As the “fiscal cliff” approaches, it’s no surprise that America’s politicians are still tussling over what tax rates to set. Any change to the tax code is likely to make somebody worse off—especially if the aim is to increase overall revenue, as in the US today. But you’d think economists could offer a more objective opinion on how best to tax, or at least narrow the options. In fact, economics is less helpful than you’d think.
The textbook tenets of good tax policy are simplicity, fairness and efficiency. Economists, presumably feeling they have no special expertise on the first two points, have tended to concentrate on what makes a tax code efficient—in other words, on how to raise any given amount of revenue with the least collateral damage to the economy. But that’s a difficult question and the list of clear recommendations from the academy is pretty thin.
Certainly, low rates applied to a broad tax base are bound to distort economic decisions less than high rates applied to a narrow base. But how much such distortions matter in the first place is harder to say. For instance, do high income-tax rates discourage people from working more (because the rewards will be smaller) and thus retard economic growth to any significant extent? It’s an open question.
For sure, there’s a pretty strong consensus that raising the top rate would not reduce effort and/or increase tax avoidance so much that revenues would actually fall. The revenue-maximizing income tax rate has been estimated for the US at 50%-70% or even higher. A top rate of 39.6%, which will take effect if the “fiscal cliff” kicks in at the end of this year and George Bush’s tax cuts expire, is much less than the rate at which the famous Laffer curve turns down. (Although as Bruce Bartlett discusses, raising the top rate to 50% in the UK in 2010 caused people to take more measures to avoid the higher tax than policy-makers had expected.)
In any case, “what rate will maximize revenues?” is the wrong question. Increasing the top rate to 39.6% might increase government revenues in the short term, but could still reduce GDP growth. And if higher taxes reduce growth, then in the long run they will reduce revenues as well, because the tax base will be smaller than it would otherwise have been.
The vital link between taxes and growth is hard to define. For one thing, it depends on what the taxes are used for—or, equivalently, on which expenditures get cut if taxes are reduced. A tax cut that increases public borrowing might lower growth by crowding out private investment. A tax cut that’s financed by reducing wasteful spending—on farm subsidies, say—would likely be good for growth.
Do people work less when income taxes rise? Researchers who’ve looked at the immediate effects tend to conclude they’re small (pdf). This isn’t so surprising. Workers typically aren’t in a position to fine-tune the hours they work according to their tax rate. At the so-called “intensive margin”, the effects look insubstantial. But studies that look at the “extensive margin”—the decision whether to work or not, as when a couple decides whether to send both earners out into the labor market or one—find a further effect.
Again, the question of short term and long term complicates the analysis. Higher taxes may influence decisions that affect incomes and growth over years and decades—whether to go to college, whether to study for a two-year or four-year degree, whether to prefer occupations that make fewer or more demands, when to retire. Given time, in other words, labor supply can respond in many different ways. Elusive as all these connections may be, it’s suggestive, at the very least, that Europeans with their higher taxes work fewer hours than Americans (pdf).
In short, what an efficient tax system would look like is much harder to say than you’d think. Here’s another example: the notion that investment income should be taxed more lightly than earnings.
This has long been a standard principle of efficient tax design. The basic idea is that taxing investment income—the return on savings—distorts the choice in favor of consumption now and against consumption in future. In theory, the aim should be to tax consumption at the same rate whenever it happens, which you can do by taxing labor income but not savings.
On the face of it, this is a rationale for tax-sheltered savings vehicles, and (fairness permitting) for low taxes on dividends and capital gains. It’s also why many would-be tax reformers advocate shifting the tax base away from income (whether wages or other kinds) and towards consumption, such as sales or value-added taxes.
But what the prejudice against taxing capital forgets is inheritance, as Thomas Piketty and Emmanuel Saez, foremost among today’s public-finance scholars, have recently shown (pdf). If wages were the only source of wealth, as standard models assume, it would make sense to tax earnings and not savings. Allow for inheritance, however, and an efficient tax system ought to tax bequests. Under certain plausible assumptions, taxing someone’s capital over her lifetime is a good substitute for taxing bequests appropriately. Contrary to the usual thinking, Piketty and Saez conclude, considerations like these “can account for the actual structure and mix of inheritance and capital taxation”.
Put it this way: Efficiency is a can of worms.
Corollary: In tax design, the axiom of simplicity has been underrated.
The main benefit of a simple tax code, to an accountant’s way of thinking, is that it’s less trouble to comply with. But this may be the least of its advantages. It’s commonly supposed that simplicity militates against fairness (for instance, a flat sales tax is simple but regressive), but the opposite is more likely to be true.
A simple tax code has a better than average chance of being fair, because people of similar means will be more likely to pay similar amounts of tax: Simplicity favors horizontal equity. Tax incidence (who bears the burden) will be easier to grasp and anomalies easier to see. Voters will demand fairness if the code is transparent enough for them to make the judgment.
The complications in the current US tax code make that next to impossible. Taxpayers with similar incomes pay very different rates. Complexity makes it harder to gauge how progressive the system is as well, because the schedule of marginal rates has little to do with the taxes people actually pay. (A top rate of 90% doesn’t make the system more progressive if nobody falls into that band.) And avoidance is harder in a simple system, so people who can afford to hire smart advisers get less benefit from doing so.
In theory, simplicity may be something you trade off against fairness. In practice, they’re likely to be complements.
A simple tax code also has efficiency advantages. If the aim is to minimize distortions, a system that treats different kinds of income and spending the same has a head start. As just noted, a simple tax code also makes avoidance harder—so, other things being equal, the same amount of revenue can be collected with lower rates. That’s another efficiency bonus.
No doubt getting to a simpler system will be difficult. Capping itemized deductions—a first step toward eliminating them—makes sense, but today’s code is encrusted with preferences and other complications for a reason: The beneficiaries have fought for them, and won’t yield them up without a struggle. Nonetheless, simple should be the goal. Succeed in that, and greater fairness and efficiency will likely follow.
