Top Income Shares and Budget Deficits Santo Milasi∗ University of Rome “Tor Vergata” October, 2012

Abstract

During the last thirty years most OECD countries have accumulated large public debts. The same period has been characterized by a considerable increase in the concentration of income at the top of the distribution and by substantial cuts to taxation imposed on high incomes. The paper argues that the concentration of income at the top of the distribution, along with a decreasing taxation imposed on high incomes, may have affected OECD countries' fiscal performances in recent decades. Using a panel of 17 OECD countries between 1975 and 2005, the paper presents the first reported evidence of a positive relationship between the top 1 percent income share and budget deficits. Moreover, the disaggregated analysis of the budget components suggests that such result is due to a negative relationship between the concentration of income at the top and budget revenues. JEL Classifications:[ D31; E62; H20; H62] Keywords: Budget Deficits, Budget Revenues, Income Inequality, Top Income Shares, Top Marginal Tax Rate

∗ Ph.D student in Economic Theory and Institutions, [email protected], Faculty of Economics – Department of Economics, Law and Institutions. I thank my Ph.D. Supervisor Prof. Robert J. Waldmann. His guidance, comments and feedback helped me to greatly improve this research.

1. Introduction Since the early 1970s many OECD countries have accumulated large public debt. However, in some countries, the increase in the stock of public debt has been modest, while in others, the debt to GDP ratios have reached levels even greater than 100 percent. What are the reasons for these differences? Historically, explanations for the dynamics of public debt have been connected to the prediction of the tax-smoothing theory advanced by Barro (1979). According to this theory, in order to minimize distortions, governments should keep tax rates relatively constant over time, allowing deficits to accumulate during recessions, when tax revenues fall, and compensate with surpluses during expansions, when tax revenues recover. However, well this theory explains observed budget deficits before, and immediately after, the 1973 oil shock, while it is difficult to reconcile it with the observed fiscal performances across OECD countries in the last thirty years. Indeed, despite the similarities in the economic fundamentals and the similar economic shocks experienced, the magnitude and the persistence of budget deficits greatly varied across OECD countries. Since it seemed hard to explain such large cross country differences using traditional economic arguments alone, during the late '80s and the '90s many theoretical and empirical studies, mainly related to politico-institutional factors, have tried to give a positive explanation to the observed deviations from the tax-smoothing theory. Nonetheless, why countries tend to frequently run budget deficits and why they behave differently from each other, are still open questions. Public debt is not the only dimension along which developed countries have differed in the last thirty years. Interestingly, the last three decades have also been characterized by two additional dynamics: the increase in the concentration of income at the top of the distribution; and the reduction of tax progressivity. After an historical decline, the share of income held by the top income groups increased constantly since the early 1970s in many developed countries. However, while in some countries like France, Germany, Denmark the increase in the concentration of income at the top has been quite modest; in other countries, especially the Anglophone ones, the income held by the top income group almost tripled with respect to its level in the 1970. At the same time,

many OECD countries have considerably cut taxation of high incomes. Indeed, while in the 1970s top tax rates varied from 40%, for Spain and Switzerland to around 75% for Sweden, the United Kingdom, or the United States; currently there is a number of countries clustering around 40% with no country whose top tax rates are above 60%. This work argues that the concentration of income may help to explain international differences in fiscal performances observed in the last thirty years. Indeed, to the extent that top income shares represent a large part of the tax base, contributing to about a third of total tax revenues in many countries, neither the importance of their tax capacity nor the reduced progressivity of the tax system can be ignored in budgetary terms. The empirical analysis developed in this work find evidence suggesting that the concentration of income at the top may be important in explaining OECD countries' fiscal performances over the last thirty years. In particular, in a panel of 17 countries over the period 1975-2005 this work provides a preliminary evidence of a positive relationship between budget deficits and the top 1 percent income share. Results suggest that the concentration of income at the top exerts its detrimental effect on public balances mainly via the revenue side. The remainder of this paper is organized as follows. Section 2 presents briefly the empirical literature dealing with the relationship between budget deficit and income inequality. Section 3 introduces a conceptual framework to explain why the concentration of income at the top coupled with low top tax rates may be important in budgetary terms. Section 4 introduces the empirical model and the data used. Section 5 presents the results, paying specific attention to the effect that the concentration of income at the top have on expenditures and revenues separately. Section 6 proposes some robustness checks. Section 7 concludes.

