• Regulated pension funds.

• Recognised clearing organisations (paragraph 142, CP
  3).
  6.12.3 The conditions imposed in paragraph 141 ensures that these
  specific repo-style transactions are, being of a short-term nature, low risk. Subject to the conditions being met, the BSD will exercise this discretion and allow a zero haircut for the following core market participants:

• Central government and the central bank of the RSA.

• Banks in the RSA eligible for a risk weight of 50 per cent.

• Other financial companies eligible for a risk weight of 20
  per cent.

• Recognised clearing organisations.
  Question:

• Do you agree with our proposed treatment for short-term repo￾style transactions? Please substantiate your response.

• If not, please propose an alternative treatment and motivate your
  recommendations.
  6.12.4 Sovereign guarantees - As specified in paragraph 28, a lower
  risk weight may be applied at national discretion to a bank’s exposure to the sovereign (or central bank) where the bank is incorporated and when the exposure is denominated in domestic currency and funded in that currency. National authorities may extend this treatment to portions of claims guaranteed by the sovereign (or central bank) when the guarantee is denominated in the domestic currency and the exposure is funded in that currency. In some cases, a claim

may be covered by a guarantee that is counter-guaranteed by a sovereign, provided certain explicit conditions are met.
6.12.5 In line with the BSD’s decision to exercise its discretion to
apply a lower risk weight (0 per cent) to banks’ exposure to their sovereign or central banks, the BSD’s will similarly exercise its discretion with regard to sovereign or central bank guarantees.

7 Credit Risk – Internal Ratings Based (“IRB”) Approach General comments on the IRB Approach
7.1 The IRB approach introduces a significant departure from the 1988
Basel Accord (on which the current capital adequacy framework is based). The IRB approach will allow banks to rely on their own internal estimates of risk components in determining the capital requirement for a given exposure. The risk components include measures of probability of default (“PD”), loss given default (“LGD”), the exposure at default (“EAD”) and effective maturity (“M”)2 .
7.2 QIS 3 was a first opportunity for both banks and the BSD to work on
the IRB proposals and to determine the overall impact on banks’ capital-adequacy levels. QIS 3, however, was done on a best -efforts basis, because banks are not yet expected to be fully compliant with the minimum requirements of Basel II. Therefore, the results are inconclusive, and much work remains for the successful implementation of the IRB approach.
7.3 Banks have made some progress with the design and use of rating
systems, as part of their overall credit-risk management. Banks, however, are far from meeting the minimum standards required by the new Accord. Data constraints, for example, are an area that further hampers the development and use of rating systems. Full compliance with IRB requirements may well extend beyond the common implementation date. It may be more prudent to delay moving to the IRB approach than to rush for compliance on the common implementation date. Since the base requirement is the Simplified

Standardised approach, banks can essentially set their own timelines for implementation of the IRB approach. Competitive implications, however, may drive banks to implement the IRB approach sooner. The BSD encourages such market-driven incentives, provided the implementation of the IRB approach is done prudently. The BSD’s preparation for developing a supervisory framework for IRB banks will largely be informed by the work done by banks.
7.4 Owing to the evolving nature of the IRB implementation process, the
BSD will not fully consult on the IRB approach at this stage. The BSD will continue to hold in-depth discussions with banks and to consult (through ongoing consultative papers) with all relevant stakeholders as further progress is made. This section, therefore, addresses only some areas of the IRB approach.
7.5 Some banks in South Africa have expressed their desire for a flexible
approach to implementation, citing the approach proposed by the British Bankers’ Association (“BBA”) and the London Investment Banking Association (“Libya”) as the appropriate strategy for South Africa. Their proposal seeks to postpone the adoption of the advanced IRB approach to a later date than the common implementation date of December 2006. The suggested approach is therefore to allow banks to operate under the 1988 Accord until 2010, when British banks should be ready to adopt the Advanced IRB approach. The aim is to reduce the cost of compliance and to ensure a smooth transition to Basel II.
7.6 The BSD has considered the above proposal suggested by banks,
and consider it to be inappropriate for South Africa owing to the following reasons, inter alia:

