Chapter 4 Risk Management in the Credit Portfolio

Transcription

1 Chapter 4 Risk Management in the Credit Portfolio This chapter begins with important issues in credit risk management. It discusses some of the important aspects of credit risk assessment focusing on the internal credit rating system. Finally, the chapter presents methodology and important features of computer based rating model that was developed during the study. 4.1 Introduction One of the major risks that banks must measure, monitor and manage is credit or default risk. It is mainly associated with the loan repayments and it is the possibility of losses associated with changes in the credit profile of borrowers or counterparties. Most simply put, credit risk is 'the potential a bank borrower or counterparty will fail to meet its obligations in accordance with agreed terms'. (BCBS, 2000b). Such losses may result into outright default or losses from changes in portfolio value arising from actual or perceived deterioration in the credit quality short of default. Credit risk is also the risk of a decline in the credit standing of an obligor of the issuer of a bond or stock. Such decline need not get converted into default, but it implies increased possibility of default. 4.2 View of the Credit Risk As described by Bessis (2002) view of Credit risk differs from banking portfolio and trading portfolio. Some of the financial products where credit defaults can take place are loans given to the retail customers, small or medium (SME) businesses or large borrowers. In the banking portfolio risk of default of a small number of important customers can generate higher losses which may potentially lead to insolvency of a financial institution. Such defaults can be of different types such as, a) default in payment obligation, b) restructuring of loan due to major deterioration of the credit standing of the borrower, or c) bankruptcies. A limited inability of borrowers that could be resolved in a short period such as three months need not turn out to be a default. However, restructuring is very close to default as borrower may not be able to fulfill his obligations unless his funding structure changes. 73

2 For the majority of banks loans given to individuals or corporate firms are the most obvious source of credit risk; however banks may face credit risk in trading portfolio i.e. in security market, foreign exchange transactions, financial futures, swaps, etc. Unlike the loans in banking book, the credit risk of traded debts is indicated by the agencies' ratings, assessing the quality of public debt issues or through changes in the value of stock market. One of the important characteristics of trading assets is that there is no need to hold these assets until the deterioration in them turn up into effective losses. It is possible to sell these assets in the market at a lower value in which case loss may be minimized to the extent of difference between pre and post default prices of those assets. However, there may be number of practical difficulties in this case such as whether it will be easy to sell these instruments when there are signs of deterioration, if at all they are sold at what discount rate, etc. 4.3 The Drivers of the Credit Risk For the credit risk both in the banking book and the trading book, it is essential for a bank to identify, measure, monitor and control the same and hold adequate capital against these risks. Moreover, for credit risk management to be effective, bank must monitor the major drivers of credit risk namely, exposure, likelihood of default and the possible recoveries in case of default. Exposure: The first important driver of credit risk is the 'exposure' which is the amount at risk with the counterparty. It is the 'quantity' of risk. A sound credit risk management imposes limits on exposures by firm, industry or region. Default and Migration: This is the second important dimension of the credit risk and it is necessary for a bank to find out the 'probability of default' as well as 'migration'. The probability that a firm migrates from one risk class to any another is migration and default is the final state of migration. When default happens the entire exposure (minus recoveries) is at risk. Migration state does not necessarily trigger any loss in the book value, though due to this default probability changes. A sound credit risk management practice is expected to find out these probabilities which requires maintaining historical data of a bori'ower. 74

3 Recovery Risk: Though lending is first and foremost dependent on the credit standing of borrowers, one cannot ignore the value of guarantees as they provide some protection in case of default. Such guarantees include collaterals or assets that are seized by the lender in case of default or the third party protections such as credit insurance or credit derivatives or other contractual agreements (covenants) between borrower and lender. However, one major difficulty in recoveries is that it involves legal procedures resulting into significant expenses and major lapse of time. History suggests that "...collateral values and recovery rates on corporate defaults can be volatile and they (recovery rates) tend to go down just when number of defaults goes up..." (Altman, Resti, Sironi, 2004). Hence a prudent credit risk management practice must administer the relation between default rates and recovery rates. 4.4 Credit Risk Assessment As losses arising out of the credit portfolio continue to be the leading source of problems in banks world-wide, it is necessary for banks and regulators to draw useful lessons from the past incidences and adopt effective credit risk assessment framework. The banking supervisors world-wide have promoted sound credit risk assessment policies because failure to identify deterioration in credit quality in timely manner may cause financial losses to banking system, finally leading to unstable financial markets and economy. Supervisors expect a bank's credit risk assessment and policies to be consistent with prudential guidelines and accounting frameworks. Based on past experience, it has been observed that significant cause of bank failures is the poor credit quality and the credit risk assessment. Hence it is important to have adequate credit risk assessment policies and procedures for timely recognition and measurement of loan losses and capital requirements. Some of the important functions that banks must carry out in credit risk assessment are as follows: 1. In the assessment stage bank should scrutinize - possible repayment transactions that the borrower can do. Also it should check for collateral and guarantees to help minimize the risk. However, the collateral cannot be a substitute for comprehensive 75

