Decision Models in Consumer Lending in the Context of Economic Uncertainty

Rajaratnam, Kanshukan, Systems Engineering - School of Engineering and Applied Science, University of Virginia
Beling, Peter, Department of Systems and Information Engineering, University of Virginia

Credit scores are the primary vehicle for assessing the risk of loan applicants. Scores are mapped to the likelihood of an applicant defaulting or becoming seriously delinquent on the loan within a pre-determined period. The default probabilities are used to determine the profitability and the capital requirement for each borrower. Scorecards are built on historical data that are aggregated across many years and hence, possibly across many economic cycles. However, there is evidence in literature that default rates should be considered conditional on current and future economic conditions. This research focusses on improving decision making in retail credit through consideration of future economic conditions. The fundamental issue that we address is that the performance of a acquisition decision policy may be dependent on prevailing economic conditions during the loan period, and yet the policy must be specified and implemented before the loan period and hence before the economic environment is known with certainty.

We addressed this research opportunity in four ways. Firstly, we develop methods for incorporating forecasts of future economic conditions into acquisition decisions for scored retail credit and loan portfolios. We suppose that a portfolio manager is faced with two possible future economic scenarios, each characterized by a known probability of occurrence and by known performance functions that give expected profit and volume. We show that, despite the uncertainty of performance induced by economic conditions, every efficient policy consists of a single cutoff score, provided the expected profit and volume performance curves in each scenario are concave.

Secondly, we prove that misestimating regulatory capital requirements in either direction results in a negative impact on profit. A source of misestimation is due to errors in forecasts of future economic scenarios, resulting in differences between the reserve amount and the amount required under the realized economic condition. Thirdly, we develop methods for incorporating forecasts of future economic conditions into acquisition decisions for a portfolio manager faced with capital constraints and costs.

Finally, we give consideration to decisions by borrowers faced with a sequence of credit offers. From the definition of adverse selection in static lending models, we show that homogenous borrowers take-up offers at different instances of time when faced with a sequence of loan offers. We postulate that bounded rationality and diverse decision heuristics used by consumers drive the decisions they make about credit offers. Under that postulate, we show how observation of early decisions in a sequence can be informative about later decisions and can, when coupled with a type of adverse selection, also inform credit risk during the period of account performance

PHD (Doctor of Philosophy)
Consumer lending, Credit scores, Decision models
All rights reserved (no additional license for public reuse)
Issued Date: