Projecting Longitudinal Marriage Patterns for Long-Run Policy Analysis: Technical Paper 2002-2

October 1, 2002
Working Paper
A description of the method used in CBO's main Social Security model to project an individual's marital status.


Josh O’Harra and John Sabelhaus

Imagine a longitudinal micro data file that contains individual earnings along with basic demographic variables such as age, sex, education, marital status, and spouse’s characteristics for a representative future sample of the population. Such a data set would be invaluable for analyzing solvency and distributional questions about Social Security and other long-run policy issues, because the longitudinal data file would have all the information needed to tabulate taxes paid and benefits received under any set of program rules. The goal of dynamic micro-simulation modeling is to produce a longitudinal data set like the one described above. This paper describes one set of building blocks in dynamic micro-simulation: marital transition equations that predict first marriage, divorce, and remarriage for individuals over time.