May 2, 2005
Amy Rehder Harris, John Sabelhaus, and Almudena Sevilla-Sanz
The Congressional Budget Office Long-Term (CBOLT) model uses dynamic micro-simulation to analyze Social Security policy. The version of CBOLT currently being used to analyze policy for the Congress incorporates micro behavioral effects insofar as agents alter their timing of initial claiming of Old Age Insurance (OAI) worker benefits when benefits change, and that has a direct impact on government outlays and a feedback on the macro economy through changes in labor supply. However, the change in benefit claim age is only one of three behavioral responses that could be considered in Social Security analysis–the other two are labor supply (before or after claim age) and saving behavior. This paper develops a structural life-cycle model in which agents make choices over all three margins. Because the structural model is developed using the same stochastic processes in CBOLT’s micro-simulation, the state-dependent behavioral “rules” obtained from solving the life-cycle model can be used to determine behavior in a CBOLT baseline or reform-analysis simulation.