Microsimulation Modelling
for the Social Sciences
(Essex, 2019)

Part of the Essex Summer School in Social Science Data Analysis, 22 July - 2 August 2019 (two week course, 35 hours). 

Lecurers: Matteo Richiardi, Diego Collando, Katrin Gasior.

Course content

The course covers the following topics:

• General introduction to microsimulation modelling in the social sciences.

• Static non-behavioural tax-benefit microsimulation: General structure, and an introduction to the EUROMOD simulation platform.

• Behavioural static labour supply models using EUROMOD: Specification options, estimation issues, applications to the microsimulation of reforms.

• Dynamic microsimulation models: General structure, estimation issues, and an introduction to NETLOGO.

• Agent-based models: General structure and their relationship to dynamic microsimulation, estimation issues.

Course objectives

The course aims at giving participants a comprehensive overview of microsimulation modelling for the social sciences. Participants will gain knowledge of the different modelling approaches, and learn how to build and estimate their own microsimulation models, using state-of-the-art microsimulation tools. 

Course pre-requisites

• Interest in distributional analysis, complexity and the role of heterogeneity

• Basic knowledge of the definition of disposable household income, equivalence scales, at-risk-of-poverty and Gini

• Knowledge of regression analysis

• Basic knowledge of general equilibrium theory

Syllabus

Day 1: Overview of the course and Introduction to static tax-benefit models

We will present the course outline and will introduce the first type of microsimulation models that we are going to cover in the first part of the course: static tax-benefit models. We will give an overview of different models and their use for single country as well as for comparative research. We will finish the first day with installing the open source EUROMOD software which we are going to use in the practical lessons over the next days.

Suggested reading:

1. O'Donoghue, C. (2019). Practical microsimulation modelling, Ch. 1. Oxford University Press.

2. O'Donoghue, C., Loughrey, J. & Harding, A., 2014. Static Models. In C. O'Donoghue, ed. Handbook of Microsimulation Modelling. Emerald Group Publishing Limited, pp. 47-75.

3. Figari, F,, Paulus, A. & Sutherland, H., 2015. Microsimulation and Policy Analysis. In A. B. Atkinson & F. Bourguignon, eds. Handbook of Income Distribution. Oxford: Elsevier B.V,, pp. 2141-2221.

4. Sutherland, H., 2014. Multi-Country Microsimulation. In C. O'Donoghue, ed. Handbook of Microsimulation Modelling. Emerald Group Publishing Limited, pp. 77-106.

Day 2: Introduction to EUROMOD

The second day will focus on the multi-country microsimulation model EUROMOD. EUROMOD is the tax-benefit model for the European Union and open source software. We will spend the second day introducing the user interface and the main features of the model. The theoretical part will be accompanied by hands-on exercises. This will enable participants to look behind the ‘black box’ of microsimulation models and help them understand how they work in practise. Given the open source nature of the model and software, participants will be able to use EUROMOD after the course and to develop their own research questions.

Suggested reading:

1. Sutherland, H. & Figari, F,, 2013. EUROMOD: the European Union tax-benefit microsimulation model. International Journal of Microsimulation, 6(1), pp. 4-26. 

2. Sutherland, H. (2018). Quality Assessment Of Microsimulation Models: The Case Of EUROMOD. International Journal of Microsimulation, 11(1), pp. 198-223. 

Day 3: Using microsimulation models for distributional analysis and impact assessment of policy reforms

Day 3 will focus on specific applications of tax-benefit models. We will introduce different possibilities of using EUROMOD and will also generate our own analysis with hands-on exercises.

Suggested reading:

1. De Agostini, P, Paulus, A. & Tasseva, I. (2015). The effect of tax-benefit changes on the income distribution in 2008-2014, EUROMOD Working Paper Series 11/15

2. Figari, F., Paulus, A. & Sutherland, H. (2015). Microsimulation and Policy Analysis. In A. B. Atkinson & F. Bourguignon, eds. Handbook of Income Distribution. Oxford: Elsevier B.V., pp. 2141-2221.

