Decision Making Under Uncertainty

Many of the decisions we face are made complicated by having uncertain consequences: how should I set my inventory when I don’t know what demand will be, should I refinance my mortgage when rates might go lower, how big a bet shall I make in a new business, and so on. Evidently a proper examination of these problems requires thinking through the possible outcomes of any decision, an understanding of how likely the various possible outcomes are, and a system for weighing the pros and cons of the alternatives. The course will provide tools for handling all three of these aspects. The reality is that formal analyses are usually carried out when you are preparing a recommendation for someone else. This course will teach you how to do that, but even more importantly, it will teach you how to think about risky situations so that it becomes part of your instinctive approach to problems. 

The first half of the course is about techniques for appraising specific problems. It consists of a somewhat traditional approach to the subject of decision analysis. We begin with estimating probabilities through subjective belief (personal experience and judgement), using experts, and using data. We will cover regression analysis, updating beliefs based on evidence, evaluating hypotheses, and the use of scenarios. We will cover the evaluation of alternatives using decision trees, simulation and preference analysis.

The second half of the course is about using systematic strategies to reduce risk. We will practice estimating exposure (e.g. how much will you lose/gain if interest rates go up), learn hedging techniques to reduce exposure (for example by insurance or using the futures market), the merits of diversification and the use of options. The final few classes applies the techniques we have covered to the analysis of personal risks. We will consider both rational and psychological aspects of risk taking.

The course is about techniques and approaches. You don’t have to be a quant jock to learn a lot from the course, but you shouldn’t take the course if you don’t enjoy using spreadsheets. Some students will have seen decision trees and forecasting before. For these students there will be some duplication but I believe the course has much more to offer. Most of the sessions will be based on exercises or short cases in order that we focus on the underlying ideas, but the situations we discuss will be based on real problems that did happen to businesses or individuals.