Consumption Tax

Those who favor eliminating taxes on capital income often couch the issue in terms of whether the government should tax income or consumption. A consumption tax—also known as an expenditures tax, consumed-income tax, or cash-flow tax—is a tax on what people spend instead of what they earn. Or, as economists Alvin Rabushka and Robert Hall put it, taxing income taxes what people contribute to the economy, while taxing consumption taxes what they take out. The United States already taxes both income and consumption, of course. For nearly a century now, the principal federal tax on individuals has been the personal income tax, which falls on both labor income (wages and salaries) and capital income (interest, dividends, and capital gains). Meanwhile, states and localities raise a large share of their revenue through sales taxes, which are taxes on consumption. The federal government also has a smattering of consumption taxes, such as the excise tax on gasoline.

Only a few advocates of a consumption tax want the federal government to replace or supplement the income tax with a national sales tax or a value-added tax (VAT) like the ones in Europe. In theory, a VAT is a tax on the difference between what a producer pays for raw materials and labor and what the producer charges for finished goods. Hence the term “value added.” In practice, VAT taxes generally are applied much like sales taxes in the United States, with the government collecting a fixed percentage of the full pre-VAT selling price of a good rather than of the true value added.

Instead of moving to a VAT, most consumption tax advocates want to modify the income tax to eliminate taxes on interest, dividends, and capital gains. Achieving that goal can be done directly by cutting taxes on capital income to zero, as President Bush tried to do in the case of dividends. Alternatively, we could achieve the same result indirectly in two other ways. One is granting unlimited tax deductions for contributions to savings plans and taxing withdrawals, just as is done now with conventional individual retirement accounts (IRAs) and 401(k) plans, but without penalties for withdrawals before age fifty-nine and a half. The second is the Roth IRA structure, in which an individual gets no deduction for savings contributions but pays no taxes on withdrawals. Both conventional IRA and Roth IRA treatments are mathematically identical with a zero tax rate on capital income, as is shown below.

Why should capital income be treated differently than wages and salaries? The economic reason is very different from the political rhetoric. It is not to favor the rich at the expense of the working man. Nor is the principal reason to eliminate the double taxation of profits, first at the corporate level and again at the individual level as dividends or capital gains. Proponents of a consumption tax argue that it is superior to an income tax because it achieves what tax economists call “temporal neutrality.” A tax is neutral (or “efficient”) if it does not alter spending habits or behavior patterns from what they would be in a tax-free world, and thus does not distort the allocation of resources. No tax is completely neutral, because taxing any activity will cause people to do less of it and more of other things. For instance, the income tax creates a “tax wedge” between the value of a person’s labor (the pretax wages employers are willing to pay) and what the person receives (after-tax income). As a consequence, people work less—and choose more leisure—than they would in a world with no taxes.

The case for a consumption tax is that the tax wedge created by taxing capital income does enormous long-term damage to the economy. Taxing interest, dividends, and capital gains penalizes thrift by taxing away part of the return to saving. The unavoidable result is less saving than society would choose in the absence of any taxes. The social value of saving is the market interest rate that borrowers are willing to pay for the use of resources now. Economists are confident that this is the value to society because it is a market price that reflects the desires of the various savers and borrowers. If each potential saver could collect the market interest rate, the result would be an optimal amount of saving—that is, an optimal division of resources between current consumption and future consumption. “Optimal” in this sense refers to the amount of saving that individuals, deciding freely on the basis of market prices, would choose to do on their own, rather than the amount of saving that a politician, social planner, or economist thinks they ought to do.

Market interest rates effectively pay people to defer consumption into the future (i.e., to save). Because the tax wedge reduces those payments, people inevitably will choose less future consumption (saving) and more current consumption. This harms the economy because less saving results in less investment, less innovation, slower growth, and lower future living standards than would be enjoyed without a tax on saving. Future consumption is reduced by both the extra current consumption and the forgone returns that greater saving would otherwise have produced. Some of this loss is a deadweight loss to society, that is, a loss to some that is a benefit to no one. Eliminating taxes on capital income would eliminate the tax wedge on saving, and total saving would be much closer to the optimal amount. The tax system would be “temporally” neutral in the sense that it would not affect the choice between current consumption and future consumption (saving).

To see how the tax wedge works, first consider how the income tax affects a person with $10,000 of pretax income. Assume, for simplicity, that the only tax bracket is the 35 percent top bracket in the 2003 tax law, that the pretax market interest rate on bonds is 5 percent, and that expected inflation is zero. Under the income tax, the individual pays $3,500 in taxes and can consume $6,500 of goods and services now or invest $6,500 in bonds paying 5 percent. If he saves the money, he will receive $325 in interest after one year, on which he pays $113.75 in taxes, leaving after-tax investment income of $211.25. Taxes reduce the interest rate he receives from 5 percent to just 3.25 percent. His potential consumption at the end of year one is $6,711.25.

