The Paper Tiger Problem
You sign the agreement. You hold the ceremony. Foreign ministers shake hands in front of flags, someone takes a photograph, and the communiqué uses the word "historic" at least twice. Then, eighteen months later, the country doing the thing the treaty was supposed to stop is still doing it, and the treaty's monitoring body has issued a strongly worded report that nobody will read.
This is not an edge case. It is the dominant experience of international treaty-making, and pretending otherwise is the first mistake most explainers make.
So what actually determines whether a compliance mechanism has real teeth? The honest answer is a cluster of design choices, political conditions, and economic entanglements that most treaty drafters either underestimate or deliberately leave vague to get signatures in the first place. The vagueness is often the price of getting everyone in the room. It is also the seed of the treaty's eventual irrelevance.
The Three Things That Actually Matter
Strip away the legal language and you are left with three concrete variables: verification, consequences, and the cost of exit.
Verification first. A compliance mechanism is only as strong as its ability to know when a violation has occurred. The Chemical Weapons Convention built in something relatively rare: an independent technical secretariat with the right to conduct both scheduled and short-notice inspections at declared and undeclared sites. That specificity matters enormously. Compare it to most environmental accords, where self-reporting is the primary data source. When a state is both the defendant and the witness, the epistemics are compromised from the start. The Montreal Protocol, which phased out ozone-depleting substances with genuine effect, relied partly on independent atmospheric monitoring. Satellite data, third-party scientific bodies, and cross-border trade records created a verification layer that did not depend on the goodwill of the party being watched. That is not a minor technical detail. It is the whole ballgame.
Consequences second. The World Trade Organization's dispute settlement system is probably the most studied example of a compliance mechanism with structural teeth. When a panel rules against a member state, the injured party can be authorised to impose retaliatory tariffs calibrated to the value of the trade harm. Not a fine paid to an abstract body. A direct, quantifiable economic cost inflicted by a specific trading partner, the kind of thing that lands on a finance minister's desk with a number attached. The mechanism converts a legal ruling into a commercial weapon, and commercial weapons are something finance ministries understand in their bones. Contrast that with the human rights treaty bodies of the United Nations system: the Committee Against Torture, the Human Rights Committee, and their siblings. They produce Concluding Observations. They express concern. They request follow-up information. The state reads the report, notes the concern, and carries on, because there is no retaliatory tariff, no suspended membership, no seized asset. The consequence is reputational, and reputational pressure works only on governments that care about their reputation with the specific audience doing the pressuring.
Cost of exit third. A treaty a state can leave by filing a notification and waiting ninety days is a fundamentally different instrument from one where exit triggers automatic sanctions, debt acceleration, or loss of market access. The Paris Agreement on climate has a low exit cost by design, because that was the only way to achieve near-universal membership. The trade-off is explicit: breadth for depth. The Nuclear Non-Proliferation Treaty has a withdrawal clause too, but the geopolitical cost of invoking it has historically been high enough to function as a deterrent in its own right. One state's withdrawal announcement was treated by major powers as a crisis event, not a routine administrative act. That reaction is itself a compliance mechanism, one that lives entirely outside the treaty text.
A Tale of Two Signatories
Consider a scenario that captures how these variables play out. Two mid-sized economies, call them State A and State B, both ratify the same fisheries conservation agreement covering a shared maritime zone. The agreement sets catch limits, requires quarterly reporting, and establishes a joint commission with the power to recommend sanctions.
State A is heavily dependent on export markets controlled by the other signatories. Its fishing industry represents about four percent of GDP, but its processed seafood exports represent twelve percent of total export revenue, almost all of it going to treaty partners. State B exports primarily to non-signatory markets and has a domestic fishing lobby with direct seats in its parliament.
Five years in, both states have exceeded their catch limits. The joint commission investigates and makes identical findings. State A's government, facing the credible prospect of losing preferential market access, quietly introduces new licensing restrictions and reduces its fleet's operating days. The numbers come back into compliance within two seasons. State B's government thanks the commission for its report and continues as before, because the threatened consequence, a recommendation to other signatories, translates into no material cost given its export profile.
Same treaty. Same text. Radically different outcomes. The mechanism did not change. The economic entanglement did.
What People Get Wrong About Soft Law
The standard critique of treaties with weak enforcement is that they are mere symbolism, feel-good diplomacy for domestic consumption. That critique is too simple, and it needs to die as a reflex.
Soft compliance mechanisms can create norm-hardening effects over time. Think of it less like a lock on a door and more like water finding its way through limestone: nothing dramatic happens for years, then suddenly there are caves. When states repeatedly submit to review processes, accept the legitimacy of the monitoring body, and adjust domestic law to meet reporting requirements, they construct a legal and institutional reality that makes future defection more costly. The UN Convention on the Rights of the Child has no enforcement mechanism worth the name. It has also been incorporated into the domestic law of more countries than almost any other international instrument, and courts in dozens of jurisdictions cite it in rulings. The compliance pathway ran through domestic law, not international sanction.
Still, this is not an argument for designing toothless mechanisms on purpose. It is an argument for intellectual honesty about what you are building. A treaty that relies on norm diffusion and domestic legal incorporation operates on a timeline of decades and requires a functioning civil society and independent judiciary in the target state. If neither condition holds, the soft mechanism produces nothing at all.
The catch: most treaty negotiations happen under political pressure to produce a signed document, which means the enforcement architecture gets traded away in the final hours. Verification schedules get loosened. Sanction triggers get softened. Exit clauses get widened. The diplomats go home with a success. The compliance body inherits the wreckage.
The Honest Measure
Ask yourself, before you read a single word of the legal text: what does a violation cost the violator, in concrete terms, paid to a specific counterparty, within a defined timeframe?
If the answer is a report, a review, a dialogue, or an expression of concern, you are looking at a mechanism designed to produce the appearance of accountability. If the answer is market access, export licences, debt terms, or membership rights in an institution the state genuinely needs, you are looking at something with real structural weight behind it.
The rest, the preambles about shared values, the references to customary international law, the elaborate monitoring committees, is scaffolding. Useful scaffolding, sometimes. But the building stands or falls on what it costs to cheat, and whether anyone with real power is watching when you do. Those two conditions, cost and witness, are what separate a treaty from a press release, and no amount of ceremonial pen-swapping changes the arithmetic.