Governments Bear Much of the Responsibility for Rising Inequality—But Also Have the Tools to Reduce It

A new study tracks the influence of public policy on inequality and argues that the main rise in inequality has been a result of the decline in public sectors in the OECD countries.



Income inequality has been at the forefront of public debates in the US and around the world in recent years, fueling voter dissatisfaction in liberal democracies and contributing to the rise of far right populists. Inequality has become a particular area of interest for economists, as numerous studies show a dramatic rise in income and wealth inequalities since the 1980s.


Research in economics reacts both to the real world around us and to the new tools available to us. As global inequality increased significantly, new tools of research in neoclassical economics have been developed, mainly the use of models of heterogeneous agents. Economists have begun to analyze the effects of the distribution of income on other macroeconomic variables, like aggregate output and technical change. They have also begun to study what determines the distribution of income and why it differs so much both across countries and over time.


In our recent paper “Inequality and Public Policy,” we explore why inequality differs across countries and over time. The paper focuses on the role of public policies, namely the policies of the central governments, local governments, and the other public institutions that together form the public sector. We discuss both budgetary and non-budgetary policies. Since most of the effects of public policy on the distribution of income are straightforward and there is no need for fancy models to understand them, the paper is empirical. Our empirical work focuses on OECD countries, since data collection on income and policies in the most developed countries is of higher quality. We also examine a wider set of countries and find that our results are robust.


Focusing on the effects of policy on inequality does not mean that we ignore other determinants of inequality, such as the level of development, globalization, and technical change. Nevertheless, our research shows that the effect of public policies is not only statistically significant, but is also quite large. One way to see the effect is to examine the difference between the distribution of economic income, which is determined in the labor market, and disposable income after receiving subsidies and paying direct taxes, mainly income tax. The gap between the two types of income will be due to public policies.


We usually measure inequality by the Gini coefficient, which goes from 0 (full equality) to 1 (extreme inequality, where all income is concentrated in the hands of a single person). The mean of the Gini coefficient of market income in the OECD countries in 2011 was 0.47, while the mean of the Gini coefficient of disposable income in these countries in 2011 was 0.34. It is a large decline in inequality. Thus, direct taxes and subsidies, which are progressive, create a significant redistribution and lead to a large reduction in inequality. Furthermore, the variability of the Gini of market income in OECD countries in that year, measured by standard deviation, was 0.048, while the variability of the Gini of disposable income was 0.062. So, differences in redistribution by governments explain a big part of the differences in inequality across countries.


The distinction between market income and disposable income plays an important role in our paper. In the first part, we analyze how fiscal policy affects the redistribution from market to disposable income by treating inequality of market income as given. In the second part, we examine how public policies affect inequality in market income through different channels and mechanisms.


The paper begins the empirical analysis by showing that the redistribution of income through direct taxes and transfers is significant and large. It does so by using a regression analysis in which the explanatory variables are market income inequality, direct taxes as percent of GDP, and subsidies as percent of GDP, while the explained variable is inequality of disposable income. This analysis shows that both taxes and subsidies have a significant and strong effect.


Direct taxes and welfare transfers are actually both the results of public intervention in the economy. As government intervenes more and supplies more public services, it has higher expenditures and needs to raise taxes, including direct taxes. The degree of public intervention also affects subsidies, since welfare benefits are an important channel of intervention in the economy.


In light of these results, we raise a conjecture that inequality of disposable income should decrease if the government increases its intervention in the economy. To test this conjecture, the paper measures public intervention by public expenditures as percent of GDP.


Indeed, the analysis shows that the share of public expenditures in GDP has a negative, significant, and large effect on inequality. For a given level of inequality of market income, a rise of public expenditures by an additional percent of GDP reduces the Gini coefficient of disposable income by 0.35 percentage points. Since the size of public expenditures across the OECD countries varies from 35 to 55 percent of GDP, this variable can therefore explain variability of the Gini coefficient of 7 percentage points across these countries. The Gini of disposable income in the OECD countries varies from 0.25 to 0.4. Hence, changes in public intervention can explain half or even more of these differences.


