Week 7 Reading
Sensitivity & Value of Information
This content is under development and subject to change.
Assigned Readings
Murphy et al. (1985) – “Repetitive Decision Making and the Value of Forecasts in the Cost-Loss Ratio Situation.” Introduces the cost-loss framework for quantifying the economic value of weather forecasts, extending from a static to a dynamic (multi-period) setting.
Runge et al. (2011) – “Which Uncertainty?” Uses expected value of perfect information (EVPI) and partial EVPI to identify which uncertainties most impede management of endangered whooping cranes. A clear worked example of VOI applied to real decision-making.
Discussion Questions
In the Runge et al. case study, “restore meadows” is the best strategy under uncertainty, but it is not the best strategy under any individual hypothesis. How is this possible? What does this imply about decision-making under uncertainty?
The partial EVPI analysis shows that resolving the “black fly” hypothesis captures 54% of the total information value. What does this mean for how the wildlife refuge should allocate its monitoring budget?
In Murphy et al.’s dynamic cost-loss model, the decision threshold changes from \(C/L\) (static) to \(C/(L-C)\) (two-period). Why does anticipating the future raise the threshold for protection? What’s the intuition?
Murphy et al. show that the value of imperfect forecasts is always non-negative (Blackwell’s theorem). Can you think of a real-world situation where acting on a forecast made things worse? How do you reconcile this with the theorem?