Microsimulation for demography

Abstract

Background: Microsimulation consists of a set of techniques for estimating characteristics and modelling change in populations of individuals.

Aims: To demonstrate how microsimulation can be used by demographers who want to undertake population estimates and projections.

Data and methods: We use data from the 2011 United Kingdom (UK) Census of population to create a synthetic population by age, sex and ethnic group. Static and dynamic microsimulations and the visualisation of results are undertaken using the statistical package R. The code and data used in the static and dynamic microsimulation are available via a GitHub repository.

Results: A synthetic population in 2011 by age, sex and ethnicity was produced for the East London Borough of Tower Hamlets, estimated from two Census tables. A population projection was produced for each of these age, sex and ethnicity groups to 2021. We used a projection of the Bangladeshi population to visualise population growth by Middle-layer Super Output Area (MSOA) and to produce a population pyramid of estimates in 2021.

Conclusions: We argue that microsimulation is an adaptable technique which is well suited to demography, for both population estimation and projection. Although our example is applied to the East London Borough of Tower Hamlets, the approach could be readily applied in Australia, or any other country.