The New Basel Accord and Credit Risk Management

The New Basel A ccord and Credit Risk M anagement Ram Pratap Sinha † The p ap er d iscusses in brief the imp licatio ns Basel II regard ing assessment of credit risk in the commercial banking sector under both the standardized approach and the foundation and advanced Internal Rating Based A p p ro ach. The p ap er also p ro v id es a brief rev iew o f so me o f the po pular cred it risk mo d els and d iscusses the impo rtant issues relating to the integratio n o f po rtfo lio cred it risk mo d els w ith the risk b u c ket ru le o f BC BS ( Basel C o m m ittee o n Banking Supervisio n). Finally, the paper pro vid es a brief o verview o f the RBI initiatives regard ing migratio n to Basel II in the Ind ian co ntext. Introduction : financial instruments like acceptances, inter-bank transactions, trade financing, foreign exchange transactions, financial futures, sw aps, bonds, equities, options and in guarantees and settlement o f transactio ns. The go al o f cred it risk management is to maximize a bank's risk- ad ju sted rate o f retu rn b y maintaining credit risk exposure w ithin co ntro l levels. The ad vent o f capital ad equacy no rms has ad d ed ano ther dimension to the process of credit risk management. Banks need to manage the cred it risk bo th fro m ho listic and individual perspective and they should have the ability to identify, measure, monitor and control credit risk, as w ell as, to ensure that they hold adequate capital against these risks. C red it risk c an b e d ef ined as the p o tential that a bank's bo rro w er o r c o u nterp arty m ay fail to m eet its obligations. It is the possibility of losses associated w ith diminution in the credit quality of borrow ers or counterparties. In a bank's portfolio, losses arise from o utright d efault d ue to inability o r unw illingness of counterparty to meet c o m m itm ents relativ e to lend ing , trading, settlement and other financial transactions. For the commercial banks, loans are the prime source of credit risk since the major share of interest income of a commercial bank comes from fund based activities. However, there are other sources of credit risk on account of both on and off balance sheet exposures. Commercial banks are exposed to counter party risk in various † Dr. Ram Pratap Sinha is an A sso ciate Pro fesso r o f Eco no mics, Go vernment Co llege o f Engineering A nd Leather Techno lo gy, Salt Lake, Ko lkata-700098 MANAGEMENT AND LABOUR STUDIES 337 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement ad equacy w as usually measured in terms of gearing ratios, all balance sheet assets w ere g iv en equal w eig hting regardless of their underlying risk and o f f b alanc e sheet item s w ere no t considered for the computation of capital adequacy. Scope and Organisation of the Paper : The objective of the paper is to discuss the fundamental issues in credit risk management in the context of Basel II norms for capital adequacy and the best practices o bserved in the d evelo ped country commercial banking sector. The paper pro ceeds as fo llo w s: Sectio n 1 d isc u sses the Basel C o m m ittee stipulations relating to assessment of credit risk for the purpose of determining capital adequacy. Section 2 provides an overview of credit risk and credit risk mo d eling. Sectio n 3 d iscusses so me im p o rtant issu es relating to the co mp utatio n o f reg ulato ry cap ital. Section 4 discusses the RBI policy on cred it risk m anag em ent. Sectio n 5 concludes. Base II on Credit Risk : The 1988 Basel A ccord proposed that cap ital ad equacy o f banks is to be m easu red in resp ec t o f their risk w eighted asset exposures. For this, the Committee considered credit risk only. In this c o ntext, tw o p o ints need elaboration: (i) the constituents of capital (ii) the risk w eights corresponding to various asset categories. (a) Section 1 : Basel Committee on Credit Risk A ssessment The Basel Co mmittee co ncluded that bank capital for regulatory / supervisory purposes is to include tw o tiers. The tier I c ap ital co mprises o f equity capital and published reserves from post tax retained earnings. The tier II capital includ es und isclo sed reserv es rev alu atio n reserv es, g eneral p ro v isio ns / g eneral lo an lo ss reserv es, hy b rid d eb t c ap ital instruments and subordinated term d ebt fo r co mputatio n o f capital g o o d w ill and inv estm ents in subsidiaries (engages in banking and financial activities which are not consolidated national systems) are to be deducted. The definition of capital is unchanged under Basel II. H isto rically , bank reg ulato rs hav e co nsid ered the attainment o f capital adequacy by commercial banks as one of the most important ingredients for the p reserv atio ns o f b ank saf ety and so und ness. Ev en in the nineteenth century, such safety considerations led to the imposition of various forms of regulatory standards on the commercial banking system. Inter alia these included an unlimited liability on bank ow ners in Scotland w hich continued up to 1862 and a d o uble liability system in the United States since 1863 w hich implied that bank share holders could lose twice on their bank investment. In the tw entieth century extend ed / unlimited liability systems came to be rep lac ed b y m inim u m c ap ital requirements and capital to liability / asset stand ard s. H o w ev er, c ap ital MANAGEMENT AND LABOUR STUDIES The Constituents of Capital : (b) The Risk W eights Corresponding to V arious A sset Categories : The Basel Accord (I) concluded that 338 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement the various balance sheet and off b alanc e sheet exp o su res o f c o m m erc ial b anks are to b e m u ltip lied b y risk w eig hts proposed by BCBS. The Committee used o nly five risk w eights 0%, 10%, 20%, 50% & 100%. The 1988 Accord concentrated only on credit risk and country transfer risk. In respect of off balance sheet items, the exposures are to be multiplied by sp ecified cred it co nv