How International Environmental Treaties Succeed Or Fail

The Montreal Protocol is routinely held up as the gold standard of international environmental cooperation. It deserves to be. But the reasons it worked are more specific and more replicable than most accounts suggest.

The science was clear and specific. In 1974, Mario Molina and F. Sherwood Rowland published their paper in Nature identifying the mechanism by which CFCs destroy stratospheric ozone. In 1985, Joe Farman and colleagues published their discovery of the Antarctic ozone hole — a dramatic, measurable, visually alarming phenomenon. The science wasn't contested for long. The mechanism was well-understood, the measurements were replicable, and the cause was traceable to a specific set of industrial chemicals.

This matters because the clarity of the science removed the most effective delaying tactic: manufactured doubt. There was an early industry campaign to discredit the science — DuPont initially pushed back hard — but the ozone hole was so dramatic and so clearly linked to CFCs that the doubt campaign collapsed faster than it would for a more diffuse problem.

The number of relevant industry players was small. Global CFC production was dominated by a handful of large chemical companies — DuPont, ICI, Atochem, and a few others. When it became clear that regulation was coming, these companies pivoted to developing substitutes rather than fighting the regulation indefinitely. DuPont filed patents for HFC alternatives and came to support the Montreal Protocol because it would force their competitors to buy the replacements DuPont was developing. Market concentration made industry cooperation easier.

Compare this to fossil fuels. The oil and gas industry involves thousands of companies across every country, employing tens of millions of people, with revenue that exceeds the GDP of most nations. The political and economic inertia is orders of magnitude greater.

The costs were manageable and the benefits were distributed. The global cost of phasing out CFCs was estimated at roughly $235 billion over the life of the protocol. That sounds enormous until you realize the avoided costs — in health care, agricultural losses, and ecosystem damage from ozone depletion — were estimated at $2.2 trillion. The cost-benefit ratio was roughly 10:1. Every nation benefited from ozone protection. No nation benefited from continued ozone depletion. The incentive alignment was clean.

The Multilateral Fund solved the equity problem. Developing nations had a legitimate grievance: rich countries had used CFCs to build their economies, and now they were being told to stop. The Multilateral Fund, established under the 1990 London Amendment, provided financial and technical assistance to developing countries for their transition. Over $4 billion has been disbursed since 1991. China, the world's largest CFC producer, received substantial funding to convert its refrigeration and manufacturing sectors.

This was not charity. It was a deal. Rich countries paid for the transition because they needed developing countries to participate for the treaty to work. Self-interest aligned with equity. That alignment was designed, not accidental.

The compliance mechanism worked. Parties report annually on their production, import, and export of controlled substances. The Implementation Committee reviews compliance. Non-compliant parties can face trade measures — a ban on trading controlled substances with other parties. The trade provisions created an economic incentive for participation that went beyond moral suasion. If you weren't in the treaty, you couldn't trade. Almost every nation found it cheaper to comply than to be excluded.

The result. CFC-12 production (the most common ozone-depleting substance) fell from over 400,000 metric tons in 1988 to near zero by 2010. Total consumption of ozone-depleting substances has decreased by over 99 percent since the protocol's peak baseline. Atmospheric concentrations of most controlled substances are declining. The ozone layer is projected to return to 1980 levels by 2066 over Antarctica, by 2045 over the Arctic, and by 2040 over the rest of the world.

As a bonus: the Kigali Amendment (2016) added HFCs — the replacement chemicals, which turned out to be potent greenhouse gases — to the phase-down schedule. If fully implemented, the Kigali Amendment alone is projected to avoid up to 0.5 degrees Celsius of global warming by 2100. The Montreal Protocol may end up doing more for climate change than any climate treaty.

Kyoto was adopted in 1997 at COP3 in Japan. It established binding emissions reduction targets for 37 industrialized countries (Annex I parties) for the period 2008-2012, with an average reduction of 5.2 percent below 1990 levels.

The US rejection. The Byrd-Hagel Resolution, passed unanimously by the US Senate in July 1997 — five months before Kyoto was finalized — declared that the US should not sign any protocol that mandated emissions reductions for developed countries without comparable commitments from developing countries. The Clinton administration signed Kyoto anyway but never submitted it for ratification. The George W. Bush administration formally withdrew in 2001, calling the treaty "fatally flawed."