To cope with the problem of potential reverse causality, we use the tool of “Instrumental Variables”: variables that are not supposed to affect inequality directly, but affect public expenditures. If we observe that they are correlated with inequality, it means that there is a direct effect to public expenditures on inequality. One can claim that the regression analysis described above does not measure well the effect of public policy on inequality, since causality can go both ways—public policy not only can affect inequality, but can be affected by it as well. Indeed, some economists claimed in the 1980s that one of the main reasons for the rise of government during the 20th century was high inequality.


To rule out the possibility of reverse causality, we use two instrumental variables: the size of public debt (which affects expenditures through interest payments) and the share in population of homeless due to natural disasters. These two variables do not have any direct link to inequality, but they do affect public expenditures. Our regressions show that these two variables have a significant effect on inequality of disposable income. We conclude that this effect must work through the mechanism of public expenditures. We also shows that the effect implied by these instrumental variables is very close to the effect of public expenditures when measured directly, which rules out the possibility of a bias.


In the second stage of our analysis, we turn to examine the effects of public policies on the distribution of market income. There are very few studies on this topic, and this analysis is quite preliminary. The paper identifies four policy tools that can affect this distribution in theory. The first is public expenditures on education, which has a significant effect on the supply of human capital and through it on the distribution of market income. The second policy is public investment in infrastructure and in public transportation, which reduces costs of mobility from the center to the periphery and thus reduces inequality of market incomes. The third public policy that affects market incomes is hiring by the public sector, which is a large employer. The fourth public policy is regulation of the labor market by the government, which also affects market incomes.


In the paper, we note that the effect of public education on inequality might not be monotonic. Consider an initial situation where most of the population has secondary education. In that scenario, inequality is low. Once people begin to acquire higher education, inequality rises. Nevertheless, at some point inequality should start to decline because when everyone acquires higher education inequality will be low again. We conjecture that the effect of public education on inequality of market income is not monotonic.


We then test how the four interventions by the public sector affect the distribution of market income across countries and over time. The tests use panel regressions, which also control for other variables that reflect common explanations for inequality, like economic development, globalization, and technical change. These regression tests find that the four public policies affect inequality as expected, although the effect is not always significant. First, the effect of public education is indeed not monotonic. Increasing public education up to 6 percent of GDP reduces inequality, while beyond 6 percent it increases inequality. Since most OECD countries spend less than 6 percent of GDP on education, public education tends to reduce inequality.


The effect of the other public policies on inequality of market income is as expected. Public employment increases inequality, and our explanation is that the share of high skilled jobs in the public sector is larger than in the private sector. Investment in infrastructure reduces inequality, as anticipated. The effect of labor market regulation on inequality is not significant in the OECD countries, but it reduces inequality significantly in a larger set of countries beyond the OECD.


Finally, we use the results of our analysis to estimate which variables have contributed most to the recent rise in inequality in the OECD countries. We find that the main rise in inequality has been a result of the decline in public sectors in the OECD countries. We focus our analysis on 14 countries in the OECD that have full data for the entire period. Public expenditures as a share of GDP in these countries decreased significantly from 52.7 percent of GDP in 1996 to 46.8 percent of GDP by the year 2015. This significant decline should account for a rise in the Gini of disposable income of 2 percentage points, which is the actual rise over these years.


Our paper therefore demonstrates, by using data mainly from the OECD countries but also from additional countries, that governments bear a significant responsibility for their countries’ levels of inequality. They can affect it mainly through more or less redistribution, by changing the size of public expenditures. Countries that have higher expenditures, which lead to higher taxes, have lower inequality. We can therefore interpret our study as presenting a trade-off to governments: higher expenditures and lower inequality, or lower taxes and higher inequality. The choice is theirs.


Michele Battisti is an Associate Professor at the University of Palermo. Joseph Zeira is a Professor of Economics at the Hebrew University of Jerusalem and the current president of the Israeli Economic Association.


Disclaimer: The ProMarket blog is dedicated to discussing how competition tends to be subverted by special interests. The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty. For more information, please visit ProMarket Blog Policy