The Montreal Protocol worked because it was designed to: it shut
non-members out of the trade, paid poorer countries to switch, and let its
members tighten the rules as the science improved. Ozone-depleting
chemicals are 99% gone and the Antarctic recovery is now measurable, but the ozone
layer needs about forty more years, and some gases still slip past the
treaty.
Ninety-nine percent of the ozone destroyers are gone
The Montreal Protocol worked because it was built to be joined and hard to stay out of: it cut off trade with countries that refused to sign, paid poorer countries to switch, and let its members tighten the rules as the science improved. Since it was signed in 1987, the world has phased out 99% of the chemicals that were thinning the ozone layer, roughly 1.8 million tonnes once each chemical is weighted by how much ozone it destroys10.
The ozone layer has started to heal as a result, and scientists can now show that the treaty is the cause14. It is slow, though: the hole over Antarctica isn’t expected to close until around 206612.
Those chemicals were in ordinary things. Chlorofluorocarbons (CFCs) and their relatives cooled fridges and air conditioners, blew the foam in cushions and insulation, and pushed the spray out of aerosol cans, so getting rid of them meant redesigning a lot of everyday products.
That's the share the world has already phased out under the Montreal Protocol
The ozone hole is closing, slowly
Cutting the chemicals was the fast part. The chlorine already up there lingers for decades, so the total reaching the stratosphere fell by only 11.5% between its 1993 peak and 202012. Over Antarctica, where the damage is worst, the ozone-destroying gases were 28% of the way back to their 1980 level at the start of 2024. Over the mid-latitudes they were 55% of the way13.
In 2025 a study that looked for the treaty’s chemical fingerprint in the ozone record found strong statistical and physical evidence that the cuts are what is driving the Antarctic recovery, not a lucky run of weather14.
The recovery is not even, though. Outside the polar regions, ozone in the lower stratosphere has shown no sign of coming back yet, even though models say it should12. And inside the Antarctic hole, October ozone in the middle stratosphere has fallen 26% since 2004, which researchers link to changes in the winds rather than to chemicals15. That doesn’t undo the Antarctic result, but it suggests the treaty’s success is proven where the damage was worst and still unconfirmed across much of the rest of the globe.
How long until the ozone hole over Antarctica has healed.
A treaty built so that joining paid
The popular story is that the treaty worked because industry had a replacement ready. It didn’t. When the Protocol was signed in September 1987, the leading replacement refrigerant, HFC-134a, existed only in laboratory quantities, and the first commercial plants did not open until 199018. The rules came three years before the substitutes.
The treaty relied on design instead, and its first lever was trade. From 1 January 1990, every member had to ban imports of the controlled chemicals from any country outside it19. A country that stayed out to keep making CFCs would soon have had nobody to sell them to.
He added that everybody joined in the end, and the restrictions never had to be used20.
Poorer countries got ten extra years to comply if they used less than 0.3 kilograms per person19. In 1990 the members also set up a Multilateral Fund to pay the extra costs of switching. By 2025 donors had pledged $5.14 billion, and the fund had approved $4.49 billion for projects that phased out 510,094 tonnes of ozone-depleting chemicals, weighted by their potency1.
A ratchet let the treaty keep up with the science. Expert panels report to the members at least every four years, and on the strength of those reports the members can speed up the phase-out by a two-thirds vote. The change then binds every member six months later, without going back to national parliaments19.
That ratchet did most of the work. The 1987 text only required CFC use to be halved21, and later rounds tightened it into the near-total phase-out the world has today10.
The chemists, diplomats and companies behind it
In 1974 two chemists, Mario Molina and Sherwood Rowland, warned in the journal Nature that CFCs could destroy ozone high in the atmosphere, work that later won them the 1995 Nobel Prize in Chemistry, shared with Paul Crutzen22. In May 1985 Joe Farman, Brian Gardiner and Jonathan Shanklin of the British Antarctic Survey published the measurement that made the threat visible. Over their Halley station, the ozone layer had thinned to about two-thirds of its thickness in earlier decades23.
Mostafa Tolba, who ran the UN Environment Programme for 17 years from 1975, drove the talks from the Vienna Convention of 1985, the framework treaty Montreal was built on, to the Protocol two years later24.
The controls they won went well beyond what the big producers wanted. In April 1987 DuPont and many European governments favoured a 20% cut and the US Environmental Protection Agency 95%, and the treaty settled on 50%21.
Industry came round for reasons that had little to do with patents. DuPont did hold a 1980 patent on a way to make HFC-134a26, but its officials later said the expiry of old patents played no part in their decision27. The draw was money: the new chemicals were expected to sell for five to ten times the price of CFCs27, and only a global treaty could guarantee them a market. In September 1986 an industry coalition of about 500 producers and users announced it would accept international controls25.
Much of the switch needed no new chemical at all. Different product designs, recycling and plugging leaks replaced up to 85% of ozone-depleting uses, and most of those fixes were never patented18.
Two-thirds of the ozone gone by 2065, without the treaty
Without the treaty, 17% of the planet’s ozone would have been gone by 2020 and two-thirds by 2065, according to a NASA-led study that assumed CFC production kept growing 3% a year28. By then, across the northern mid-latitudes (30 to 50 degrees north, a band that takes in Washington, Rome and Beijing), the midsummer UV index would pass 30, and unprotected skin would burn in about five minutes instead of fifteen28.