ersio n factors so that they can be converted into their credit equivalents and then appropriate risk w eights are to b e ap p lied . The A c c o rd stip u lated that b anks w ill b e required to maintain a minimum of 8% capital in relation to their risk w eighted assets (RW A ). Tier I capital sho uld be at least 4% o f RWA. the BIS proposed a new capital adequacy frame w ork (1999) also know n as Basel II. The characteristic feature of Basel II is that it uses a three p illar ap p ro ach consisting of: The Basel I system came to be criticized by co ncerned parties o n three majo r grounds: It gave an equal risk w eighting to all corporate credits irrespective of the differences in their underlying credit risks. ii) It f ailed to rec o g niz e that b y u nd ertaking c red it p o rtf o lio d iv ersificatio n banks can hav e potential capital savings iii) It led to extensive regulatory capital arb itrag e w hic h ad d s to the riskiness of bank asset portfolios b) a sup erv iso ry rev iew p illar to ensure that the bank's capital is aligned to its actual risk profile and c) a m arket d isc ip line p illar to enhance the role of other market p artic ip ants in ensu ring that ap p ro p riate cap ital is held by prescribing greater disclosure. (a) M inimum Capital Requirement : In so far as credit risk is concerned, commercial banks under Basel II have tw o options available: Basel II : The N ew Capital A dequacy Framework Given the experience w ith Basel (1988), MANAGEMENT AND LABOUR STUDIES a minimum capital requirement pillar Under Pillar I, the BCBS replaced the one size fits-all framew ork set o ut in the 1988 A cco rd w ith a variety of options. A s per Basel II, c o m m erc ial b anks, w ith authorization of their supervisor, can choose form among a number o f o p tio ns d ep end ing o n the c o m p lexity of their risk management. W eaknesses of Basel I : i) a) 339 i) Less complex banks can adopt a stand ard iz ed ap p ro ac h b u ild ing u p o n the 1988 A ccord and introducing the u se o f external c red it assessments. ii) Banks p o ssessing m o re advanced risk management capabilities (with the ability to meet v igo ro us sup erv iso ry stand ard s) c an f o llo w an Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement internal rating s b ased ap p ro ac h. Und er this ap p ro ach, so me o f the key elements of credit risk, such as the probability of default of the b o rro w er, w ill b e estim ated internally b y a bank. specific risk pro file and co ntro l environment. This internal process could then be subject to supervisory rev iew and interv entio n w here appropriate (c) Finally, in the New A cco rd, the Committee laid dow n disclosure req u irem ents and recommendations which will allow m arket p artic ip ants to assess critical information describing the risk profile and capital adequacy of banks. The proposals thus contain more detailed guidance on three aspects: the disclosure of capital struc ture, risk exp o sures and capital adequacy. The committee also proposed an exp lic it c ap ital c harg e f o r operational risk which refers to the risk o f d irec t / ind irec t lo ss resulting from in adequate or failed internal p ro c ess, p eo p le and systems or from external events. Three alternatives w ere suggests for the computation of operational risk i) (b) a stand ard iz ed ap p ro ac h w hich relies on industry w ise loss data ii) internal m easu rem ent approach w hich uses bank's ow n loss data multiplied by a formula for expected loss and iii) lo ss d istributio n ap p ro ach w hich allow s the bank to use its ow n probability analysis. T he S tand ard iz ed A p p roach f or M easurement of Credit Risk : A s indicated earlier, one alternative for the commercial banks under Basel I is to measure credit risk in a standardized method on the basis of external credit assessm ents. The stand ard iz ed approach, w hile having a less complex fo rmat, attempts to align regulato ry capital requirements by pro vid ing a w ider category of risk w eights and a w id er rec o g nitio n o f c red it risk mitigation techniques. Supervisory Review Process : The Basel II A ccord has affirmed the importance of the supervisory rev iew p ro c ess as a c ritic al component to the minimum capital req u irem ents. The A c c o rd , therefore, laid dow n procedures through w hich supervisors ensure that each banks has sound internal p ro cess in p lace to assess the ad equacy o f its cap ital and set targ ets f o r c ap ital that are co mmensurate w ith the bank's MANAGEMENT AND LABOUR STUDIES M arket Discipline : Risk W eights Under Basel II : The new BIS proposal envisaged a major chang e in resp ect o f risk rating o f co rp o rate lo ans. A s p er Basel I, all co rp o rate lo ans hav e the same risk weight of 100%. Under the new proposal there are three stages. i) 340 For corporate borrowers rated AAA Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement to A - by authorized rating agencies (like S&P, Mo o dy's etc) the risk w eight is 20% ii) (d) sho uld d isclo se qualitative and q u antitativ e inf o rm atio n. Q ualitativ e d isclo sures enable u sers to c o m p are assessm ent m etho d s and id entify relativ e strengths/ w eaknesses. Thus it is essential to have information such as the definition of default, the time ho riz o n, and the targ et o f the assessm ent. Q u antitativ e disclosures provide information on the actual default rates experienced in eac h assessm ent c lass and inf o rm atio n o n c red it q u ality migration. i.e. the likelihood of an A A A g rad e instru m ent b eing downgraded to AA etc over time. For the borrow ers rated below B, the risk w eight increases to 150%. External Credit A ssessment : The d eterm inatio n o f risk w eig hts corresponding to any security under the stand ard ized ap p ro ach d ep end s o n external agency provided credit rating. C o nseq u ently , the so u nd ness and reliability of the agencies performing the assessments are critical to the effective functioning of the new system. In view o f this, Basel