What happens when functions that arrived as one package come apart

Article 7 of 8  ·  Series I of III  ·  Published 23 September 2026  ·  Analysis → Forecast → Recommendations

How this series measures things. Every article applies the same three questions to its subject. Concentration: how many genuinely independent alternatives exist, once shared upstream origins are traced rather than counted. Criticality: what stops if this fails, and how quickly. Substitution time: how long until an alternative actually functions. This article turns the instrument on the arrangement most people have never examined: the set of functions their state provides as a single package.

1. The Signal

Money is the paradigm case of something a state provides. It is also the function that has come apart fastest, and the numbers are no longer small.

The stablecoin market stood at roughly 317 billion dollars in early April 2026 and around 322 billion in May. On-chain transaction volume for 2025 exceeded 33 trillion dollars — more than the annual payment volume of Visa's network.

That is a payment rail operating at systemic scale, denominated mostly in dollars, issued by parties that are not central banks.

The response from states is the part worth attention, because it was not prohibition.

The European Union's stablecoin provisions have applied since June 2024, with the broader service-provider regime live from December 2024 and the transitional period ending on 1 July 2026, after which authorisation became mandatory and unauthorised issuers face removal from EU markets. The United States passed the GENIUS Act in July 2025, creating a federal licence for payment stablecoin issuers. Hong Kong's Stablecoins Ordinance took effect in August 2025 and the first two licences were granted in April 2026. Brazil's central bank published its authorisation regime in November 2025, effective February 2026. Japan revised its Payment Services Act. Singapore, the United Arab Emirates and Australia have licensing regimes in force or in draft.

The frameworks converge on the same requirements: full reserves in high-quality assets, licensed issuance, segregation, redemption rights, periodic disclosure, and a prohibition on paying interest to holders.

So the state did not reclaim the function. It licensed it. The activity is performed by someone else, under conditions the state sets, and the state's role moved from provider to permissioner.

One further figure shows how far the arrangement has travelled. A single issuer's attestation for the final quarter of 2024 reported around 113 billion dollars of direct and indirect United States Treasury exposure, which on the published tables of foreign official holdings would place it among the twenty largest sovereign holders of US government debt.

A private provider of a state function, holding the sovereign's own debt at sovereign scale, licensed by that sovereign to do it.

And in several countries the sequence ran the other way entirely. Dollar stablecoin adoption in Argentina, Turkey, Nigeria and parts of Southeast Asia arrived ahead of any domestic regime, meaning the population unbundled the currency function from its own state before that state had a position on the matter.

2. The Mechanism

The modern state arrives as a package. Security, currency, courts, identity, registries, education, infrastructure, licensing, welfare, the enforcement of contracts, and the definition of who counts as a member.

The package is so familiar that its composition looks natural, and it is not. Those functions do not belong together for any logical reason. They were bundled because bundling them was, for several centuries, the only economically viable way to deliver any of them.

Why the bundle formed

Three constraints did the work, and all three were about cost.

—   Coordination cost. Establishing a common currency, a common law and a common measure of anything required an authority that could compel adoption over a territory. Voluntary adoption of a standard across a large population was not achievable at acceptable cost.

—   Information cost. Knowing who someone was, what they owned and whether they had complied required records, and records required a bureaucracy with physical reach. Only an entity with territorial presence could maintain them.

—   Enforcement cost. A judgement is worth nothing without the capacity to enforce it, and enforcement requires physical presence and a monopoly on legitimate coercion within a defined area.

Given those three, delivering the functions separately was not an option anyone declined. It was an arrangement nobody could afford.

The bundle was therefore a solution to a cost structure, not an expression of what belongs together. When the cost structure changes, the bundle has no internal reason to hold, and it comes apart along the seams where the original constraint has weakened rather than uniformly.

What changed

Coordination cost fell first. A standard can now propagate by adoption rather than by compulsion, and a payment rail with 33 trillion dollars of annual volume was assembled without anyone being required to use it.