4 assessment of the borrower or counterparty nor can it compensate for sufficient information. Primarily, bank must examine borrower's repayment capacity. 2. While assessing the potential credits it is important for banks to make necessary provisions for expected losses and to hold adequate capital to absorb the unexpected losses. The assessment of credit should not be based only on present conditions, but "bank should take into consideration potential future changes in economic conditions... and should assess their credit risk exposure under stressful condition." (BCBS, 2000b). While deriving the capital adequacy and loss provisions, it is important for banks to take into consideration what could possibly go wrong with individual credits as well as within various credit portfolios. Scenario analysis, stress testing and analysis of correlation of various risks could be some useful mechanisms to assess the potential future problems. 3. At this stage, it is necessary for a bank to establish a system that reliably classifies loans on the basis of credit risk. Such systems will help the banks in valuating loans and determining appropriate loan provisions. Banks should use loan grading systems for larger loans whereas smaller loans may be graded on the basis of payment delinquency status. "A well-structured loan grading system is an important tool in differentiating the degree of credit risk in the various credit exposures of a bank. This allows a more accurate determination of the overall characteristics of the loan portfolio." (BCBS, 2005). World-wide, supervisors and other financial organisations including Basel Committee have published guidelines on this subject which spell out in detail the factors that should be taken into account for loan grading. Using this information, banks must devise their internal grading framework that will help them to determine groups of loans that would be collectively assessed for the loan loss measurement. Such grading systems should also be updated after taking a periodic review to maintain accurate risk grades. All credits should be at least reviewed annually to reasonably assure that the risk grades are precise and up-to-date. Large, complex loans or loans with higher risk category should be reviewed more frequently. Credit risk assessment models should be used by banks in various aspects of credit risk measurement process including credit scoring, measurement of risk at both the individual transaction level and at the 76

5 portfolio level. Such models should also consider the impact of changes and other loan-related variables such as default probability, loss in case of default, exposure at default, collateral value and migration probabilities. Banks should also apply back-testing on such models to establish internal tolerance limits and differences between estimated and actual outcomes. 4. A sound credit risk assessment process also includes developing a comprehensive information system to monitor the quality of its loan portfolio. The information generated from such a system helps the top management in determining adequate level of capital and the amount of loan provisions that the bank should be holding. Such systems should be capable of bringing out exposures that are approaching to their risk limits. They should also help the management in identifying concentrations* of risks within the credit portfolio. Such a system can provide routine reports such as the details of irregular accounts, overdue interest or installments, etc. In addition to such routine reports, banks can also use such systems to generate information which is analytical in nature, e.g. the report showing historical information of default classification for various industries. This would facilitate the management to have strong indication of industry default rates. The main objective of such credit information system is, to enhance the value of the data by providing accurate and timely information thereby contributing to better planning of the credit portfolio. 5. It is important for sound risk assessment practice to be based on published guidelines or regulative rules. However, the same should be supplemented with experienced credit judgement of the senior management of the bank. Loss experience in the past and corresponding data may or may not be available or relevant to current circumstances; in such a case, senior management's experienced judgement should be used to estimate the amount of impairment loss. The management should also consider any current factors such as changes in economic and business conditions while doing the risk assessment exercise. However, the *Concentrations can occur when bank offers higher level of credits to single counterparty, group of connected counterparties, a particular industry or a geographic region 77