2 Basel Committee on Banking Supervision (April 2003) “Consultative Document: The New Basel Capital Accord” Bank for International Settlements

• The BSD consider it to be important to adopt an evolutionary
  approach, reaping benefits along the way, as banks progress towards the most advanced approach. An evolutionary approach ensures that banks gain the appropriate benefit of a reduction in capital requirements, if they are able to quantify and demonstrate the lower risk within their business, as well as to comply with the minimum requirements. The BSD, therefore, places value on migration from the Standardised approach to the Foundation IRB approach and, then, to the Advanced approach to credit risk. This moves away from the notion of “all or nothing”, and appropriate recognition is given to banks of differing levels of sophistication.
• The BSD encourages a close working relationship with banks as
  they progress towards the more advanced approaches available under Basel II. In recognising the complexity of implementing the IRB approach, the BSD will adopt a flexible approach, allowing, for example, partial implementation during the transition phase. If the transitory (that is, Standardised and Foundation IRB) are part of the regulatory framework, the BSD will have the opportunity to work closely with banks through its regular supervisory processes.
• The base requirement for Basel II compliance in South Africa is
  the Simplified Standardised approach. Therefore, it is left entirely up to the bank to decide on the opportune timing for implementation and adoption of the IRB approach. Accordingly, some banks may well be compliant with the IRB approach only at the end of 2009. Should any bank, however, wish to avail itself of the IRB approaches on or immediately after the common implementation date, the regulatory point of view is that such approaches should be made available to the bank. This will ensure that banks adopting the more sophisticated approaches gain a competitive advantage in terms of, inter alia, preferential pricing, reduced capital requirements and a more efficient credit-

risk management process. The BSD believes that a healthy competitive environment is favourable for the maintenance of financial stability. Consultation
7.7 As already mentioned, the BSD will adopt a consultative approach to
the implementation of Basel II. The BSD will inform its implementation initiatives by learning from the vast amount of work that banks are already doing towards preparation for compliance with the standards of the IRB approach. Consultation will be in the form of, inter alia:

• In-depth discussions at the AIF, resulting in the different
  committees making appropriate recommendations to the BSD.
• Bilateral discussions with individual banks.
• Pilot studies involving various banks from time to time.
• Obtaining information through detailed questionnaires, which will
  also allow banks to asses themselves – in effect, the BSD will be request banks to carry out self-assessments in various areas to understand their level of preparedness for Basel II.
• Benchmarking of the progress of peer banks, and giving
  appropriate feedback to banks, that may be swaying out of line with industry norms or who may require further guidance.
• Consultation with other supervisory bodies and other
  internationally active banks in developed countries, in order to inform the BSD’s implementation efforts.
• Discussions with various consultants and other bodies that may
  be developing solutions to assist banks in their implementation initiatives.

7.8 In this first consultative paper, the BSD is providing its initial thoughts
on the following areas only:

• Adoption of the IRB approach across assets classes –that is,
  recognition of partial use.
• Guidance on transitional arrangements.
• Discussion of the BSD’s initial thinking on IRB rating systems, in
  particular, the use of models.
  The views expressed here merely reflect the BSD’s initial thinking on the subject and should not be construed as a final regulatory requirement. The BSD’s views are aimed at eliciting comments from banks and to obtain appropriate recommendations when banks may have better alternative ideas. The BSD comments are furthermore based only on the provisions of CP 3, which may change as the Basel Committee continues its work in addressing concerns raised by banks about CP 3. The BSD will accordingly change its views and requirements to reflect changes introduced in the final Accord, due in June 2004. Partial use of the IRB approach
  7.9 Paragraphs 225 to 231 of CP 3 deal with the adoption of the IRB
  approach across asset classes.
  7.10 Once a bank adopts an IRB approach for part of its holdings, it is
  expected to extend the approach across the entire banking group. The Basel Committee recognizes, however, that, for many banks, it may not be practicable, for various reasons, to implement the IRB approach across all material asset classes and business units at the same time. Furthermore, once on IRB, data limitations may mean that banks can meet the standards for the use of own estimates of

