Altman Z-score
The Z-score formula for predicting bankruptcy was published in 1968 by Edward I. Altman, who was, at the time, an Assistant Professor of Finance at New York University. The formula may be used to predict the probability that a firm will go into bankruptcy within two years. The Altman Z-score (and all similar models) have the following characteristics:
1.It was developed by using a statistical methodology to distinguish between two groups of companies – bankrupt and non-bankrupt. These groups were of equal size.
2.The statistical methodology created two groups (or grades). Companies scoring below a “cutoff point” were classified as bad. Those scoring above the “cutoff point” were classified as good. However, there were classification errors, and some companies were misclassified. The cutoff point between the two groups resulted in:
a) Some bankrupt companies having scores above the cutoff point.
b) Some non-bankrupts having scores below the cutoff point.
3.Altman (and some other model builders) identified 3 zones on the scale of his published model:
a) Good, i.e., no bad companies were classified as good.
b) Bad, i.e., no good companies were classified as bad.
c) Zone of Ignorance, i.e., an area between the good and bad companies which included all the test sample classification errors.
4.The result is what appears to be a simple three signal result: red, yellow and green lights.
How ZETA® credit system addresses these problems:
1.It provides context for a score in several ways:
a) Reports provide percentile rankings relative to the universe of public companies and relative to the public companies in like SIC codes.
b) Reports provide an indication of how the company fits in the public bond rating scale.
c) Reports provide estimates of default probabilities associated with scores.
2.The probability of default can be used in various financial modeling applications:
a) Estimating the appropriate risk premium in commercial lending.
b) Estimating the appropriate discount in publicly traded bonds.
c) Modeling the credit limit in trade credit applications.
d) Calculating the incremental expected cost associated with individual suppliers.
3.These characteristics are independent of the assumptions that were made in the original test sample and amount to extensive holdout experience.
4.Because we understand the spectrum of credit quality, we know how to customize decision rules such as:
a) “Accept only if the company is investment grade.”
b) “Reject the customer if the expected default rate is greater than 5%.”
c) “Find a replacement supplier if the expected cost (due to possible failure) is 20% higher than the contractual cost.”