6 experienced judgements should be exercised subject to established loan policy of the bank. Also, all the reasonable assumptions should be supported by adequate documentation. 6. When the bank determines probability of default, and loan losses bank may arrive at a single amount or a range of possible amounts. If more than one amount is determined a bank should recognise an impairment loss equal to best estimates within the range after considering all relevant information at the measurement date. 7. Finally, it is for the bank's board of directors and senior management to ensure that the bank has appropriate credit risk assessment process and effective internal control to determine adequate capital and loss provisions in accordance with the stated policy of the bank. Though primarily it is bank's responsibility to maintain an effective credit risk assessment system, supervisors should periodically evaluate the effectiveness of a bank's credit risk policies and practices for assessing loan quality. Supervisors should be satisfied with "the quality of a bank's loan review system for identifying, classifying, monitoring and addressing loans with credit quality problem in a timely manner" (BCBS, 2005). It is necessary for them to obtain necessary information through periodic regulatory reporting or doing on-site examinations. Supervisors must ensure that while establishing the loan loss provisions and the capital adequacy, banks have taken into consideration factors such as updated valuation of collaterals or cash flow predictions in present economic conditions. Proper estimates of loan losses in an individual as well as group of loans improve over time when substantial information is accumulated regarding the factors affecting the repayment prospects of the borrower. For this purpose, the supervisors should ensure that banks are maintaining information systems and controls for identifying, monitoring and controlling problems in credit quality in timely manner. Finally, if supervisor finds significant deficiencies in credit assessment are not addressed by the bank in timely manner it should be reflected in higher capital requirements for the bank. 78

7 4.5 Measuring Credit Risk: A Rating Based Approach Lending being one of the oldest functions of banics they have been dealing with procedures of measuring risks in the same. Loosely speaking the credit risk measurement attempts to obtain some measure of the dispersion of possible future outcomes (losses). Its main aim is not only to find out possibility of losses but also to quantify the losses as well. "... In practice the focus of risk measurement is on downside outcomes, rather than upside outcomes, so that measures of risk tend to focus likelihood of losses, rather than characterizing the entire distribution of possible future outcomes. (BCBS, 2002) Thus, finding out the possible loss in case of default is one of the major tasks of credit risk management. One of the important approaches to credit risk measurement is the credit rating system. Rating systems measure the extent of credit risk and differentiate individual credits and groups of credits by the risk they pose. The main focus of a rating process is to test the fundamental credit strength of a borrower. Rating categories, typically either labeled as 'AAA' or numbered (e.g. I to 8) represent the underlying measure of credit risk. It allows measurement and management of credit risk consistently throughout the life cycle of credit so that a bank can alter its exposure with respect to risk. Rating is based on criteria which include both qualitative assessment of the borrower's credit standing and quantitative variables mostly, financial variables. Rating can be of two types: External Rating and Internal Rating. External Rating:. External ratings are assessments of credit standing of a debt issue that serves the needs of the investors to have a third party view on the credit risk of the debt. External rating may not represent rating of the issuer, though the rating assigned to senior unsecured debt is closely associated to issuer's rating. Smaller debts may default without a default of issuer. External rating is done by the rating agencies. Some of the rating agencies at international level are, Moody's, Standard and Poor's(S & P), Flitch, etc. Similarly, CRISIL, ICRA, CARE, Flitch are the rating agencies in India. Such agencies carry out rating for debt issues of large listed companies such as corporate firms, banks and sovereign borrowers. 79

8 Internal Rating: Internal rating systems are developed to meet the individual bank's own requirement. Internal ratings are the ranks assigned to the borrowers or facilities for classifying their risks. Unlike external ratings, the firms that are internally rated are the counterparties of bank such as small and medium enterprises. Generally, banks do not make rating information public, even the firm being rated is typically not informed about its current rating. As described by Deventer, Outram (2003) a bank should use internal rating for following purposes: 1. Acceptance or Rejection of new transactions: Rating can be used as an index of whether a new transaction should be accepted or rejected. 2. Monitoring of Credit Quality: The migration of the credit ratings can be used to judge the overall quality of the credit portfolio. 3. Allocation of Resources: If a particular credit is found to deteriorate in its rating, bank can deploy more resources for monitoring the credit quality of that loan. 4. Loan Pricing: Rating system helps in loan pricing. The price decided should be sufficient to compensate the risk of earning and capital. 5. Adequacy of Loan Loss Reserves and Capital: Internal Rating can give the distribution of portfolio quality which can be used to judge the adequacy of reserves for loan losses and provision for loss in respective loan period. Also, it can be used in measuring the capital adequacy of the bank. Rating serves as the foundation for new supervisory regulations such as Basel II. Particularly, Basel II emphasizes the use of internal rating systems for evaluating capital requirements and provisions for loan losses. In fact, one of the limitations of earlier Basel I accord was that, it did not take into consideration bank's own internal risk management techniques while calculating capital requirements. All loans were implicitly required 8% capital regardless of the borrower's external credit ratings or collateral offered. This resulted into mispricing of credit risk and retaining more risky loans in credit portfolio. This problem was overcome in Basel II by introducing concepts of internal ratmg systems that would help in doing risk-based pricing of a loan. For measuring credit risk, Basel II 80