II p laced the o nus o f rec o g niz ing external c red it rating agencies o n the resp ectiv e natio nal banking sector supervisors. In view of this, the rating agencies should fulfill the follow ing criteria: (a) (e) O bjectivity : The methodology of cred it rating must be rigo ro us, systematic, amenable to validation based on historical experience and responsive to changes in financial condition. (b) In the simplified Standardized Approach the bro ad BCBS sug g estio n fo r the alignment of risk weights corresponding to the rating grades assigned by the external credit rating agency (in respect of claims on sovereigns/ central banks and claims on banking and securities firms0are p ro v id ed in Table 1. The stand ard risk w eig ht fo r claims o n c o rp o rate ( inc lu d ing c laim s o n insurance companies) is 100%. Independence : A cred it rating Rating T ransp arency : The ind ivid ual assessments must be av ailable to bo th d o mestic and f o reig n b anks w ith o u t any discrimination. MANAGEMENT AND LABOUR STUDIES Resources and Credibility : A n external c red it rating ag enc y should have sufficient resources to carry o ut cred it rating o f high quality. Risk W eights Corresponding to the External Credit Rating Agency Provided Rating Grades : agency sho uld be ind ep end ent f ro m p o litic al o r ec o no m ic pressures that may influence the rating. (c) Disclosure : A credit rating agency 341 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement Table1 : Exposure Rating G rades and Risk W eights (the Simplified Standardized A pproach) A sset Class Risk Grades 1 2 3 4 to 6 7 Claims on sovereigns and central banks 0% 20% 50% 100% 150% Claims on banks and securities firms 20% 50% 100% 100% 150% BCBS (2003) : The N ew Basel Capital A ccord. (ii) T he Internal Approach : Computation of Risk W eights : Rating Based The computation of risk w eights under the IRB ap p ro ach d ep end s o n the exp ec ted lo ss f ro m an exp o su re. Expected loss (EL) = {Expected default frequency} x exposure at default (EA D) x loss given default (LGD). Expected default frequency (EDF) is the probability that the borrower will default. Loss given default (LGD) is the percentage of the loan at default that is expected to be lost in case of default by the borrow er.LGD is a function of the seniority of the loan and the type, quantity and quality of the collateral. The Internal Rating Based (IRB) A p p ro ac h allo w s c o m m erc ial banks to use their internal rating sy stem s su b jec t to reg u lato ry approval and periodic validation. A s per the IRB approach, banks must classify their exposures into the fo llo w ing bro ad classes o f assets: (a) corporate, (b) sovereign, (c) bank, (d) retail, and (e) equity. Each of the asset classes covered under the IRB framework has three key elements: (a) For most of the asset classes, the BCBS has m ad e av ailab le tw o b ro ad approaches under IRB: a foundation and an advanced. Under the foundation IRB ap p ro ach, banks p ro v id e their o w n estimates of EDF. How ever, other risk components are provided by the market reg u lato r. Und er the ad v anc ed approach, banks provide more of their ow n estimates of EDF, LGD and EA D, subject to meeting minimum standards. In both cases, banks must always use the risk w eight functions provided for the p u rp o se of d eriv ing c ap ital req u irem ents. The f u ll su ite o f approaches is described below : Risk com p onents : The estim ates o f risk f ac to rs p ro v id ed by banks (so m e estim ates are, ho w ev er, provided by the supervisor). (b) Risk weight functions : The functions through w hich the risk c o m p o nents are transf o rm ed into risk w eig hted assets f o r the purpose of determining capital requirements. (c) M inimum requirements : The minimum standards that a bank must fulfill for using the IRB approach for a given asset class. MANAGEMENT AND LABOUR STUDIES 342 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement (a) (b) Corporate, Sovereign, and Bank Exposures : Section 2 : A Primer on Credit Risk and Credit Risk M odeling : Under the foundation approach, banks must p ro v id e their o w n estimates of EDF associated w ith each of their borrow er grades, but must use supervisory estimates for the other relevant risk components. The o ther risk co mp o nents are LGD, EA D. Under the advanced appro aches, banks may pro vide their ow n estimates of EDF, LGD and EAD. A s w e have indicated in the beginning, from a commercial bank's point of view, credit risk is the risk o f default and change in the credit quality of issuers of sec u rities, c o u nter- p arties and intermediaries, to whom the bank has an exposure. Inter alia, the fo llo w ing are the key components of credit risk: (i) that a bank will not receive the cash flows or assets to which it is entitled because a party w ith w hich the co mmercial bank has a bilateral contract defaults on one or more obligations (e.g. principal / interest payment obligation) The Specialized Lending categories : There is an exception to this general rule for the five sub-classes of assets identified as Specialized Lending (SL).These fiv e sub-classes are Pro ject Finance, Object Finance, Co m m o d ities Finance, Inco m e Producing Real Estate and High Volatility Commercial Real Estate. Banks that d o no t m eet the requirements for the estimation of PD under the corporate foundation approach for their SL assets w ill be required to map their internal risk g rad es to f iv e su p erv iso ry c ateg o ries, eac h o f w hic h is asso ciated w ith a sp ecific risk w eight. This version is termed the su p erv iso ry slo tting c riteria ap p ro ach.. Banks that meet the requirements for the estimation of PD w ill b e ab le to u se the foundation approach to corporate exposures to derive risk weights for all classes of SL exposures except HVCRE. A t natio nal d iscretio n, banks meeting the requirements for HVCRE exposure will be able to use a foundation approach. MANAGEMENT AND LABOUR STUDIES D irect D efault Risk : It is the