Information cost fell next, and Article 4 examined the consequences: verifying identity, entitlement and history became something that can be done at distance and at scale, which is why the same governments building identity infrastructure are building it to be portable across borders.

Enforcement cost has not fallen at all. Compelling a person or seizing an asset still requires physical presence and a territorial monopoly, and no technology has changed that.

Two of the three constraints collapsed and one did not, which predicts exactly where the bundle splits.

Applying the instrument

The three questions produce an unusual result here, and the unusual part is the reason the whole subject is under-examined.

By this series' own instrument, the bundled state is the single most concentrated dependency in an ordinary life, and almost nobody has ever assessed it as one. It is invisible in the way that water is invisible to a fish — not hidden, simply never separated from the background.

Unbundling changes that reading, and not in one direction. Concentration falls, which is a real improvement. What replaces it is examined in section 4, and it is not straightforwardly better.

3. Subtheme One — Where the Bundle Splits

The split line is not arbitrary and it is not about importance. It runs between functions that can be performed at distance and functions that require the capacity to compel.

The state as licensor rather than provider

The most common misreading is that unbundling means the state loses. In money it clearly did not. Every major jurisdiction responded to a payment rail it did not build by licensing it, on conditions, with reserve requirements, redemption obligations, disclosure duties and a bar on paying interest.

That last condition is the most revealing. Prohibiting interest on payment stablecoins protects bank deposits from outflow, which is a monetary policy objective delivered through the licensing of a private instrument. The state stopped providing the rail and retained the ability to shape what the rail may do.

This is authority reorganised rather than lost, and Article 4 explains who can execute it. Licensing requires the trained judgement to assess an issuer, the legal basis to impose conditions, and the capacity to supervise — all scarce. A state without them does not become a licensor. It becomes a jurisdiction where the function is performed under someone else's licence.

Which is what is happening in the third group

Where a domestic regime arrived before adoption, the state licenses. Where adoption arrived first — in Argentina, Turkey, Nigeria and parts of Southeast Asia, where dollar-denominated stablecoins were taken up ahead of any domestic framework — the population unbundled the currency function unilaterally, and the state's remaining options are narrower and worse.

The pattern from Article 4 repeats exactly. States that could acquire the capacity became licensors. States that could not became territories where a function is performed under rules set elsewhere, for holders whose recourse runs to a foreign regulator.

What the coordination shows

The convergence of frameworks across the European Union, the United States, Hong Kong, Brazil, Japan, Singapore and the Gulf on the same core requirements is a genuine achievement and an incomplete one. A financial stability review in October 2025 found limited full alignment on capital treatment, risk management and cross-border cooperation, which is the honest position: the substance converged and the edges did not.

For a firm, this means operating across jurisdictions is not a matter of building a product but of holding a stack of licences, each with its own reserve rules, and reconciling requirements that agree in principle and differ in application.

4. Subtheme Two — The Risk Profile Inverts

Unbundling reduces concentration, which by every argument in this series is an improvement. It also changes what kind of dependency you hold, and the change runs the other way.

What the bundle actually provided

The bundled state was a concentrated dependency with three properties that are easy to overlook because they were never itemised.

—   Obligation. The provider had duties to you that it did not choose and cannot unilaterally revise — due process, non-discrimination, a legal basis for any refusal.

—   Recourse. When it failed, a defined route existed to contest the failure, and that route was itself provided by the same entity under an obligation to hear you.

—   Continuity. The provider could not exit your market, discontinue the service line, or decide you were no longer commercially interesting.

None of those three is a feature of a commercial provider, and none of them is unusual to lose — they were simply never priced, because they arrived attached to something nobody chose.

The inversion

An unbundled position looks better on concentration and worse on everything else.

The honest reading is that the trade is real in both directions. You have swapped a concentrated dependency with obligations for a diversified one without them, and which is better depends entirely on whether the diversification is genuine.

And frequently it is not

This is where the series instrument earns its place. Nominal diversification with a single upstream origin is the failure this whole series has described in supply chains, in institutions, in measurement and in our own forecast record. It reappears here in a personal form.

