Analysing tax-benefit reforms with nonparametric methods

Static tax-benefit microsimulation models (MSM) are widely used and reputed tools for public policy analysis but a more than careful use of them is compelling. This paper focuses on the analysis of MSM output, suggesting to use nonparametric kernel methods rather than most often used histograms for the analysis of income distributions, and nonparametric regressions in place of histograms by income deciles for identifying gainers and losers. Although slightly less intuitive, nonparametric methods are robust and can be easily tested. A robust and testable analysis of MSM output is crucial as MSM are often used also for tax-benefit policy design. A MSM is used for the analysis of 1998 Italian personal income tax reform. Although not a radical reform, nonparametric methods allowed for detection of the role of the tax system for the emergence of bimodality of disposable income distribution, dramatically different impact depending on some family types, differential impact of losses and gains at different levels of income, and a more robust understanding of the main components of the reform.