Pigovian Taxes and Subsidies: Putting a Price on What the Market Ignores

In 1920, the British economist Arthur Cecil Pigou published a deceptively simple idea that still organizes most of modern environmental policy. If a factory's smoke imposes costs on its neighbors that the factory never pays, Pigou argued, the government should make it pay: by levying a tax exactly equal to the damage. Do that, and the factory's books would finally reflect the true cost of its smoke, and it would cut production to the level society actually wants. No need to micromanage the factory or ban anything. Just fix the price.

A century later, that instrument bears his name. A Pigovian tax is a tax on an activity equal to the external cost it imposes on third parties; a Pigovian subsidy is its mirror, a payment equal to an external benefit. Both do the same fundamental job: they put a price on something the market was ignoring, so that private incentives line up with social welfare. As the biographical entry on Pigou notes¹, his framework of taxing activities that generate negative externalities and subsidizing those that generate positive ones became the foundation for the economics of environmental and social policy.

The idea in one sentence

The whole theory reduces to a single target. Recall that a negative externality opens a gap between private cost and social cost, and the market overproduces by exactly the amount that gap distorts. A Pigovian tax is engineered to be the size of that gap. Set the per-unit tax equal to the per-unit external harm, and the producer's private cost rises until it equals the social cost. From that moment, the producer's own profit-seeking calculation (keep producing while it pays) lands precisely on the efficient quantity. The tax doesn't fight the profit motive; it redirects it.

This is why economists prize Pigovian taxes over blunt instruments like outright bans or rigid output limits. A ban assumes the activity has no acceptable level. A quota assumes the regulator knows the right quantity for every firm. A Pigovian tax assumes only that the regulator can estimate the harm per unit, and then lets each firm decide for itself how much to cut, in whatever way is cheapest for it. Firms that can reduce pollution cheaply cut a lot; firms that can't cut pay the tax. The reductions happen where they are least costly, which is exactly what minimizes the total cost of cleaning up.

A carbon tax, worked line by line

Nothing makes this concrete like running the numbers, so take the most-discussed Pigovian tax in the world: a tax on carbon dioxide emissions.

The starting point is the social cost of carbon: the estimated dollar value of the long-run damage caused by emitting one additional ton of CO2, covering climate impacts on agriculture, health, property, and productivity. Estimates vary widely with assumptions, but suppose for this example the figure is $50 per ton of CO2. That is the external harm each ton inflicts on the world, and therefore, in Pigou's logic, the size the tax should be.

Now watch it flow through real decisions:

• Gasoline. Burning one gallon of gasoline releases about 0.0089 metric tons of CO2. At $50 per ton, the Pigovian tax on a gallon is roughly 0.0089 x $50 = $0.44 per gallon. A driver filling a 15-gallon tank pays about $6.60 extra. That surcharge is not arbitrary revenue: it is the driver finally paying for the climate damage the fill-up causes.
• Coal power. A coal-fired power plant emits roughly 1 ton of CO2 per megawatt-hour of electricity. At $50 per ton, the tax adds about $50 per MWh to coal-fired generation. Natural gas, emitting roughly half as much CO2 per MWh, picks up only about $25. Wind and solar, emitting almost none in operation, pay almost nothing.
• The margin moves. Here is the engine. Before the tax, suppose coal generated power at $40/MWh and natural gas at $45/MWh: coal wins on price, so the grid burns coal. After the tax, coal costs $40 + $50 = $90/MWh while gas costs $45 + $25 = $70/MWh. Now gas is cheaper, and generation shifts toward it. Push the comparison to a renewable at $60/MWh with near-zero tax, and the renewable beats both. The tax didn't ban coal; it changed which option is cheapest at the margin, and the market reallocated on its own.

That last point is the magic of the design. No regulator ordered the grid to switch fuels. The grid switched because the tax made the dirty option genuinely more expensive, expensive by exactly the amount of harm it causes. As Resources for the Future explains in its carbon-pricing primer³, this is the central appeal of carbon pricing: it harnesses the cost-minimizing behavior of millions of firms and households to cut emissions wherever cutting is cheapest, rather than dictating specific technologies from above.

Subsidies: the same tool, sign reversed

For positive externalities, the Pigovian instrument simply flips. If installing rooftop solar confers benefits on others (cleaner air, reduced grid strain, accelerated learning that lowers everyone's future costs), the homeowner captures only part of the value and underinvests. A subsidy equal to the external benefit closes the gap. The U.S. has used exactly this logic in the federal residential clean-energy credit, which the IRS describes as covering 30% of the cost of qualifying solar and other clean-energy installations⁴, a Pigovian subsidy that pays households for the spillover benefits their installations create.

The symmetry is exact. A tax adds the missing cost so that harmful activities shrink to the efficient level; a subsidy adds the missing benefit so that beneficial activities grow to it. Both repair a broken price.

A profitable side effect, and a warning about it

Pigovian taxes also raise revenue, and that revenue can replace other, more distorting taxes: a so-called "double dividend." Tax carbon and cut income taxes by the same amount, and you have discouraged something harmful (pollution) instead of something useful (work). This is genuinely attractive. But it carries a trap: once a government depends on the revenue, it acquires an interest in the activity continuing, which can blunt the tax's whole purpose. The point of a Pigovian tax is to shrink the harmful activity, which shrinks the revenue. A tax designed to fund the budget and a tax designed to fix an externality are pulling in opposite directions, and honest policy keeps them distinct.

Where the idea gets hard

Pigou's framework is elegant in theory and genuinely difficult in practice, and the difficulty is almost entirely one problem: you have to know the size of the external cost. Set the carbon tax too low and you barely dent emissions; set it too high and you choke off activity that was worth more than the harm it caused. The social cost of carbon is estimated through complex models loaded with contested assumptions, including how to value harm to future generations, how to handle deep uncertainty, and how to price impacts in poorer countries. Reasonable economists put it anywhere from under $20 to over $200 per ton. The tax is only as good as that estimate.

There are further frictions. Externalities are often hard to attribute to specific sources, so the tax may hit a proxy (gallons of fuel) rather than the true harm (emissions in a particular location). Taxes can fall hardest on lower-income households, who spend a larger share of income on energy, which is why serious carbon-tax proposals usually pair the tax with rebates. And political economy intrudes: the industries facing the tax lobby hard against it, while the diffuse beneficiaries rarely organize.

None of these undermine the core logic, they qualify it. Pigou's insight remains one of the most powerful in economics precisely because it solves the externality problem without pretending the regulator is omniscient about how each firm should respond. It asks the regulator to estimate one number (the harm per unit), and then trusts the market to do the rest. When that number can be estimated even roughly, a Pigovian tax beats the alternatives, and a century of environmental policy, from sulfur-dioxide trading to carbon pricing⁵, is built on the strength of that bet.