← Back to the map

Net Zero Cost Benefit

What are the benefits of net zero?

Scroll down ↓

The benefit of climate policy is the climate damage it avoids.

Climate economists have spent more than 40 years estimating climate damages.

Each dot is one such estimate.

A few studies find that climate change can be a net gain (for example, fewer cold deaths and better crop growth).

But most find that climate change is a net loss.

One summary of these studies comes from Richard Tol, one of the world's most cited climate economists.

His study sums up all the estimates in one curve: the warmer it gets, the worse off we are.

Another summary comes from William Nordhaus, the only climate economist to win the Nobel Prize.

Nordhaus adds extra costs for tipping points and for damages the studies leave out.

With current policies, the world warms by about 2.9°C by 2100.

That means an annual cost in 2100 of 1.8–2.8% of world GDP.

This site uses the average of the two damage curves.

That means a damage equivalent to 2.3% of world GDP in 2100.

If the whole world goes net zero, warming in 2100 is about 1.5°C instead.

The benefit is the difference: 1.6% of world GDP in 2100.

That is about $9 trillion a year.

You may have heard of much higher estimates. Some are included in Nordhaus's summary.

The best known of the others was retracted. Others have serious problems, explained in the notes.

Notes
  1. Damage curves. Tol combines 69 estimates from 39 studies, published from 1979 to 2021, into one central curve, a weighted average of several curve shapes (Tol 2024); values here come from the author's own workbook. Nordhaus's summary (Barrage and Nordhaus 2024) tabulates 56 estimates: 1.6% of output lost at 3°C, plus 1% for tipping points such as the melting of the Greenland ice sheet (Dietz et al. 2021), plus 0.5% for damages the studies leave out, in total 3.1% at 3°C. All three grow with the square of warming, so at the 2.86°C of current policies in 2100 they are 1.45%, 0.91% and 0.45%, 2.8% in total. The two summaries share many of the same studies. The dots show Tol's estimates first and then Nordhaus's; dots outside the chart (such as Burke et al. 2015) are not drawn. Note that the two summaries place some of the same studies at slightly different temperatures. Nordhaus converts each study's warming to pre-industrial, mostly by adding 0.4°C (Barrage and Nordhaus 2024); Tol takes the studies' numbers as given (Howard and Sterner 2017 discuss the issue). Shifting Tol's curve the same way would change the benefit very little (from 1.55% to 1.50% of GDP).
  2. Who they are. Richard Tol (University of Sussex) has about 44,000 citations and an h-index of 112 in OpenAlex (2026). William Nordhaus (Yale, 1941–2026) received the 2018 Nobel Memorial Prize in Economic Sciences "for integrating climate change into long-run macroeconomic analysis" (Nobel Prize Outreach 2018).
  3. Temperatures. Warming above pre-industrial, from the MAGICC 7.5.3 climate model: current policies give 2.86°C in 2100 (the interpolated scenario between two AIM scenarios, matched to CMIP7 Medium, the UN Climate Panel's new medium scenario, which itself gives 2.87°C; see Temperature); with all countries at net zero by 2050, 1.48°C (intermediate air pollution).
  4. The benefit. At 2.86°C, damages are 1.8% of GDP (Tol) and 2.8% (Nordhaus), on average 2.3%. At 1.48°C, they are 0.74% and 0.75%, on average 0.75%. The difference, 1.55% of world GDP in 2100, is the avoided damage. With world GDP in 2100 of about $548 trillion (AIM's projection for current policies, 2026 dollars), that is about $8.5 trillion a year. For other years, the same calculation is done with each year's temperatures and the projected size of the world economy. The benefit is global: it accrues to everyone, not just to the countries that cut emissions. Added up over 2027–2100 and turned into today's money with the same discount rate as the costs (Nordhaus's discount rate; see Costs), world net zero brings $24 trillion of benefits.
  5. How much do the choices matter? The result depends on two choices: the damage estimate and the discount rate. The table shows the global climate benefit per dollar of cost, 2027–2100, with each damage summary alone and with lower discount rates, down to no discounting at all, the most generous assumption. For the world and the major economies in the table, the benefit stays well below the cost in every case.