2. Literature review Even if there are some recent papers studying the effects that the income share of the super-rich may have on other outcomes, none of them focus on its effect on budget deficits. Moreover, these studies are rarely interested in the concentration of income on its own, but they mainly use top

income shares as a proxy for inequality across the distribution in studying some "inequality and..." relationship1. On the other hand, there exists a huge literature on budget deficit determinants. Most of this literature focuses on politico-institutional factors as crucial factors for understanding budget deficits2: political variables, such as the ideological characteristics or the degree of fragmentation of the government, and institutional factors, like certain aspects of the process of budget determination, have been the most common explanations for the observed OECD countries' fiscal performances. Despite the existence of such extensive literature, the relationship between the concentration of income at the top of the distribution and the budget deficit has not been investigated so far. Thus, a strictly related literature on this topic does not exist. Nonetheless, there are a number of theoretical and empirical contributions which study the relationship between income inequality and budget deficit. Considering that top income shares may also proxy for income inequality across the distribution; some of arguments relating income inequality to budget deficits are common to ones relating budget deficits to the concentration of income at the top. Hence, this work initially draws on the "inequality and deficit" literature as a starting point to study the relationship between the concentration of income at the top and budget deficits. The first study relating income inequality to public debt is Berg and Sachs (1988). In a cross-section of low and middle income countries, they find that higher income inequality is significantly associated with a higher probability of debt rescheduling. Their main hypotheses is that a high degree of income inequality delays debt rescheduling because it undermines the political stability and the political effectiveness needed for successful macroeconomic management. Berg and Sachs (1988) argue that high inequality in income leads to the polarization of the society into conflicting groups, such polarization may translate into coups, fruitless alternation of power between politicians or into "bribing activity" to hold the political power. All these sources of political instability reduce timely fiscal adjustments. 1

For example, Leigh and Posso (2007), analyzing the relationship between inequality and savings, find no consistent evidence that lagged top income shares have any significant impact on future savings rates. Andrews, Jencks and Leigh (2011), using top income shares, find no systematic relationship between inequality and growth over the period 1920-1999. 2 See, for example, Alesina and Perotti (1995) for a literature survey.

Woo (2003a) develop a theoretical model which relates income inequality, social polarization and fiscal performances along the industrialization process. The model predicts that high initial income inequality together with a large sectoral income gap during industrialization, lead to the polarization of sector preferences for different types of government spending. Such polarization in the society affects policymakers' preferences which polarize supporting high public spending for their preferred sector, so leading the government to run large deficits. Woo (2003a) (2003b) provide evidence in support of these theoretical predictions. The author shows that, in a panel of respectively 90 and 57 countries over two decades (1970-1979,1980-1990), the countries which experienced the greatest fiscal instability are those with highly polarized economic societies, measured by indicators of income inequality, like the Gini coefficient. However, Woo finds that the strength of the influence of social polarization on fiscal performances depends on how public institutions work. When institutions are able to contrast possible conflicts of interest over government spending, the detrimental effect of income inequality and, thus of social polarization, on fiscal outcomes might be less important. In particular Woo (2003a) underline how stronger budget institutions and stringent budget rules seem to be the most effective way to mediate the detrimental effect that high income inequality may have fiscal discipline. Larch (2012), in a sample of 30 middle-income and advanced countries over the period 1960-2008, finds very weak evidence of a direct effect of income inequality on budget balances. However, Larch suggests that, inequality does not affect budget balance directly; it might influence a country's fiscal performance by interacting with other political or economic variables. In particular, Larch (2012) finds that, although right-wing government is positively associated with budget surplus, the interaction between income inequality and right-wing government has a negative effect on budget surplus. This would suggest that, since high levels of income inequality increase the pressure for public spending, the tendency of right wing government to be fiscally disciplined is weakened when income inequality is high. Moreover, Larch (2012) finds that the beneficial effect of economic growth on budget balance is undermined by a high degree of income inequality. According to the

author this result come from the fact that high income inequality creates pressure on the government to spend extra public revenues deriving from economic growth for redistribution, instead of pursuing balanced budgets. Azzimonti et al (2012), based on the common trend shown by public debt, financial market liberalization and income inequality after the 1975, propose a theoretical model which relates the dynamics of these three variables. In particular, Azzimonti et al (2012) argue that government debt increases when both capital markets are liberalized and income inequality increases. The model predicts that high income inequality, represented by the entrepreneurs-workers dichotomy, increases the demand for government borrowing. Indeed, according to Azzimonti et al (2012) issuing public bonds would be beneficial both for entrepreneurs, since public bonds provide safe assets available for consumption smoothing, and for workers, since the interest rate on public debt is low, and through the government debt, they can borrow cheaply. According to their model, the relationship between inequality and public debt holds independently of the international regime of capital markets. Interestingly, they argue that, when countries are highly financially integrated, the increase in income inequality in some countries may cause the increase in government borrowing in all countries. In addition, across a sample of 22 OECD countries from 1973 to 2005, Azzimonti et al (2012) find empirical evidence in support of their theoretical predictions that inequality, financial integration and rising government debt move together3.