LGD and EAD for some but not all of their asset classes/business units at the same time.
7.11 The BSD realises that banks may not be able to apply consistent
treatment to exposures in certain emerging-market countries owing to various constraints, as well as the inability of that specific emerging market country to supervise IRB banks. In such cases, banks may be forced to apply the Standardise (or Simplified Standardised Approach) to their portfolio of exposures in that country. The BSD will take cognisance of such constraints in determining compliance with these provisions of the new Accord. The BSD will allow, subject to explicit supervisory approval, exemption from the IRB approach for some exposures in non-significant business units and asset classes that are immaterial in terms of their size and perceived risk profile, or when the IRB approach cannot be implemented owing to certain constraints. This exemption is limited to a maximum of 20 per cent of the total worldwide assets of a bank. Capital requirements for such operations will be determined according to the Standardised (or Simplified Standardised) approach. Under the Pillar II supervisory- review process, the BSD determine the adequacy of the capital levels for such exposures in line with the inherent risk taken. The BSD may require an additional capital charge for such exposures, if it considers the capital-adequacy level to be inappropriate for the risks taken.
7.12 Banks, however, are encouraged to migrate exposures (referred to
above) that are permanently exempted to the IRB approach, as soon as the constraints no longer exist, or when they are able to apply the IRB approach methodology to such exposures. Consistent treatment of exposures is in line with prudent risk-management principles and is therefore encouraged.
7.13 The BSD is aware that application of the IRB approach will be simpler
for some portfolios and may pose difficulty for other portfolios. Portfolios for which application of the IRB approach may pose

difficulties might include middle-market corporates (in respect of which default data may be limited), and high-quality, low-default corporates, in respect of which insufficient default events (and therefore, a small number of statistically insignificant data points) make it difficult to model the risk parameters for such exposures. The BSD will accordingly allow banks to adopt a phased rollout of the IRB approach across different portfolios. All material asset classes and business units, however, will have to be on the IRB approach within three years of the common implementation date. Therefore, a bank must have fully adopted the IRB approach by 31 December 2009 for all material asset classes and business units, unless the BSD grants permission to extend this period on good grounds demonstrated by the bank through an explicit approval process.
7.14 Banks are required to produce a detailed implementation plan,
specifying to which extent and when they intend to roll out IRB approaches across significant asset classes (or subclasses in the case of retail) and business units over time. The plan should be exacting, yet realistic, and has to be agreed with the BSD. It should be driven by the practicality and feasibility of moving to more advanced approaches, and should not be motivated by a desire to adopt a Pillar I approach to minimises a bank’s capital charge (paragraph 227, CP 3).
7.15 All potential IRB banks are accordingly requested to submit a detailed
IRB implementation plan to the BSD by no later than 31 December
2004. The plan should as a minimum contain the following:

• An indication of the approach (Foundation or Advanced
  approach) and a motivation for adopting the specific approach.

• A detailed rollout plan, as discussed above.

• Specific details of a corporate-governance framework in place to
  ensure appropriate oversight over the implementation process, as well as after implementation.

• Some form of cost/benefit analysis.

• Any potential constraint and plan to overcome such constraints.

• Risks within the implementation process and risk-mitigation
  strategies for managing or mitigating such risks.

• Resource planning.

• The likely qualitative and quantitative impact of implementing the
  IRB approach.

• Detailed processes to ensure data integrity of input into rating
  systems, as well as process to validate the same.
  Transitional arrangements
  7.16 Proposed IRB banks will be required to comply with the “parallel-run”
  process discussed above. They will accordingly be required to calculate their capital-adequacy requirement under both the current and the Basel II regulatory framework. The parallel process commences on 1 January 2006. and ends on 31 December 2006. The BSD will regard this process during the parallel run, formal capital requirements will still be based on the current regulatory framework.
  7.17 The transition period starts on the date of implementation and will last
  for a period of three years from that date. During the transition period, the following minimum requirements will be relaxed (paragraph 233, CP 3).
  7.18 For corporate, sovereign and bank exposures under the foundation
  approach, the requirement that, regardless of the data source, banks use at least five years of data to estimate the PD will be relaxed to a