9 Accord has suggested three approaches viz. the Standardized Approach, the Foundation IRB Approach, the Advanced Approach Internal rating system should be based on qualitative and quantitative information regarding the borrower's credit-worthiness. Also, it should be able to calculate both borrower's rating and the individual facility's rating that he has availed from the bank. Borrower's rating means a category of creditworthiness to which borrowers are assigned on the basis of a specified and distinct set of rating criteria, from which estimates of default probability are derived. Facility rating means a category of loss severity in the event of default to which transactions are assigned on the basis of a specified and distinct set of rating criteria. While assigning ratings to a borrower, banks should consider important factors such as, borrower's historical and projected capacity to generate cash to repay a debt and support other cash requirements, the capital structure and the likelihood that unforeseen circumstances could exhaust the borrower's capital cushion, the quality of earnings, the quality and timeliness of information about the borrower, the degree of operating leverage, the borrower's ability to gain additional funding through access to debt and equity markets, management's skill, the borrower's position within the industry and its future prospects, etc. Similarly, for facility rating, banks should check the presence of third-party support, the maturity of the transaction, the structure and lending purposes of the transaction which influence positively or negatively the strength and quality of the credit. These may refer to the status of borrower, priority of security, any covenants attached to a facility etc. Thus, the internal rating system allows the bank to assign lower risk weights to higher rated loans and therefore lower capital requirement. On the other hand, loans with low rating would carry higher risk weights and higher capital requirement Techniques for Modeling Credit Risk Rating A number of techniques have been in use in modeling of ratings and default probabilities. Some of them are discussed below,

10 1. Altman's Z-score model: This is a predicting model created by Edward Altman in his pioneering work in The model is mainly used for consumer loans and applies to both rating and default probability. It finds out a score based on five different weighted financial ratios and the scoring technique is based on discriminant analysis. It suggests two different Z-scores for public limited companies and private companies. Depending on the value of the score, the firm is more or less likely to default. 2. Logit - Probit Models: These models can predict binary events such as default or non-default, also they can scale the probability that such events occur. Simple regression models are also used to find default frequencies. However, when the relation between rating and financial variables such as operating profit, leverage, etc. is non-linear logit - probit models are applied. Moody's RiscCalc model is based on this technique. 3. Artificial Neural Network Technique: In 1980s computerized expert systems were commercially applied in various disciplines where a set of rules (suggested by an expert) were stored as a computer program to analyze information. Developing computerized expert system is largely dependent on acquisition of human expert's knowledge. Since this a time-consuming and error-prone task, alternative systems are used which use induction techniques to infer the human experts' decision process by studying their decision process. Such systems are Artificial Neural Network based systems which can simulate human learning process for making predictions. They can learn the nature of relationship between input and output by sampling information sets again and again. ANN can accommodate interdependencies and non-linear relationship between default probabilities and other key attributes of the firm. Also, they can incorporate subjective, nonquantifiable information into credit approval decisions. However, a major disadvantage of neural networks is they lack transparency. Since the internal structure of network is hidden from the user, it is difficult to duplicate the system using the same input. Also, since ANN does not reveal intermediate steps, it is not possible to know the relative importance of the key financial variables. ANN are generally used to forecast the corporate bond rating. 82