risk (ii) Credit Quality M igration Risk : It is the risk that chang es in the possibility of a future default by a borrow er w ill adversely affect the present value of the contract w ith the borrower today (iii) Indirect Credit or Spread Risk : risk d u e to m arket p erc ep tio n o f inc reased risk ( i.e., p erhap s because of the business cycle or p erceiv ed cred it w o rthiness in relatio n to o ther m arket participants) (iv) Settlement Risk : risk arising from the lag betw een the v alue and settlem ent d ates o f Sec urities transactions (v) Sovereign Risk : risk of exposure to lo sses d ue to the d ecreasing v alu e o f f o reig n assets o r increasing v alue o f o bligatio ns denominated in foreign currencies 343 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement (vi) C oncentration Risk : risk o f exchange rates, unemp lo yment rates etc. Thus default is closely related to market movements - often m arket p articip ants anticip ate forthcoming credit events before they actually hap p en. In fact, default is a special case of rating d o w ngrad e (by rating agencies) w here the c red it q u ality has deteriorated to the point so that the borrow er can not service any more the debt obligations. A n adequate credit risk model should be able to capture both migration risk and d efault risk in a co nsistent & integrated framework. increased exposure to losses due to concentration of investments in a g eo g rap hic al area o r o ther economic sector (vii) Counterparty Risk : risk of changes in values of contingent assets and liabilities (i.e., such as sw aps that are not otherw ise reflected in the balance sheet). Credit Risk M odeling Credit risk model seeks to determine the present value o f lo an expo sures and fixed income instruments and thereby tries to q u antif y the risk o f no nrealization of promised cash flows. Inter alia, cred it risk mo d els perfo rm the follow ing important functions: (a) Credit Risk M odeling : The Alternative A pproaches : (i) Risk A ssessment and Rating : Credit quality migration approach c o nsid ers c red it risk to b e a functio n o f the p ro bability o f transition from one credit quality to another (inclusive of default), w ithin a given time frame (usually one year). The probabilities in turn, depend on rating agency provided data i.e. they are average historical transitio n frequencies. Presently three models are described very briefly : The Asymptotic Single Risk Factor Model Credit Metrics, and Credit Risk+. C red it risk m o d els enab le assessment o f risk po sitio n o f a particular borrow er and thereby enable the rating o f bo rro w ers/ b o rro w ing p ro g ram m es. Su c h rating provides signal to the market participants and greatly facilitates the decision making such as credit approval or having exposure to a particular creditor ship security or not. It is to b e no ted here that comprehensive modeling of credit risk requires full integratio n o f market and cred it risk. This is because counterparty default by a borrow er depends on (i) borrow er specific factors and (ii) changes in market and economic conditions as reflected by changes in interest rates, the sto ck market ind ices, MANAGEMENT AND LABOUR STUDIES The Credit Q uality M igration Approach : (a) The Asymptotic Single Risk Factor M odel of Portfolio Credit Risk: The A symptotic Single Risk Factor mo d el o f p o rtfo lio cred it risk intro d uced by Vasicek (1991) postulates that a borrower defaults w hen the value of its assets falls 344 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement below some threshold. In addition, the m o d el assu m es that asset v alues are d riv en b y a sing le common factor: over a one year time horizon i.e. the maximum possible loss in the value of the portfolio during the period. The model accommodates seven credit quality states. The change in the portfolio value depends on the mo vements in the values o f the individual instruments contained in the p o rtf o lio w hile the m o v em ents d ep end o n c red it quality migration, the underlying process is that of Merton (1974). Vit =βiM t + √(1 - βi2)Z it w here: V it is the value of assets of borrow er i at time t; Mt and Z it d eno te the c o m m o n and idiosyncratic factors, respectively; and β i [–1, 1] is the bo rro w erspecific coefficient for the common f ac to r. The c o m m o n and id io sy nc ratic f ac to rs are ind ep end ent o f each o ther and scaled to random variables w ith mean 0 and variance 1.8 Thus, the asset return correlation betw een borrow ers i and j is given by. βi β j. Merton identified that a lender is effectively w riting a put option on the assets of the borrow ing firm w hereas shareholders hold the call option. If the value of the firm g below a certain threshold value, the owners will put the firm to the debtholders. Consequently, a borrower can be expected to default when the value of its assets falls below some cut o ff lev el. The cut o ff lev el d ep end s o n the v alue o f firm's liabilities. To illustrate this point, consider a firm i having asset value V it at time t, and an outstanding stock of debt, D it. Under the Merton m o d el d ef au lt o c c u rs at the maturity date of the debt, t + n, if the firm's assets, Vi, t + n, < Di, t + n (face value of the debt at that time). In the A SRF model, the probability d istributio n o f d efault lo sses is derived as follow s: Let the ind icato r K it equal 1 if borrower i is in default at time t and 0 o therw ise. Co nditio nal o n the value of the common factor, the expectation of the indicator equals E(Kit/ M t) = Prob(Vit<F-1(PD it)/ M t) = H{(F-1(PD it) - βiMt)/ √(1-βi2)} w here PD it is the unco nd itio nal probability that borrow er i is in default at time t; the cumulative distribution function (CDF) of Z it is denoted by H(·); and the CDF of Vit is F(·), implying that the default threshold equals F–1(PD it). (b) Operationalisation of the Merton framew o rk requires assumptio n about the distribution function of firm's asset value over time. The Credit