minimum requirements of two years of data. The requirement will increase by one year for each of the three years of transition.
7.19 For retail exposures, the requirement that, regardless of the data
source, banks use at least five years of data to estimate loss characteristics will be relaxed to a minimum requirement of two years. The requirement will increase by one year for each of the three years of transition.
7.20 The stringent requirements for the Advanced IRB approach will not be
relaxed, and banks will have to demonstrate that they meet these standards prior to being given the approval to adopt this approach.
7.21 For corporate, sovereign, bank and retail exposures, the BSD will
relax the requirement that a bank demonstrate that it has been using a rating system that was broadly in line with the minimum requirements for at least three years prior to qualification. The requirement will increase by one year for each of the three years of transition. This decision is based on an understanding that compliance with the use test will require significant change processes in the bank in order to transform the organisation from managing credit risk only qualitatively (using expert judgment system) to quantitatively (using statistical models underpinned by a vast amount of historical data) and qualitatively. IRB ratings systems
7.22 The area of IRB ratings systems is a complex and much work remains
to be done by banks and the BSD. The BSD plans to conduct assessments of banks, and IRB readiness during the latter half of
2004. This process will inform the BSD of the methods, processes,
controls and data collection and IT systems that support the assessment of credit risk by banks. The BSD is currently developing

a detailed and systematic methodology to carry out the bank - readiness assessments since work with the banks will allow the BSD to develop detailed review procedures (as part of the supervisory review process) to ensure that banks’ systems and controls are adequate to serve as the basis for the IRB approach.
7.23 To be eligible for an IRB approach, a bank has to demonstrate to its
supervisor that it meets the IRB requirement, both at the outset and on an ongoing basis (paragraph 354, CP 3). If a bank intents to adopt the IRB approach on the common implementation date, it should already have put in place some form of rating system, to provide a meaningful assessment of borrower and transaction characteristics; as well as a meaningful differentiation of risk and reasonably accurate and consistent quantitative estimates of risk.
7.24 The onus is on a bank to demonstrate that its IRB rating system
complies with Basel II requirements.
7.25 The BSD encourages banks to submit a detailed plan of how they
intend to implement an IRB-compliant credit-rating system. Prior to approving a rating system, the BSD would ideally like to work closely with banks for an extended period of time, in order, for example, to determine whether the models being utilised to assess credit risk are indeed producing accurate and consistent results. The BSD therefore, suggests that banks commence this consultation process as soon as possible. IRB model approval and validation
7.26 Many potential IRB banks have expressed a desire to learn of the
BSD’s supervisory expectation of the models that such banks intend to use or are already using as part of their credit-risk management. Basel II minimum requirements clearly set out the supervisory

expectation from banks. The BSD will add to these requirements and provide additional guidance on an ongoing basis, as it works closely with banks during the implementation phase.
7.27 In terms of paragraph 379 of CP 3, credit-scoring models and other
mechanical procedures are permissible as the primary or partial basis of rating assignment and may play a role in the estimation of loss characteristics. Sufficient human judgment and human oversight is necessary to ensure that all relevant information, including that which is outside the scope of the model, is taken into consideration and that the model is used appropriately.

• The burden is on a bank to satisfy its supervisor that a model or
  procedure has good predictive power and that regulatory capital requirement will not be distorted as a result of its use. The variables that are input into the model must form a reasonable set of predictors. The model must be accurate, on average, across the range of borrower or facilities to which the bank is exposed, and there must be no known material biases.

• A bank has to have in place a process for vetting data inputs into
  a statistical-default or loss -prediction model, including an assessment of the accuracy, completeness and appropriateness of the data specific to the assignment of an approved rating.

• When model results are combined with human judgment, the
  judgment must take into account all relevant information not considered by the model. The bank has to have written guidance, describing how human judgment and model results are to be combined.

• A bank has to have procedures for human review of model￾based rating assignments. Such procedures should focus on
  finding and limiting errors associated with known model weaknesses and must also include credible, ongoing efforts to improve the model’s performance.