11 4. Merton Model of Default: This model is built on option theoretic approach to default. It is based on a consideration that equity holders have the option to sell the firm's assets rather than repay the debt if the asset value deteriorates. The holder can exercise the 'put' option under these circumstances Advanced Credit Rating Models in Practice 1. KMV Model: The original Merton model has been extended by KMV Corporation (Credit Monitor) to predict expected default frequency of a firm. For this purpose, it uses the information on stock prices and the capital structure of a firm. It is based on an assumption that a firm will default if its asset value falls below a certain value (default point). Using the standard deviations of asset's value from a default point and historical default rates it finds out expected default frequency. 2. Credit Metrics: In 1997, J P Morgan published the first document which set a benchmark in the area of portfolio risk management. It gives a method for estimating the distribution of value of assets in a portfolio subject to changes in credit quality of individual borrower where a portfolio consists of different standalone assets represented by a series of future cash flows each of them would have. This model uses simulation technique to estimate the value of assets. 3. Credit Risk+ : This was introduced by Credit Suisse Financial Products. This is based on a portfolio approach to modeling risk of default. It is a statistical model that takes into account information relating to size and maturity of an exposure and the credit quality and risk of an obligor. This model can also take into account background factors such as state of the economy. 4.6 Credit Risk Management Processes in Indian Banks Most of the banks in India have structured their governance framework for ensuring effective credit risk management. The overall responsibility for credit risk management remains with the Board of Directors of banks. It is important that the board sets a right credit culture and risk appetite of the bank for lending activities. "Senior management must establish a comfort zone for risk taking and ensure that people within the organization understand it and remain with it." (Caouette, Altman, Narayanan, 1998:23) The board is 83

12 also responsible for approving policy for credit pricing, credit exposures /concentration limits for industry, country, etc. conforming to regulatory guidelines. The credit policy is prepared by the Risk Management Committee (RMC) comprising of the Head of the bank, other directors, general managers, etc. This committee also sets the important guidelines for a risk rating system and setting the link between the rating scale and broad pricing bands. As per RBI guidelines, the bank's board approves bank's benchmark Prime Lending Rate (PLR). After this, the risk management department is responsible for deciding the 'Credit Spread' based on credit rating model (if any) adopted by the bank. The borrower segmentation and loan product offerings are governed by the business strategy decided by the top management of the bank. In India generally, the borrowers are divided into following segments, Commercial and Institutional Banking (C&IB) segment (corporate customers) Small and Medium Enterprises (SME) Enterprises Retail Segment (such as housing loan, auto loan, education loan) The banks generally offer a wide range of loan products which includes, 1) Credit Products such as working capital (fund-based, non-fund based), mediumterm loans for 3 to 5 years for business expansion, R&D activities, etc, long term loans or project finance for new industry, off-balance sheet items such as LC, guarantee, bills, etc. 2) Transactional products such as cash management, documentary credit 3) Asset Products such as home loans, consumer finance, personal loans, etc. Any credit proposal originating from the Branch generally involves a rigorous appraisal process before it is recommended for sanction by designated authorities. The first step in this is to ensure that there are no regulatory restrictions on the borrower for availing the loans, the purpose of loan and the security offered. The manager is also expected to conduct in-depth analysis as regards to the technical/commercial/financial/managerial scrutiny of the proposal (due diligence). In case of new business it is important to check financial projections given by the customer and his managerial and entrepreneurial capabilities. It is at this stage, the bank undertakes a credit risk rating exercise for 84

13 evaluating credit risk of borrowal accounts at the facility level and at the borrower level both. As a general practice, the rating exercise is carried out at loan-originating branch, Regional/Zonal Office or the Head Office depending on under whose purview (delegated power) the loan size falls. (Some Indian banks, however, follow a practice of doing the rating exercise at centralized level i.e. at the Head Office and then communicate the results to the loan-originating branch.) The activity diagram given in Figure 4.1 explains the activities involved during sanctioning and disbursement phase of a loan cycle. As shown in the diagram if the loan size is within the sanctioning power of the head of that unit of a bank (such as branch) the rating is carried out at that unit. Otherwise the loan proposal is sent to the higher level in the hierarchy where rating will be done. If the score (or grade) is found to be above the cut-off level set by the bank, the proposal is short-listed for further processing before the final disbursement is done. 85