3. Why do Top Income Shares matter for Budget Deficits? The traditional argument in the "Inequality and Deficits" literature surveyed in the previous section, suggests that high levels of income inequality, raising the pressure for redistribution lead to an increase in public spending that, if unmatched with higher revenues, is translated into budget deficits. Since some recent studies suggest that top income shares may also proxy for broader 3

Azzimonti et al (2012) differs from the other papers described above in many aspects. First, they study the relationship between the change in inequality and the change in public debt, while all the other studies on this topic deal with levels of income inequality and budget deficit. Moreover, they use the yearly change in public debt instead of the more common budget deficit as dependent variable.

measures of income inequality across the distribution4, the arguments used to explain the relationship between income inequality and budget deficits could also be applied to explain the relationship between the concentration of income at the top and budget deficits. However, this work argues that the concentration of income at the top of the distribution itself, not only as a proxy of income inequality across the distribution, may be important in explaining countries' fiscal performances. Indeed, to the extent that top incomes represent a large part of the tax base, contributing to about a third of total tax revenues in many countries, their importance cannot be ignored in budgetary terms. Moreover, besides top incomes' tax capacity, what matters in budgetary terms is also how much such capacity is taxed. In light of the substantial cuts to top marginal taxation experienced in many OECD countries in recent decades, it is interesting to analyze if such cuts may have also concurred in affecting OECD countries' fiscal performances. In order to understand if the high concentration of income at the top, along with a decreasing taxation of high incomes, may have affected OECD countries' public balances in the last thirty years; three questions are crucial: What share of national income does the top income group actually have? What has happened to taxation of high incomes over the last thirty years? What does the evidence suggest about how top incomes respond to changes in taxation? In what follows, particular attention is given to the relationship between the concentrations of income at the top and cuts to top marginal taxation, and to their potential effect on countries' public balances. First of all, the relevance of top income shares lies in their huge "tax capacity". The OECD last report "Divided we stand: Why inequality keep raising" (2011) claims: "...the growing share of income going to top earners means that they now have a greater capacity to pay taxes. Governments may consider raising marginal tax rates on income...” Graph 1 highlights the considerable level of total pre-tax income held by the 10 percent of the population. However, the level of income held by the top 10 percent income group greatly varies from one country to another: in some countries the share of national income held by

4 Leigh (2007), comparing top 1 percent and 10 percent income shares with the Gini coefficient, finds a positive and robust relationship between the series. However, Leigh himself concludes: "In summary, top income shares are far from perfect as a measure the distribution of income across society"

the top 10 percent income group is even higher than 40 percent; in others, it is around the 20 percent. According to the OECD, the top 10 percent of earners contribute to about a third of total tax revenues in many countries: 28% in France, 31% in Germany and 42% in Italy.

Graph 1: Top 10 Percent Income Share

Source: “World Top Income Database” (2011)

The whole picture becomes even more impressive looking at the top percentile income share. Graph 2 shows that the top 1 percent income share has increased considerably in almost all OECD countries. Interestingly, virtually all the increase in the top decile income share observed in Graph 1, is due to the huge increase in the income share of the top percentile. Atkinson, Piketty and Saez (2011) show that in the United States the share of total income going to the top 1 percent of income earners has increased dramatically from 9 percent in 1970 to 23.5 percent in 2007. Other English speaking countries like, United Kingdom, Canada, Australia, New Zealand and Ireland, have also experienced a substantial increase in the concentration of income at the top; in all these countries the current top 1 percent income share is bigger than 10 percent. Instead, in other countries, like France, Denmark and Netherlands, the top 1 percent income share is lower and it has been quite stable over time.