Metrics framework assumes that a borrow er firm's asset value V it follow s a standard geometric Brow nian motion i.e. M etrics (1996) w as d ev elo p ed by the Risk Metrics Group. In the case of credit metrics, the objective is to determine the value at risk of the credit portfolio C red it MANAGEMENT AND LABOUR STUDIES V it = V 0 e {(µ- 62/ 2)t + 6 √tZt} W here Z t is a stand ard W iener 345 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement process: Z t~N(0,1). µ and σ2 are the m ean & v arianc e o f the instantaneous rate of return on the assets o f the firm d V t/ V t . The dynamics of Vt is described by dVt/ V t = µ d t+ σd w t w here W t is a standard Brow nian motion and √ t Z t ≡ W t – W o is no rm ally d istributed w ith z ero m ean & v arianc e eq u al to t. V t is lo g normally distributed with expected value at time t, E(V t) = Vo e µ t. factor loadings (Wi1, ... ,Wik) which measure the sensitivity of borrower i to each of the risk factors. Thus: Pi(x) = Pm(i) (x1Wi1 + x 2Wi2 + ... + xkWik) Where Pm(i) is the unconditional default probability for a grade m borrow er and the x's are positively valued w ith mean one. (ii) The c red it q u ality m ig ratio n approach depends on tw o critical but unrealistic assumptions. First, all borrow ers included in the same rating class have the same default p ro bability . Seco nd , the actual d efault rate and the histo rical average default rates are identical. The applicability of credit quality migration approaches is therefore open to questions. In the Credit Metrics framework, the risk-return relatio nship in the co ntext o f firm's assets may be described as: Ri = Ω i Xi + Yωi W here R i is a latent v ariab le asso c iated w ith b o rro w er i representing the return from the asset. Ri depends on tw o sets of factors: the borrow er specific risk factors represented by Xi and the systematic risk factors represented by Y. Ωi and ωi are the relative factor loadings. A borrow er defaults if Ri < D i (some threshold value). (b) There are tw o im p o rtant methodologies for assessing firm sp ecific d efault rate-the K M V M odel (d evelo ped by the KMV Corporation) and Credit Portfolio V iew (developed by Mckinsey). C red it Risk + : C red it Risk + ( d ev elo p ed b y C red it Su isse Financial Products)is a model of d efault risk. Each bo rro w er has only tw o possible end of period states, default and non-default. In the event o f d efault, the lend er suffers a loss of fixed size, this is the lend er' s exp o su re to the borrower. In the KM V mo d el, the actual probability of default (the Expected Default Frequency) is derived for each borrow er. The probability of d efault in the KMV mo d el is a f u nc tio n o f the f irm ' s c ap ital structure, the volatility of the asset returns and the current asset value. Expected Default Frequencies can be view ed as a cardinal ranking of borrow ers relative to default risk. In the Credit Risk+ framew ork, the conditional probability of draw ing a default for borrower i depends on: (a) the rating grade m(i) of borrower i (b) the realisation of risk factors x (= x 1, ..., x k) and (c) the vector of MANAGEMENT AND LABOUR STUDIES Structural Credit Risk M odels: Credit Portfolio View, developed by Mckinsey, is a multifactor model which attempts to simulate the joint 346 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement conditional distribution of default and migratio n p ro babilities fo r various rating groups in different ind u stries, f o r eac h c o u ntry , conditional on the value of macro ec o no m ic f ac to rs like the unemplo yment rate, the rate o f grow th in GDP, the level of long term interest rates, foreign exchange rates, government expenditures and the aggregate savings rate. The calibration of the model requires that the user haves reliable default data for each country, and possibly for each industry sector within each country. information around the rest of the portfolio. How ever, in any credit risk model, the risk contribution an individual exposure depends not only on the exposure itself but also on w hether the exposure is large relative to the rest. Go rd y (2003) sho w ed that if the f o llo w ing tw o c o nd itio ns are fulfilled , then a cred it risk mo d el can p ro d uce risk co ntributio ns that behave as risk bucket capital ru l e s : ( i) There is o nly o ne systematic risk factor that drives the performance of all obligors. (ii) no exposure in the portfolio accounts individually for a significant share of portfolio risk. Section 3 : Computation Of Regulatory Capital From Credit Risk M odels: Some Important Issues (i) A s it has been mentioned earlier, Credit risk models can be usefully applied for computation of regulatory capital. The Basel C o m m ittee o n Banking Supervision (1999) undertook a detailed study of how internal credit risk models c an b e u tiliz ed f o r d eterm ining reg ulato ry cap ital. The co m m ittee acknow ledged that a carefully specified credit risk model can provide a better measure of portfolio credit risk. However, there are several difficulties regarding the application of credit risk models for the computation of regulatory capital. (a) Finger (1999) developed a model w here the assets of each obligor depend o n o ne co mmo n market index and a term w hich is peculiar to individual obligors. A i = f (Z, µi) Where A i is the value of assets of borrower i, Z = the common market index, µi is the idiosyncratic factor corresponding to borrower i. Z and i are independent standard normal variables. In the tw o state case, where the borrower either defaults o r no t (but d o es no t experience rating migratio n), o ne has, w ith respect to each borrow er, a default threshold αi such that borrow er i d efaults if A i < α i . The d efault p ro b ab ility w ith resp ec t to borrow er i is related to the default threshold in the follow ing manner : p i = φ ( αi), so that αi = φ-1 (p i). Integration of Portfolio Credit Risk M odels W ith Basel Risk Bucket Capital Rules : The BIS Capital Accord risk bucket capital rule demands that capital assigned to an individual exposure is based only on the characteristics o f the exp o su re and no t o n MANAGEMENT AND LABOUR STUDIES Treatment Of Systematic Risk : 347 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement (ii) parameter n* is given by the inverse of the portfolio Herfindahl Index. 