14 : Customer : Branch :RO/ZO :HO 1 Submit Application Receive Application Initial Processing Size<Sancl.im Check Loansizc It Loansize> SancPowei > Receive Application Process with Rating Model Check lx)ansize II' Loansize>Sanc Power Si/e<SancPower Receive Application Yes No No ^ Yes Process Application Score > Cutoff Any Additional Requirement No Process with Rating Model Yes o Any Yes Score > Cutoff Additional Requirement Process with Rating Model Score > No A CutotT Yes y \ "\-^ Yes Any Additional Requirement Receive Application l^ Disburse ^. Send Back 1^ Receive ADPlication No Process Application No Process Application Figure 4.1: An Activity Diagram Showing the Processing of Loan Application 86

15 The post-disbursement phase involves supervision and follow-up procedures. In this stage the bank ensures timely receipt of stock / asset statements at stipulated frequency. It also carries out thorough scrutiny of the financial data received from the borrower. Conduct of the borrower's account with the bank is checked and if any irregularities are found the corrective measures are taken through reporting to proper authorities. Few more aspects of monitoring the conduct of account are conducting stock audits and credit audits. In stock audit bank carries out comparison of two consecutive stock statements, verification of sales and purchases done by the borrower, etc. Credit audit examines compliance with exact sanction and post-sanction procedures laid down by the bank. It also conducts reviews of the quality of entire portfolio, regulatory compliance, adequacy of documents, etc Issues in Credit Risk Management Processes faced by Indian Banks One of the important indicators of the health of the banking industry is its level of Non- Performing Assets (NPAs). In the recent past Indian banks are relatively better placed as compared to the past. A few banks have even managed to reduce their NPA levels below 1 %. (e.g. Oriental Bank of Commerce was a zero NPA bank before GTB merger) However, banks cannot remain complacent with this situation. With the rising interest rate scenario since early 2004, most Indian banks have not only suffered from a sharp drop in the 'treasury income', but also had to make massive provision due to massive fall in bond prices. Income from investment book is no longer the driver of profit position of banks. Moreover, with the growth in economic activities in the country, there is a perceptible mix in the lending activities of Indian banks. Given the perspective, it is imperative that banks take every precaution to maximize their interest income in the loan portfolio. This calls for strong credit granting and monitoring process. One of the major issues faced by Indian banks in their credit process is to establish appropriate framework for credit monitoring. More than 80% (in terms of size) of loans are being sanctioned at the Head office and other controlling offices. However, each of the account needs to be dealt with and monitored at the branch level. Though RO/ZOs are responsible for monitoring, the primary responsibility lies with the branch manager, "...the pressure of day-to-day operations load at the branches, branch manager could do little justice in monitoring the accounts, and today at best his/her extent of "monitoring' is 87

16 restricted only to 'passing cheques' in such accounts." (Saha, Subramanian, 1999). This brings out a need that monitoring of high-value accounts should be viewed as a responsibility of the bank and not only the branch. In Basel II era, where rating migration will a play a key role, monitoring migration would hold key to ensure effective risk management in credit portfolio. One more issue that is faced by Indian banks is that of the management of concentration risk. To overcome this problem bank should be in a position to find out industry wise exposures and default ratings. This will allow the bank to manage the concentration risk in the credit portfolio, by diversifying across the geographical area, industry and type of loan. In addition to above-mentioned points, many banks are not using risk-sensitive pricing. This requires appropriate risk classification of the borrower done using a rating model which is based on comprehensive information of a borrower. Banks also should note that rating exercise does not stop at the disbursement stage. As shown in the following Figure 4.2, throughout the life cycle of credit based on the conduct of loan account rating needs to be updated. Loan Origination Initial Rating Rating Update Loan Completion Borrower data Rating data Transaction data of repayments ^ 7 Building a Database Figure 4.2: Credit Life Cycle Source : Bank of Japan CRM Report,

17 Given a background and issues related to credit portfolio of Indian banks a computer based rating model was developed which is discussed in the following write-up. 4.7 Computer Based Credit Rating Model for Corporate Borrowers Step 1: Understanding the Requirements of Model As a first step of the exercise, it was necessary to understand from the experts various components /sub-components of the rating process, relative importance of those components in doing rating, formulation (such as formulation of various financial ratios). This was done by conducting discussions with concerned subject experts in NIBM. The model is based on two main categories of input. The first category comprises of information of each of the borrower such as his financial or business details. The second category includes the information required such as scoring method, weights, cut-off scores, etc. that will be used in finding out the final score of the borrower account. The Figure 4.3 shows the type of broad categories of information and flow of information that is taking place in the rating exercise. As it can be seen, the user feeds the model the necessary details of each risk category. Then the model uses the master database to assign scores. It also uses the weights to find out the final score and then the rating of the borrower. The database for scores and weights is created with the consensus of the Credit Risk Management department of a bank. Finally, the model can also generate some useful reports needed by the top management. 89

