LABSim
The LABSim model, developed for the project “Investigating economic insecurity through microsimulation” funded by the Institute for the Analysis of Public Policies (INAPP), an Italian government agency. The aim of the project is to develop a rich dynamic microsimulation model of individual life course events to facilitate research on economic insecurity at the individual and household level and investigate its determinants. The key innovation is the linkage of the dynamic microsimulation with a static tax-benefit model, EUROMOD, which allows ex-post evaluation of the policies put in place in the aftermath of the Great Recession in terms of their impact on economic insecurity, as well as ex-ante evaluation of hypothetical policies aimed at reducing economic insecurity.
By Patryk Bronka and Matteo Richiardi (2020). Estimates by Cara Booker and Zhechun He (2020). Earlier contributions by Ross Richardson (2018).
The model receives as input a representative sample of the population for each simulated country, called the initial population. The initial population is evolved through time according to a number of processes shown in the diagram above and described in more detail in the paragraphs below. These processes can be estimated on the same, or different, datasets as the initial population, but require longitudinal data. To take into account the effects of tax and benefit policies, gross incomes are transformed to net incomes by statistically matching simulated households with best fitting households in the EUROMOD output data, for whom EUROMOD calculates the gross-to-net income ratio for a given year, tax-benefit regime (more information here).
The microsimulation is composed of six different modules: (i) Demography, (ii) Education, (iii) Health, (iv) Household composition, (v) Non-labour income, and (vi) Labour supply, as shown by circles in the diagram above. Each module is in turn composed of different processes or sub-modules, for example ageing process in the demographic module, or a wage setting process in the labour supply module. In each period, agents first go through the ageing process, followed by the population alignment process, which adjusts the population structure to official projections by gender, region, and age. Then, the education module determines if individuals should remain or re-enter education. Students are not at risk of work until the next period and do not enter the labour supply module. Individuals who leave continuous education for the first time have their level of education determined and can become employed. The health module calculates an individual continuous health score, and evaluates whether the individual is long-term sick or disabled (in which case, he / she is not at risk of work). Next, in the household composition module, adult children who still live with their parents decide whether to leave the parental home, couples are formed using either a parametric matching procedure based on potential wage and age differentials between prospective partners, or on a bi-proportional iterative scaling procedure to reproduce the distribution of matches observed in the data, given individual characteristics. The fertility process determines whether females in couples give birth to a child. Fertility is modelled at the individual level, with an option of alignment to the fertility rate implied by the population projections. Finally, individuals enter the labour supply module, in which a) their potential wage is calculated using a Heckman-corrected wage equation, b) the closest matching EUROMOD household (in terms of a number of keys relevant for the tax and benefit policies, such as health, number of children, region, and age) is selected and the net household income for any given gross (market) income is calculated, c) the utility-maximising choice of number of hours of work supplied by the members of the household is determined using a structural random utility labour supply model (RUM), whose parameters are estimated on the EUROMOD input data. That determines the household’s actual disposable income.
The source code for the model is open source and publicly available from GitHub – click here. Please note that input data required to run the model is not freely available and requires permission from the data owner – please contact us for further details.