Alternatively, if the individual puts the $6,500 in a Roth IRA and buys the same bonds, he gets the $325 in interest payments tax free and has $6,825 of potential consumption at the end of year one (assuming there are no early withdrawal penalties). The result is the same with a conventional IRA. In that case, he invests the entire $10,000 and pays no tax immediately. A year later he has $10,500—the $10,000 principal and 5 percent in interest income. If he withdraws the entire amount, he pays taxes of 35 percent on all of it, or $3,675. He is left with $6,825, the same as with the Roth IRA.

Despite its allure of eliminating the bias against saving, a true consumption tax runs into fervent opposition from some, mostly liberal, economists. As noted, their principal objection is that the greatest direct benefits of a consumption tax would go to high-income individuals. Since they are in higher tax brackets, high-income households get a greater dollar benefit from deducting savings (traditional IRA) or having after-tax contributions accumulate tax-free income (Roth IRA). In addition, high-income households have a greater ability and propensity to save, and thus are more likely to take advantage of opportunities for tax-free capital income. The counterpoints to that argument are two. First, those who pay the most in taxes inevitably will get the greatest dollar benefit from tax reductions. Second, the economic benefits from greater saving—more innovation and greater GDP growth—would be distributed to everyone in the form of a faster increase in real incomes, including wages.

The one objection to a consumption tax based on pure economics is that it would require a higher tax rate in order to raise the same revenue as an income tax that includes capital income. For this reason, a consumption tax would be less neutral between work and leisure than the current income tax. This would cause people to work less, and would increase the deadweight loss from the tax wedge on labor income. Advocates of a consumption tax maintain that the gains from additional saving and investment would greatly outweigh the losses from less work effort, though it is impossible to know with certainty whether that is correct. However, it is worth noting that the “flat tax” proposed by Alvin Rabushka and Robert Hall, which is actually a consumed-income tax, calls for a tax rate of just 19 percent.

The practical objection to a consumption tax used to be that it is too complicated to monitor the amounts that people save or dissave each year. But that actually can be done quite easily, as several decades of experience with IRAs, 401(k)s, and other special savings vehicles have shown. Moving to a complete consumption tax system for the individual tax code would entail little more than allowing universal, unlimited IRAs for everyone and doing away with penalties for early withdrawals. Individuals could continue with the IRAs and 401(k)s they currently have or roll them over into the new plans.

The new unlimited IRAs could take the form of conventional or Roth IRAs. The only genuine difference between the two—a major one—is the impact on federal tax receipts. The present values of tax revenues under the two systems are essentially the same, but the timing is radically different. Under conventional IRAs, the federal treasury takes a big hit today because contributions are removed from the tax base, but future tax revenues get a boost when withdrawals are taxed. Under Roth IRAs, current tax revenues are largely unaffected, but future revenues take the hit because withdrawals are not taxed at all. Assuming that Congress could somehow achieve true fiscal discipline, it could borrow to fund deficits caused today by conventional IRAs and repay the loans with tax revenues from future IRA withdrawals. In the real world, the Roth approach has much more appeal in Washington because it produces greater revenues today.

Both conventional and Roth IRAs differ from the Bush administration’s approach of directly cutting tax rates on capital income. The difference comes in the treatment of existing savings and wealth. The Bush tax cut conferred an enormous windfall on owners of existing wealth because it reduced taxes they had anticipated paying on assets they had already purchased. In other words, the cut changed the rules in their favor after the game had begun. (Owners of existing wealth actually got a second windfall in the form of an increase in stock prices in response to the reduction in dividend taxes; they would have received that type of windfall under any approach to reducing taxes on capital income.)

The IRA approach, in contrast, confers the benefit of zero taxes only on new saving. Under conventional IRAs, all income—from labor or capital—could be invested without going through the tax turnstile and then taxed when withdrawn. Under Roth IRAs, after-tax income from any source—labor or capital—could be invested in plans that earn tax-free capital income. Both IRA arrangements put all income on an equal tax footing, achieve temporal neutrality by cutting taxes on returns to all new saving to zero, and reduce the cost to the treasury by limiting the zero tax to returns on new saving. The Bush approach confers a significant tax advantage to existing wealth over and above the initial windfalls. Someone with $100,000 of capital income (other than interest) pays, at most, $15,000 in taxes and can invest $85,000. In contrast, an individual with high labor income has to pay $35,000 in taxes and then has only $65,000 to invest.

In 2001, the Bush administration recommended creation of lifetime savings accounts and retirement savings accounts; both are Roth-type plans with contributions of after-tax dollars. Bush proposed contribution limits of $7,500 per individual to each account, or $30,000 for a couple with no children. The administration reintroduced both plans in 2003 and 2004, but with the contribution limits reduced to $5,000 ($20,000 for the couple without children). Both proposals failed because they were opposed by liberals and by lobbies with strong vested interests in the current tax code. The lobbyists included insurance companies, which benefit from special tax treatment on what are called deferred annuities, and investment firms that administer 401(k) plans. Both feared that their plans would be eclipsed by opportunities for direct, tax-free saving by individuals. Taxation of capital income is sure to remain a highly contentious issue, and attainment of a true consumption tax is sure to remain a goal of savingsminded tax reformers.