 |
 |
|||
 |
 |
|||
|
|
|||
|
|
|||
 |
A Flat Tax Would Produce Explosive U.S. Economic Growth Jude Wanniski & David Goldman with Jay Turner and Evan Kalimtgis March 31, 1992 |
 |
Executive Summary: A comparison of marginal tax rates and economic growth among the world's largest 23 economies provides striking evidence that radical tax reform would unleash an enormous economic boom in the United States. The world's fastest-growing economy, Hong Kong, also has the lowest marginal tax rates on capital and labor, and the simplest (or "flattest") rate schedule in the world. Remarkably, statistical analysis shows that the differences in marginal tax rates among these 23 economies account for between 80 and 90 percent of their differences in economic performance.
Our results support the arguments in favor of a low, flat tax in the United States. Critics of flat tax proposals, such as Democratic presidential candidate Jerry Brown's, argue that the tax would penalize lower-income taxpayers while failing to raise adequate revenues. Even on a static basis, there are good arguments that a flat tax can be made fair to lower-income families. The experience of the 1980s, when the proportion of total income tax revenues paid by the top 1% of taxpayers rose to 27% in 1989 from only 18% in 1981, strongly suggests that a flat tax such as Brown proposes would broaden the tax base and raise additional revenues. It is possible to quibble with some of the details of his proposal, but that is an appropriate exercise at a much later stage of discussion. The purpose of a presidential election campaign is to broach important differences in economic philosophy and matters of broad strategy.
By far the most powerful argument in favor of the flat tax is that it will promote higher economic growth. Previous studies have argued from historical data that lower taxes foster higher growth. The work we present employs far more recent data than previous studies and examines a more comprehensive measure of taxation. Most importantly, our study examines the effect of marginal tax rates – the tax burden on added income – rather than the average tax rate. Our results show that it is not the average tax burden but the marginal tax rate which determines growth rates – a conclusion which strongly supports the adoption of some form of flat tax.
A Flat Tax Would Produce Explosive U.S. Economic Growth
Democratic presidential candidate Jerry Brown has done the country a service by focusing national debate on a flat tax, a proposal which has won support across a broad spectrum of academic economists. The international tax comparison presented in this report provides strong evidence that a flat tax, such as Brown's proposal, would promote faster economic growth in the United States. Evidence is overwhelming that lower taxes correspond to faster growth on a global basis.
Regarding the merits of the flat tax approach to lowering tax rates, former Governor Brown is correct to emphasize that the more than 4,000 pages of the tax code reflect the accretion of the claims of special interests over many decades. Eliminating virtually all special treatment for taxpayers in favor of a low, easily-collectible tax would generate enormous benefits for the U.S. economy as a whole. First, the barriers to small business, for whom tax compliance is a substantial expense, would be reduced; because small business is the largest generator of new jobs, the economy would benefit. Second, the tax base would broaden, as a result of reduced incentives to avoid taxation. Third, the activities of the thousands of America's best-trained professionals in the legal and accounting professions would be redirected towards creating new enterprise rather than avoiding taxation. As Professor Paul Samuelson, a liberal Democrat, writes in the 13th edition of his textbook, Economics, the flat tax "would put thousands of tax lawyers out of business. The economy would become more efficient as people spent less time worrying about the impact of their actions on taxes and more time worrying about production of aircraft and computers and generation of innovation."
Objections to the flat tax plan boil down to two. The first is that a 13% income tax (with deductions for mortgage payments, rent, and charitable contributions) and a 13% national sales tax would add approximately $200 billion to the deficit. The second is that poorer taxpayers would lose and richer taxpayers would gain. Both these objections, though, refer to the details of Gov. Brown's plan, rather than to the concept itself, and bear easy refutation. The Brown program represents a conceptual outline, a statement of economic philosophy, appropriate for debate during a presidential election campaign. At such time as Governor Brown were elected President, his administration and Congress would hammer out the details.
