Abstract

Large microsimulation models, such as EUROMOD, are valuable tools to assess potential policy reforms for both academics and policy makers. Yet these models are often complex and difficult to build as the knowledge required are scattered across many disciplines or even uncodified. When developing a simulation model with real-world applications, there are many considerations that are largely ignored when experimenting with a proof-of-concept or a small-scale academic model. Issues such as architecture, data, complexity, links between models and validations can be challenging in almost all microsimulation projects. This session introduces some common algorithms and practices in developing a large-scale microsimulation model, covering popular methods in incorporating spatial and other external information into data, demographic and labour market simulation, alignment, links with CGE/DSGE and others. It also shares the practices in developing complex models, including architecture design, version control, debugging and validations. The session targets researchers who currently work on, or intend to work on complex simulation models. Bio Dr Jinjing Li is an Associate Professor at the National Social and Economic Modelling Centre (NATSEM) at the University of Canberra, Australia. His main research interest is policy modelling and evaluation using microsimulation techniques, with a focus on simulating the shifts in income distribution and the behavioural response due to changes in the economic environment and public policies. Dr Li developed a range of economic simulation models for different organisations and collaborated with many government agencies, think tanks and international organisations around the world. He currently leads multiple Australian policy microsimulation projects on taxation and health.