Three Essays On Income Tax Progressivity PDF Download

"Income redistribution is one of the primary concerns for policy makers and economists. Among the countries in the Organization for Economic Co-operation and Development (OECD), the degree of income redistribution (measured by the percentage decrease in the income Gini coefficients between the before and after taxes/transfers) ranges from 5% (Chile) to 49% (Ireland). Understanding and comparing redistribution policies across countries in a unified framework is not an easy task. However, recent developments in quantitative general equilibrium heterogeneous-agents models allow us to address several issues. In this dissertation I study three issues about the redistribution polices using a state-of-the-art quantitative general equilibrium model. Chapter 1 uncovers Pareto weights that justify the current progressivity of income taxes in 32 OECD countries. Chapter 2 shows that the current tax rate in the U.S. can be close to political equilibrium under an ex-ante differences in earnings ability and income-dependent voting behaviors. Chapter 3 finds and explains the negative relationship between economic outlook and income redistribution. In Chapter 1, we develop a model that reproduces income distribution and redistribution policies in 32 OECD countries. The individual income tax schedule is assumed to follow a log-linear tax function, which is widely used in the literature (Heathcote et al., 2016). According to our model, the optimal tax progressivity under the equal-weight utilitarian social welfare function varies from 0.21 (South Korea) to 0.41 (Ireland), and the corresponding optimal redistribution ranges between 20% (South Korea) and 37% (Ireland). For 22 countries, mostly European countries, the current progressivity is higher than optimal. In the other 10 countries, including the U.S., the optimal progressivity is higher than the current one. In our model the optimal tax progressivity is favored by the majority of the population in almost all OECD countries. Then, why does the current (suboptimal) tax rate prevail? The society's choice for redistribution may differ from the equal-weight utilitarian welfare function (Weinzierl, 2014; Heathcote and Tsujiyama, 2016), or can be affected by various factors such as the externality of public expenditure (Heathcote et al., 2016), and the preference heterogeneity (Lockwood and Weinzierl, 2015). In this chapter we ask a rather simple positive question within the utilitarian framework: what are the weights in the social welfare function that justify the current tax progressivity as optimal? We interpret these relative weights in the social welfare function as broadly representing each society's preferences for redistribution and political arrangement. According to our calculations, in Sweden, the average Pareto weight on the richest 20% of the population is only 0.53, whereas that on the poorest 20% is 1.74. By contrast, in Chile, the Pareto weight on the richest 20% is 2.65, whereas that on the poorest 20% is a mere 0.15. In the U.S. that on the richest 20% is 1.45 and that on the poorest 20% is 0.60. We also compare our social weights to those from Lockwood and Weinzierl (2016), who extend Mirrleesian (1971) framework to uncover weights. To our knowledge, this is the first study that compares how societies aggregate individual preferences over redistributive policies, and does so across a large set of countries. The utilitarian social welfare function often predicts that the optimal income tax rate in the U.S. is much higher than the current rate (e.g., Piketty and Saez, 2013). In Chapter 2, we focus on the interaction of ex-ante heterogeneity in household earnings and income-dependent turnout rates. While the relationship between each factor and income redistribution has been reported by many studies (Benabou and Ok, 2001; Charite et al., 2015, Mahler, 2008), quantitatively neither effect alone is large enough to explain the current tax rate. However, the interaction of the two magnifies the effect on redistribution, political equilibrium can be close to the current tax rate. More specifically, we construct three model economies: no ex-ante heterogeneity (NH), small ex-ante heterogeneity (SH), and large ex-ante heterogeneity (LH). All three economies match the overall income dispersion (Gini coefficient) in the data, but the share of ex-ante productivity (ability) and ex-post productivity (shocks) is different. According to our estimates following Guvenen (2009), 31% (SH) and 57% (LH) of wage dispersions are driven by ex-ante productivity. In the NH, by design, all wage dispersions are from ex-post productivity. For tractability, a flat tax rate and a lump-sum transfer are assumed in this chapter. The current tax rates in the three economies are set to 24% from the U.S. data. According to our model, the optimal tax rates under an equal-weight utilitarian social welfare criterion are similar in all three economies: 37% (NH), 38% (SH) and 37%. These high optimal tax rates are consistent with a majority of literature based on a utilitarian social welfare function (e.g., Piketty and Saez, 2013; Heathcote and Tsujiyama, 2016). The tax rates chosen by a simple majority rule are 37% (NH), 37%(SH), and 34% (LH), still much higher than the current rate. However, once we introduce increasing voter turnout rates with income, as in the data (Mahler, 2008), the political equilibrium vastly differs across the three economies. The tax rates chosen by effective voting are 35% (NH), 33% (SH), and 27% (LH). In LH, where income dispersion is driven mainly by ex-ante productivity, the insurance benefit from a heavy tax-and-transfer policy diminishes, and high-ability households are more against strong redistribution. If their turnout rates are higher, a relatively low tax rate can become a political equilibrium, which is close to the current tax rate. In Chapter 3, I find a new relationship between the economic outlook and redistribution among 33 OECD countries between 1996 and 2010, using the historical forecasts in the World Economic Outlook and the Standardized World Income Inequality Database. A one percentage point decrease in expected growth is associated with a 0.005 point and 0.9% increase in the income Gini before taxes and transfers. To examine this relationship I introduce labor-augmenting technology into my model at the cost of assuming a simple tax structure (linear tax and lump-sum transfer). The current tax rate (21.8%) and labor-augmenting productivity growth (3%) are chosen to match the U.S. economy before the Great Recession. Then, after an unanticipated productivity slowdown, the productivity growth decreases to 1%. Once productivity slows down, households save more to prepare for lost consumption in the future. As the capital-to-output ratio increases, the interest rate goes down from 4% to 1.7%. As seen in previous chapters, explaining the current tax rate is still disputed. Leaving this question to other studies, this chapter focuses on the effect of a productivity slowdown. More specifically, social weights that justify the current tax rates are derived, and, given these weights, the optimal tax rate under the low-growth regime is calculated. While all households save more against productivity slowdown, poor households, who are close to borrowing constraints, have more difficulty in increasing their savings. Hence, higher tax rates (23.6%) and more transfers can enhance social welfare under the low-growth regime. This relationship between expected growth and redistribution is similar to my empirical estimates. A general equilibrium effect from increased capital plays an important role. If interest rates are fixed, private savings are more effective against a productivity slowdown, since households can continue to save at the same rate. In this economy the optimal tax rate under the low-growth regime is much lower than the current rate."--Pages v-viii.