The point is not to produce a purely flat tax for some ideological reason, but to reduce the 4,000 pages of tax code to perhaps 40. Hong Kong, the world's fastest-growing economy, has the system closest to a pure flat tax. Hong Kong's maximum 15% tax on income, though, does not apply to incomes less than $HK50,000 ($6,500), a middle-class income by that country's standards. Lower incomes are taxed at 2% to 12%. Hong Kong also taxes corporate profits at 16.5%, slightly above the rate for personal income. Professor Alvin Rabushka of Stanford University, the author of the standard book on the subject, proposes that the flat tax not apply to family incomes of less than $16,000. An appropriate threshold for a flat tax, or reduced tax rates for lower incomes, must be determined. The guideline should be that no one should suffer a tax increase as a result of the flat tax.
Regarding the loss of revenues, critics, such as Citizens for Tax Justice, assume that the introduction of a flat tax will not widen the tax base, nor generate additional economic growth and, hence, increase revenues. The experience of the past decade, though, argues strongly for a dynamic as opposed to a static approach to revenue estimates. Upper-income taxpayers, whose income includes substantial earnings from capital, have enormous discretion about the amount of income they report. Lower taxes encourage high-income taxpayers to declare more income rather than shelter it. As a result of the reduction in the maximum marginal tax rate from 70% in 1981 to 28% in 1986, the top 1% of U.S. taxpayers paid 27% of all income taxes in 1989, against only 18% in 1981.
An international comparison of taxes rates and economic growth provides strong evidence that lower taxes foster stronger economic growth. Differences in marginal tax rates among most of the world's 23 richest economies account for between 80% and 90% of differing growth rates during the years 1984-89. This conclusion derives from a study employing the most comprehensive measure of marginal tax rates available to date. Using the "Doing Business"1 guides compiled by Price Waterhouse, we calculated the combined effect of income, sales, local, wealth, payroll, and capital gains taxes. The 23 countries were selected on the basis of availability of comparable tax data. Together, they represent almost the entire Organization for Economic Cooperation and Development and most of the leading Newly Industrialized Countries (NICs), that is, most of the world economy.
These facts are worth examining in a world in which marginal tax rates have risen in most major industrial countries during the past year. Last year's budget agreement, as well as hefty increases at the local level, contributed to the U.S. recession. Germany imposed an income-tax surcharge, among other increases, last summer. Japan and Italy have increased their capital gains tax.
The evidence is overwhelming that low marginal tax rates promote growth, while high rates inhibit it, even though there is no perfect methodology for calculating the total effect of marginal tax rates on a consistent basis. Each national case is different; a weighted average of tax rates may ignore important traps in the tax code. After all, one can still drown in a river that is on average only two feet deep. The United States, for example, has the lowest marginal tax rates on labor in the industrial world, given low income tax top rates and the absence of a national value-added tax, but the highest effective tax rate on capital gains. We have presented our case elsewhere that this glaring anomaly in the U.S. tax code is the chief culprit in the present recession. It is also not strictly fair to argue that lower marginal tax rates as such caused higher growth throughout the world, although the statistical correlation between low taxes and high growth is extremely high. Low marginal tax rates have been associated in most countries with a "policy mix" including relatively stable management of the national currency and close attention to regulatory incentives and disincentives. As a study of the sources of growth during the 1980s, therefore, these results must be regarded as provisional, pending a comprehensive treatment of the economic policy mix in the world's leading economies.
One noteworthy fact is that Hong Kong, with the closest to a flat tax of any important economy, showed aggregate GNP growth of 97% between 1984 and 1989, almost three times as much growth as the next strongest economy.