18 Borrower's Profile Financial Risk Details Set Weights User Industry Risk Details ^ ' 1 r,r,r.' Credit Risk Rating Model Set Industry Scores Credit Dept I \. Management Risk Details Set Scores for Fin. Ratios Business Risk Details Set Cut-Offs Project Details Reports Figure 4.3: Information Flow in the Rating Exercise The overall hierarchy of risks maintained in the model is shown in the following diagram Overall Credit Risk Risk Scoring: A Bottom up Approach < Risk Categories Risk Parameters V Risk Factors Figure 4.4: Risk Hierarchy a. Risk Category: These are the categories that can affect company's performance. Such categories can be internal as Financial, Business, Management, Facility or external such as the Industry to which the company belongs to. b. Risk Parameters: The detailed parameters within each risk category, e.g. Under Financial risk parameters such as quality of financial statements, past financial or cash flow adequacy are important c. Risk Factors: They are the constituents of each risk parameter. E.g. In this step it was decided that in 'Past Financials' which financial ratios are other financial 90

19 indicators should be considered. For each risk factor values are either computed or entered by the user. d. Scores: Depending on the value in risk factor, a score will be calculated on the scale of 1 to 6, 1 being lowest and 6 being highest. e. Weights: Depending on the relative importance weight is assigned to each risk factor/parameter/category that will be used to compute the final rating/score. The broad risk categories can be described as follows, Financial Risk : This is the most important aspect of the model. The main input to this risk category is the CMA data that is collected by the bank from their potential borrowers. Using CMA data bank can review the past financial position (if company is old) of the borrower and the cash flow adequacy for the future based on the projections made by the borrower. Under this category quality of borrower's financial statement is also tested. Business Risk: In this category parameters such as latest turnover of the borrower, his market position, growth in his business in last financial year, etc. are tested. Management Evaluation: For rating this category factors used are track record of the borrower, number of years of experience, corporate governance, etc. Security/Conduct of Account: In this option, mainly, security/collateral and borrower's past payment records are tested for rating. Project Risk Evaluation (in case of Term Loans) : In this, the project under consideration is evaluated based on certain financial ratios and repayment period for the project. The rating model can be used for rating the facilities such as working capital fund based, working capital non-fund based and term loan. The overall rating for the borrower can also be found out by calculating the weighted average of the rating of individual facility in which the respective sanctioned amounts are used as weights. Though the ultimate objective is to find out facility rating or borrower's overall rating it is possible to find rating of individual risk category. This is because the model takes modular approach in arriving at final rating where each module represents each risk category. 91

20 Step II : Designing the database and selecting proper tools After having understood the overall rating methodology in the first step the next step was to design a proper database for the same. Considering the fact that the model would work right from the branches to the Head Office level, it was necessary to select a platform that is commonly available at all places. Hence it was decided that the system would be developed in Visual Basic with its native Jet database engine at the back-end so that it can be distributed at ease without a need of specific database management system at the back end. Step III: Incorporation of important features In this stage the actual program development/coding took place. Some of the important features that were incorporated in the model development were as follows, 1. The financial details of the borrowers: Reserve Bank of India has mandated banks a certain way of analyzing the balance sheets of borrowers. The borrowers are required to submit their balance sheet data in a specific format called CMA (Credit Monitoring Arrangement) data format. This data is generally submitted in MS Excel format and the same data would be used by the rating model. Hence, an interface is given in the system to retrieve this Excel based CMA data so that user need not re-enter the data again. Also, a facility to enter the financial details is given in case user does not maintain CMA data in Excel. 2. The Qualitative and Quantitative Scoring: Some of the risk factors are quantitative in nature such as various financial ratios or 'Percentage of Net Profits achieved to projections made last year'. For such risk factors scores are maintained in the master files from where system would pick up the right score depending on the value in that risk factor. However for qualitative factors (such as 'Quality of Financial Statements' or 'Quality of Management Personnel') user is provided a Scoring Guide which he can read before entering the score for that risk factor. 3. The scores, weights and cut-offs: As indicated earlier the master data is maintained in a system to store scores of various risk factors, and weights which are used for various risk parameters and risk categories. The model is kept flexible to 92

