Microsimulation
Microsimulation is a type of computer-aided simulation . Microsimulation is often used to assess the effects of measures before they are implemented in the real world.
Most of the things and situations around us are made up of several individual components. If, in a simulation, what is to be examined is broken down into its components and the simulation uses models of these components and not a model of the whole that consists of them, for example individual gas atoms are simulated to simulate the movement of a gas cloud or individual vehicles are simulated to simulate the flow of traffic from a "microscopic simulation" and thus distinguishes such models from a macro simulation .
Microsimulation in economics and the social sciences
Here microsimulations are used at different points, whereby the same term can be very different types of simulations.
Heuristic microsimulation
The heuristic or strategic simulation is mainly used in management. Most of the time, these are macro simulations in the sense of cybernetic control loops, often in the vicinity of simulation games , but microsimulations can also be used. In traffic physics, heuristic microsimulations are used intensively; here the simulated system consists of individual road users. Characteristic of the heuristic simulation is the largely missing connection to data and the emphasis on interaction effects between micro-units. In sociology such micro-simulations have been used since the 1970s. Robert Axelrod's computer tournaments have become known in the field of game theory research , but methods of statistical physics and, today, especially computer science ( multi-agent systems (MAS) or multi-agent simulation ) are also used.
Economic microsimulation
In the field of empirical economic research, microsimulation models are used for a variety of applications, in particular for questions of reforms of the tax or transfer system, transport planning or health policy. The advantage of microsimulations, in contrast to general equilibrium models, is that, in the ideal case, the full heterogeneity of the unit under investigation in a population, e.g. B. in terms of income, housing situation or household composition is mapped. Tax reforms, for example, can have very different effects on individuals' incomes, depending on household composition, age or employment status. This level of detail cannot be achieved in macro models. On the other hand, macro models are more capable of mapping feedback in other sectors of the economy, for example in the form of wage or price changes. So-called micro-macro models, which allow feedback between the two levels, form an intermediate path.
Economic microsimulations are based on representative statistical surveys such as the microcensus , the SOEP or the EVS . On the one hand, a database must represent the population in a sufficiently representative manner and, on the other hand, contain sufficient information on the individual observation units. For this reason, despite the large number of cases, administrative data is only suitable for special applications. The income tax statistics, for example, only include households subject to income tax, for which little demographic information is given. From a scientific point of view, a link between administrative and survey data appears to be desirable. However, there are privacy concerns.
Economic microsimulation models can be differentiated with regard to their treatment of the time factor. Static models only have two states (before and after), while dynamic models represent several periods. Furthermore, a distinction is made between models with and without behavioral components. The latter assume that individuals have a behavioral calculation (usually in the form of maximizing utility) with regard to their work, savings or consumption behavior. On the basis of estimated parameters of a utility function, it can be analyzed how the behavior of individuals changes when behavioral determinants (e.g. income) change. With the help of case weights, these individual behavioral changes can be aggregated to the total population in a next step.
The users of economic microsimulations to estimate the effects of reform are in particular ministries at the state and federal level as well as the European Commission.
Methods
There are a number of methods for implementing microsimulation: Cellular automata have their origins in mathematics or computer science. The transition from cellular machines to discrete multi-agent simulations is sometimes fluid, the name sometimes simply depends on the subject. The finite element method is used intensively in engineering.
Further examples
FHP model , evacuation simulation , Nagel-Schreckenberg model , VISSIM , lattice theory , MikroSim (FOR 2559)
Individual evidence
- ↑ Figari, F., A. Paulus, H. Sutherland (2014). Microsimulation and Policy Analysis . In: Handbook of Income Distribution. Ed. by AB Atkinson and F. Bourguignon. Vol. 2. Elsevier-North Holland. Chap. 25th
- ^ Li, J., C. O'Donoghue, G. Dekkers (2014). Dynamic Models. In: Handbook of Microsimulation Modeling. Ed. by C. O'Donogue, pp. 305-343.
- ↑ MikroSim - DFG FOR 2559. Retrieved on May 8, 2019 (German).