Evaluation of a decision support system for credit management decisions

This study was conducted to evaluate the effectiveness of a decision support system (DSS) for credit management. This study formed a part of a larger initiative to assess the effectiveness of IT-based credit management processes at the State Bank of India (SBI). Such a study was necessitated since credit appraisal has emerged to become a critical sub-function in Indian banks in view of growing incidence of non-performing assets. The DSS that we assessed was a credit appraisal system developed in Quattro Pro® at SBI. This system helps in the analysis of balance sheets, calculation of financial ratios, cash flow analysis, future projections, sensitivity analysis and risk evaluation as per SBI norms. We used a strong quasi-experimental design, called the Solomon's four-group design, for our assessment. In our experiment, managers of SBI who attended training programs at the SBI training college, were the subjects. The experiment consisted of measurements that were taken as pre- and post-tests. An experimental intervention was applied between the pre-tests and the post-tests. The intervention, or stimulus, consisted of DSS training and use. There were four groups in our experiment. The stimulus remained constant as we took care to ensure that the course contents as well as the instructors remained the same during the course of our experiment. Two were experimental groups and two were the control groups. All four groups underwent training in credit management between the pre- and post-tests. Results from our research show that while the DSS is effective, improvements need to be made in the methodology to assess such improvements. Moreover, such assessment frameworks, while being adequate from a DSS-centric viewpoint, do not respond to the assessment of a DSS in an organizational setting. In our concluding section, we have discussed how this evaluative framework can be strengthened to initiate an activity that will allow the long-term, and possibly the only meaningful, evaluation framework for such a system.