Non-Binding Agreements Now Govern More Than Treaties Do
The serious criticism of any voluntary international framework is that it has no teeth. No enforcement, no penalties, no court. Governments sign, praise it, and go home.
The criticism is accurate and mostly beside the point, because binding treaties have become close to impossible to conclude among the states that matter, and the soft instruments that replaced them turn out to do a surprising amount of governing.
Look at what’s happened to ratification. A treaty requires a legislature, and legislatures in most large democracies now have narrow majorities, short attention, and a strong incentive to treat any international commitment as a sovereignty concession. The US Senate needs sixty-seven votes for a treaty and hasn’t reliably produced them for anything contested in decades. So the binding instrument has become the rare exception, reserved for arms control and trade, and even those get renegotiated or abandoned more freely than they used to be.
What fills the space is a family of things with unglamorous names. Principles. Codes of conduct. Memoranda. Standards published by bodies most people have never heard of. They bind nobody and they shape behaviour anyway, through four mechanisms that are worth separating.
Procurement is the first and the most powerful. A voluntary standard becomes mandatory the moment a large buyer writes it into a contract. No legislature votes. A compliance department reads a clause and the supply chain reorganises around it.
Domestic transposition is the second. A government that endorsed a principle abroad has handed its own regulators a ready-made justification, and agencies write rules against that text without any further international step. The soft instrument becomes hard law one jurisdiction at a time, in national forms that vary considerably.
Focal points are the third. Firms operating in twenty markets want one internal policy, so they build against the strictest or the most widely endorsed text and apply it everywhere. That’s how the EU’s data rules reached companies with no European operations, and it’s how any sufficiently endorsed framework spreads without an enforcement mechanism.
Coalition definition is the fourth, and it’s the one that gets underestimated. Endorsing a framework is how a state signals which technology stack, which standards bodies and which vendors it’s aligning with. Adoption is cheap, since there’s nothing to enforce, and it comes bundled with market access and commercial relationships. That bundle is the actual product being sold.
Which is why the interesting metric for any non-binding agreement isn’t the enforcement clause. It’s adoption count, procurement uptake, and whether a major regulator has started citing the text in rulemaking.
The failure mode is real and should be stated. Soft law is unaccountable in a way treaties aren’t. There’s no ratification debate, no public record of what was traded, and no obvious mechanism to withdraw. Bodies with narrow membership set standards that reach far beyond it. The parties best equipped to shape the drafting are the firms being regulated, since they employ the people who understand the subject matter, and the resulting text tends to reflect that.
Both things are true at once. This is a weaker form of governance than a treaty, and it’s the form that actually operates. The last decade of AI, data and climate policy has been written in it almost exclusively.
Anyone waiting for the binding version is waiting for an instrument that the domestic politics of the major powers no longer permit.