21 allow the weights, cut-offs depending on the bank's interpretation of them. However, any unwarranted changes in this data would lead to incorrect rating results. Hence, only system administrator is given a privilege to make any changes in these important master files. 4. Business/Industry Risk Categories: A large number of risk factors could fall under these two risk categories. The system provides flexibility of selecting the relevant factors pertaining to an industry. The Economic Cell of the bank is expected to decide and freeze the most relevant factors for each industry and review and update the same periodically using internal industry intelligence or using external industry databases. Once the factors are finalized, the Economic Cell is expected to convey the same to the credit rating officers at the branch and the regional level. 5. Borrowers with Multidivisional Companies; A borrower may have more than one company. A model can incorporate three different business lines for a single borrower. For such borrowers the model computes scores of industry risk and business risk separately for each division. The final scores for industry risk and business risk are the weighted average of the individual industry score and business score, on the basis of division wise latest turnover. 6. Import/Export Facility: As stated earlier the model may work at all the places such as branch, Regional/Zonal Offices and the Head Office in a bank. However, the top management may be interested in having the overall (portfolio level) reports reflecting the data from all the branches, regional offices, etc. Hence, it was felt necessary to have the facility of exporting the data (from various points at which borrower rating has been done) and importing the data (where consolidation is needed). Thus, a facility was given in the model where data can be exported from a branch to a regional office and then from a regional office to the head office. The import facility was provided also at regional and head office level. 93

22 4.7.1 Important Reports* 1. Transition Matrix This report provides the profile of credit quahty changes or migrations that have taken place for the selected credit portfolio between any two years. It is a summary of how all the rated borrowers have migrated within various standard account categories as well as to default category in selected two years. This report generates two transition matrix tables: a) Number of transitions and defaults between year 1 and year 2: This table represents the actual number of accounts that have migrated (or remained there) between the two categories. b) Transition and Default Probabilities between year 1 and year 2: This table represents the percentage of total accounts in the first selected year that have migrated between different categories and to the default category in the second selected year. 2. Industry wise Rating Distribution by number of companies This report generates the distribution of the number of rated borrower accounts over all the rating grades for each industry and for the entire credit portfolio. It also generates the Industry wise default probabilities (i.e. the number of 'D' rated borrowers as a percentage of total borrowers in each industry). This report is important in getting an idea about industry concentration risk in the credit portfolio. 3. Industry wise Rating Distribution by the loan amount sanctioned This report is similar to earlier report. Only difference is it generates the distribution of the 'Sanctioned Limits' of the rated borrower accounts over all the rating grades for each industry and for the entire credit portfolio. It also generates the industry wise default probabilities in terms of exposure, i.e. exposure of D rated borrowers as a percentage of total exposure of borrowers in each industry. Sample Output reports/screens are enclosed in Annexure P 94

23 4. Company's Year wise Overall Rating In a year a borrower is likely to avail more than one facility. By default the model computes the rating for each of the facility such as working capital (fund based or nonfund based) or term loan. If the user wants to see overall rating of the borrower, he can do so by generate this report. 5. Stressed Accounts The system provides a facility to store cut-off values of various financial indicators such as Current ratio, interest coverage ratio, etc. If any of such cut-offs are violated by the borrower, the same can be seen in the form of report on 'Stressed Accounts'. This report gives year wise details of violated cut-offs by each borrower. 4.8 Conclusion This chapter discussed the important aspects of credit risk assessment and role of internal credit rating model in the same. The internal credit rating system embodies financial institution's viewpoint towards evaluation of credit risk. Though many Indian banks have been using computer based rating systems to score the borrower, they lack robustness and consistency. The rating model designed and developed in this study is based on comprehensive qualitative and quantitative risk parameters. It provides competitive advantage to the banks as they would now be in a better position not only to price exposures but also monitor high-value accounts efficiently both at branch level and central office level. The model also facilitates to build 'rating history' of borrowers that can track generation of default and transition probabilities. It provides a facility to monitor industry concentrations, supervise high-value accounts and have a continuous view on the overall portfolio quality. 95