Three payment providers can settle through one rail. Two currencies can be backed by the same reserve asset. Several services can be licensed by one regulator, so that a single supervisory decision reaches all of them at once. A portable credential is portable only within the framework that recognises it.

So the useful question about an unbundled position is not how many providers you use. It is how many independent points of failure sit behind them.

The Personal Dependency Map

The practical instrument, and the one to do before anything else in section 12. For each function the bundle used to deliver, record four things.

—   Who provides it now — which may be a state, and which state, or a private party under whose licence.

—   What obligation they owe you — statutory, contractual, or none, and whether it can be revised without your agreement.

—   What recourse exists if it fails — a defined appeal, a regulator, a court in which jurisdiction, or nothing.

—   What sits upstream — the rail, the reserve, the regulator, the identity framework that this provider itself depends on.

The fourth column is where the surprises are. Most people who have unbundled deliberately have done so provider by provider, and have never traced the upstream. The map is worth two hours and is almost never made, because there is no moment at which anyone is prompted to make it.

5. What Most Analysis Gets Wrong

That unbundling means the state is weakening

In the clearest case it did the opposite. Faced with a payment rail at systemic scale that it had not built, every major jurisdiction acquired authority over it inside two years, on conditions it wrote. The function left the state as provider and returned as licensed activity, which is a different arrangement and not obviously a weaker one.

That it is driven by ideology

It is driven by cost structure. Coordination and information costs fell and enforcement cost did not, and the bundle is splitting along exactly that line regardless of anyone's preferences about it. In the third group of countries it happened without any policy at all, because populations adopted a dollar instrument faster than their governments formed a position.

That the unbundled position is freer

It is less concentrated and less protected. The statutory obligations, the defined recourse and the inability of the provider to exit were real and were never separately valued. Trading them for optionality can be a good trade and it is a trade, and the party who thinks they got something for nothing has not read the terms of service.

That this is available to everyone

Unbundling is easier the more you already have. It requires documentation, a jurisdiction that permits it, technical capability, and enough assets that a provider wants your business. The person with the most concentrated dependency on the bundled state is generally the person least able to unbundle from it, which means the population that would benefit most from alternatives is the one for which they are least available.

6. Base, Stress and Extreme

Four paths, with our probability assessment and the condition that would falsify each. Probabilities sum to one hundred.

The second path is the one that will be least discussed and matters most. Unbundling by the capable removes the users who cross-subsidised universal provision, and the residual service is left to those who could not leave. That dynamic is well documented in every previously unbundled utility and there is no reason to expect an exception here.

7. Forecast — One Year, to mid-2027

Licensing regimes multiply faster than they align

Probability 0.70  ·  Confidence: Medium-High

We expect further jurisdictions to adopt stablecoin licensing on the converged template, and we expect cross-border alignment on capital treatment, supervision and recognition of foreign issuers to remain incomplete.

The evidence for the second half is already on the record. A financial stability review in late 2025 found limited full alignment across jurisdictions, and the United States statutory deadline of 18 July 2026 for coordinated final rules passed without a coordinated package from the six agencies involved. That is not unusual: agencies missed roughly forty percent of the statutory deadlines in the comparable post-2010 financial reform.

Second-order effect. A firm operating across borders is not building a product but assembling a stack of licences whose requirements agree in principle and differ in application. That favours large incumbents over new entrants, in a sector whose original argument was disintermediation.

What would weaken it. A mutual recognition arrangement between two or more major jurisdictions allowing an issuer licensed in one to operate in another without separate authorisation.

8. Forecast — Three Years, to 2029

The licensing template extends to a second unbundled function

Probability 0.60  ·  Confidence: Medium

We expect the pattern established in money — private provision, state licensing on conditions, convergence on a common template — to be applied to at least one further unbundled function. Identity credential provision, cross-border credential recognition and automated dispute resolution are the candidates.

The mechanism is precedent. Once a jurisdiction has built the supervisory apparatus to license a private provider of a formerly public function, applying it to the next function is a marginal cost rather than a new capability, and the argument has already been won once.

