Decomposing choice under uncertainty

What is it about?

When facing choice under uncertainty, such as ordering inventory or deciding on capacity, one needs a forecast, as well as an understanding about different risks and outcomes in order to be able to make a good decision. The paper examines whether decomposing an uncertain choice in such a way improves decision making, and shows that this is only the case if the uncertainty of the underlying decision problem is not too big, and if the cost of making a choice that is too low ex post is higher than the cost of making a choice that is too high.

Featured Image

Why is it important?

Making decisions under uncertainty is an almost daily activity for any manager; it makes intuitive sense to decompose such decisions by separately deciding on forecasts and service levels. We show in this paper that such a decomposition only works well under specific circumstances - managers can easily end up being more biased in their decision making if they attempt to decompose the decision. These kind of decisions had not been examined in the experimental literature in a decomposed fashion before; this is the first paper to do so. This approach also allows us to more clearly examine the components of a newsvendor decision (point forecasts, uncertainty estimates and service levels), and to point out behavioral biases explicitly in these components.

Perspectives