A distinguishing feature of the present study is that it relates growth to marginal, rather than average, tax rates. Previous studies, e.g., Gerald W. Scully's paper on "Tax Rates, Tax Revenues, and Economic Growth" (National Center for Policy Analysis: Houston, March 1991) have argued from historical data that lower average tax rates foster higher growth. Most importantly, our studies examine the effect of marginal, rather than averate, tax rates. No statistical measurement we employed revealed a high degree of correspondence between the average tax rate, i.e., total tax revenues as a percentage of national income, and the growth rate of national income among any large sample of national economies. Using a sample of over 100 national economies, the correlation is almost nil. In the sample of 23 countries discussed in this study, the correlation between economic growth and average tax rates is only 10%. Nor is there any correlation among any large sample of national economies between average tax rates and marginal tax rates. It is frequently the case that the citizens of a country will elect to pay a high proportion of their income in taxes, while maintaining relatively low tax rates at the margin. The strong correlation between growth and marginal, rather than average, tax rates, reinforces economic common sense: incentives to labor and capital take effect at the margin, rather than being spread across the entire income stream. It is the percentage of the last dollar, mark, or yen earned, which the production factor is permitted to retain after taxes, that determines whether that last increment of effort is remunerative.* * * * *
The results of our study are summarized in the following table, which compares weighted GNP growth and marginal tax rates in the countries studied:
Real GNP Growth % Marginal Tax
1984-89 Weighted Total
Hong Kong 97 15.9
Chile 35 53.6
Singapore 33.49 48.4
Japan 24.75 58.6
Brazil 24 46.2
Spain 24 62.9
Colombia 23.5 46.2
Canada 22 52.9
Australia 21 54.5
Finland 21 60.9
UK 19.85 58.6
US 18 51.7
Italy 17 63.0
Ireland 16.8 66.6
Switzerland 16 52.4
France 15 66.6
Austria 14 62.8
Belgium 14 68.8
Germany 13.6 64.0
Netherlands 12 70.1
Sweden 12 72.1
Norway 11.85 67.4
Greece 11 66.7
Denmark 8 80.2
Since the change of national income over time represents the compound growth due to the factors which cause it, a curvilinear correlation is the most appropriate means of evaluating the effect of marginal tax rates on growth. A curvilinear correlation shows an r2 of .91, whereas a linear correlation shows an r2 of .76. In plotting the results for the 23 countries against a vertical axis showing GNP growth and a horizontal axis showing marginal tax rates, the results of the curvilinear correlation are that the correspondence between the two variables is 91%.
Marginal tax rates were first calculated by tabulating the tax burden on labor and capital, and then combining these into a weighted index. The top marginal tax on labor was calculated by adding together the top marginal tax rate on personal income at federal and local levels, wealth taxes, social security taxes, and sales or VAT taxes. (The wealth tax was multiplied by three to equate a tax on assets to taxes on income.)Taxes on Labor
Top Marginal Local Tax Wealth Tax Social Security Tax Sales Tax
Tax Rate on Labor
Australia 47 0 0 1.25 3
Austria 50 0 0 8.75 20
Belgium 55 5 0 5.43 19
Brazil 10 1.25 0 6.75 12.5
Canada 29 16 0 1.70 7
Chile 35 0 0 2.60 18
Colombia 30 0 0 5.00 12
Denmark 68 0 1 0.03 22
Finland 39 19 0 1.70 18
France 39 0 1.5 18.00 19
Germany 45 0 0 6.11 14
Greece 50 0 0 7.00 18
Hong Kong 15 0 0 0.00 0.5
Ireland 53 0 0 3.00 20
Italy 50 4 0 4.75 19
Japan 50 5 0 6.00 3
Netherlands 60 0 0.8 8.00 18.5
Norway 23.5 26.5 0 8.50 20
Singapore 33 0 0 27.00 0
Spain 56 0 2 1.10 12
Sweden 20 30 3 0.00 20
Switzerland 13 32 0.3 4.40 8
UK 40 0 0 8.00 15
US 31 7 0 7.19 5Taxes on capital were calculated by adding together the inflation-adjusted capital gains tax, employers' social security contributions, and the top corporate tax rate. The capital gains tax was assigned a weight of 40% of the other taxes, since it is a smaller portion of corporate income than ordinary income.