• A bank has to have a regular cycle of model validation including
  monitoring of model performance and stability, review of model relationships and testing of model outputs against outcomes.
  7.28 Banks have to document, in writing, their rating systems’ design and
  operational details (paragraph 380, CP 3). Banks are encourage to start submitting such documentation to the BSD for review. Such documentation will assist the BSD in conducting the bank-readiness reviews.
  7.29 If a bank employs statistical models in the rating processes, the bank
  has to document its methodologies and submit such documentation to the BSD for review. If a bank intends to use a model obtained from a third party, the onus is on the bank and the vendor to satisfy the BSD that the minimum requirements are met. The BSD therefore, encourages banks to meet with it together with their model vendor, to discuss how they intend to comply with the minimum requirements.
  7.30 The BSD does not consider it appropriate to approve model vendors,
  owing to the systemic implication of such approval. The use of an acceptable model acquired from a vendor is no guarantee that a banks’ rating system will be approved by the BSD. The model is merely one component of the rating system, and banks have to comply with all requirements of the rating -system design for which the new Accord provides.
  7.31 Over the past six months, the BSD has undertaken extensive
  theoretical work in the area of model approval and validation. The extent of the work was addressed briefly in Chapter 3. The BSD will continue this work, and, through extensive consultation, will devise an appropriate model-approval and -validation methodology. This process will also be informed by the work undertaken at the Basel Committee level. The Accord Implementation Group has established an IRB validation subgroup, which will deliberate on this subject and,

possibly, produce guidelines for supervisors. Since the BSD forms
part of this group, progress made by the group will be communicated
on a regular basis, preferably through the AIF.
7.32 The BSD’s work thus far has confirmed its initial intention to collect
credit-risk data from banks - on a regular basis - during the transitory implementation phase. Initial thinking points to a need for data collection for the following reasons, among many others:

• A growing body of evidence suggests that there is a need for
  reference data for validation of credit-risk models by supervisors. By collecting PD data, for example, the BSD would contribute towards a database that it may require in future.
• Regular data collection will inform the BSD of the
  implementation progress of banks.
• Data will assist the BSD in guiding banks – the BSD will perform
  a regular benchmarking exercise between the IRB banks and will give confidential feedback to banks , so that they can refocus their efforts if they are straying away from what is becoming the industry norm.
• Data will help the BSD with developing a supervisory review
  process for the supervision of banks’ credit-risk management processes.
  7.33 The BSD therefore, suggests that it collect relevant credit-risk data on
  a quarterly (or monthly) basis from potential IRB banks as from 1 June 2004. At this stage, the BSD will require banks to provide PD data for a given set of corporate exposures. The BSD will do further work on the exact specification for data collection and provide the specification to banks before June 2004. The BSD has received a data specification from some banks, but would like to consult more broadly.

Questions:

• Do banks agree with the BSD’s intention to collect certain credit-risk data
  on a regular basis? Please substantiate your response.
• Banks are requested to provide the BSD with their thinking on which data
  the supervisor should be collecting and how this data should be used to assess banks’ credit-risk management processes, and to validate credit￾risk models.
• Banks are requested to provide the BSD with the type of data that they are
  currently collecting and any validation work that they may be doing.
• Banks are requested to provide any further information it may deem
  relevant in guiding BSD’s work in this area.

8 Operational Risk
8.1 The quantitative approach to the supervision of operational risk is a
new concept for the BSD, which will continue to work with banks in order to devise a supervisory review process for the supervision of this risk.
8.2 This chapter consults on a few aspects of operational risk and is not
intended to be a comprehensive consultative process. The BSD will develop further guidance and supervisory expectation pertaining to operational risk, in due course.
8.3 In the interim, banks are encouraged to keep the BSD informed of any
new development pertaining to the measurement and management of operational risk in their organisation.
8.4 The BSD supports the Basel Committee’s decision to levy a capital
charge for operational risk. The importance of operational-risk management has been demonstrated by bank failures such as that of Barings Bank. The additional capital charge for operational risk will act as a buffer against losses that may not have been covered in the past and which may have left depositors vulnerable. The additional capital charge will further ensure that the board of directors and senior management of banks devote the appropriate level of attention to this risk, which, in some cases, may have been neglected in the past. Alternate standardised approach
8.5 The Alternate Standardised Approach (“ASA”) to operational risk
introduces a volume-based charge for the corporate and retail business lines, as opposed to a charge based on gross income.