Graph 2: Top 1 Percent Income Share

Source: “World Top Income Database” (2011)

In fact, as Atkinson, Piketty and Saez (2011) note, despite the average U.S. federal individual income tax rate of top percentile tax filers was 22.4 percent, the top percentile paid 40.4 percent of total federal individual income taxes in 2007. The situation is not different in United Kingdom, where the current share of total income taxes paid by top 1 percent is the 28%. Diamond and Saez (2011) argue that if the average U.S. federal income tax rate on the top percentile were increased from 22.4 percent (as of 2007) to 29.4 percent, tax revenues, net of any behavioral response, would raise by 1 percentage point of GDP. Thus, considering such high concentration of income at the top, the potential top income groups' contributions to tax revenues cannot be ignored by governments in shaping their taxation and budgetary policies. Moreover, top income's tax capacity becomes even more important in light of the fact that in most of OECD countries the increase in such capacity has not been followed by a similar increase in the tax capacity of the rest of the population. Indeed, many recent studies show that in the last twenty years many OECD countries have experienced an unequal distribution of gains deriving from economic growth. The OECD's report "Divided we stand: why inequality keeps raising" (2011) highlights this issue : "The earnings of the richest 10% of employees have taken off rapidly, relative to the poorest 10% in most cases. And those top

earners have been moving away from the middle earners faster than the lowest earners, extending the gap between the top and the increasingly squeezed middle-class. The largest gains were reaped by the top 1% and in some countries by an even smaller group: the top 0.1% of earners". Fitoussi and Saraceno (2010), analyze the difference between the average growth of real income for each quintile and the growth of average income. They show that, in the last two decades, there has been a process of redistribution from the four bottom quintiles to the fifth in many OECD countries. Moreover, Fitoussi and Saraceno (2010) note that, over the last twenty years, the fifth quintile is the only one which shows an increase greater than that of average real income. Atkinson, Piketty and Saez (2011) show that U.S. real income per family grew at a modest 1.2 percent annual rate from 1976 to 2007. However, when excluding the top 1 percent, they show that the average real income of the bottom 99 percent grew at an annual rate of only 0.6 percent, which implies that the top 1 percent captured 58 percent of the whole. However, studying the relationship between the concentration of income at the top and the budget deficits cannot abstract from analyzing to what extent top incomes are taxed. While it is clear that the concentration of income at the top is strongly increased over the last thirty years, what about taxation at the top? In fact, after several decades into the postwar period with quite high degrees of income tax progressivity, several OECD countries have experienced falling tax rates in the last three decades. The reduction of tax progressivity has been mainly characterized by the huge decline in top marginal tax rates. Across the OECD countries the average top marginal tax rate fell by nearly 11 percentage points. Indeed, Graph 3 clearly shows the decreasing pattern of top marginal income tax rates in most of the OECD countries. In the late 1970s top tax rates varied from the 40% for Spain and Switzerland to the 75% for Sweden or the United Kingdom. The United States had quite high top tax rates too, around 70%. Currently, top tax rates are much lower than they were in the late 1970s, with no country above 60% and many countries around 40%. In addition, Piketty and Saez (2006) show that the decline in progressivity is not confined to income

taxation only. When also considering the distributional elements of estate and property taxes, they argue that the rich pay today much lower taxes than they did in 1970.

Graph 3:Top Marginal Income Tax Rates

Source: Data from Piketty, Saez, Stantcheva (2011)

However, concluding that such cuts to top marginal taxation may have negatively affected OECD countries' public balances via lower tax revenues is not straightforward at all. Indeed, in evaluating the net effect on public balances of cuts to top marginal taxation one has to take into account that such cuts, by providing greater incentive to work or invest, might foster economic growth. Thus, cuts to top tax rates could not only spur more economic activity, but also help pay for themselves since positive behavioral responses would more than offset the mechanical loss of revenues. This argument is central to the so called "supply-side" economic theory. In sum, according to this theory, although a higher taxation on top incomes might potentially generate higher public revenues to spend in redistribution or fiscal adjustments, it would give rise, instead, to behavioral responses in order to minimize tax payments, so hampering growth and reducing tax revenues. How individual tax rates influence behavioral responses among the rich is one of the most studied topics in the economics of taxation. Lindsey (1987) was one of the first to point out that the

