Objective Evaluation for Agricultural Lending

One of the more significant changes in the variables. Variables found to be significant U. S. agricultural industry in recent years has were: total debt divided by total assets, reabeen the increased use of credit to finance prosonable farm value, total liabilities, marital duction and capital expenditures. Since 1970, status, family living expense as a percentage of outstanding farm debt has more than doubled, total farm expense, and current liabilities dirising at an average annual rate of 9.3 percent. vided by current assets. Statistical analysis However, because net farm income has not insuggested the function should correctly creased as fast, the debt burden for farm classify 85 percent of the loans. operators has become relatively higher (MeliJohnson and Hagan evaluated the financial char and Waldheger). position and progress of acceptable and probThis increase in debt load has made financial lem borrowers from three central and northevaluation more difficult for lenders. Narrow western Missouri Production Credit Associaprofit margins and increased average loan size tions. A linear discriminant model was dehave made financial institutions more aware of veloped containing three significant ratio varithe need to determine, for their loan portfolio, ables: loan repayment made plus marketable how borrower and agricultural business inventory divided by loan repayment characteristics relate to debt repayment abilanticipated (called a repayment index), current ity and loan quality. assets divided by current liabilities, and total A study was designed to develop an obdebt divided by the total assets. The model jective credit evaluation technique based on correctly classified 62 percent of the loans loan repayment ability characteristics of farm analyzed. This model has been used by the Fedborrowers. Such a technique would aid lenders eral Intermediate Credit Bank of St. Louis to in discriminating between borrowers who assist in classifying loans for the Sixth Farm would be acceptable and those who would posCredit District. sibly turn out to be weak and have problems in Dunn and Frey developed a credit scoring repayment. model using acceptable and problem loans from Production Credit Association loan appliPREVIOUS RESEARCH cations for the cash grain area of central Illinois. The data were taken from applications of Numerous methods have been devised to borrowers who were new loan applicants beevaluate relative financial and personal charactween 1964 and 1968 and who still had loans teristics of borrowers; however, the most comoutstanding in 1971. Multiple discriminant monly used and widely accepted technique has analysis was used to determine that there was been multiple discriminant analysis. Discrimino significant difference between data from difnant analysis has been applied to the classificaferent years; however, differences between any tion of agriculturally related loans by Bauer possible subgroups, of data were not conand Jordan, Johnson and Hagan, and Dunn sidered. Stepwise discriminant analysis was and Frey. then used to determine the variables and their Bauer and Jordan collected data on good and respective coefficients that most significantly problem loans for the period 1958-69 from two distinguished between acceptable and problem eastern Tennessee Production Credit Associaloans. The final model contained four signifitions. No attempt was made to identify subcant variables: total liabilities divided by sets of the data that might have affected the total assets, amount of credit life insurance, overall effectiveness of the analysis. Stepwise amount of note divided by net cash farm inregression analysis was used to find the most come, and acres owned. The model correctly significant variables and multiple discriminant classified 90 percent of the acceptable loans analysis was used to find coefficients for these and 60 percent of the problem loans.