3. Jara, X. H. & Sutherland, H. (2014). The implications of an EMU unemployment insurance scheme for supporting incomes. EUROMOD Working Paper Series 5/14. 

4. Salanauskaite, L. & Verbist, G. (2013). Is the neighbour's grass greener? Comparing family support in Lithuania and four other New Member States. Journal of European Social Policy, 23(3), pp.315-331.

Day 4: Work incentives

The final session on static-models will move towards behavioural components. One specific application of a tax-benefit model is the calculation of work incentives. Indicators at the intensive margin such as Marginal Effective Tax Rates measure to what extent tax-benefit systems incentivise to increase working hours or to earn more. Indicators on the extensive margin such as Participation Tax Rates measure to what extent systems incentivise to take up work. We will introduce the idea of work incentives in general and show how to calculate Marginal Effective Tax Rates using EUROMOD.

Suggested reading:

1. Adam, S., Brewer, M. & Shephard, A. (2006). The poverty trade-off: work incentives and income redistribution in Britain, IFS. 

2. Jara, H.X., Gasior, K. & Makovec, M. (2016). Low incentives to work at the extensive and intensive margin in selected EU countries, EUROMOD working paper series 3/17.

3. Matsaganis, M. & Figari, F, (2016). 

4. Pirttila, J. & Selin, H. (2011). Tax Policy and Employment: How Does the Swedish System Fare? CESifo Working Paper Series 3355.

Day 5: Introduction to labour supply modelling

Day 5 will introduce the static (one period) and dynamic (multi-period) microeconomics frameworks of labour supply. We will give an overview of the application of the static framework to evaluate the ex-ante and ex-post impact of tax-benefit reforms, and the role of EUROMOD in these applications. During the rest of the day we will centre on static structural models (and will come back to static non-structural models on day 7). We will describe some alternative modelling options and focus our attention on discrete choice (DC) models based on the random utility maximisation (RUM) approach. We will then derive an empirical specification for such a model and estimate it.

Suggested reading:

1. Aaberge, R., & Colombino, U. (2018). Structural Labour supply Models and Microsimulation. International Journal of Microsimulation 11(1), pp.162-197. 

2. O'Donoghue, C. (2019). Practical microsimulation modelling, Ch. 5. Oxford University Press.

Day 6: Behavioural microsimulation of tax-benefit reforms

Day 6 will focus on the use DC-RUM models in behavioural microsimulation of tax-benefit reforms. We will first go through the issue of utility calibration to place individuals in their observed discretised hours. Subsequently we will show how we can estimate income distribution measures (such as poverty, inequality, etc.). This is not straight forward because the DC-RUM approach does not identify a particular level of hours for each individual after a reform but produces a probability distribution over the discrete options of hours. We will finally (ex-ante) evaluate a tax-benefit reform.

Suggested reading:

1. Creedy, J., & Kalb, G. (2005). Discrete hours labour supply modelling: specification, estimation and simulation. Journal of Economic Surveys, 19(5), pp. 697-734.

Day 7: Non-structural labour supply modelling

We will derive a reduced-form econometric specification from the static framework to assess the impact of actual tax-benefit reforms. We will show how synthetic, simulated and group-level instrumental variables can be used to deal with the endogeneity between labour supply and Marginal Effective Tax Rates (METR) and income. We will finish this day by estimating a reduced-form model relating METR and hours worked.

Suggested reading:

1. Blundell, R., Duncan, A., & Meghir, C. (1998). Estimating labor supply responses using tax reforms. Econometrica, pp. 827-861.

2. Cutler, D. M., & Gruber, J. (1996). Does public insurance crowd out private insurance? The Quarterly Journal of Economics, 111(2), pp. 391-430.

3. Gruber, J., & Saez, E. (2002). The elasticity of taxable income: evidence and implications. Journal of Public Economics, 84(1), pp. 1-32.

