Files in this item



application/pdf9314894.pdf (6MB)Restricted to U of Illinois
(no description provided)PDF


Title:Performance-based credit scoring model
Author(s):Khoju, Madhab Raj
Doctoral Committee Chair(s):Barry, Peter J.
Department / Program:Agricultural and Consumer Economics
Discipline:Agricultural Economics
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Economics, Agricultural
Economics, Finance
Abstract:The repayment ability of the borrower and, therefore, the loan performance should be contingent on actual performance of the financed business because other sources of income may not be reliable as evidenced by recent economic recession in USA. Accordingly, credit scoring models should basically focus on evaluating the lender's judgements about a borrower's future business performance. Similarly, the validity of these models should be based on the accuracy in predicting borrower's actual business performance in future. This requires a time series data for the same sample borrowers so that the data for the preceding years are used as the estimating sample and the data for the following year are used as the hold-out sample. The estimating and hold-out samples in existing credit scoring studies, however, are constituted of cross section observations on different borrowers. Moreover, these studies make no references to the performances of the businesses.
This study proposes and develops a performance based credit scoring model (PBCSM) that can be used to assess the probability of loan repayment based on potential business performance. The PBCSM was first estimated and validated using time-series of 74 grain farms in central Illinois. The data from 1986 to 1989 were used as an estimating sample while the 1990 data were used a hold-out sample. A farm was identified as acceptable (problem) if the repayment ability was higher (lower) than the loan repayment obligations. Because the repayment ability of a business is the result of solvency position, liquidity, realized profitability and operating efficiency, these criteria were examined as the potential predictors of business performance.
The relationship of predictor variables to business performance is conditional on the favourable or unfavourable economic environment that generated the estimating sample. Since the economic environment differs across years, lenders need to update the PBCSM. However, updating requires significant resources. Accordingly, the PBCSM was also estimated from a data base generated for a representative grain farm (data base generation involves much less resources than collecting actual data) and its usefulness was evaluated by comparing it to the PBCSM estimated from pooled time series.
The estimated performance based credit scoring models indicate that three financial ratios--rate of return on equity, equity to asset ratio, and operating efficiency ratio--are the significant predictors of business performance and, therefore, loan performance. These results hold for both sets of data used for model estimation--actual time series for a sample of grain farms as well as the generated data for a representative grain farm. The estimated models correctly predicted 78.38 and 77.78 percent of the business performances in the hold-out samples of actual time series and generated data, respectively. These predictive accuracies did not differ significantly from the predictive accuracies of existing credit scoring models in the literature.
Issue Date:1993
Rights Information:Copyright 1993 Khoju, Madhab Raj
Date Available in IDEALS:2011-05-07
Identifier in Online Catalog:AAI9314894
OCLC Identifier:(UMI)AAI9314894

This item appears in the following Collection(s)

Item Statistics