8.6 The QIS 3 exercise demonstrated that smaller banks lending in higher
risk markets are severely penalised by the higher margins levied for the added risk taken. The higher margin income in some cases acts as a buffer against portfolio losses. An operational-risk charge based on gross margin will unnecessarily impose a higher penalty on such banks owing to the higher gross income derived from the higher margins. An increase in capital requirements of up to a 40 percent resulted for some smaller banks. The ASA was introduced to mitigate this unintended consequence.
8.7 The BSD will allow a bank to adopt the ASA for corporate and retail
business lines, provided the bank is able to satisfy the BSD that this alternate approach provides an improved basis, by, for example, avoiding double counting of risk (paragraph 615, footnote 91, CP 3). Partial use
8.8 In responses to the BSD’s survey (Banks Act Circular 14/2003), many
banks expressed their desire to adopt the partial-use option for operational risk. Under the partial-use provisions, a bank must calculate the operational risk charge for some business lines by using one approach and might use another approach for the rest of the business lines.
8.9 The AMA approach is new and will develop in an evolutionary
manner. Accordingly, the BSD is amenable to banks adopting a partial-use approach to the implementation of operational-risk requirements in terms of Basel II standards.
8.10 Banks, however, will be required to discuss their plans for partial use
with the BSD. They will have to demonstrate the difficulty of using a consistent approach across all business lines. Banks should not

make use of the partial-use provisions in order to minimise operational-risk capital requirements. Questions:

• Banks are encouraged to provide a detailed rollout plan for the
  implementation of the Basel II operational-risk standards. The plan should indicate whether the partial-use option will be exercised, and details of the phased approach to implementation and the likely impact on the organisation, as well as capital requirements, shall be provided.
• Do you have any suggestion on how the partial use of the AMA
  approach can best be rolled out – for example, should the partial-use requirement be applied across business lines or legal entities? Data requirements
  8.11 For the AMA, banks have to track internal loss data according to
  certain specific criteria. The tracking of internal-loss event data is an essential prerequisite to the development and functioning of a credible operational-risk measurement system (paragraph 630, CP 3).
  8.12 Internally generated operational-risk measures used for regulatory
  capital purposes have to be based on a minimum five-year observation period of internal losses, whether the internal loss data are used directly to build the loss measure or to validate it (paragraph 632, CP 3).
  8.13 A bank’s operational-risk measurement system has to use relevant
  external data (either public data and/or pooled industry data), especially when there is reason to believe that the bank is exposed to infrequent, yet potentially severe, losses (paragraph 634, CP 3).

8.14 Accordingly, the BSD encourages banks to commence collecting loss
data from internal and external sources. Owing to confidentiality issues, banks have been finding it difficult to establish infrastructure in order to pool loss data. The BSD believes that pooling of data is crucial and would be willing to facilitate discussions, through the AIF, between banks in order to initiate this data-collection process.

9 Future Plans
9.1 This chapter sets out the BSD’s preliminary implementation plans for
2004.
Resource planning
9.2 The BSD has done extensive work on developing a strategic plan for
the implementation of Basel II. The BSD has also done some thinking on the work streams that will be required and the projects that have to be initiated.
9.3 The BSD plans to recruit specialised staff over the coming months to
support the implementation process, and to assist the supervisory function in the supervisory review process under Basel II
9.4 If necessary, the BSD will consider the possibility of employing
external consultants to assist it in some highly specialised areas of implementation, Accord implementation forum
9.5 The AIF proved to be successful with the banks in 2003, since it gave
them an opportunity to discuss implementation ideas and issues freely and openly with both the BSD and their peers.
9.6 The BSD will resume the AIF meetings in February 2004. A detailed
meeting schedule is appended as Annexure B.

