Dynamic Rawlsian Policy

A well-known time-inconsistency problem hinders optimal decision-making when policymakers are constrained in their pesent choices by expectations of future outcomes.  The time-inconsistency problem is caused by differences in the preferences of policymakers who exist at different points in time.  Adapting the arguments of Rawls (1971), we propose that these differences can be eliminated if policy is set from behind a 'veil of ignorance', without knowledge of when the policy will be implemented.  We set up a well-defined choice problem that captures this normative perspective.  The policies that it generates have a number of appealing properties.