Pattern of Direct Taxation in Affluent Democracies

In this paper, we investigate the political-economy of tax progressivity. We take as a starting point the following stylized fact: there is a negative correlation between the average tax rate and the concentration of taxation on wealthier households at the country level (Verbist and Figari 2014, Guillaud, Olckers and Zemmour 2017). We conjecture the existence of a common pattern of taxation in OECD countries: the average effective tax rate borne by households is a linear function of their rank in the income distribution, with the same coefficient for each country. Put differently, the increase in the tax rate from one quantile to another is constant along the income distribution, and of the same size among OECD countries. We interpret this as a political compromise addressing the demand for a minimum progressivity of the tax system in times of tax competition. We test this conjecture using LIS micro-data complemented by imputed data on social security contributions. Effective tax rates are recovered at the household level, including personal income tax, employer and employee social contributions. Our sample is made of 20 OECD countries, over the period 1999 to 2013. This conjecture has the following implications. Once we observe the tax rate of the median income and the pre-tax income distribution, we can predict with a good accuracy the general tax schedule of the country. This includes the average tax rate of any percentile of income in the population, and the overall progressivity of the tax system (concentration of taxes and marginal tax rate).