Day 8: Dynamic microsimulation models

We will provide an introduction to dynamic (multi-period) microsimulation models and their applications. With respect to static (one-period) models, dynamic models require to update the characteristics of the population at each period. Moreover, as projections extend to a longer time horizon, the issue of general equilibrium feedback becomes more relevant. We will also provide an overview of the simulation platform JAS-mine.

Suggested reading:

1. Dekkers G., & O'Donoghue C. (2018). Increasing the impact of dynamic microsimulation models. International Journal of Microsimulation 11(1), pp. 61-96. 

2. Li J., O'Donoghue, C. (2013). A survey of dynamic microsimulation models: uses, model structure and methodology. International Journal of Microsimulation, 6(2), pp. 3-55.

3. Richiardi, M., & Richardson, R. (2017). JAS-mine: A New Platform for Microsimulation and Agent-Based Modelling. International Journal of Microsimulation, 10(1), pp. 106-134. 

4. Richiardi, M., & Richardson, R. (2016). Agent-based Computational Demography and Microsimulation using JAS-mine. In: Grow A, van Bavel J. Agent-Based Modeling in Population Studies, Springer, Berlin.

Day 9: Agent-based models: An introduction

We will provide an introduction to agent-based models, their features, historical development and applications. We will introduce the NETLOGO simulation platform, and guide the students through the creation of their first agent-based models.

Suggested reading:

1. Billari, F., & Prskawetz, A. (2003). Agent-Based Computational Demography. Using Simulation to Improve Our Understanding of Demographic Behaviour. Springer.

2. Cederman, L.-E. (2001). Agent-based modeling in political science. The Political Methodologist, 10(1), pp. 16-22.

3. Epstein, J. (2006). Generative Social Science: Studies in Agent-Based Computational Modeling. Princeton University Press.

4. Epstein, J, & Axtell, R. (1996). Growing Artificial Societies: Social Science from the Bottom Up. The MIT Press.

5. Gilbert, N., & Hamill, L. (2015). Agent-based modelling in Economics. Wiley.

6. Macy, Willer (2002). From Factors to Actors: Computational Sociology and Agent-Based Modeling. Annual Review of Sociology, 28, pp. 143-166.

7. Miller, J., & Page, S. (2006). Complex Adaptive Systems: An Introduction to Computational Models of Social Life. Princeton University Press.

8. Richiardi, M. (2013). The missing link: AB models and dynamic microsimulation. In: Leitner S., Wall, F. (eds). Artificial Economics and Self Organization. Springer, Lecture Notes in Economics and Mathematical Systems, vol. 669.

Day 10: Agent-based models: Specification

In the final day of the course, we will focus on the specificities of agent-based models vis-a-vis traditional economic models, in particular with respect to agents’ behaviour (expectations, bounded rationality, learning) and equilibrium properties. We will then describe through a simple example how agent-based models can be estimated, and point to the recent literature on the topic.

Suggested reading:

1. Delli Gatti, D., Gallegati, M., Fagiolo, G., Richiardi, M., & Russo, A. (2018). Agent-based Models in Economics: A Toolkit. Cambridge University Press.

2. Messier (2017). The Code is the Model. International Journal of Microsimulation, 10(3), pp. 184-201.

3. Richiardi (2013). The missing link: AB models and dynamic microsimulation. In: Leitner S, Wall, F (eds). Artificial Economics and Self Organization. Agent-Based Approaches to Economics and Social Systems. Springer, Lecture Notes in Economics and Mathematical Systems, vol. 669, Berlin.

4. Richiardi (2016). The Future of Agent-Based Modelling. Eastern Economic Journal, 43(2), pp. 271-287.

5. Richiardi (2017). The Code and the Model. International Journal of Microsimulation, 10(3), pp. 204-208. 

6. Tang (2017). The Code is the Model, Sometimes. International Journal of Microsimulation, 10(3), pp. 202-203.