top rate cut in the Economic Recovery Tax Act of 1981, which lowered U.S. top tax rate from 70% to 50%, coincided with a very large increase in the share of income reported by the top 1 percent of the income distribution. Lindsey estimates the elasticity of taxable income with respect to the netof-tax share to be between 1.6 and 1.8. Feldstein (1995a), studying the 1986 Tax Reform Act, which lowered the U.S. top marginal rate from 50 percent to 28 percent, finds that the percentage increase of income was much higher for high-incomes, who benefited the strongest cut. Feldstein estimates that, among high earners, the elasticity of taxable income with respect to the net-of-tax rate is very high, between 1 and 3. Moreover, Feldstein (1995b, 1999) argued that the Omnibus Budget Reconciliation Act of 1993 (OBRA93), which instead raised marginal tax rates mainly on high incomes, would have distorted the economic activity without raising relevant tax revenues. Feldstein claimed that the withdrawal of the rate increases could have generated positive behavioral responses which would have actually increased tax revenues. However, as argued by Stiglitz (2004), the OBRA93 experience showed that raising taxes on the rich did not have such negative effects. Indeed, contrary to some studies' prediction, in those years U.S government observed a substantial increase in budget revenues which favored public debt reduction, without experiencing slower growth rates. Giertz (2009) notes: "For OBRA93 the median response was zero, with an interquartile range from -0.5 to 1 percent of GDP. It is noteworthy that half of public economists surveyed thought that raising marginal tax rates for the highest income groups (in 1993) would not result in decreased steady-state GDP". Although OBRA93 do not seem to have generated strong behavioral responses, as already noted, previous studies on the 1981 and 1986 U.S. tax reforms reported instead quite high elasticities of taxable income between 1 and 3. One of the first studies which cast doubts on such large top incomes’ elasticities is Goolsbee (2000). Goolsbee shows that the efficiency losses derived from high marginal tax levels were limited, and that earlier findings overstated them. According to Goolsbee, such overestimation is mainly due to the fact that researchers have not taken into account that top incomes may shift income to capital gains in order to minimize taxation and that they may

adjust the timing of the realization of their income and capital gains. Thus, Goolsbee suggests that the strong behavioral responses found by previous research would be mainly due to responses which do not affect GDP growth. More recent studies have greatly improved the analysis of the variety of ways in which top incomes respond to taxation and the efficiency implications associated with these behaviors5. However, a common feature of these recent contributions is that they all report top incomes' elasticity close to 1 or even lower6; much smaller than the earliest estimates by Feldstein (1995a) and Lindsey (1987). Among these recent works, a number of studies suggest that cuts to top taxation seem to have uniquely increased the income of the rich without leading to better growth performances. Therefore, they argue that top tax rates observed in many OECD countries, not only in the United States, in the last thirty years could be much lower that the optimal ones. Piketty, Saez and Stantcheva (2011) show that in the last thirty years, countries like the United States or the United Kingdom, which heavily cut their top tax rates, did not experience significantly higher economic growth than other countries like Germany or France which keep top taxation quite stable. Analyzing top income and top tax rate data in 18 OECD countries; they find that there is a strong correlation between cuts in top tax rates and increases in top 1 percent income shares since 19757; while there is no clear relationship between cuts to top marginal taxation and economic growth. According to the authors this result suggests that the supply-side effect, according to which a cut in top rates translates into additional economic activity among upper incomes, higher top income shares and higher economic growth do not seem plausible. Instead, according to the authors, the hypotheses consistent with the “bargaining model” seem to be more convincing. According to this model, gains at the top come at the expense of lower income earners. In the bargaining scenario, a cut in top tax rates generates a "trickle-up" transfer from lower to upper incomes with an increase in top income shares, but no additional economic activity. In light of their results, Piketty,

5

Slemrod (2002) presents a precise analysis of the ways in which people respond to changes in taxation. Auten, Carroll, and Gee (2008) find that top incomes’ ETI equals to 1.09. Saez (2004) reports an ETI of 0.62 for the top 1 percent and of 1.09 for the top 0.01 percent. 7 A number of other recent studies provide evidence in this sense. For example, Atkinson and Leigh (2007b) and Roine, Vlachos, and Waldenström (2010), combining data for several countries and using a multivariate approach, show that there is a strong negative relationship between top marginal tax rates and top income shares. 6