9.7 The first Steering Committee meeting will review the composition of
the various subcommittees, and various amendments may be necessary to ensure that all stakeholders are fairly represented. This step is important because some smaller banks have expressed a concern that they have been involved in the process thus far. Bank-readiness review
9.8 From June 2004, the BSD will conduct indepth reviews of
banks’ readiness for Basel II implementation.
9.9 The BSD will provide detailed guidelines of what banks can expect
during this review process.
9.10 The BSD will form a specialised team to carry out these reviews.
9.11 As discussed above, the BSD requests that banks submit, as soon as
possible, their detailed implementation plans, which will inform the BSD’s review work in the latter part of 2004.
9.12 The BSD will also develop a detailed data specification and, if banks
are agreeable, start collecting data as soon as possible. High-quality data are critical for the formulation of effective internal risk assessments. Banks may need to implement substantial changes to their internal systems to prepare for appropriate data collection and revised reporting requirements. These changes may require integration and modification of systems and new software. The BSD encourages banks to consider their data needs and to comprehend fully the techniques they will need to use to derive appropriate estimates of risk-based data. The BSD would also encourage private initiatives/processes for credit information sharing and for assessing comparability of pooled data with internal bank experience.

9.13 Bank-readiness reviews may include, inter alia:

• The identification of a subset of banks to engage in a
  comprehensive review, focusing on each bank’s internal practices, readiness for Basel II and key implementation challenges.
• Requesting banks to conduct informal self-assessments of their
  readiness based on minimum requirements.
• Reviews through targeted visits, separately from the normal on￾site examination of a bank. Such an approach would distinguish
  the nature of these exploratory visits from formal examinations, and encourage fact-finding and an exchange of views between banks and the BSD.
• Structuring of a review to focus on a specific area. For example,
  an IRB review might focus on (a) the rating system structure, (b) quantification methods, (c) data and informational technology, (d) controls and oversight mechanisms and (e) validation. Questions:
• Do you have any comment on our proposed approach to carry out
  bank-readiness reviews? Please substantiate your response.
• Please provide any other suggestion that may enhance the process?
  Fourth quantitative impact study
  9.14 On the one hand QIS 3 proved to be an invaluable process for the
  eight participating banks, in that it allowed them to spearhead the Basel II implementation process in their organizations. The resulting gap analysis allowed banks to plan implementation. On the other hand, the QIS 3 exercise placed great strain on the scarce resources of banks.

9.15 Banks excluded from the QIS 3 exercise may have been deprived of
the associated benefits, and may feel that they are at a competitive disadvantage.
9.16 Subsequent to the release of the final Basel II document, BSD will
undertake a fourth quantitative impact study, requiring all banks to participate. The quantitative impact study is likely to take place during the last quarter of 2004 or early 2005. This exercise will provide BSD with the opportunity to test many of the issues arising, and the results will guide its decision making progress. Owing to its resource intensive nature, and given the tight deadlines towards implementation, some of the eight participating banks may feel that they want to postpone such a study until the parallel run proposed by the Basel Committee. The BSD may consider such a proposal after it has had an opportunity to discuss the issue with the relevant bank. Further consultation and guidelines
9.17 This paper is the first of a series of consultations that the BSD will
undertake with banks during the implementation process.
9.18 As various issues are discussed and resolved, the BSD will promptly
develop guidelines, so that banks can continue with their implementation initiatives. IRB approval and validation
9.19 The BSD will continue work already commenced in the area of IRB
approval and validation. The BSD’s representation on the global IRB validation sub-group will largely inform its developmental work on approving and validating IRB ratings systems.

9.20 As already discussed, the BSD has undertaken a benchmarking study
involving a few banks. The BSD will communicate its findings with the relevant bank in details during the bank readiness assessment process. During this process the BSD may undertake various validation studies on banks’ portfolios. Examples of such validations studies include sensitivity analysis of models, back testing, replication studies, etc. Development of a regulatory framework
9.21 The BSD is considering the drafting of additional guidelines for banks
and supervisory staff. The guidelines may include the qualifying criteria for models or rating-system approval. Supervisory staff guidelines may include a detailed specification of review processes, policies and procedures.
9.22 Basel II will imply a complete overhaul of the current capital adequacy
framework. The BSD will work to develop a broad framework for a Basel II-compliant capital adequacy framework. The first meeting of the AIF Regulatory Framework Subcommittee will debate and discuss such a framework in the form of a workshop. Subsequently, the BSD will putt in place the appropriate building blocks to develop a Basel II￾compliant bank legislative framework. A good starting point is to develop legislation for the Standardised approach to credit risk and the Basic Indicator Approach to operational risk. Because the overall framework of the former approach does not differ significantly from the current framework, the BSD believes that this part of the new framework can be developed timeously and released to the industry, for comment. The BSD will also develop appropriate regulatory reports for these approaches. The BSD later work will deal with the development of a framework for the IRB approach.

