Summary of the AER Paper

A WELFARE ANALYSIS.

of Policies Impacting Climate Change

∼ ∼ ∼

Governments spend trillions on climate policy every year — but which dollars go furthest? We evaluated 96 climate-related tax and spending programs using a unified economic framework, and found that the conventional wisdom gets several things wrong.

BY HAHN · HENDREN · METCALFE · SPRUNG-KEYSER

Part 1 · The Framework

One yardstick for every policy

Most climate policy debates are fragmented. Analysts compare carbon taxes using cost-per-ton metrics. Clean energy advocates cite job creation. Economists debate fiscal multipliers. These conversations rarely speak to one another, making it nearly impossible to say which policies are genuinely worth pursuing.

This paper applies a single unified metric — the marginal value of public funds (MVPF) — across the full range of climate policy. The MVPF asks: for every dollar the government spends net of revenues, how much value do the beneficiaries receive in return?

"An MVPF above one means a policy creates more welfare than it costs government. The higher the MVPF, the better the policy compares to simply giving money back to taxpayers." — Hahn, Hendren, Metcalfe & Sprung-Keyser (2026)

We also extend the MVPF framework in an important new direction: incorporating learning-by-doing spillovers. When government subsidies for clean energy technologies drive down costs through scale and experience, future buyers benefit even without a subsidy — and that future benefit belongs on the ledger.

⚙️

Novel Contribution

The learning-by-doing extension

Prior MVPF analyses treated climate subsidies as one-time transfers. But clean energy subsidies do something more: they push industries down the cost curve. Every wind turbine built today makes the next one cheaper to manufacture. This paper is the first to systematically incorporate these dynamic spillovers into the MVPF framework — a methodological innovation that significantly changes the ranking of production-side subsidies relative to demand-side policies.

Interactive Policy Explorer

Where does each dollar go furthest?

Each dot is one of the 96 policies. Hover for details. Subsidies and taxes are shown separately — the direction of "better" flips between them. Values are illustrative; see the paper for complete estimates.

The Sample

What counts as a climate policy?

The 96 programs span five broad categories — from blunt fiscal instruments to targeted behavioral interventions.

Clean Energy Subsidies

Transportation Programs

Carbon Taxes & Cap-and-Trade

Energy Efficiency Programs

Conservation Nudges

Part 2 · The Findings

Clean energy production wins

Of all 96 policies examined, subsidies for the production of clean energy — particularly the wind production tax credit — deliver more welfare per dollar than anything else in our sample. Including EV subsidies, which receive far more political attention.

The reason comes down to learning-by-doing. When the wind industry scales up because of a production credit, turbine costs fall for everyone — not just today's buyers. That cost spillover, accumulated across the global energy transition, is substantial. When you put it on the ledger, production subsidies look dramatically more efficient than they do in conventional analyses.

Nudges work — when geography cooperates

Conservation nudges — programs that send households reports comparing their electricity use to their neighbors' — have large MVPFs, but only in the right places. The electricity savings these programs generate only reduce emissions if the local grid is dirty enough to mean that less consumption translates into fewer fossil fuels burned.

Target a nudge at a household running on coal-heavy power, and the MVPF is very high. Target the same household on a nearly carbon-free grid, and the climate benefit shrinks nearly to zero. Location is everything.

"Conservation nudges have large MVPFs when targeting regions with dirty grids." — From the paper's abstract

Try it · Interactive

How clean is your grid? Drag to find out.

🟢 Clean grid
(hydro, nuclear, wind) ⚫ Dirty grid
(coal, gas heavy)

Nudge program MVPF

0.6

Low — limited climate benefit

🟢

On a clean grid, electricity savings produce almost no emission reductions. The nudge has limited climate value here.

Carbon pricing is underrated

Fuel taxes and cap-and-trade programs are politically unpopular — they put a visible price on something people pay for every week. But the analysis shows they are among the most efficient instruments available, both as climate tools and as revenue raisers.

A well-designed fuel tax or carbon market raises substantial government revenue while internalizing the costs that emissions impose on society. By the MVPF metric, these programs are highly efficient — meaning the welfare generated per dollar of net government cost compares favorably to many popular subsidy programs.

Two metrics, sometimes different answers

Alongside the MVPF analysis, we also calculate traditional cost-per-ton-of-CO₂ estimates for each policy. These two frameworks are asking subtly different questions, and they don't always agree on which policies to favor.

Cost-per-ton asks: how cheaply can we eliminate a unit of emissions? MVPF asks: how much value does this program create per dollar of net government cost? A policy can score well on one metric and poorly on the other — which is exactly why the choice of yardstick matters so much for policy design.

Metric	What it measures	Strengths	Limitations
Cost per ton of CO₂	Government expenditure per unit of emissions reduced	✓ Simple to compute ✓ Widely understood	✗ Ignores distributional effects ✗ Ignores fiscal costs
MVPF	Welfare created per dollar of net government cost	✓ Captures fiscal incidence ✓ Comparable across policy types	✗ Harder to estimate ✗ Less familiar to policymakers

The paper shows cases where cost-per-ton rankings and MVPF rankings diverge — sometimes sharply. A subsidy that looks expensive per ton can still have a high MVPF if it is well-targeted to high-value beneficiaries. Conversely, a cheap-per-ton policy might score poorly on welfare if most of its fiscal cost benefits those who would have acted anyway (low additionality).

🌬️

Wind beats EV subsidies

Production-side subsidies for wind and solar have higher MVPFs than vehicle subsidies, driven by learning-by-doing spillovers.

📍

Nudges depend on the grid

Conservation nudges are highly effective in coal-heavy regions — but provide little climate benefit where electricity is already clean.

⚖️

Carbon pricing is efficient

Fuel taxes and cap-and-trade are among the most efficient policy instruments — even if they are politically costly to implement.

🔄

Cost-per-ton tells a different story

Traditional cost-per-ton estimates and MVPF rankings sometimes diverge, highlighting why the metric choice matters enormously for policy.

Part 3 · Policy Implications

What should governments do differently?

01 Shift subsidies toward production, not purchase. Wind and solar production credits outperform EV purchase subsidies — partly because production-side policies drive learning-by-doing spillovers that benefit future adopters at no additional fiscal cost.
02 Target nudges geographically. Conservation nudge programs are cost-effective only where the electricity grid is carbon-intensive. Deploying them on clean grids wastes program funds on emission reductions that don't materialize.
03 Take carbon pricing seriously. Despite their political difficulty, fuel taxes and cap-and-trade programs are among the most efficient instruments available. Policymakers who rule them out are leaving substantial welfare gains on the table.
04 Use the right yardstick. Cost-per-ton estimates and MVPF rankings can tell different stories about which policy to prioritize. Analysts and policymakers should be explicit about which question they are trying to answer before choosing a metric.