Tax on Consumption Isn't Quite What You Think

07/01/1997
Summary of working paper 5832

...although a consumption tax would reduce the percent of income paid by the highest-income households substantially it would not reduce it as much as is often thought.

Critics often claim that consumption taxes would shift the tax burden away from high-income households because, under a consumption tax system, wages would be taxed fully while income from capital would not be taxed, and high-income households receive a large share of their income from capital. Now a recent NBER study casts new light on this issue. In "Distributional Implications of Introducing a Broad-Based Consumption Tax" (NBER Working Paper No. 5832,) William Gentry and Glenn Hubbard first challenge the assumption that a consumption tax would leave all income from capital untaxed. Then they estimate the impact of alternative tax systems on the distribution of taxes paid. They find that the critics of consumption taxes are only partially correct; the shift in tax burdens would not be quite as large as are sometimes claimed.

Earnings on capital, the authors note, include not just the return to a riskless investment, but also a premium for risk, economic profits, and simple luck. Relative to a pure income tax, only the portion of capital earnings attributable to a riskless investment would be free of tax under a consumption tax. Thus, even a consumption tax would tax much capital income.

Gentry and Hubbard assume a hypothetical federal tax system that gives a $15,000 exemption for a single person and a $25,000 exemption for married households, and taxes all income beyond those levels at a flat 25 percent rate. Using data from 1989, they then compare the effects on tax burden of including various forms of income in the tax base.

With the tax base as defined by the current income tax system, they find, the top 5 percent of households (those with incomes above $136,263 in 1989 dollars) would have paid 56 percent of all federal income taxes. Shifting to a consumption tax under the conventional assumption that no capital income is taxed would reduce the share of taxes paid by that same high income group to 46 percent. But assuming, as Gentry and Hubbard do, that a consumption tax exempts only the riskless return to capital, the percent of taxes paid by the top 5 percent of taxpayers only falls to 50 percent. In other words, although a consumption tax would reduce the percent of income paid by the highest-income households substantially, it would not reduce it as much as is often thought.

Of course, annual income is not the same as wealth. So Gentry and Hubbard also consider the impact of the consumption tax on the percent of taxes paid by wealthy households, as distinct from high-income households. With the same hypothetical tax system they describe and with the tax base as defined by the current income tax system, the wealthiest 5 percent of households (those with a net worth greater than $545,000 in 1989) would have paid 48 percent of all federal income taxes. Shifting to a consumption tax under the conventional assumption that no capital income is taxed would reduce the share of taxes paid by that group to 30 percent. Assuming, as Gentry and Hubbard do, that a consumption tax would exempt only the riskless return to capital, the percent of taxes paid by the wealthiest 5 percent of households would only fall to 37 percent. (DRH)