Regressivity-Reducing VAT Reforms

A concern about a more extensive use of the Value Added Tax (VAT) in national tax systems often arises both from its impact on aggregate consumption and its alleged regressivity over income. Yet, the empirical evidence on this latter issue is still narrow mainly due to the lack of joint data on income and expenditures with enough detail to account for commodity-specific tax rates. After discussing relevant problems in the measurement of VAT incidence over current income - which are likely to cause severe upward bias in the estimated regressivity - the paper aims at analysing the distributional implications of different VAT structures. In a framework of marginal tax reforms, relying on the concept of Gini elasticity (Yitzhaki, 1983), a general methodology is proposed to analyse and improve the distributional profile of VAT over income. Using a static microsimulationmodel(EGaLiTe),the methodology is applied on a comprehensive dataset of expenditures and incomes obtained by a statistical matching of two different sources representative of the Italian population. It is shown that an alternative allocation of goods among existing rates could mitigate the regressive profile of the tax over income, and that a properly designed two-rate setting could even improve the distributional outcome compared with the current setting. Finally, behavioural responses to tax-driven price changes are also simulated in order to assess the potential impact of the proposed reforms on aggregate expenditures.