Mate Matching for Microsimulation Models: Technical Paper 2002-3

Working Paper
November 1, 2002

Kevin Perese

In dynamic microsimulation models, the process of determining who marries whom in the annual marriage market has important implications for numerous other processes modeled. There are two broad methodologies used in most microsimulation models to match individuals with each other–a stable marriage algorithm approach or a stochastic framework approach. This paper reviews those two approaches and summarizes the mate-matching techniques used in other dynamic microsimulation models. Building upon the methods used in other models, this paper presents an extension to the stochastic mate-matching approach. This new technique accurately replicates the joint distribution of spousal characteristics (differences in age, education, and earnings) observed in survey data. This technique is used in the Congressional Budget Office’s Long-Term (CBOLT) microsimulation model.