Chapter 11: Demographic Models

Demographic microsimulation is at once compellingly simple and devilishly complex. It is simple because it seeks to simulate familiar human demographic processes at the individual level: people are born, they may marry, reproduce, migrate, and then they die. Its complexity, on the other hand, arises from the nature of those familiar behaviors. In order to simulate a demographic event, we must specify not only the age, sex, marital status, and parity specific rates at which the event occurs to members of each identified population subgroup, but also for some events, a set of decision rules which may involve more than one individual and/or a stochastic mechanism for adjusting individual level heterogeneity possibly in a heritable manner. In this chapter, I use the term demographic microsimulation to refer to dynamic stochastic microsimulation programs wherein the unit of observation is the individual, and the main purpose is to create simulated populations with kinship networks in order to answer questions of interest to demographers and other social scientists. Although many economic or population related questions can be addressed using static models (in which time plays no role) or deterministic models (in which only one outcome of the simulation is possible), simulating kinship requires that uncertainty of events and the passage of time be taken seriously.¹