Social Science Computer Review
Volume 29, No. 1
Special Issue on Microsimulation
Ronald E. Anderson and Chantal Hicks, Editors
Highlights of contemporary microsimulation / Ronald E. Anderson & Chantal Hicks
Abstract: Microsimulation is a family of computer simulations that all depend upon individual level data as input to one or more processes which are aggregated into macro-level outcomes. A popular application of microsimulation is automobile traffic systems where the individual records are vehicles. Persons constitute the individuals of most policy-oriented microsimulation models built by economists, sociologists, geographers, demographers, and other policy analysts. While the field is over 50 years old, in the past two decades the volume of policy microsimulation models has skyrocketed. The papers in this Special Issue reflect the exciting developments and findings of these models.
This paper introduces microsimulation by presenting its main underlying ideas as well as its main strengths and drawbacks. Microsimulation is currently experiencing a boom which is driven by three main forces. The first is the increased demand of policymakers for detailed projections and models able to assess distributional and long-term sustainability issues of social security systems. The second is the emergence of new research paradigms with an increased emphasis on individuals within their context, studied from a longitudinal, multi-level perspective. The third concerns technological advances, providing not only the necessary computer power but also the programming tools for model development, accessible to scientists without specialized programming skills. While static microsimulation models are established tools for policy analysis, dynamic microsimulation has yet to find its way into the methodological toolbox of mainstream social scientists—but the prospects are promising.
Primary care in an aging society: Building and testing a microsimulation model for policy purposes / Janet Pearson, Roy Lay-Yee, Peter Davis, David O’Sullivan, Martin von Randow, Ngaire Kerse, & Sanat Pradhan
Abstract: We describe the development of a microsimulation model of primary medical care in New Zealand for 2002 and demonstrate its ability to test the impact of demographic ageing, community support and practitioner repertoire. Micro-level data were drawn from four sources: two iterations of the New Zealand Health Survey (NZHS 1996/7 and 2002/3); a national survey of ambulatory care in New Zealand (NPMCS 2001/2); and the Australian National Health Survey (ANHS 1995). Data from the New Zealand surveys were statistically matched to create a representative synthetic base-file of over 13,000 individuals. Probabilities of health experiences and general practitioner (GP) use derived from the ANHS, and of GP activity derived from the NPMCS were applied via a Monte Carlo process to create health histories for the individuals in the base-file. Final health care outcomes simulated - the number of visits in a year, the distribution of health conditions, and GP activity levels – were validated against external benchmarks. Policy-relevant scenarios were demonstrated by a forward projection to 2021, and by implementing counterfactuals on key attributes of the synthetic population. The results showed little change in model-predicted health care outcomes. There is potential for this approach to address policy purposes.
Microsimulations for poverty and inequality in Mexico using parameters from a CGE model / Araceli Ortega Díaz
Abstract: In the current paper we use microsimulations to analyze poverty and inequality through the accumulative impact on wages of sequential changes in the labor market unemployment rate (U), employment structure (S), wages by economic sector (W1), average labor income (W2), and employment structure by skill level (M). These effects are based on the parameters obtained from a Computable General Equilibrium Model, with parameters allowing for simulations from 2003 to 2015. The results show that a yearly increment in the real wages will reduce extreme poverty by half. The changes in the employment structure and wages by economic sector reduce poverty in this model because workers move from the agricultural sector to the service sector, implicitly allowing workers to move from sectors with lower wages to higher wages sectors. Nevertheless, an increase in wage inequality is related to changes in the employment structure by skill level as more skilled workers obtain better wages than unskilled employees.
Household projection and its application to health/long-term care expenditures in Japan using INAHSIM-II / Tetsuo Fukawa
Abstract: By using a micro-simulation model named INAHSIM, we conducted a household projection in Japan for the period of 2010-2050. INAHSIM-II specifically means that the initial population is created by using the INAHSIM model itself. The model produces such outputs as a) the number of the elderly according to dependency and/or living situations, b) the relative parents/children ratio taking into account the number of brothers and sisters, and c) a one-year transition matrix by household type. As an application of the model, a projection of health and long-term care expenditures is made for the years 2010-2050 in Japan.