9.23 The BSD will also be informed by work carried out by supervisors in
other jurisdictions.

APPENDIX A: Structure of the Accord Implementation Forum
Accord Implementation Forum
Steering Committee*
Risk
Management
Sub-Committee
Pillar I
Disclosure Sub￾Committee
Pillar III
Economic Impact
Sub-Committee
Regulatory
Framework Sub￾Committee
Credit Risk
Group
Market Risk
Group

• Steering Committee :

• Errol Kruger (Chairman)
• Jay Tikam (Vice Chairman)
• Other regulatory bodies (TBD)
• Banking Council x1
• Standard Bank x1
• Nedcor x1
• Firstrand x1
• Absa x1
• Investec x1
• Gensec x1
• Marriot x1
• Abil x1
• SAICA x1
• Foreign branches x2
• Rep offices x1
  16 + Members
  Operational
  Group
  South African Reserve Bank

APPENDIX B: Schedule of AIF meetings during 2004
Bank Supervision Department
South African Reserve Bank
ACCORD IMPLEMENTATION FORUM
MEETINGS FOR 2004
DATE TIME VENUE SUB
COMMITTEE
CONTACT
DETAILS
19 FEB 10:00-
12:00
DISCLOSURE JOAN
(011) 8065903
27 FEB 10:00-
12:00
SARB REGULATORY ELRICA
(012) 313 3733
04 MARCH 10:00-
12:00
SARB STEERING
COMMITTEE
ELRICA
(012) 313 3733
09 MARCH 14:00-
16:00
INVESTEC RISK
MANAGEMENT
PAULINE
(011) 286 7945
31 MARCH 10:00-
12:00
FIRSTRAND ECONOMIC
IMPACT
DOLORES
(011) 282 1610
05 MAY 10:00-
12:00
DISCLOSURE JOAN
(011) 8065903
14 MAY 10:00-
12:00
SARB REGULATORY ELRICA
(012) 313 3733
19 MAY 13:00-
15:00
FIRSTRAND ECONOMIC
IMPACT
DOLORES
(011) 282 1610
28 MAY 10:00-
12:00
SARB STEERING
COMMITTEE
ELRICA

DATE TIME VENUE SUB
COMMITTEE
CONTACT
DETAILS
12:00 COMMITTEE (012) 313 3733
09 JUNE 10:00-
12:00
INVESTEC RISK
MANAGEMENT
PAULINE
(011) 286 7945
04 AUG 10:00-
12:00
DISCLOSURE JOAN
(011) 806 5903
11 AUG 10:00-
12:00
SARB STEERING
COMMITTEE
ELRICA
(012) 313 3733
18 AUG 10:00-
12:00
SARB ECONOMIC
IMPACT
ELRICA
(012) 313 3733
25 AUG 10:00-
12:00
NEDCOR REGULATORY LANDIE
(011) 2941150
01 SEP 10:00-
12:00
SARB RISK
MANAGEMENT
ELRICA
(012) 313 3733
03 NOV 10:00-
12:00
DISCLOSURE JOAN
(011) 806 5903
08 NOV 10:00-
12:00
FIRSTRAND ECONOMIC
IMPACT
DOLORES
(011) 282 1610
11 NOV 10:00-
12:00
SARB STEERING
COMMITTEE
ELRICA
(012) 313 3733
15 NOV 14:00-
16:00
INVESTEC RISK
MANAGEMENT
PAULINE
(011) 286 7945
24 NOV 10:00-
12:00
NEDCOR REGULATORY LANDIE
(011) 2941150

• PLEASE TAKE NOTE:

• Confirm your attendance with the institution hosting the meeting.
• Some of the meetings might have to be changed during the course of the
  year. In such cases, a notification would be sent out in advance to all members of the forum.