Embedded Files
M8-02 lecture
M8-02 lecture
Classification defined
Classification defined
Classification is a predictive analytics technique used to separate or classify a sample into one or more groups.
Bankruptcy classification
Bankruptcy classification
In 1968, Edward Altman published a study that classified companies based on a bankruptcy risk score he developed, called Altman's Z. His score is still used today. The idea is that certain common business ratios, such as profitability, could be used identify firms with a risk of going bankrupt.
Calculation
Calculation
The calculation of the Altman Z score is based on 5 factors.
These five factors are aggregated into a single score through the following formula.
Altman’s Z-Score = 1.2*X1 + 1.4*X2 + 3.3*X3 + 0.6*X4 + 1.0*X5
Altman’s Z-Score = 1.2*X1 + 1.4*X2 + 3.3*X3 + 0.6*X4 + 1.0*X5
Loan extension classification
Loan extension classification
An important decision for lenders is whether to extend a loan to a potential borrower. Banks and other lenders evaluate borrowers based on their credit worthiness to decide whether they will extend a loan to them. What are some variables that can be used to classify potential borrowers?
Credit history
Credit history
What is the borrower's credit score?
How much total outstanding debt does the borrower already have?
Has the borrower declared bankruptcy in the last 3-5 years?
Income
Income
What is the borrower's annual income?
Loan specifics
Loan specifics
What is the purpose of the loan?
Fraud/no fraud classification
Fraud/no fraud classification
Both internal and external auditors are interested in predicting whether fraud is present in the financial statements. In a fraud/no fraud classification, an auditor may use factors to help classify companies into two groups:
Firms that likely committed financial statement fraud.
Firms that are not likely to have committed financial statement fraud.
In a famous study, Messod Beneish (1999) developed a model to help identify firms that have a high likelihood of financial statement fraud.
Beneish found that 8 factors help predict fraud.
Increases in receivables compared to last year.
Decline in gross margin as a percent of sales.
Decrease in noncurrent assets relative to current assets.
Increase in sales growth in the current period.
Descrease in depreciation expense.
Decrease in selling, general, and adminstrative expenses.
Increase in debt.
Higher total accruals relative to total assets.
Similar to Altman, Beneish aggregated these factors into a single fraud prediction score called Beneish's M-Score.
Page updated
Google Sites
Report abuse