That scenario rests on a growth guess, and the latest scientific assessment calls a similar one extreme12. Two other lines of evidence don’t need it. Replaying the real weather of recent decades with and without the treaty, one model found the Antarctic hole would have been 40% larger by 2013, and a deep ozone hole would have opened over the Arctic in 201129. The fingerprint study ties the recovery that is actually under way to the cuts.
The health estimates are large, and they cover only the United States. The US Environmental Protection Agency estimates that for Americans born between 1890 and 2100, the treaty will prevent about 443 million skin cancers, 2.3 million skin cancer deaths and 63 million cases of cataracts. Counting only illness between 1980 and 2100, the figures are 104 million, 550,000 and 15 million30. The agency itself warns that such estimates are extremely sensitive to assumptions30, and nobody has run the same numbers for the whole world.
A climate treaty in all but name
CFCs are also powerful greenhouse gases, so the phase-out was a climate measure too, and a big one. By 2010 it was cutting emissions by about 8 billion tonnes of CO2-equivalent a year (other gases counted as the amount of CO2 that would cause the same warming). The Kyoto Protocol’s first targets aimed for roughly 1 billion, and even after subtracting the warming caused by the substitute gases, Montreal’s cut was far bigger31. The UN puts the total avoided between 1990 and 2010 at 135 billion tonnes10.
That's how much the Montreal Protocol had cut yearly greenhouse gas emissions by 2010, before counting the warming from the substitute gases
The latest assessment estimates the phase-out will spare the planet 0.5 to 1 °C of warming by mid-century, although it measures that against the same extreme scenario of runaway CFC growth12. A 2021 study found a quieter benefit. In a world without the treaty, UV damage to plants would have left 325 to 690 billion fewer tonnes of carbon stored in plants and soils by the end of the century, and 0.5 to 1 °C more warming32.
For scale, the carbon dioxide the world is still putting into the air.
Why Montreal may not work twice
Montreal is often held up as proof that the world could fix climate change the same way, and its admirers include a UN Secretary-General.
The strongest objection is that it was an easier problem. The legal scholar Cass Sunstein calculated that for the United States, the treaty’s benefits dwarfed its costs so thoroughly that acting alone would have paid off, which was never true of the Kyoto climate treaty33.
A family of industrial chemicals made by a handful of companies is not the same as the fossil fuels that run most of the economy. The geographer Mike Hulme has cautioned that Montreal offered “a false optimism” for exactly that reason21.
Where the comparison holds is in how the treaty acted. Benedick told the US Senate in 2005 that it imposed real short-term costs to guard against dangers that still rested on scientific theory rather than proof25. The trade rules, the fund and the ratchet were negotiated, and any treaty could copy them. That suggests the part of Montreal that could travel is its design, even if the easy economics behind it cannot.
The Kigali Amendment, and the gases still outside the treaty
The Kigali Amendment, agreed in 2016, extends the treaty to hydrofluorocarbons (HFCs), the gases that replaced CFCs. They spare the ozone layer, but many are potent greenhouse gases, and without controls they were on course to equal 9 to 19% of the world’s CO2 emissions by 205034. It requires 18 HFCs to be cut by more than 80% over 30 years10, and 174 countries had ratified it by August 202635.
Full compliance would avoid 0.3 to 0.5 °C of warming by 210012. So far, though, HFC emissions are still rising: they grew 18% between 2016 and 202012.
Two ozone-depleting gases sit outside the treaty altogether. Nitrous oxide is now the largest ozone-depleting emission: in 2020, human emissions of it did more than twice the ozone damage of all the CFCs emitted that year12. And dichloromethane, an industrial solvent that breaks down within months, is still growing and already takes a small, steady bite out of the global ozone layer12.
CFCs are still leaking, and one cheat was caught
Some banned gases are still getting out. In 2018 atmospheric scientists noticed that CFC-11, banned for years, was declining only half as fast as it should. Emissions had risen by about 13,000 tonnes a year since 2012, while reported production was close to zero2. The source was traced mostly to eastern China. Once the finding went public, emissions there fell back to pre-2013 levels by 20194, and global emissions dropped 26% in a single year3.
Other leaks are legal, and harder to close. Five CFCs are still allowed as raw materials for making other chemicals, and their emissions rose from 1.6 to 4.2 thousand tonnes a year, weighted by potency, between 2010 and 20205. HFC-23, a by-product of making a common refrigerant, should have fallen 87% after 2014; instead, 2018 emissions were the highest on record7.
The CFC-11 cheat was caught by the network of instruments that watches the air, and several of the satellite instruments in it are due to retire within a few years12.
Forty more years to 2066
Scientists’ next full assessment of the ozone layer is due at the end of 2026, and it will update the recovery dates. If nothing goes badly wrong, the Antarctic hole should close around 2066, almost eighty years after the treaty was signed12.
Some new pressures could push that date out. If rocket launches reach about 2,000 a year, they could thin the global ozone layer by 0.29% by 20308, and smoke from Australia’s bushfires of 2019 and 2020 depleted 3 to 5% of the ozone over the southern mid-latitudes9. Neither is large yet, but both are growing sources that the treaty was not written for.
The harder tests look political. An HFC phase-down runs into the 2040s, the monitoring network needs new satellites, and nitrous oxide is outside the treaty entirely. Our reading is that the chemistry is largely done, and whether the recovery finishes on schedule now depends on keeping that machinery funded and watched.