Failure Prediction Model on Performance of Agricultural Cooperatives' Rice Mills in Northern Thailand

Abstract

This study shows the development of a logit and discriminant failure prediction model to prediction failure non-failure in the cooperatives' rice mills of Northern Thailand. The objective of the study is to provide a failure prediction model developed for the the cooperatives' rice mills of Northern Thailand using financial data and efficiency of management which is derived from data envelopment analysis (DEA). The theoretical model developed for the study include five constructs that indicate success or failure in cooperatives' rice mills. The five constructs are capital (C), asset (A), management (M), earning (E) and liquidity (L) position-hence the term CAMEL rating is used in the study. Prediction results show that the logit failure prediction model predicted failure more accurately than discriminant failure prediction model. The logit failure prediction model accuracy rate of 95.77% compare to 93.75% of discriminant failure prediction model. Policy recommendation from this study was to set up the failure prediction model and early warning system for the agricultural cooperatives' rice mills in northern thailand by using the Logit failure prediction model due to higher accuracy and better result of early warning system. For the setting of the failure prediction model it was recommended that it should consist of total operating income to total operating expense ratio, total operating income to current assets ratio and efficiency of management (variable returns to scale).