The Corrective

Every tax in this volume so far has been a necessary evil: it raises revenue, but at the cost of the deadweight loss, the trades it destroys, the surplus it forgoes. The art of taxation, on this understanding, is to minimize the damage — to raise the revenue with the smallest loss. But there is one tax that escapes this logic entirely, that does good rather than harm, whose deadweight loss is not a cost to be minimized but a benefit, because the trades it destroys are trades that should not have happened. This is the corrective, or Pigouvian, tax — the tax on a harm — and it is the great exception to everything the volume has so far argued.

The corrective tax answers a failure the volume on the externality diagnosed. Some actions impose costs on others that the actor does not bear — the factory whose smoke fouls a neighborhood, the driver whose car adds to congestion, the emitter whose carbon warms the world. These are negative externalities: costs spilled onto third parties, outside any price. And because the actor does not pay these costs, he does too much of the harmful thing — the factory makes too much smoke, the driver drives too much, the emitter emits too much — for he weighs only his private cost, not the larger social cost his action imposes. The market, which coordinates beautifully when each bears the cost of his own acts, fails here, overproducing the harm, because the price the actor faces is too low: it omits the cost borne by others. The result is a wedge of the opposite kind from a tax's — a gap not between buyer and seller but between private cost and social cost, with the actor's private cost below the true social cost, and the harmful activity expanded into the gap.¹

Arthur Pigou saw the remedy: tax the harm by exactly the cost it imposes on others, and the gap closes. If the factory is made to pay, in tax, a sum equal to the damage its smoke does to its neighbors, then the cost it omitted is restored to its reckoning: its private cost, plus the tax, now equals the true social cost, and it will produce just the amount of smoke that is worth its full cost to society — no more. The tax does not distort the factory's decision; it corrects it, making the factory weigh the whole of what its action costs, the part borne by others as well as the part borne by itself. The Pigouvian tax thus aligns private incentive with social cost — the same alignment the market achieves on its own when there are no externalities, here restored by the tax where the externality had broken it. And it sets the right amount of the harm: not zero, for some smoke may be worth its cost if the factory's product is valuable enough to justify the damage, but the amount at which the marginal benefit of the activity just equals its marginal social cost. The tax is the price the market failed to charge — the cost of the externality, levied at last on the one who imposes it.²

Here is the inversion that sets the corrective tax apart from every other. An ordinary tax destroys trades that should have happened — mutually beneficial exchanges, killed by the wedge — and that destruction is the deadweight loss, the cost of the tax. The corrective tax destroys trades too: it raises the cost of the harmful activity and so reduces it. But the trades it destroys are trades that should not have happened — units of smoke, of congestion, of carbon, whose true social cost exceeded their benefit, and which existed only because the actor was not paying their full cost. To destroy these trades is not a loss but a gain: society is better off without them, for they were destroying more value, in harm to others, than they created, in benefit to the actor. So the corrective tax has a deadweight loss that is negative — it improves efficiency rather than reducing it, mending the market's overproduction of the harm. Where the ordinary tax buys revenue at the price of lost good trades, the corrective tax is paid for by the prevention of bad ones; its distortion is the correction of a prior distortion, and the wedge it drives closes a gap rather than opening one.³

The corrective tax offers, then, a kind of free lunch — or rather a double one. First, it corrects the externality: it reduces the harm to the efficient level, mending the market failure, a benefit in itself worth having even if the revenue were thrown away. Second, it raises revenue — and revenue that can be used to reduce other, distorting taxes. This is the double dividend: the corrective tax both improves the allocation of resources, by curbing the harm, and yields money that can replace the deadweight loss of an ordinary tax, by funding a cut in, say, the income tax. A tax on carbon, set at the social cost of emissions, would curb warming to the efficient degree and raise revenue to cut the taxes on work and saving — taxing a bad to spare the taxing of goods. This is why economists, across the spectrum, so widely favor corrective taxes: they are the rare instrument that raises revenue while improving, rather than harming, the economy's efficiency — the one tax that taxes what we want less of, and so should be the first tax levied and the last repealed. The deep maxim of the corrective tax reverses the whole logic of the others: tax bads, not goods.⁴

The corrective tax is not without difficulty, and its limits return us to the volume's theme. To set it correctly, one must know the size of the external cost — the damage the smoke does, the harm of a ton of carbon — and this is hard to measure, so the tax is set by estimate, and may be too high or too low. Its incidence, moreover, obeys the same logic as any tax: the corrective tax, like all taxes, shifts, and may settle on parties other than the polluter — the carbon tax passed forward to consumers in higher prices, the burden landing where elasticities send it, not necessarily on the emitter the tax meant to discipline. And there are harms too diffuse, too hard to attribute, for any tax to reach. But the corrective tax remains the great exception and the great lesson: that a tax need not be a drag on the economy, that the wedge can mend as well as wound, that the instrument we have treated as a necessary evil can, rightly aimed, be a positive good. With this exception noted, the volume can turn to its deepest question, which the corrective tax has only sharpened. For even the corrective tax shifts — its burden, like every tax's, settling not where the law aims it but where the market sends it. Who, in the end, bears a tax — any tax — and what does the answer reveal about the power of the state that levies it? The next plate is the flagship: the burden.⁵

The corrective tax is the great exception to the economics of taxation: the tax on a harm, which corrects rather than distorts. Where the externality leads the market to overproduce a harm — smoke, congestion, carbon — because the actor does not bear its full social cost, the Pigouvian tax restores the missing cost, taxing the harm by the damage it does, until the actor reckons the whole of what his action costs and does only the amount worth its full cost. Such a tax has a negative deadweight loss: it destroys not good trades but bad ones, improving efficiency rather than reducing it; and it pays a double dividend, mending the market and raising revenue to spare the taxing of goods. Tax bads, not goods, is its maxim, reversing the logic of every other tax. Yet the corrective tax, for all its virtue, shifts its burden like any other, settling where the market sends it rather than where the law aims. And that returns the volume to its central question, now ripe for its deepest answer: who really bears a tax, and what does the answer tell us about the state that levies it? The next plate is the burden.