NATIONAL CENTER FOR POLICY ANALYSIS
HOME / DONATE / ONE LEVEL UP / ABOUT NCPA / CONTACT

Tax Rates, Tax Revenues And Economic Growth

Appendix A

Data on taxes and gross domestic product are available from International Monetary Fund government statistics. The dependent variable in the statistical estimation is the logarithm of per capita tax revenue in 1980 dollars. The independent variables are the logarithm of per capita domestic product in 1980 dollars, the tax rate, the tax rate squared, and the difference between the tax rate and expenditures as a share of gross domestic product. This last variable captures some of the effect of future tax liability. The sample size is 103 countries. All of the variables in the estimation are very highly significant (see Table A-1). The coefficients of interest are tax variables. The coefficient of the tax rate on the logarithm of tax revenue (with its standard error in parentheses) is 0.1036 (0.0019). The coefficient of the tax rate squared on the logarithm of tax revenue is -0.0012 (0.00004). Thus, as tax rates rise, tax revenue also rises but at a diminishing rate. The maximum rate is solved by dividing the coefficient of the tax rate on tax revenue by two times the coefficient of the tax rate squared on revenue. This yields 43.2 = .1036/.0024. Thus the tax revenue maximizing tax rate is 43.2 percent. This seems rather high, but a number of countries in the world have taxes near or above this rate.

There also exist tax maximizing rates for various types of taxes. Using a similar empirical specification, we find that the tax maximizing rate for the income tax, sales tax and trade taxes expressed as a fraction of GDP are 22.5, 12.5 and 13.2 percent, respectively. A revenue maximizing government will set all tax rates to their tax maximizing levels.

The growth rate in the estimation of the relationship between tax rates and economic growth is the compound growth rate in per capita real domestic product over the period 1960 to 1980. It is calculated from the data on per capita real gross domestic product from Summers and Heston.1 The independent variables are the growth rate in real capital per head over the same period, real gross domestic product per capita in 1980, the difference between taxes and expenditures as a share of GDP, the tax rate and the tax rate squared.

Note that the growth rate is over the period 1960 to 1980 and the tax rate is for 1980. There is a sound theoretical reason for using the end state tax rate. The period was one in which the size of government grew. The growth in government is viewed as a growth in tax liability. Presumably, people make behavioral adjustments in expectation of future tax liability.

Real gross domestic product per capita in 1980 is in the equation to adjust for convergence of growth rates. According to neoclassical economic theory, high income (capital intensive) economies grow more slowly than low income (labor intensive) economies because the marginal returns to capital are low in the former and high in the latter. In theory, capital flows to its highest valued use. In fact, capital probably has been flowing, from the high-risk, less-developed world to the safer havens in the West. Contrary to the predictions of neoclassical theory, the high per capita income countries are growing at the faster rate.

All of the variables in the estimation are highly significant (the regression equation is presented in Table A-1). The variables of interest here are the tax variables. The coefficient of the tax rate on the growth rate is 0.1339 (0.0688). The coefficient of the tax rate squared on the growth rate is -0.0035 (.000089). The growth maximizing tax rate is 19.3 ( = .1339/.0069) percent.

The growth maximizing tax rates for various types of taxes were also calculated. The empirical specification essentially is the same as reported above. The growth maximizing tax rates for the income tax, sales tax and trade taxes are 11.9, 4.6 and 9.4 percent, respectively. Note that, like aggregate taxes, the growth maximizing tax rates are about half that of the revenue maximizing tax rates.

1Robert Summers and Alan Heston, "Improved International Comparisons of Real Product and Its Composition: 1950-1980," Review of Income and Wealth, 30, June 1984, pp. 207-262.


Home |  Support Us |  All Issues |  Social Security |  Debate Central |  Contact Us
Dallas Headquarters: 12770 Coit Rd., Suite 800 - Dallas, TX 75251-1339 - 972/386-6272 - Fax 972/386-0924
Washington Office: 601 Pennsylvania Avenue NW, Suite 900 South Building, Washington, DC 20004 - 202/220-3082 - Fax 202/220-3096
© 2001 NCPA