Finance & Development, March 2015, Vol. 52, No. 1

Ruud De Mooij and Michael Keen

PDF version

It is hard to design a fair and efficient revenue system

Tax policy is often guided by simple rules of thumb. Sometimes, they are strikingly right. But sometimes they can be dangerously misleading.

There is an adage, for example, that “an old tax is a good tax.” That may be true for, say, the property tax. But taxes on windows and beards are long gone, import tariffs are in decline, and new levies, such as the value-added tax, have gained ground. Designing a good tax system requires more than a good slogan.

We dealt before with the basic principles of taxation (see “Back to Basics” in the December 2014 F&D). Now we apply them to some central and current tax policy debates.

Personal income

The great appeal of the personal income tax is that it taxes people on an indicator of their ability to pay, collecting progressively more from those with higher incomes. But the indicator is imperfect, because the government cannot be sure whether a high income results from intrinsic talent or luck—which will not be affected by taxation—or hard work and creativity—which might be. Taxing income might not only discourage effort (not just hours worked, but also, for example, entrepreneurial activity and striving for promotion), but can also give rise to tax avoidance and evasion.

The design of the personal income tax, therefore, revolves around a fundamental trade-off: progressive taxes support equity objectives, but can reduce efficiency. As long as people have differing views on what is equitable, there will never be universal agreement on the best tax schedule. But careful theory and empirical evidence have illuminated key considerations.

There is, for instance, the need to consider not only the personal income tax but all taxes and all income support measures—such as the Earned Income Tax Credit in the United States, which gives money to low-wage workers in amounts gradually reduced as income increases. Income support is simply negative income taxation and, when withdrawn as income rises, acts just like a tax on that additional income.

There is indeed a strong case for subsidizing earnings of low-wage workers, because their willingness to work is particularly sensitive to tax, and it is cheaper to ensure their well-being when they are working. But while the average tax rate at the bottom will therefore typically be negative, the effective marginal rate—the additional tax paid (or benefit not received) when income rises by one dollar—should be positive. Otherwise the subsidy will extend to all taxpayers, including those who do not need it. Targeting income support at the poorest limits the revenue cost of earnings subsidies and can be consistent with efficient redistribution, even though it may create high effective marginal rates for the poorest.

The proper tax rate structure for high-income earners has always been contentious. Many have concluded that the best-off could be taxed at marginal rates of 60 percent or more without leading to reduced effort or avoidance or evasion large enough to cause the amount of tax they pay to fall. If raising revenue were the only concern, that would be fine. But well-off taxpayers would suffer, which presumably matters for overall social welfare. Moreover, some analysts believe that the calculations underlying the optimal marginal rate fail to capture adverse effects on entrepreneurship.

In broad qualitative terms, the optimal marginal rate structure should thus have a U shape—starting high to recoup support to the very poorest, falling to preserve incentives for the people in the middle, and finally rising to secure revenue from the better-off. This runs counter to the idea that marginal rates should always increase with income, but is consistent with the more basic notion that the average rate should increase with income. All this, however, still leaves plenty of room for debate on the precise shape of that U.

Capital income conundrum

Capital income—interest, dividends, and capital gains—is in most countries received largely by the better-off. High taxes on capital income (or on the underlying wealth) are therefore often viewed as a good way to address inequities. But theory offers further perspectives on this issue.

Capital income enables consumption in the future. Taxing it increases the cost in terms of consumption forgone today. Prudent people who prefer to postpone consumption (or transfer it to their heirs) will be taxed more than those who do not. Some see this as violating horizontal equity (the principle that those who are in all relevant respects identical should be treated the same) on the grounds that time preference is not a legitimate basis on which to differentiate tax liabilities. Moreover, by discouraging saving, a tax on capital income may create relatively large deadweight losses (those incurred from transferring resources out of the private sector).

What all this implies is intensely debated among public finance economists. At one extreme is the view that because it distorts behavior so much, the optimal tax on capital income is zero, with redistribution better achieved by a progressive tax on labor income alone. At the opposite extreme is the view that labor and capital income should be taxed identically—for many years the most popular view. Neither view is on entirely firm theoretical grounds. What has become clear, however, is that the desirable tax rate on capital income, even if not zero, may well differ from that on labor income—not least because capital is more mobile internationally, making it harder to tax without driving the base abroad. Many countries now employ some form of dual income tax, taxing capital income separately from labor income, and at a relatively low rate.

Corporate tax controversies

The notion of tax incidence—who ultimately bears the real burden of a tax—is key when it comes to business taxation—and can lead to the surprising conclusion that much of the incidence might fall on workers. Take an economy that is small in world capital markets, and so must take as given the after-tax rate of return on investment: investors will move their capital abroad if they earn less than this rate. If a country now taxes the returns that investors earn there, the before-tax rate of return will have to rise by enough to leave the after-tax return unchanged. An outflow of capital is then required. But that outflow leads to a lower domestic capital-labor ratio, which reduces labor productivity—and, in turn, wages. So workers, not shareholders, bear the real incidence of the corporate income tax. Because it is more efficient to tax workers directly than indirectly through the corporate tax, the optimal corporate income tax for such an economy is zero.

But there are important qualifications.

First, normal capital returns (the minimum return required by investors) should be distinguished from above-normal returns—called “rents.” Unlike normal returns, rents that are specific to a particular country can be taxed without affecting investment (think, for instance, of natural resource rents). The traditional corporate income tax, however, is not a rent tax because it taxes all returns to equity, both normal and above normal. It can be turned into a rent tax, though—for instance by allowing companies to reduce their taxable income through a deduction for normal equity returns. Some countries have moved in this direction.

Second, practical considerations loom large. The corporate income tax, for example, effectively taxes earnings that companies retain, which are hard to tax at the personal level. Similarly, if there were no corporate tax, small businesses could escape tax by incorporating and labeling their earnings capital income. Moreover, in many developing economies it is relatively easy to collect taxes from a few large companies.

Taxing consumption?

A uniform tax on consumption—applying the same rate to all goods and services—is broadly equivalent to a uniform tax on wage and profit income. It simply operates on the other side of an individual’s budget, so its distortions on the labor market should be similar too. Because income taxes fit better with the principle that people should be taxed on their ability to pay, why should governments tax consumption at all?

There are practical reasons to do so: taxing both income and consumption reduces compliance risk by diversifying the government’s revenue base. But there are also more fundamental justifications, such as taxing particular types of consumption to address externalities, which are effects, good or bad, on those not involved in the underlying transaction—pollution, for example. Such taxes could also address other problem behaviors, such as drinking and smoking. Another reason is that differential rates might help reduce tax-induced disincentives to work. Empirically, however, it has proved hard to identify elements of rate differentiation that are warranted on such efficiency grounds—perhaps with a few exceptions, such as child care services.

Many feel that necessities such as food should be taxed at especially low rates because the poor spend a large proportion of their income on them. But this is a costly way to pursue equity because while the poor spend a larger proportion of their income on necessities, the rich spend a larger absolute amount and so benefit most from a low rate. Almost all advanced economies, and many others too, should have devices better suited to pursuing their fairness goals—such as income-related transfers or other forms of cash support to the neediest, or public support for housing, health care, and basic education.

There is a fair degree of professional consensus that a uniform, broad-based consumption tax is a sensible benchmark for good policy—with little convincing rationale (externalities aside) for rate differentiation. This is one simple rule of thumb that gives good, practicable advice—but it rests on quite detailed empirical and theoretical reasoning. Policymakers must be wary of the many that don’t. ■