    This siteTol onlyNordhaus only3% a year2% a yearNo discountingNordhaus, no discounting
    World9¢7¢12¢11¢13¢17¢23¢
    OECD12¢8¢16¢14¢17¢24¢32¢
    United States15¢10¢21¢18¢22¢32¢43¢
    EU2714¢9¢19¢16¢20¢28¢39¢
    Germany14¢9¢19¢17¢20¢28¢38¢
    United Kingdom11¢7¢14¢12¢15¢21¢28¢
    Japan9¢6¢12¢11¢13¢17¢23¢
    China10¢7¢14¢13¢16¢27¢36¢
    India11¢7¢15¢13¢15¢21¢28¢
    For a few poor countries with high emissions for their income, such as Pakistan, Nepal and Chad, the benefit can exceed the cost under the most generous assumptions. That is because the cost model does not model them separately: each country gets its region's cost as a share of its GDP, while its benefit follows its emissions. These countries emit two to five times more per dollar of GDP than their regions, mostly from livestock, oil and gas, so their net zero would probably cost much more than a GDP share suggests. As a rough illustration, if their cost is scaled up by how much more they emit per dollar than their region, they get 15–35¢ back per dollar, in line with other countries. The full table for every country is in the download.
  6. Cleaner air. Standard cost-benefit analyses of climate policy, including Nordhaus's Nobel-winning work, count only the climate damage avoided. Net zero would also cut air pollution. But remember that much of that pollution falls anyway: as countries get richer, they clean up their air. Under current policies, world sulfur emissions fall by about a third by 2050 and two-thirds by 2100. Even in AIM's 2020 policies, where CO₂ emissions keep rising, sulfur falls by almost 60% by 2050. Net zero cuts this somewhat further. The air-pollution paths come from CMIP7 and Sun et al. (2021), which, split by country on this site, start from 2025 levels fitted to historical data, not from AIM, the cost model, which has its own. They agree on how much more net zero cuts: about half of the remaining sulfur pollution in 2050 and 2100 (AIM: 60% and 50%). An estimate of the extra health gain from net zero, with each life-year saved valued as in Robinson et al. (2019), gives a side benefit of roughly 0.1–0.2% of world GDP. While not trivial, it does not change the comparison in any decade from 2030 to 2100: net zero costs 7.2% of GDP in 2050 and still about 4.3% in 2100, while the climate and health benefits together are about 0.3–0.4% of GDP in 2050 and under 2% in 2100. Crucially, much of the health gain could be had far more cheaply with targeted pollution controls, such as scrubbers, filters and vehicle standards.
  7. After 2100. This site reports costs and benefits until 2100, as most climate analysis does. Climate damages keep going after 2100, but so do the costs: net zero must be kept up indefinitely, and in 2100 it still costs about 4.3% of GDP, against climate and health benefits of under 2%. If the trends of 2080–2100 simply continued, benefits might overtake costs sometime in the 22nd or 23rd century, but only after more than a century of much larger costs. Nordhaus's DICE model, which counts damages for centuries, finds in its latest version that the optimal policy keeps global CO₂ emissions only modestly below today's in 2050 and lets warming reach about 2.6°C by 2100 (Barrage and Nordhaus 2024).
  8. Higher estimates. Burke, Hsiang and Miguel (2015) estimate a 23% loss at 5.3°C; Nordhaus's summary includes it at half weight. The best known higher estimate, by Kotz, Levermann and Wenz (2024), found 19% lower income by 2050. It was retracted in 2025 because its errors were too large to correct (retraction notice). Studies of this kind estimate how growth moved in years that happened to be warmer or cooler, then assume the effect repeats and compounds every year to 2100. Re-examinations find that a few unusual observations drive much of the result (Curtin and Burgess 2026; Barker 2024), that reasonable alternative choices give anything from large losses to gains (Newell, Prest, and Sexton 2021), and that the global estimate of 12% lower GDP per degree depends heavily on a few years (Goulet Coulombe and Klieber 2025).
References