The US rejection was devastating. The world's largest emitter at the time (China overtook the US in total emissions around 2006) simply opted out. And because Kyoto only covered Annex I countries — exempting developing nations including China and India — its actual coverage of global emissions was limited to roughly 30 percent of the world total. You can't solve a planetary problem with a treaty that covers less than a third of the emissions.

The structural flaw: binary burden-sharing. Kyoto divided the world into two categories: countries with binding obligations and countries without. This reflected a legitimate equity principle — the developed world caused the bulk of historical emissions and should lead the response. But it created a political impossibility. Domestic politics in the US, Australia, and Canada made it untenable for politicians to accept binding constraints while economic competitors (China, India, Brazil) faced none. The binary structure invited defection.

What it did accomplish. Kyoto wasn't useless. The EU met its collective target (an 8 percent reduction from 1990 levels). The Clean Development Mechanism created a framework for carbon offset projects in developing countries. The Emissions Trading System in the EU — the world's first major carbon market — was built to comply with Kyoto obligations. The administrative architecture of emissions accounting, reporting, and verification was established and remains in use.

But the bottom line: global emissions rose roughly 60 percent between 1990 and 2020. Kyoto did not bend the emissions curve.

Paris (2015) learned from Kyoto's failures and made a deliberate trade: it exchanged binding top-down targets for universal participation through voluntary bottom-up pledges.

The architecture. Each nation submits a Nationally Determined Contribution specifying what it plans to do about emissions. There's no penalty for missing targets. The agreement includes a "ratchet mechanism" requiring nations to submit new, more ambitious NDCs every five years. The goal — limiting warming to well below 2 degrees, pursuing 1.5 degrees — is collective, but the commitments are individual.

What it got right. Universal participation. 196 parties have signed. Unlike Kyoto, Paris covers all major emitters including China, India, and (intermittently) the United States. The inclusive design made it politically viable. No country had to accept constraints its competitors avoided.

What it got wrong. Ambition and enforcement. The sum of current NDCs, even if fully implemented, puts the world on track for roughly 2.5 to 2.8 degrees of warming. Many nations are not on track to meet even their insufficient pledges. There's no enforcement mechanism. The ratchet is supposed to increase ambition over time, but there's nothing to compel it beyond diplomatic pressure.

The Trump administration withdrew the US from Paris in 2017, the Biden administration rejoined in 2021, and the second Trump administration signaled withdrawal again in 2025. The political instability of US participation undermines the agreement's credibility. It's hard to build a durable global framework when the world's second-largest emitter changes its mind every four years.

The fundamental tension. Paris was designed to be politically achievable. Kyoto was designed to be scientifically adequate (for its signatories). Neither succeeded at both. The next framework needs to solve the same problem the Montreal Protocol solved: make compliance economically rational for every participant, with funded transitions and meaningful consequences for defection.

The CBD, opened for signature at the 1992 Earth Summit in Rio, is the least discussed and most comprehensively failed of the major environmental treaties.

In 2010, the CBD adopted the Aichi Biodiversity Targets: 20 goals to be achieved by 2020. They included protecting 17 percent of terrestrial areas and 10 percent of marine areas, halving the rate of habitat loss, preventing extinctions, and eliminating harmful subsidies.

The 2020 Global Biodiversity Outlook found that zero of the 20 targets were fully met. Six were partially achieved. Fourteen were missed entirely. Species extinction rates remain 100 to 1,000 times above the background rate. 75 percent of land surface has been significantly altered by human activity. 66 percent of the marine environment has been significantly altered. One million species are estimated to be at risk of extinction.

The Kunming-Montreal Global Biodiversity Framework, adopted in December 2022, set new targets for 2030 — including protecting 30 percent of land and sea areas (the "30x30" target). Whether this fares better than Aichi remains to be seen, but the structural problems that sank Aichi remain.

Why biodiversity treaties fail. Biodiversity loss is the most diffuse of all environmental problems. It has no single cause (unlike CFCs), no single sector (unlike energy), and no simple metric (unlike parts per million of CO2). It's driven by agriculture, urbanization, pollution, invasive species, overexploitation, and climate change simultaneously. There's no equivalent of a CFC substitute — you can't swap a destroyed ecosystem for a technological replacement.

