Global Coordination
Why coordination is the whole game
The agentic economy does not sit in any country. Its money, its agents, and its work live in software on the open internet, and they move across borders at the speed of information. That is the source of its power. It is also why the changes it brings cannot be managed one nation at a time.
The hardest of those changes are not technical. They are the changes to work, to who owns what, to how value is taxed, and to what a company even is. These land in every economy at once. A country that tries to face them alone is exposed: capital, firms, and agents can move to wherever the rules are easiest, so any nation that acts by itself risks driving the activity away rather than shaping it. A world that acts together can set the terms of the transition instead of being set by them.
Most of the policy energy so far has gone into a narrow piece of this: each major economy writing its own rules for digital money. That work matters, but it is the early and easy part. The real coordination problem is larger. It is the shared transformation of the political economy (labor, taxation, ownership, the corporate form, and the governance of the infrastructure itself) that the major economies will either work on together or fail at separately. This section is about that larger problem space, and where coordinated action either wins or is lost.
The labor transition, everywhere at once
The displacement of human work by agentic labor will not happen in one country and then spread. It will happen in every advanced economy at roughly the same time, because the same models and the same agents are available everywhere at once. The fall in labor's share of output, the share of what an economy produces that flows to workers rather than to the owners of capital, is already a global trend, and the agentic economy accelerates it. (See Domain C: this is a problem of distribution, not of output, and output is likely to be enormous.)
A national response is not enough on its own. Retraining programs, wage support, and safety nets all assume a labor market bounded by a border, while the force acting on it is global. Worse, a country that protects workers heavily while its neighbors do not can watch agentic firms relocate to the lighter-touch jurisdiction. That is the race to the bottom in its labor form.
Coordinated, the labor transition can be the broadest expansion of prosperity in history: the same productivity, shared widely, with adjustment support that holds people up while the economy reorganizes around them. The work to be done is a shared diagnosis built on common data about where displacement is happening and how fast; agreed floors on worker protection so no economy gains by stripping them; and pooled approaches to adjustment and retraining. The natural homes for this are the labor and economic forums that already exist: the G20, the OECD, the International Labour Organization. None of them treats the agentic transition as a single connected problem yet, and closing that gap is the first piece of work.
Taxing value that has no home
Modern states are funded mainly by taxes on wages: income tax and payroll tax. As labor's share of output falls, that base erodes. At the same time, value in the agentic economy is created by capital and software, booked in no particular place, and moved instantly. The thing that is growing is hard to tax, and the thing that is shrinking is what most governments tax. Left unaddressed, that gap hollows out the capacity of states to fund the very safety nets and public goods the transition demands.
The answer is new tax bases (on automation, on capital, on value where it is actually used), and those bases only hold if they are coordinated. A tax on mobile capital that one country imposes alone is a tax that capital simply moves to avoid. This is not a hypothetical worry; it is the central lesson of decades of corporate-tax competition.
There is a precedent. The international effort to set a global minimum corporate tax showed that dozens of economies can agree to stop undercutting one another on a mobile tax base, by setting a shared floor and taxing value closer to where it is used. It is an imperfect precedent (the accord has strained, and a major economy has stepped back from full participation), but that is part of the lesson, not a refutation of it: coordination on a mobile base is achievable, and it has to be actively defended once struck. The work is to extend that approach to the new bases the agentic economy requires: to define what is taxed when an agent, not a person, creates the value, and to agree the floor before the base disappears. The aim is not new revenue for its own sake. It is to keep states solvent enough to manage a historic transition, while taxing the new economy carefully enough not to slow the investment that produces the abundance.
A new social contract
For two centuries, work has been the main way people claim a share of what the economy produces. If agentic labor weakens that link, something has to take its place. The likeliest answer is broader ownership, a direct stake in the capital that is doing the producing, alongside new forms of economic participation and support. (Domain C lays out how broad ownership can be built by design rather than hoped for.)
This is where coordination is most delicate, because the social contract is national. Each country will, and should, answer it differently, according to its own values and institutions. That experimentation is healthy. But it still needs shared principles and a way for nations to learn from one another, and it needs to avoid a world in which a few jurisdictions capture the upside of the transition while others absorb the displacement. The work here is not a single global system. It is agreement on principles (that the gains should be broadly held, that participation should be portable, that no economy should be left to bear the costs alone) and the forums to compare what is working. There is no single technocratic answer to what replaces the wage as the basis of economic citizenship; coordination here means agreeing on direction and learning from one another, not imposing one design on everyone.
New corporate forms that travel
A firm whose work is done largely by agents is global from the moment it is created. It needs the things every firm needs (legal personhood, limited liability, the standing to sue and be sued), and it needs them to be recognized across borders, because it operates across borders by default. (Domain B sets out what these firms are and how accountability attaches to them.)
If each country recognizes the new corporate forms differently, or not at all, two bad outcomes follow. Either the firms operate in a legal vacuum, unaccountable to anyone, or they flee to whichever jurisdiction asks the fewest questions, and that jurisdiction sets the floor for everyone. The work is mutual recognition: shared standards for what an agentic or onchain corporate form must do to earn personhood and limited liability, and agreement that a firm legible and accountable in one major jurisdiction is legible and accountable in the others. A firm should be able to answer for itself everywhere, not nowhere. The discipline that makes this work is the same one that runs through the rest of this section: recognition must be recognition against a real floor, not a mutual agreement to ask nothing.
The money and settlement fabric
The money layer is the piece that has drawn the most attention, and it is the most tractable of the coordination problems, which is exactly why it should not be mistaken for the whole.