1 (Σxi)2 (Tex) n* = — = —— = ——2 H Σxi2 Σxi Treatment of The G ranularity Problem : In actuality bank portfolios are not fine g rained . Ev en banks and financial institutio ns w ith v ery large asset base have geographical and industrial concentrations. The present Basel Capital A ccord takes care of the problem of individual b o rro w er c o nc entratio n ( the granularity problem) through the granularity ad d o n charge. The granularity charge acco unts fo r difference betw een the industry average and the particular portfolio in question. The granularity charge can be either a charge or an offset depending on whether the portfolio in q u estio n is m o re o r less co ncentrated than the stand ard industry portfolio w hich is used as the b enc hm ark b y the Basel Committee. ( N o te: Exi is the exp o su re to borrower i.) If the exposure adjusted by GSF and n* is greater than 4% of the portfolio risk w eig hted assets, then the granularity charge will be positive, otherw ise the charge w ill actually be an o ffset to acco unt fo r the portfolio's relative diversification in size. The intro d uctio n o f granularity c harg es rep resents a m ajo r d ep arture fro m the risk bucket capital approach in the sense that it is derived from the entire portfolio and not from individual positions. In the risk bucket rule, computation of capital charge for a prospective new position is relatively simple because capital is computed on a stand -alo ne basis. In the new system, the problem is solved by assum ing that the g ranularity charge is a homogenous function of the position sizes. Consequently, it is po ssible to d eco mpo se the granularity charge as a sum over all the positions : The computation of the granularity charge in the Basel Accord is based on the Herfindahl Index. The granularity adds on charge is given by the formula : G = Tex. GSF – 0.04 RWA n* w here Tex d eno tes the to tal exposure, RWA denotes the total risk w eig hted assets, GSF(granularity scaling factor) is an emp irically d eriv ed scaling f ac to r w hic h d ep end s o n the p o rtf o lio av erag e d ef au lt probability, recovery rate and risk sensitivity. n* is the size o f the id ealized ho mo geno us po rtfo lio that represents an equivalent risk to the ac tu al p o rtf o lio . The MANAGEMENT AND LABOUR STUDIES G = Σx i ( ∂G/ ∂ Ex i) (1) w here ∂ G/ ∂ Exi = GSF {(2 Exi/ Tex)_(1/ n*)} – 0.004RW i Here RW i is the risk w eight fo r exp o sure i. To assess the true regulatory capital corresponding to position i, one needs to add the position's share of the granularity charge from (1) to the base capital 348 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement computed by using the risk bucket rule. defaults. To further illustrate this point further, let us invoke the riskreturn relationship in the standard Credit Metrics framew ork : Ri = Ω i Xi + Yωi and borrow er i defaults if Ri < Di. Now suppose there are n rating grades. Suppose the initial value of the asset is A i and after one period, the value is V ij (for a credit quality migration to grade j). If w e are interested in measuring the d o w nsid e risk o nly then the Exp ected Lo ss (after o ne tim e p erio d ) in the mark to market framew ork is: EL = ΣVij Pij – A i. In the event of a non-homogeneous portfolio, the granularity charge can be determined by using a tw o step procedure. In the first stage, it is essential to m ap the actual p o rtf o lio to a ho m o g eneo u s comparable portfolio by matching moments of the loss distribution. Gordy (2003) developed a matching procedure based on five moments: (i) expo sure-w eighted expected default rate, (ii) expected portfolio lo ss rate, (iii) co ntrib utio n o f idiosyncratic default risk to loss v arianc e, ( iv ) c o ntrib u tio n o f idiosyncratic recovery risk to loss variance and (v) contribution of systemic risk to loss variance. In the sec o nd stag e, the g ranu larity c harg e is d eterm ined f o r the constructed portfolio and the same is then ap p lied fo r the actual portfolio. (b) Section 4 : The RBI Policy on Credit Risk M anagement : According to the stand taken by the RBI, credit risk management should receive the prime importance at the bank level. A bo ut 60% o f the to tal business risk f ac ed b y c o m m erc ial b anks is contributed by credit risk. As such credit risk related problems emerge from factors like the absence of credit standards, poor p o rtfo lio risk manag ement and the inability to assess the imp act o f an economic dow nsw ing. A sound credit risk m anag em ent sy stem sho u ld therefore enable the commercial banks to identify problems on a real time basis and take ap p ro p riate c o rrec tiv e measures. Integration of Credit Risk M odels W ith M ark to M arket Accounting : The intro d uctio n o f p rud ential accounting standards necessitate that commercial banks and other financial institutions follow mark to market valuation of securities inc lu d ed in their p o rtf o lio . However, integration of the concept of credit risk w ith that of mark to market accounting is not an easy proposition. For example, in the credit quality migration framework, potential loss is identified w hen there is a rating d o w ng rad e. How ever, no loss actually arises until and unless the bo rro w er MANAGEMENT AND LABOUR STUDIES In A pril 1992, the Reserve Bank of India introduced a risk asset ratio system for both Indian and foreign banks operating in India as a capital adequacy measure in accordance with the Capital Adequacy Norms prescribed by Basel Committee. Su b seq u ently , the RBI issu ed its Guideline on Credit Risk Management in October 2002. The salient features of 349 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement