Commercial Bank Management Exam #2- Risk Models

Problems associated with discriminant model:

-Only considers two extreme cases (default/no default) -No reason to expect that the weights in a credit scoring model will be constant long-term; sensitivity to variable weights -Ignores hard to quantify factors, including business cycle effects and reputation -Database of defaulted loans is not available to benchmark the model

Logit Model

-Overcomes weakness of the linear probability model by restricting the estimated range of default probabilities from the linear regression model to lie between 0 and 1 - Quality of credit scoring models have improved, providing positive impact on controlling write-offs and default

Borrower-Specific Factors

-include reputation, leverage, volatility of earnings, and collateral

Terms and meaning of Altman's Discriminant Function

A Z score of less than 1.81 should be considered a high default risk; between 1.81 and 2.99, an indeterminant default risk; and greater than 2.99, a low default risk. -X1= Working capital/total assets ratio -X2= Retained earnings/total assets ratio -X3= EBIT/total assets ratio -X4= Market value equity/ book value of total liabilities -X5= Sales/total assets ratio

Level of Interest Rate

A higher level of interest rates may lead to a higher default rates, so lenders are more reluctant to lend under such conditions.

Leverage

A measure of the existing debt of the borrower; the larger the debt, the higher the risk premium

Reputation

Based on lending history of the borrower; better reputation implies a lower risk premium

2 types of Qualitative Factors

Borrower-Specific Factors Market Specific Factors

Market Specific Factors (2 of them)

Business Cycle Level of Interest Rates

Quantitative Models

Credit scoring models are mathematical models that: - use observed loan applicant's characteristics to calculate a score representing the applicant's probability of default or -Sort borrowers into different default risk classes

Collateral

If collateral is offered, the risk premium is lower

Business Cycle

Lenders are less likely to lend if a recession is forecasted

Major Weakness of Linear Probability Model

Major weakness is that estimated probabilities of default can often lie outside of the [0,1] interval -Since superior statistical techniques are readily available, there is rarely justification for employing linear probability models

Estimating the probability of repayment using linear credit scoring model example: Example 10-2: Suppose there were two factors influencing the past default behavior of borrowers; the leverage or debt-equity ratio (D/E) and the sales-asset ratio (S/A). Based on the past repayment experience, the linear model is: PD = 0.5 (D/E) - 0.0525 (S/A) Assume a prospective borrower has a D/E = 0.3 and an S/A = 2.0. Its expected probability of default (PD) is

PD = 0.5 (0.3) - 0.0525(2.0) = 0.045 or 4.5%

Borrower Specific Factors (4 of them)

Reputation Leverage Volatility of Earnings Collateral

Volatility of Earnings

The more stable the earnings, the lower the risk premium

Altman's Discriminant Function Example: Example 10-3: Suppose that the financial ratios of a potential borrowing firm take the following values: X1= 0.2 X1 indicates the firm is reasonably liquid X2= 0 X2is zero and X3 is negative, indicating X3= -0.20 the firm has had negative earnings/losses X4= 0.10 X4indicates the firm is highly leveraged X5= 2.0 X5 firm is maintaining its sales volume

Z=1.2(0.2)+ 1.4(0)+ 3.3(-0.2)+ 0.6(0.10) + 1.0(2.0) Z= 0.24 + 0 - 0.66 + 0.06 + 2.0 Z=1.64 With a Z score less than 1.81 (high risk), the Bank should not make a loan to this borrower until it improves its earnings.

Qualitative Models

consider borrower specific factors as well as market, or systematic, factors

Market Specific Factors

include business cycle and interest rate levels

Linear Probability Model

use past data such as financial ratios as inputs into a model to explain repayment experience on old loans.

Linear Discriminant Model

§Discriminate models divide borrowers into high or low default risk classes contingent on their observed characteristics. §Similar tolinear probability models, linear discriminate models use past data as inputs to explain repayment experience. §The relative importance of the factors used in explaining past repayment performance then forecasts whether the loan falls into the high or low default class.

Quantitative Models By selecting and combining different economic and financial characteristics, the bank may be able to:

§Numerically establish which factors are important in explaining default risk §Evaluate the relative degree or importance of these factors §Improve the pricing of default risk §Be better able to screen out bad loan applicants §Be in a better position to calculate any reserves needed to meet expected future loan losses

Risk models can be broken down into 2 groups:

§Qualitative §Quantitative

Altman's Discriminant Function

§Used for publicly traded manufacturing firms. §The indicator variable Z is an overall measure of the default risk classification of a commercial borrower. §The Z depends on the values of various financial ratios of the borrower and the weighted importance of these ratios based on past observed experience of default. Z=1.2X1+ 1.4X2+ 3.3X3+ 0.6X4+ 1.0X5