Capital Gains Tax Adjusted for Inflation Social Security Tax (Employers) Top Corporate Rate
Australia 50.76 0.00 39
Austria 0.00 8.75 30
Belgium 0.00 5.43 40
Brazil 63.50 11.40 33
Canada 20.60 0.00 43.5
Chile 39.20 0.00 35
Colombia 37.80 0.00 30
Denmark 52.47 0.02 40
Finland 31.76 0.00 23
France 16.57 0.00 39
Germany 35.70 6.11 42
Greece 30.48 0.00 42
Hong Kong 0.00 0.00 16.5
Ireland 31.28 0.00 40
Italy 0.00 4.75 36
Japan 26.26 0.00 40
Netherlands 0.00 0.00 42
Norway 0.00 0.00 50.8
Singapore 0.00 0.00 31
Spain 21.49 1.10 35
Sweden .96 19.00 30
Switzerland 10.30 4.40 35
UK 42.21 0.00 35
US 31.96 7.19 34
The totals for labor and capital were then combined to obtain a weighted total for the top marginal tax rate:
TAX ON LABOR TAX ON CAPITAL WEIGHTED TOTAL
Australia 51.3 59.3 54.5
Austria 78.8 38.8 62.8
Belgium 84.4 45.4 68.8
Brazil 30.5 69.8 46.2
Canada 53.7 51.7 52.9
Chile 55.6 50.7 53.6
Colombia 47.0 45.1 46.2
Denmark 93.0 61.0 80.2
Finland 77.7 35.7 60.9
France 80.5 45.6 66.6
Germany 65.1 62.4 64.0
Greece 75.0 54.2 66.7
Hong Kong 15.5 16.5 15.9
Ireland 76.0 52.5 66.6
Italy 77.8 40.8 63.0
Japan 64.0 50.5 58.6
Netherlands 88.9 42.0 70.1
Norway 78.5 50.8 67.4
Singapore 60.0 31.0 48.4
Spain 75.1 44.7 62.9
Sweden 79.0 61.8 72.1
Switzerland 58.3 43.5 52.4
UK 63.0 51.9 58.6
US 50.2 54.0 51.7The weighting system by which the total marginal tax rate was estimated is shown in the table above. Taxes on personal and corporate income, wealth, payrolls, and sales are each weighted equally at one. The wealth tax, though, is multiplied by a factor of three, assuming that taxes on wealth (i.e., capital assets) are the equivalent of an equal tax on a much larger amount of income. Capital gains is weighted at 0.4. In the final tabulation, labor is given a weight of 0.3, and capital a weight of 0.2, reflecting the relatively larger amount of labor income compared to capital income. The weighting system is summarized in the table below:
Weighting System
Weight Category
1 TOP MARGINAL
1 LOCAL
1 WEALTH
3 WEALTH MULTIPLIER
1 SOCIAL SECURITY-LABOR
1 SALES
1 SOCIAL SECURITY-CAPITAL
1 CORPORATE
0.4 CAPITAL GAINS
Totals
0.3 LABOR
0.2 CAPITAL
As a cross-check on these results, we used an alternate statistical methodology. A multilinear regression of percentage changes in GNP between 1984 and 1989 against all changes in individual taxes was conducted. This regression shows a strong negative correlation between all categories of taxes and growth. By this methodology, tax differentials between nations seem to account for almost 80% of the disparity in growth rates.
Several items are not taken into account by the model. The distribution of the burden of a given tax is not weighted to reflect the existing income dispersment in each country. This will lead to biased coefficents on the taxes, due to the fact that a change in the tax on capital has a much different effect in a country where a high proportion of GNP is derived from capital formation than in a low proportion country.
Additionally, the model does not take into account elasticities. Analysis of economic behavior on the margin is inextricably linked with elasticities. In a country with an extremely high sales tax, even a very small decrease in the tax may lead to a flood of consumption, as people may have been doing without for quite some time. We would witness then a temporary decrease in investment. These cross-elasticities between consumption and investment, as determined by the overall income level, and the relative price of each, should be included for completeness.
With these caveats, the results of the regression are as follows:
Variable Coefficent t-statistic
Top Marg. Income -.52 -3.16
Local -.72 -3.17
SS employees -1.14 -3.35
SS employers -1.12 -2.74
Sales -.58 -1.99
Top Corporate -.94 -3.27
Capital Gains -.38 -3.16
The dependent variable is the percentage change in GNP between 1984 and 1989 in the countries studied. The independent variables display a 78% correlation with the dependent variable.* * * * *