Second-order effect. Licensing is the point at which the capacity divide from Article 4 becomes a sovereignty divide. A state that can supervise sets conditions; a state that cannot has the function performed on its territory under conditions set elsewhere, for residents whose recourse is foreign.

What would weaken it. The money case remaining exceptional, with other unbundled functions either staying unregulated or being reclaimed into direct public provision.

9. Forecast — Five Years, to 2031

The rail keeps growing and the recourse gap does not close

Probability 0.50  ·  Confidence: Medium

We expect the stablecoin market to exceed one trillion dollars, and we expect that growth to arrive without a corresponding development of cross-border recourse for holders.

The two halves have different mechanisms. Growth follows utility and regulatory clarity, both of which are improving. Recourse requires cross-border cooperation between supervisors, which the same reviews describe as the least developed area, and which delivers benefits to holders in a jurisdiction other than the one paying for it.

Second-order effect. A holder in a country with no domestic regime has an instrument denominated in a foreign currency, issued under a foreign licence, with recourse to a foreign supervisor who has no obligation to them. That is the least protected position in the whole arrangement, and it is concentrated in the populations that adopted earliest and had the fewest alternatives.

What would weaken it. A cross-border compensation or recognition mechanism giving holders in non-issuing jurisdictions a defined claim, or growth stalling below the threshold.

10. Forecast — Ten Years, to 2036

A state accepts privately issued money for its own obligations

Probability 0.35  ·  Confidence: Low

The sharpest available test of how far the money function has unbundled is whether a state will accept a privately issued instrument in settlement of what is owed to it — taxes, duties or statutory fees. We put this at a little over one in three within a decade, in at least one G20 member.

The case for is that several states already accept privately issued instruments through intermediaries and that the distinction is increasingly technical. The case against is stronger: accepting private issuance for public obligations concedes the last exclusive attribute of sovereign money, and it hands the issuer a claim on the state that the state has agreed in advance to honour.

We hold this at low confidence and state the probability anyway. A forecast about the boundary of sovereignty over a decade is close to the limit of what this method can support, and putting a number on it is more useful than a paragraph describing the possibility, because a number can be wrong in public.

What would weaken it. Explicit statutory prohibitions on accepting non-sovereign instruments for public obligations, adopted in several major jurisdictions, which would settle the question before the decade is out.

11. Signals to Watch

—   Mutual recognition between supervisors, which is the difference between a licensing template and an actual cross-border market

—   Whether recourse mechanisms follow the licences. Reserve rules protect the instrument; they do not give a foreign holder anywhere to go

—   The licensing template being applied to a second function — identity, credentials or dispute resolution

—   Adoption running ahead of regimes, which is where a population unbundles unilaterally and ends up least protected

—   Terms of the residual bundle: as capable users leave, what happens to the service left for those who did not

—   Public alternatives launched to displace private provision, and whether volume actually moves

—   Concentration in what sits upstream — rails, reserve assets, licensing authorities — behind nominally separate providers

12. Recommendations — Individuals

Most people are already partially unbundled and have never mapped it. The work is not to unbundle further; it is to find out what you have and what stands behind it.

Immediate — 30 days

Make the Personal Dependency Map from section 4. For each function — income, payments, savings, identity, legal status, healthcare, records, professional standing, dispute resolution — record the provider, the obligation owed to you, the recourse if it fails, and what sits upstream of that provider.

The fourth column is the point of the exercise. Two banks that clear through the same system, two currencies backed by the same reserve asset, three services licensed by one authority — each looks like diversification in the first three columns and disappears at once in the fourth.

Build — 12 months

Where you hold an unbundled position, find out where the recourse actually runs. For a private provider of a formerly public function, this means the licence, the jurisdiction, and the supervisor — and whether that supervisor has any obligation to a holder resident elsewhere. The answer is frequently that it does not, and knowing that in advance changes how much you keep there rather than whether you use it at all.

