Investigating the Incidence of Value Added Tax on Households Income: New Evidence from Italy

Abstract

This paper analyses the incidence of the Value Added Tax (VAT) on Italian households income. To address this question, we developed a non- behavioural microsimulation model, the Value Added Tax Simulation Model (VATSIM- DF II). The goals of VATSIM- DF (II) are to estimate actual and expected VAT revenues, assess the VAT incidence on households disposable income, and simulate the distributional effects of changes in fiscal policies in Italy. Compared to existing models, the main achievements of this study are: (i) the creation of a unique original dataset, which includes tax register data; (ii) the implementation of a matching procedure, based on Tax Register data, which outperforms other data fusion strategies used in the existing literature. These innovations allow us to create a reliable and unique dataset to simulate changes in VAT and to produce results consistent with the most updated macroeconomic data. We tested our model, at current VAT legislation, and we show the VAT burden on Italian households confirming the regressivity of VAT. Finally, we analyse the effect of a revenue neutral reform, with two VAT rates, which applies the reduced VAT rate also to female and babies sanitary products. (ii) the adjustment of survey data on income and consumption to National Accounts data at the year of the simulation, which makes our estimations on VAT extremely coherent with the most updated macroeconomic data. To the best of our knowledge, there are not existing models in Italy with these features and this makes VATSIM- DF (II) a promising tool for reliable microsimulations about the distributional effects of VAT. We show that the procedure we used to create the dataset VATIC 2019, outperforms other methods used in the literature about microsimulation models in Italy. Also, we show an application of our model by testing a revenue neutral reform with two VAT rates. The reform consists only of two rates: a reduced VAT rate of 7% and an ordinary VAT rate of 20%. In addition to other basic goods, the reduced VAT rate is also applied to female sanitary towels and babies nappies. We show that this reform does not strongly affect the VAT burden on households. Conversely, it benefits low- income households reducing the VAT burden for basic goods and services (e.g. food and non-alcoholic beverages; clothing and footwear; and housing, water, gas, electricity and other fuels). However, the VAT burden on households visibly increases for restaurants and hotels, in particular for low- income households. A further extension of the VATSIM- DF (II) model, which will be described in a forthcoming paper, focuses on the behavioural effects produced by tax- shifts and changes in fiscal policies.