RBI's Credit Risk Management policy are as under: (a) So urce: R.B.I. (2002): Guid ance Note on Credit Risk Management D evel opm ent of Sound Risk M anagement Practices : (b) The RBI stip u lated that the commercial banks should develop sound procedures to ensure that all risk associated with credit facilities are fully and promptly evaluated by the related lending and credit officers. The loan policy formulated b y a b ank sho u ld c o v er the methodologies for measurement, monitoring and control of credit risk. Banks are req u ired to ev o lv e comprehensive risk rating system that serv es as a sing le p o int indication of diverse risk factors of c o u nter p arties in relatio n to lending and investment decisions. The risk grade so assigned should be reflected in the bank's pricing of lo ans. (c) In order to facilitate this, the RBI g u id eline stip u lated that the commercial banks constitute a high lev el Cred it Risk M anag ement Committee (CRMC) w hich should be headed by the Chairman / CEO / ED, and should consist of heads of Credit Department, Treasury, C red it Risk M anag em ent Department (CRMD) and the Chief Economist. Simultaneously, each bank should also set up Credit Risk Management Department (CRMD) w hic h sho u ld f u nc tio n ind ep end ently o f the C red it A dministration Department. The RBI guid eline also d etailed the functions of CRMC and CRMD. The Interval of Risk Evaluation : The evaluation of credit risk should be based on the total exposure in respect of a counterparty including investments. The portfolio quality is to be evaluated on an on going basis. Regarding off balance sheet exp o su res, the c u rrent and potential credit exposures may be measured on a daily basis (d) C om putation of C redit Risk W eights : Computation of credit risk weights fo r the d eterminatio n o f capital adequacy is an important aspect of credit risk management. Basel II introduced a system of assignment o f risk w eights o n the basis o f external or internal ratings. The RBI favors greater reliance on internal ratings based approach (subject to stand ard izatio n). This appro ach can make use o f supplementary bo rro w er info rmatio n (w hich is usually no t av ailable to cred it rating ag enc ies) and c an b e extended to unrated borrow ers in Diagram 1 : Typical Organization Structure of C redit Risk M anagement D epartment of a Commercial Bank MANAGEMENT AND LABOUR STUDIES D evelopment of Comprehensive Risk Rating System : 350 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement the unorganized sector. The RBI feels this w ould encourage banks to refine their risk assessment and m o nito ring p ro c ess thereb y facilitating credit risk management in a better w ay. In this context, the task o f the market regulato r is twofold : (i) (ii) the revised time frame, fo reign b anks o p erating in Ind ia and Ind ian banks o p erating abro ad w ere asked to meet the Basel II norms by March 31, 2008, w hile all other scheduled commercial banks w ere asked to ad here to the guidelines by March 31, 2009. to ensure the integ rity o f different banks systems and that the p aram eters are c o nsistent ac ro ss v ario u s institutions, Capital Requirement for Credit Risk : The Basel Rev ised C ap ital A d equacy Framew o rk pro vid ed three alternative approaches for the computation of capital requirement f o r c red it risk- Stand ard iz ed A p p ro ach, Fo und atio n Internal Rating Based A p p ro ac h and A dvanced Internal Rating Based A pproach. The RBI in its guideline stip ulated that d ecid ed that all c o m m erc ial b anks in Ind ia (excluding Local A rea Banks and Regional Rural Banks) shall adopt Standardized A pproach (SA ) for credit risk. to encourage banks to use their ow n internal rating systems fo r d istinguishing betw een credit quality, Implementation of Basel II : In June 2004, the RBI had issued guideline to scheduled commercial banks regarding the maintenance of market risk capital on the lines o f 'A m end m ent to the Cap ital A cco rd to Inco rp o rate M arket Risks'. In the same month, the BIS released the rev ised c ap ital adequacy framew ork w hich w as subsequently further updated and a new v ersio n w as ultim ately released in June 2006. In response to this, the RBI issued its master circular relating to the prudential norms on capital adequacy in July 2006 w hich w as further updated and a new version w as released in July 2007. Under the Standardized Approach, the credit risk is dependent on the rating assig ned by the elig ible external credit rating agencies. The Reserve Bank has identified the external cred it rating ag encies w hich meet the eligibility criteria sp ec if ied u nd er the rev ised Framew ork. Banks may rely upon the ratings assigned by the external credit rating agencies chosen by the RBI for assigning risk w eights for capital ad equacy purpo ses. The risk w eight mapping of short term rating grades of the domestic rating agencies are provided in table 1 : Initially the R.B.I had set the d ead line fo r implementatio n o f Basel II on 31st March 2007 w hich, however, had to be extended. As per MANAGEMENT AND LABOUR STUDIES 351 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement Table 2 : Basel II risk weights corresponding to the various rating grades Basel II Risk Rating A gencies W eights CARE CRISIL Fitch ICRA 20% PR1+ P1+ F1+ A 1+ 30% PR1 P1 F1 A1 50% PR2 P2 F2 A2 100% PR3 P3 F3 A3 150% PR4 & PR5 P4 &P5 F4 & F5 A4 & A5 100% Unrated Unrated Unrated Unrated The RBI guideline on Basel II stipulated that commercial banks should have the ability to assess credit risk at the portfolio lev el as w ell as at the exp o sure o r counterparty level. Commercial banks sho uld be particularly careful in the matter of identification of credit risk concentrations and ensuring that their effects are adequately assessed. This should include, inter alia, consideration of various types of dependence among exposures, incorporating the credit risk effects of extreme outcomes, stress events, and shocks to the