Additionally, the economic beneficiaries of biodiversity destruction (agribusiness, logging, fishing, real estate development) are powerful and concentrated, while the beneficiaries of biodiversity protection (future generations, rural communities, the global public) are diffuse and politically weak. The incentive structure is the reverse of what makes treaties succeed.

Extracting across all four cases, here's what distinguishes treaties that work from treaties that don't.

1. Specificity of targets. The Montreal Protocol banned specific chemicals on specific timelines with specific exemptions. The CBD set aspirational targets with vague implementation pathways. Specific beats vague every time.

2. Available substitutes. Where technological alternatives exist and are economically viable, transitions happen. Where they don't, treaties become aspirational documents. This means that treaty design and technology policy are inseparable — you need to develop the substitutes before or during the negotiation, not after.

3. Funded transitions. Developing countries won't sacrifice economic growth for environmental goals unless someone pays for the transition. The Montreal Protocol's Multilateral Fund solved this. Climate and biodiversity treaties have chronically underfunded their equivalent mechanisms. Every dollar spent on transition funding for developing nations is an investment in treaty compliance.

4. Monitoring and verification. If you can't measure it, you can't enforce it. CFC production is measurable at the factory level. CO2 emissions are measurable but harder to attribute. Biodiversity loss is measurable in aggregate but extremely difficult to track at the granular level needed for enforcement.

5. Consequences for non-compliance. Trade restrictions, financial penalties, market access conditions — these work. Moral suasion, peer pressure, and naming-and-shaming work less well. The Montreal Protocol's trade provisions made non-participation economically irrational. The Paris Agreement's absence of enforcement makes non-compliance costless.

6. Narrow scope. A treaty that tries to solve one well-defined problem is more likely to succeed than a treaty that tries to solve everything at once. This has uncomfortable implications for climate change, which is structurally a problem of everything at once. It suggests that breaking the climate problem into multiple narrower agreements — one for energy, one for agriculture, one for transport, one for industry — might be more effective than a single omnibus treaty, even though it's less elegant.

The Montreal Protocol proves that planetary cooperation on environmental problems is possible. The failure of biodiversity and the inadequacy of climate treaties prove that it isn't automatic. The determining factors are structural, not moral. We don't need better people. We need better-designed agreements.

If Law 1's premise is correct — that collective action on species-level problems is possible when humans agree — then the treaty record tells us what "agreement" actually requires. It requires specific targets, funded transitions, available technological substitutes, robust monitoring, meaningful enforcement, and a scope narrow enough to be tractable.

The next planetary agreement we need to get right is probably some framework addressing AI governance, or ocean management, or space resource allocation. Whatever it is, the blueprint is sitting there in the Montreal Protocol. The question is whether we'll use it.

1. The Protocol Comparison. Create a side-by-side comparison of the Montreal Protocol and the Paris Agreement across five dimensions: scope, targets, monitoring, enforcement, and funding. For each dimension, write one sentence about what Paris would look like if it adopted the Montreal approach.

2. The Substitute Test. Pick one major contributor to climate change — coal power, internal combustion engines, industrial agriculture, cement production. Research the available technological substitutes. How far along is each substitute? What would it cost to transition? What's the Montreal Protocol equivalent — the moment when the substitute becomes cheaper than the original?

3. The Treaty Autopsy. Pick one of the 20 Aichi Biodiversity Targets that was missed. Research what happened. Why was it missed? What structural features of the CBD made failure likely? Design a single-issue treaty that could achieve that one target using Montreal Protocol principles.

4. The Funding Exercise. The Green Climate Fund has consistently failed to reach its $100 billion annual target. Research where the shortfall is. Which countries are underpaying relative to their commitments? Calculate what each major economy's "fair share" would be using at least two different allocation methods (historical emissions, GDP, per-capita emissions). Write a one-page argument for the method you find most defensible.

5. Design Your Own Treaty. Pick a planetary-scale problem that doesn't yet have a major international treaty — microplastic pollution, antibiotic resistance, space debris, soil degradation. Design a treaty from scratch using the six principles identified in this article. Be specific about targets, timelines, monitoring, enforcement, funding, and scope. Where does the design break down? What's the hardest part to get right?