THIRD PRESENTATION Q5: since the EDF relies on market prices to predict defaults, how does KMV apply such

data to private firms? Q6: What is the difference between KMV and other approaches? Q7: How do you test the predictive ability of default models?

Answer: Q5: Moody’s KMV estimates the default probability of a firm based on Merton’s approach. It determines an EDF (expected Default Frequency) using market value and volatility of the firm, estimated from market value of its stock, the volatility of its stock and the book value of its liabilities. However, private firms does not have market price, therefore we cannot determines EDF using market price of stock. Therefore, these models must rely upon the private firms reported characteristics and accounting data. The market value of the firm is based on two values: Operating value is calculated as earnings before taxes, interest, depreciation and amortization (EBITDA) times a multiplier. Liquidating value is based upon the firm’s book liabilities.

When EBITDA is high, the mkt value approaches the operating value; when it is low, it approaches the liquidating value. The asset volatility of the firm is modelled as a function of sales size, industry and asset size, which are determined by a multivariate statistical technique. Using the market value and volatility thus estimated, the EDF is estimated for the distancefrom-default ratios on the basis of the public firm default experience.

Q6: The EDF differs from other approaches by:

Q7: Simplest test of the predictive ability of default models is to compare the prediction with the actual outcome.- It relies on the information in equity prices It does not try explicitly to be predictive EDF simply looks at the current value of the firm relative to its default point and historical volatility. Predicting default when it did not occur (For example predict the probability default is 70% when the actua probability of default is only 20%) . it is because the current value of the firm is a good prediction of future values. There are 2 possible errors: 1. There is a close connection between changes in EDF values for a given firm and changes in the equity value of the firm. If the prediction is correct. Failing to predict default when it did occur (For example predict the probability default is 20% when the actual probability of default is 70%) 2. The most distinguishing feature of the EDF measure is that it imposes such a direct connection between market values and default probabilities. EDF model is totally dependent on stock prices for its information content.