Saez and Stantcheva (2011) claim: "...socially optimal top tax rates might possibly be much higher than what is commonly assumed". In their preferred estimates8, they find that the socially optimal top tax rate would be equal to 83%. Romer and Romer (2012) make conclusions which are basically in line with the ones suggested by Piketty, Saez and Stantcheva (2011). Romer and Romer (2012), focusing on the interwar period in the United States, find that short-run labor-supply or incomeshielding effects of marginal rate changes, while clearly present, were of limited economic significance. According to the authors changes in marginal tax rates have a precisely estimated but small impact on reported taxable income. Considering elasticity of top incomes to taxation ranging from 0.19 to 0.38, Romer and Romer (2012) suggest that the revenue-maximizing rate would be 84 percent in the low elasticity case and 73 percent in the high elasticity one. Drawing on Piketty, Saez and Stantcheva (2011), this work presents some very preliminary evidence in favor of the "bargaining" model scenario: Graph 4 suggests that lower top tax rates are correlated with higher top income shares but not with higher economic growth. Panel A of Graph 4, showing a strong negative correlation between average top marginal income tax rate and average top 1 percent income share, somehow confirms the well-documented role of low top marginal tax rate in the raise of top 1 percent income share. On the other side, Panel B shows that the correlation between the average growth rate of GDP per capita and the average top marginal income tax rate is almost zero, suggesting that low top tax rates do not seem to be associated with higher economic growth. Thus, Graph 4, shows that low top marginal taxation are strongly correlated with higher top income shares but not with higher growth; encourage this work search for a positive relationship between the concentration of income at the top and budget deficits. In sum, although it is hard to draw any final conclusion about the net effect that cuts to top taxation may have on economic growth and tax revenues, and in turn on public balances, there is now considerable evidence which supports the facts that top marginal tax rates observed in recent decades could be lower than the optimal ones. To the extent that this is true, it is reasonable to wonder if the high concentration of income at the 8

In this case the overall elasticity, equal to 0.5, comes from 0.2, "supply-side elasticity", plus 0.3, "bargaining effect elasticity".

top, along with low taxation at the top, may have affected OECD countries' public balances over the last three decades.

Graph 4: Top 1 Percent Income Share, Top Marginal Tax Rates and Growth PANEL A

PANEL B

Source: Data from Piketty, Saez, Stantcheva (2011). Calculations of the author

3.1 On the Causality Direction between Top 1 percent Income Share and Top Marginal Taxation After having shown that a negative relationship between top income shares and top marginal tax rates exists, the last question is: which is the causality nexus between the two? Indeed, understanding better which is the causality nexus between top marginal tax rates and top income shares may help to give a more comprehensive analysis of their effect on countries' fiscal performances. The most common argument suggests that lower top marginal tax rates have caused higher concentration of income at the top. However, the case of reverse causality from higher concentration of income at the top to lower marginal tax rates has to be taken into account. As noted by Atkinson, Piketty and Saez (2011)9, since top incomes command a large share of national resources, they have a great incentive to invest in activities instrumental to gain tax advantages, such as tax cuts or preferential tax rates, from government. In order to shed light on the causality 9

They note that:"...There is also the possibility of reverse causality. The increases in top incomes as a result of changed executive remuneration policies may have increased political pressure for cutting top taxes. We need therefore a simultaneous, as well as multivariate, model."

direction between top income shares and top marginal tax rates, some preliminary evidence is presented. Although only suggestive, results in Table 1 seem to corroborate the hypothesis according to which lower top marginal tax rates may cause higher top 1 percent income share, and not the opposite. Indeed, column (1) shows that the coefficient on the five year lagged level of top marginal tax rates, TMTRi ,t − 5 , is highly significant in explaining Top1i ,t at time t, even controlling for the five year lagged top 1 percent income share. Moreover, column (2) shows a significant negative relationship between changes in top 1 percent income share and changes in top marginal taxation observed in the previous year. On the opposite, results do not provide support for the hypothesis that higher top 1 percent income share may cause lower top marginal taxation. Indeed, column (3) and (4) show that both the coefficient on Top1i ,t − 5 , in the TMTRit regression, and the coefficient on ∆Top1i ,t −1 , in the ∆TMTRit regression, are far from being significant.

Table 1: Top 1 Percent Income Share and Top Marginal Income Tax Rates (1) (2) (3) Top1 ∆Top1 TMTR TMTRt-5

-0.0235***

Top1t-5

(0.006) 0.9439*** (0.068)

0.0221* (0.011)

(4) ∆TMTR

0.7407***

-0.0374***

(0.082) -0.0236 (0.384)

(0.013)

-0.0127*

∆TMTRt-1

(0.007) 0.3051 (0.214)

∆Top1t-1 Constant N-countries Observations R-squared

1.9539**

0.5669

9.4975

0.3974

(0.844) 17 374

(0.542) 17 401

(6.696) 17 403

(1.163) 17 386

0.85

0.13

0.73

0.11

Note: Cluster robust standard errors in parenthesis. All regressions include year dummies. *** p