Résumé
Standard evaluations of public policies involve discounting the flow of expected net benefits at a unique discount rate. Consequently, they systematically ignore the insurance benefits of policies that hedge the aggregate risk, and the social cost of projects that raise the aggregate risk. Normative asset pricing theory recommends adjusting the discount rate to the project’s risk, but few countries have attempted to implement this complex solution. We explore the equivalent "stochastic discount factor" approach based on the expected value of its state-contingent NPV, using the relevant state-contingent Ramsey discount rate. Under our "stress discounting" approach, projects are evaluated under two polar risk-free economic scenarios, one business-as-usual scenario, and one low-probability catastrophic scenario. Inspired by the recent asset pricing literature on macro catastrophes, we show that this approach adequately values assets’ risk premia under a minimal, intuitive, and operationally simple departure from the standard risk-free approach with a unique discount rate. We carry out benchmarks to check the accuracy of this approach, then apply it to value a nuclear waste disposal.
Mots-clés
Project valuation; stochastic discount factor; rare disasters; cost-benefit analysis; social discounting;
Codes JEL
- G12: Asset Pricing • Trading Volume • Bond Interest Rates
- H43: Project Evaluation • Social Discount Rate
- Q54: Climate • Natural Disasters • Global Warming
Référence
Frédéric Cherbonnier, Christian Gollier et Aude Pommeret, « Stress discounting », TSE Working Paper, n° 26-1697, janvier 2026.
Voir aussi
Publié dans
TSE Working Paper, n° 26-1697, janvier 2026
