Sensitivity Analysis of a Tax Evasion Model Applying Automated Design of Experiments
- Cite this paper as:
- Szabó A., Gulyás L., Tóth I.J. (2009) Sensitivity Analysis of a Tax Evasion Model Applying Automated Design of Experiments. In: Lopes L.S., Lau N., Mariano P., Rocha L.M. (eds) Progress in Artificial Intelligence. EPIA 2009. Lecture Notes in Computer Science, vol 5816. Springer, Berlin, Heidelberg
Risk, audit frequency, expected utility, decision on the rate of tax evasion: probably these words occur to the reader first about tax evasion modeling. However, it can easily turn out in the real world that the ’everyday evader’ hasn’t got reliable information about the risks of evasion, about the possible amount of fine, or about the efficiency of the tax authorities. The TAXSIM agent-based tax evasion model was developed to understand the macro-level taxpayer behavior better with its help. The model and first simulation results were presented on the ESSA 2008 conference. The aim of this article is to present a sensitivity analysis of the model. We applied Design of Experiments method to reveal the main parameter-response correlations on a selected parameter domain and used two extreme parameter sets to examine on what level the contradictory factors can compensate each other.
KeywordsAgent-based computational economics agent-based simulation tax modeling tax evasion
Unable to display preview. Download preview PDF.