Hrvatsko društvo ekonomista

ARTICLES ::: AUTHORS ::: ARTICLES SEARCH

February 2007. ::: Vol.58 No.01-02

Branko Novak
Ivica Crnković

CLASSIFICATION OF BANK DEBTOR DISTRESS BASED ON OFFICIAL FINANCIAL STATEMENTS Izvorni znanstveni članak The paper presents research of models for classification of bank debtor distress based on ratios calculated from official financial statements. Basic sample consists of 168 company’s debtors of middle size bank, and 15 financial ratios were used. In research multiple discriminate analyses, logit model and multidimensional scaling method was used for classifying companies based on orderliness in payment of their liabilities. Since discriminate analysis and logit model could not successfully discriminate “good” and “medium” companies, basic sample was appropriately limited to 141 companies classified as “good” or “dubious”. All estimated models have high rate of reliability in prediction of orderliness in payment of liabilities that can express the degree of company’s efficiency or distress. Reliability of all created models was confirmed on control sample as well. This research also demonstrated that multidimensional scaling methods can be successfully used for classifying of companies based on magnitude of their business problems using appropriate financial ratios. business problems; bank debtors; multiple discriminate analysis; logit model; multidimensional scaling method. Puni tekst (Hrvatski) Str. 41 - 71 (pdf, 375.45 KB)

CLASSIFICATION OF BANK DEBTOR DISTRESS BASED ON OFFICIAL FINANCIAL STATEMENTS

Izvorni znanstveni članak

The paper presents research of models for classification of bank debtor distress based on ratios calculated from official financial statements. Basic sample consists of 168 company’s debtors of middle size bank, and 15 financial ratios were used. In research multiple discriminate analyses, logit model and multidimensional scaling method was used for classifying companies based on orderliness in payment of their liabilities. Since discriminate analysis and logit model could not successfully discriminate “good” and “medium” companies, basic sample was appropriately limited to 141 companies classified as “good” or “dubious”. All estimated models have high rate of reliability in prediction of orderliness in payment of liabilities that can express the degree of company’s efficiency or distress. Reliability of all created models was confirmed on control sample as well. This research also demonstrated that multidimensional scaling methods can be successfully used for classifying of companies based on magnitude of their business problems using appropriate financial ratios.

business problems; bank debtors; multiple discriminate analysis; logit model; multidimensional scaling method.

Puni tekst (Hrvatski) Str. 41 - 71 (pdf, 375.45 KB)