Key findings
Benefits reduce UK income inequality more than direct taxes do. Before redistribution, the highest-income fifth of individuals, on average, has an income that is 12 times as large as the poorest fifth. Adding all cash benefits and deducting direct taxes (including income and council taxes as well as employer and employee National Insurance contributions) brings this figure down to 5. Benefits account for around 80% of this reduction, while direct taxes account for 20%.
This is because benefits (as a share of income) are more concentrated at the bottom of the income distribution than direct taxes are at the top. The poorest fifth of individuals receives 54% of net (after tax and benefit) household income in benefits, while the highest-income fifth receives only 3% (15.6 times less as a share of net income). By contrast, the highest-income fifth pays just 2.7 times as much direct tax as a share of income as the poorest fifth (though in cash terms they pay 13.7 times as much). The progressivity of direct taxes is limited by council tax. Net of council tax support (which reduces liabilities for low income families), it accounts for 5% of the income of the poorest fifth and only 2% for the richest fifth. Ignoring council tax (so counting just income tax and NICs) the richest fifth pay 4.1 times as much direct tax as a share of their income as the lowest income fifth.
Indirect taxes are broadly distributionally neutral. Overall, 15% of people’s expenditure is accounted for by indirect tax (around two-thirds of which is VAT), with little variation across the income distribution. This result contrasts with some analysis (including analysis by the ONS) suggesting that indirect taxes are regressive. The difference lies in the fact that, because they are levied on expenditure, we measure the impact of indirect taxes as a share of expenditure. Indirect tax does look regressive when measured as a share of income, but this is driven by households with high spending relative to their income at a point in time –households who are very likely to be on a low income only temporarily.
Overall, we find that direct and indirect taxes are progressive. This result is in contrast to the headline finding of the ONS, which concludes in its analysis of the effects of taxes and benefits on inequality that ‘overall, taxes had a negligible effect on income inequality’. We estimate that direct taxes have over 1.6 times as much effect on inequality as the ONS estimates, but the main difference between our analyses lies in the fact that we find indirect taxes to be distributionally neutral rather than regressive, because we measure their effect as a share of expenditure, not income.
Methodological choices have a significant effect on estimates. Beyond indirect tax, our results differ from the ONS’s results because of a number of differences in methodological choices. We take steps to address survey under-reporting of top incomes, include employer National Insurance contributions as a direct tax. We account for the fact that some benefits are taxable and others untaxed. We also measure inequality between individuals, rather than between households. These choices have significant impacts on the estimated redistributive effects of taxes and benefits.