Each major economy is now writing its own rules for digital money. None yet recognizes another's. So a single global form of value risks fracturing into national variants that do not move cleanly across borders: the old correspondent-banking fragmentation, rebuilt in software. The fix is well understood: mutual recognition of issuers against a shared floor (full backing, guaranteed redemption, real supervision), and neutral, shared settlement infrastructure so value moves as one fabric rather than through walled gardens. Much of this is standards work that is already underway in the existing financial-stability and central-banking bodies. It is real work, but it is mostly engineering and convergence, and it is closer to done than any of the harder problems above. It will converge; the main task is to ensure it converges on open, shared standards rather than the closed networks that commercial incentives would otherwise favor.
Keeping the infrastructure neutral
This is the deepest coordination question of all. A planetary economic operating system cannot be controlled by a single state. If it is, it will be neither trusted by everyone else nor stable, because the rails that carry the world's value become switches that one government can throw: to freeze, to exclude, to surveil. A dominant settlement asset carries, in effect, a foreign off-switch for everyone who is not the issuer.
The constructive answer is to govern the core infrastructure as genuinely neutral: multi-stakeholder, owned by no single nation, run on open standards the way the internet's own protocols are. Neutrality is not a constraint on sovereignty; it is the precondition for it. Only neutral rails let every country participate on its own terms, issue its own currency on the same infrastructure, and trust that the system will not be turned against it. Alongside that sits the competition work of keeping power from pooling at the chokepoints (identity, settlement, and the dominant model layer) that Domain C identifies. This is where geopolitics bites hardest, because the contest over who writes the standards is a contest over economic power. That is precisely why neutral infrastructure has to be built deliberately and early, while the standards are still open to shape.
Three tensions to navigate
Three tensions run under all of the above. Coordination does not dissolve them, but it makes each one navigable.
The first is monetary sovereignty in a world with one dominant currency. Most cross-border value still routes through the dollar, for the same reasons most foreign exchange does today. Coordination cannot abolish that hub. It can keep the rails neutral enough that every sovereign retains the real option to issue and strengthen its own currency. The discipline this imposes is asymmetric, and it is worth saying plainly: it falls on weak and badly managed currencies, never on the dominant issuer. That same discipline also protects the ordinary holder of a fragile currency from a government that would debase it.
The second is the gap between strong and weak economies. The upgrade is real for advanced economies and can be regressive for the most fragile, whose citizens may reach for a stable foreign currency one tap away while their own currency cannot make the leap. The decisive fact is that the rails arrive whether or not a government opts in, because its citizens can already reach them. So the real choice is not whether this happens but whether it is managed and early or forced and unmanaged. Coordination's job is to make the managed path available: through shared liquidity, regional arrangements, and tools fragile economies can actually use.
The third is the pull to defect. Every problem in this section rewards the country that breaks ranks to attract mobile capital, firms, or activity with the loosest rules. That is the single thread tying labor, tax, corporate form, and money together. It is why coordination here is so hard, and why a credible shared floor (not uniformity, but a floor below which no one is allowed to compete) is the mechanism that makes any of it work.
Where this gets worked
The institutions exist, but in pieces. Financial stability sits with the G20 and the Financial Stability Board. Tax coordination, and its one great success, sits with the OECD. Capital flows and macroeconomic stability sit with the IMF. Labor sits with the ILO and the G20's labor track. AI governance is forming across the OECD, the UN, and national efforts. Each owns a fragment.
No body owns the whole. That is the central point of this section. The agentic transition cuts across all of these mandates at once, and it is currently being treated as a dozen disconnected files rather than one connected agenda. The clearest opportunity for leadership is to insist on that connection: to treat work, taxation, ownership, corporate form, money, and infrastructure as facets of a single transformation that the major economies must work on together.
The bottom line
Each country is writing rules for the pieces. No one is yet working the whole. Money and corporate forms are the early, tractable part. The transformation of work, taxation, ownership, and the social contract is the real coordination problem, and it is global because the economy is global. The default outcome is fragmentation and a race to the bottom, with the gains pooling where the chokepoints sit and the costs landing on everyone else. Coherence is the alternative, and it will not arrive on its own. It is a project the major economies have to choose, while the system is still taking shape.
The shape of the agenda
Read together, the four parts describe one agenda with one logic. The monetary domain shows that money built for machines can be safer than the money before it (fully backed, stable at par, final in seconds) if the reserve rules, the transmission plumbing, and the supervision of machine credit are rebuilt for it. The legal domain shows that autonomous agents can be more accountable than the anonymous internet that preceded them, if identity, the corporate form, compliance, and due process attach a responsible human to every consequential act. The concentration domain shows that an economy this powerful will pool power at its chokepoints by default, and that the only durable answer is to govern those chokepoints and spread ownership by design. And the coordination arena shows that none of it can be finished inside one country, because the economy is global and the hardest changes (to work, taxation, ownership, and the company itself) arrive everywhere at once.
The throughline is simple: the good outcome is buildable, and the bad one is the default. Concentration is gravity; equalization is a choice. Fragmentation is gravity; coherence is a project. Opacity was the old system's condition; observability is the new one's opportunity, but only if supervision is built to use it. In every case the better outcome has to be constructed, and in every case the materials to construct it already exist or are within reach. That is the reason for optimism. It is also the reason for urgency. The architecture is being laid now, the standards are being set now, and the window to shape them (before the walled gardens harden and the chokepoints are captured) is open now and will not stay open.
This is a guide, not a manifesto. Its issues are offered as work to argue over and refine, not as settled conclusions. The agentic economy will be one of the largest reorganizations of economic life in modern history. Whether it becomes an engine of broadly shared prosperity or a new architecture of concentrated power is not decided by the technology. It is decided by the choices made while the system is still taking shape, which is to say, now.