Credit Risk Models

Credit Risk Models

A credit risk model seeks to determine, directly or indirectly, the answer to the following question: Given our past experience and our assumptions about the future, what is the present value of a given loan or fixed income security? A credit risk model would also seek to determine the (quantifiable) risk that the promised cash flows will not be forthcoming. The techniques for measuring credit risk that have evolved over the last twenty years are prompted by these questions and dynamic changes in the loan market.

The increasing importance of credit risk modelling should be seen as the consequence of the following three factors:

  1. Banks are becoming increasingly quantitative in their treatment of credit risk.
  2. New markets are emerging in credit derivatives and the marketability of existing loans is increasing through securitisation/ loan sales market.
  3. Regulators are concerned to improve the current system of bank capital requirements especially as it relates to credit risk.

Credit Risk Models have assumed importance because they provide the decision maker with insight or knowledge that would not otherwise be readily available or that could be marshalled at prohibitive cost. In a marketplace where margins are fast disappearing and the pressure to lower pricing is unrelenting, models give their users a competitive edge. The credit risk models are intended to aid banks in quantifying, aggregating and managing risk across geographical and product lines. The outputs of these models also play increasingly important roles in banks’ risk management and performance measurement processes, customer profitability analysis, risk-based pricing, active portfolio management and capital structure decisions. Credit risk modeling may result in better internal risk management and may have the potential to be used in the supervisory oversight of banking organisations.

In the measurement of credit risk, models may be classified along three different dimensions: the techniques employed, the domain of applications in the credit process and the products to which they are applied.

Techniques: The following are the more commonly used techniques:

  1. Econometric Techniques such as linear and multiple discriminant analysis, multiple regression, logic analysis and probability of default, etc.
  2. Neural networks are computer-based systems that use the same data employed in the econometric techniques but arrive at the decision model using alternative implementations of a trial and error method.
  3. Optimisation models are mathematical programming techniques that discover the optimum weights for borrower and loan attributes that minimize lender error and maximise profits.
  4. Rule-based or expert systems are characterised by a set of decision rules, a knowledge base consisting of data such as industry financial ratios, and a structured inquiry process to be used by the analyst in obtaining the data on a particular borrower.
  5. Hybrid Systems In these systems simulation are driven in part by a direct causal relationship, the parameters of which are determined through estimation techniques.

Domain of application: These models are used in a variety of domains:

  1. Credit approval: Models are used on a stand alone basis or in conjunction with a judgemental override system for approving credit in the consumer lending business. The use of such models has expanded to include small business lending. They are generally not used in approving large corporate loans, but they may be one of the inputs to a decision.
  2. Credit rating determination: Quantitative models are used in deriving ‘shadow bond rating’ for unrated securities and commercial loans. These ratings in turn influence portfolio limits and other lending limits used by the institution. In some instances, the credit rating predicted by the model is used within an institution to challenge the rating assigned by the traditional credit analysis process.
  3. Credit risk models may be used to suggest the risk premia that should be charged in view of the probability of loss and the size of the loss given default. Using a mark-to-market model, an institution may evaluate the costs and benefits of holding a financial asset. Unexpected losses implied by a credit model may be used to set the capital charge in pricing.
  4. Early warning: Credit models are used to flag potential problems in the portfolio to facilitate early corrective action.
  5. Common credit language: Credit models may be used to select assets from a pool to construct a portfolio acceptable to investors at the time of asset securitisation or to achieve the minimum credit quality needed to obtain the desired credit rating. Underwriters may use such models for due diligence on the portfolio (such as a collateralized pool of commercial loans).
  6. Collection strategies: Credit models may be used in deciding on the best collection or workout strategy to pursue. If, for example, a credit model indicates that a borrower is experiencing short-term liquidity problems rather than a decline in credit fundamentals, then an appropriate workout may be devised.

Credit Risk Models: Approaches

The literature on quantitative risk modelling has two different approaches to credit risk measurement. The first approach is the development of statistical models through analysis of historical data. This approach was frequently used in the last two decades. The second type of modelling approach tries to capture distribution of the firm’s asset-value over a period of time.

The statistical approach tries to rate the firms on a discrete or continuous scale. The linear model introduced by Altman (1967), also known as the Z-score Model, separates defaulting firms from non-defaulting ones on the basis of certain financial ratios. Altman, Hartzell, and Peck (1995,1996) have modified the original Z-score model to develop a model specific to emerging markets. This model is known as the Emerging Market Scoring (EMS) model.

The second type of modelling approach tries to capture distribution of the firm’s asset-value over a period of time. This model is based on the expected default frequency (EDF) model. It calculates the asset value of a firm from the market value of its equity using an option pricing based approach that recognizes equity as a call option on the underlying asset of the firm. It tries to estimate the asset value path of the firm over a time horizon. The default risk is the probability of the estimated asset value falling below a pre-specified default point. This model is based conceptually on Merton’s (1974) contingent claim framework and has been working very well for estimating default risk in a liquid market.

Closely related to credit risk models are portfolio risk models. In the last three years, important advances have been made in modelling credit risk in lending portfolios. The new models are designed to quantify credit risk on a portfolio basis, and thus are applied at the time of diversification as well as portfolio based pricing. These models estimate the loss distribution associated with the portfolio and identify the risky components by assessing the risk contribution of each member in the portfolio.

Banks may adopt any model depending on their size, complexity, risk bearing capacity and risk appetite, etc. However, the credit risk models followed by banks should, at the least, achieve the following:

  • Result in differentiating the degree of credit risk in different credit exposures of a bank. The system could provide for transaction-based or borrower-based rating or both. It is recommended that all exposures are to be rated. Restricting risk measurement to only large sized exposures may fail to capture the portfolio risk in entirety for variety of reasons. For instance, a large sized exposure for a short time may be less risky than a small sized exposure for a long time.
  • Identify concentration in the portfolios.
  • Identify problem credits before they become NPAs.
  • Identify adequacy/ inadequacy of loan provisions.
  • Help in pricing of credit.
  • Recognise variations in macro-economic factors and a possible impact under alternative scenarios.
  • Determine the impact on profitability of transactions and relationship.