Tax-Benefit Patterns: Impact on Inequality Reduction
The present paper focuses on the economic effects of tax and benefit systems, with the aim of assessing which channel empirically achieves inequality reduction. Redistribution policies differ in their targeting (structure of social benefits), the more or less progressive taxation (structure of taxes and social security contributions) and overall generosity (level of social spending). But the observation of national institutional configurations shows that no country combines all three factors that promote greater vertical redistribution (tax progressivity, high level of spending, and high concentration of social benefits).
Based on the finest available empirical data (Luxembourg Income Study), the paper analyses how the tax structure, shape and volume of social benefits help to reduce inequalities. We are thus able to measure the fraction of inequality reduction brought about by taxes and social security contributions, and the fraction of inequality reduction owed to each type of transfers. The analysis is conducted at the country-year level for 21 OECD countries over two decades, in a comparative perspective.
Our contribution is twofold. First, the paper mobilizes recent micro data concerning not only the level but also the composition of household income for more than 20 countries in the recent period. Second, the project holds together the two channels of inequality reduction (taxes and transfers), while most studies focus on one aspect or another.
The expected outcomes are of two types: 1) Establish a dynamic typology of the tax and benefit structure by country, for identifying not only the type of redistribution profiles, but also changes within countries and country movement from one profile to another. 2) Assess empirically the model of reducing inequalities, for each country and at the household level: contribution of taxation in reducing inequalities; contribution of social transfers; performance of each national system in terms of income inequality reduction.
Context and state of the art
From an analytical viewpoint, several mechanisms might affect vertical redistribution (for a given tax rate and primary income distribution):
- The structure of mandatory contributions: other things being equal (for a given average tax rate) the more progressive the tax structure, the more it contributes to mitigating the initial level of inequality.
- The structure of social benefits: other things being equal (for a given amount of social spending), the more benefits are targeted to the poor, means-tested, and decreasing by income, the more they contribute to reducing inequalities.
- The amount of social spending: other things being equal (especially the structure of tax and benefit), the higher the budget for social protection, the greater the reduction of inequalities.
The observation of national institutional configurations shows that no country combines all possible policy tools that promote greater vertical redistribution (tax progressivity, high level of spending and high concentration of social benefits). The explanations of this phenomenon mainly stem from political sustainability and might be linked to the political construction of redistribution demand (Guillaud, 2013).
Prasad and Deng (2009) show that countries whose compulsory levy structure is the most progressive (USA and Australia) are also countries whose social spending is less developed. Moene and Wallerstein (2001), Korpi and Palme (1998), Brady and Bostic (2015), McKnight (2015) show that countries in which benefits are most concentrated are less effective in reducing inequality than countries where benefits are more universal -but more generous ("the paradox of redistribution"). Rehm (2011) as well as Françon and Zemmour (2013) show social benefits are more generous when the risk of unemployment is more homogeneous in the population.
While most studies focus on one aspect or another, our project holds together the two channels of inequality reduction: taxes and transfers. To do this, we use recent micro data that holds the composition of household income: the LIS (Luxembourg Income Study) database that provides detailed income data based on national household surveys and tax records.
Data and methodology
The LIS database identifies for a large sample of households their primary income (labour income, capital income), income after mandatory contributions, and income transfers (decomposed by social protection function –unemployment/pension/etc.). Therefore, LIS data does not provide direct information on institutions (number of threshold, official rate of taxation, social legislation), but on their empirical effects: it helps recovering the effective distribution of mandatory contributions and social security benefits on a representative sample of the population.
Our analysis is done at the country-year level. Twenty-one countries are included in the study with at least one survey wave from 2010 or later: Australia, Brazil, Canada, Colombia, Czech Republic, Denmark, Estonia, Finland, France, Germany, Ireland, Israel, Italy, Panama, Poland, Netherlands, Norway, Korea, Taiwan, USA, UK. Other countries (Sweden, Belgium, etc.) might be included but on older data (mid-2000s). Data are available over one or two decades, according to the country.
The paper aims to assess how the tax structure, shape and volume of social benefits contribute to reducing inequalities. We thus decompose our analysis in two steps.
First, to get a picture of national tax and benefit systems, we compute three indicators. (i) The degree of progressivity of the tax-benefit system (Kakwani  index). It is used to capture the marginal variation of tax/benefit according to income, and neutralizes the effect of the distribution of primary income. (ii) The degree of concentration of social benefits. It shows the concentration of transfers due to both the primary income distribution and the progressivity of the tax-benefit structure. (iii) The level of social spending (share of social spending in GDP). This helps us to provide a descriptive typology of national systems.
Second, to measure the contribution of taxes and benefits to inequality reduction, we generate inequality indicators (Gini coefficient and inter-decile ratios) for different analytical stages, located between primary income (before taxes and transfers) and disposable income. From this calculation, we can measure the fraction of reducing inequalities brought about by compulsory taxes, and the one brought about by each type of transfers. This can be done for each of the 21 countries with yearly intra-country variations, allowing cross-country as well as dynamic comparisons.