assumptions made ab o u t the p o rtf o lio and exp o su re behavior. Banks should also carefully assess concentrations in counterparty credit exposures, including counterparty credit risk exposures originating from trading in relatively illiquid markets, and determine the relative impact on bank's capital adequacy. Capital A dequacy Position of Indian C om m ercial Bank s : T he Recent Evidence Tables 3-4 present the capital adequacy ratios (for 2007-08 and 2008-09) of the State Bank and its associate banks under Basel I and Basel II .Tables 5-6,7-8 and 9-10 present the capital adequacy ratios for the nationalized banks, private sector banks and foreign banks respectively. Table 3 : CRA R (Basel I) of the SBI and Its A ssociate Banks Descriptive 2007-08 2008-09 Statistics Tier I Tier II Tier I Tier II Max 9.14 6.01 8.53 5.58 Min 6.74 4.30 6.30 4.03 Average 7.33 5.25 7.18 4.89 Standard Deviation 0.902 0.678 0.762 0.560 Source: Statistical Tables Relating to Banks in India, 2008-09. MANAGEMENT AND LABOUR STUDIES 352 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement Table 4 : CRA R (Basel II) of the SBI and Its A ssociate Banks Descriptive Statistics Tier I 2007-08 Tier II Max 7.41 6.25 9.38 6.06 Min 6.54 4.28 6.94 4.39 Average 7.04 5.48 7.97 5.43 Standard Deviation 0.342 0.796 0.888 0.605 Tier I 2008-09 Tier II Source : Statistical Tables Relating to Banks in India, 2008-09. Table 5 : CRA R (Basel I) of the Nationalized Banks Descriptive Statistics Tier I 2007-08 Tier II Max 11.29 5.49 11.28 5.51 Min 5.05 1.45 5.30 1.99 Average 7.58 4.20 7.51 4.57 Standard Deviation 1.66 1.16 1.4498 0.8129 Tier I 2008-09 Tier II Source : Statistical Tables Relating to Banks in India, 2008-09. Table 6: CRA R (Basel II) of the Nationalized Banks Descriptive Statistics Tier I 2007-08 Tier II Max 8.97 6.24 11.88 6.15 Min 4.88 4.28 6.11 2.10 Average 6.97 4.93 8.09 5.05 Standard Deviation 1.2445 0.6461 1.2529 0.9301 Tier I 2008-09 Tier II Source: Statistical Tables Relating to Banks in India, 2008-09. Table 7 : CRA R (Basel I) of the Private Sector Banks Descriptive Statistics Tier I 2007-08 Tier II Max 48.29 5.77 44.22 6.20 Min 6.10 0.47 6.45 0.00 Average 13.26 2.55 13.62 2.53 Standard Deviation 9.0628 1.6355 8.2236 1.9208 Tier I 2008-09 Tier II Source: Statistical Tables Relating to Banks in India, 2008-09. MANAGEMENT AND LABOUR STUDIES 353 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement Table 8 : CRA R (Basel II) of the Private Sector Banks Descriptive Statistics Tier I 2007-08 Tier II Max N.A. N.A. 41.69 7.10 Min N.A. N.A. 6.19 -0.25 Average N.A. N.A. 13.33 2.66 Standard Deviation N.A. N.A. 7.3465 1.9854 Tier I 2008-09 Tier II Source : Statistical Tables Relating to Banks in India, 2008-09. Table 9 : CRA R (Basel I) of the Foreign Banks Descriptive Statistics Tier I 2007-08 Tier II Max 235.82 5.43 501.34 6.37 Min 8.07 0.00 8.54 0.29 Average 51.57 1.40 75.19 1.93 Standard Deviation 57.1088 1.49049 116.6195 1.689176 Tier I 2008-09 Tier II Source : Statistical Tables Relating to Banks in India, 2008-09. Table 10 : CRA R (Basel II) of the Foreign Banks Descriptive Statistics Tier I 2007-08 Tier II Max 107.95 5.68 317.51 5.43 Min 7.24 0.11 7.43 0.00 Average 29.23 1.46 44.16 1.53 Standard Deviation 23.1355 1.377476 67.29328 1.4944 Tier I 2008-09 Tier II Source : Statistical Tables Relating to Banks in India, 2008-09. S ectio n 5 : Observations : T he relating to risk management. In this c o ntext, o ne id entif ies tw o m ajo r problem areas in the migration of Indian banks to the new regime: C o ncl ud ing The above discussion show s that the process of credit risk management has become increasingly complex with rising complexities in the fund and non-fund based activities of commercial banks. C o m m erc ial b anks need to inv est sufficient resources for meeting the in ho use and regulato ry requirements MANAGEMENT AND LABOUR STUDIES (i) Pro-cyclicality of the new capital adequacy norm : The new capital adequacy norm links capital adequacy requirement w ith the quality of borrow er rating 354 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement thereb y p ro m p ting b anks to provide lending to borrow ers w ith good rating grade. However, during eco no mic d o w nturn the rating quality is likely to decline. Thus the new reg im e w ill reinf o rc e cyclicality. (ii) credit rating o r credit histo ry is available. Inclusive grow th of the Indian economy requires that the b anking sec to r allo c ate m o re reso u rc es to the u no rg aniz ed secto r. Thus financial inclusio n poses a major challenge before the commercial banking sector as to how there can be an appropriate trade off betw een risk mitigation and universal service obligation. Presence of a Large Unorganised Sector : In Ind ia w e hav e a larg e unorganized sector for w hich no MANAGEMENT AND LABOUR STUDIES 355 Vol. 35 No. 3, August 2010 The New Basel A ccord and Credit Risk M anagement Reserve Bank of India (2007) : Master Circular on Prudential Norms on Capital A d eq u ac y , D BO D N o . BP.BC . 4/ 21.01.002 / 2007-08 Bibliography: BCBS (2003) : The New Basel Capital Accord, www.bis.org. Credit Suisse Financial Products (1997) : Credit Risk+, CSFP Website. Reserve Bank of India(2002): Guidance N o te o n Cred it Risk M anag em ent, O c to b er, D ep artm ent o f Banking Operations and Development, Central Office, Mumbai Finger C.C. (2001) : The One - Factor Credit Metrics Model In The New Basel Capital A ccord, Risk Metrics Journal, Volume 2(1), www.riskmetrics.com. Risk Metrics (1996) : Cred it Metrics Th e Te c h n i c al D o c u m e n t, w w w .riskmetrics.com. Gordy, Michael (2003), "A risk factor model foundation for ratings-based bank cap ital rules," Jo urnal o f Financial Intermediation, vol. 12, 199-232. Vasicek, Oldrich (1991), "Limiting loan lo ss p ro bability d istributio n," KMV Working Paper . Merton, R C (1974) : "On the pricing of co rpo rate d ebt: the risk structure o f interest rates", Journal of Finance, vol 29, pp 449-70. MANAGEMENT AND LABOUR STUDIES 356 Vol. 35 No. 3, August 2010