Pension Reform and Life-Course Distributions

Abstract

In order to carry out micro-level analyses of economic behavior, and of the influence of public policy over time, such as examining the redistributive impact of the tax-benefit system over the life course, it is necessary to utilize a panel dataset with many years of data. In general, such datasets are not available, either because the analysis relates to the future, as in the case of pension projections, or because existing datasets do not cover sufficiently-long time horizons. Instead, therefore, dynamic microsimulation models are used to synthetically generate a hypothetical panel. In this chapter, we discuss some of the methodological issues related to the construction of a dynamic microsimulation model. Building upon the work in Chapter 8, we develop the concept of income-generation models to model changing income distributions over time. We also introduce the concept of alignment, which allows us to calibrate the aggregate results of income-generation models to external control totals. The chapter discusses methodological issues related to dynamic modelling, such as static versus dynamic ageing, behavioural versus statistical simulation, discrete versus continuous time, open versus closed models, steady state versus forecasted projections, cohort versus population models, and validation. Methodologically, we discuss the calibration method known as alignment. From a measurement issue point of view, we describe how to quantify inter-temporal redistribution, between and within life trajectories. We conclude by undertaking a simulation exercise, modelling the distribution of income over the lifetime, and developing and applying a dynamic microsimulation model using a Chile case study.