Mandatory superannuation and self-sufficiency in retirement: An application of the Appsim Dynamic Microsimulation Model / Marcia Keegan
Abstract: One of the most significant concerns about the ageing population in Australia is the impact on pension costs. Mandatory occupational superannuation was introduced in 1993 to reduce future pension costs: a person with moderate to high levels of superannuation can provide for themselves to some extent and thus has a reduced pension entitlement.
The ability of mandatory superannuation to reduce future pension costs is best modelled by dynamic microsimulation, as this takes into account the effect of numerous factors such as disability, child-rearing, employment history and life expectancy on superannuation levels.
Abstract: We suggest a new comprehensive measure of support given through tax-benefit systems to families with children. Using microsimulation techniques, this accounts for all provisions contingent on the presence of children, while usually only gross child/family benefits are considered. We use EUROMOD, the European Union tax-benefit microsimulation model, to quantify the support for children and analyse its impact on household incomes and child poverty for 19 countries. We find that the conventional approach underestimates on average the total amount of support for children by about one fifth. Furthermore, the differences between two measures vary considerably across countries and are, therefore, critical for cross-national comparisons.
Microsimulation of the temporary housing situation following an urban disaster—Case study of an anticipated Tokyo metropolitan earthquake / Keiichi Sato
Abstract: At present, the expected occurrence of a massive earthquake in the Tokyo metropolitan area is a matter of concern for the Japanese Cabinet Office. In the event of a major urban disaster, it is unclear whether the present policy of housing assistance will work effectively. Microsimulation is a useful research approach to make preliminary proposals for concrete policy issues. To carry out a reliable microsimulation, a household’s behavior model was developed using an Internetbased questionnaire system. In addition, the microdata pertaining to rental housing and construction sites of prefabricated housing were generated. Based on these foundational works, a microsimulation model that illustrates the temporary housing situation following an urban disaster was developed. This paper discusses (1) the microsimulation outline, (2) the results of calculations, and (3) issues concerning and ways to enhance the developed model to resolve these issues.
Modeling the impact of taxes and transfers on child poverty in South Africa / Kate Wilkinson
Abstract: The research described here shows how microsimulation modelling can be used to analyse the impact of tax and transfer policy on the situation of children in South Africa. Given that the government has made explicit legislative commitments to children and that poverty rates are highest for households containing children it might be expected that children would be a priority in policy reforms. The microsimulation model SAMOD is used to simulate certain policy reforms announced in the 2008 budget, and implemented during 2008, and explore the impact of these reforms on children and other age groups across the income distribution. A budget-neutral policy reform is also considered to determine the extent to which policy could do more to support children without adversely affecting other poor groups.
A dynamic MSM with agent elements for spatial demographic forecasting / Wu, Birkin, & Rees
Abstract : Individual based models such as Microsimulation models (MSM) provide an alternative to macroscopic models in social simulation and modeling. In contrast to the traditional models where individual characteristics are often blurred or even disregarded, MSM provides the realistic disaggregated information that is often vital for modern policy problems. MSM has been extensively applied and well tested in social modeling. However, it has been criticized for being less strong in modeling interactions between individuals and individual behaviors. MSM also struggles where realistic micro-data are not available. Agent Based Models (ABM) can model the demographic process through interactions between the agents and with the environment that they live in. The combination of MSM and ABM provides a new approach to enhance complex social modeling. In our study, we attempt to provide better groundwork to facilitate policy and decision making for the UK population through a hybrid model that combines the strength of the two complementary techniques.
Factors associated with email and Internet use for health information and communications among Australians and New Zealanders / Robin Gauld
Research into the use of email and the Internet for health information and communications has predominantly hailed from North America and European Union countries. By drawing on an Australian and New Zealand survey sample (n=406) of health Internet users, this article expands the field. It explores factors associated with a series of email and Internet use behaviors including use of email to communicate with doctors, presenting Internet information to doctors, perceptions of Internet information reliability, and checking of website credentials. Regression results reveal a digital divide within the Internet using population; that those who email doctors believe this improves communications; that more frequent Internet searchers had higher odds of finding information not previously provided by their doctor, and of taking Internet information to their doctor; that those expressing difficulty communicating with doctors had higher odds of believing Internet information to be more useful; and that older respondents were more likely to perceive Internet information to be reliable, yet less likely to check website provider credentials. Implications of these findings are discussed.