Keep at least one function inside the bundle wherever the bundled version carries obligations you would want in a crisis. Statutory duties, defined appeal routes and a provider that cannot exit are worth something specific, and the moment you need them is not the moment to discover you traded them away.

Position — 3 years

Treat the residual bundle as a declining service and plan accordingly. As capable users leave, what remains is funded by fewer of them, which is the pattern from every previously unbundled utility. This is uncomfortable to say plainly and it is what the evidence supports.

Avoid. Counting providers as protection. The whole of this series reduces to one correction and it applies here: nominal alternatives with a single upstream origin are one alternative, and the upstream is where you have to look.

Why this works. You cannot change what your state provides or how it licenses. You can know which functions you have already moved, what obligation each provider owes you, and what sits behind them — and that map takes two hours and almost nobody has one.

13. Recommendations — Business

Unbundling reaches a business twice: in what it depends on, and in what it must now hold licences to do.

Immediate — 60 days

Map which formerly public functions your operations now source privately — payments, identity verification, credential checking, dispute resolution, records — and for each, record the licence it operates under and the jurisdiction supervising it. A provider's regulatory status is an operational dependency and it usually sits in no register.

Then trace upstream. Multiple providers frequently converge on one rail, one reserve asset, one supervisor or one identity framework. A single supervisory decision that reaches all of them at once is a correlated failure that looks like diversification on the vendor list.

Build — 12 months

Where you operate across jurisdictions in a licensed activity, build for licence stacking rather than for a single framework. The templates converge on substance and differ in application, and an implementation written to one jurisdiction's interpretation requires rework in each additional one. This is the substitution-time argument from Article 1 applied to regulatory architecture.

And treat regulatory deadlines in this area as indicative rather than binding. The United States statutory deadline for coordinated final rules passed without a coordinated package, and the precedent from comparable reforms is that a substantial share of such deadlines are missed. Plan against the capability arriving, not against the date.

Position — 3 years

Expect the licensing template to extend to further functions you rely on. If your business depends on a private provider of a formerly public function, assume that provider will be licensed within a few years, that the licence will impose conditions changing its product, and that the cost will pass to you.

Where you are the provider, expect the same. The unbundled position is not a permanent regulatory gap; it is the interval before the licensing regime arrives, and building as though the regime will arrive is materially cheaper than retrofitting.

Avoid. Treating an unlicensed provider's terms as equivalent to a licensed one's obligations. The difference is invisible in normal operation and total in a failure, and it is the only difference that matters on the day it matters.

Why this works. Regulatory status is a dependency with a concentration reading of one, a criticality that is total, and a substitution time measured in licensing cycles. It belongs in the same register as your single-source suppliers, and it is almost never there.

14. Recommendations — Capital

The structural observation for capital is that a formerly public function becoming a licensed private activity changes what determines returns in it, and the change is legible in advance.

Immediate — this quarter

For any exposure to a provider of a formerly public function, establish the licence, the supervising jurisdiction and the conditions attached. Reserve requirements, redemption obligations, segregation rules and prohibitions on paying interest are not compliance detail. They determine the business model, and a prohibition on paying interest to holders defines the whole revenue structure of an issuer.

Build — 12 months

Look for upstream concentration behind nominally separate positions. Providers in different jurisdictions can hold the same reserve asset, settle through the same rail, or fall under the same supervisor. Sector and geographic diversification do not diversify any of those, and this is the same correlated exposure the series has described everywhere else.

Separate two risks that are usually merged. That an instrument fails is a credit and operational question. That the licensing conditions change is a regulatory question, and in a converging international template a condition adopted in one major jurisdiction tends to appear in others within a year or two, which makes it forecastable in a way credit risk is not.

Position — 3 years

Watch the recourse gap rather than only the growth. A market can grow while the protection available to holders in non-issuing jurisdictions does not, and the second determines what a stress event does. Growth is well reported. The recourse position is in supervisory reviews that few people read.

Avoid. Reading regulatory clarity as a reduction in risk overall. Clarity reduces legal uncertainty and it also fixes the conditions under which a business operates, which caps some outcomes as firmly as it protects against others.

Why this works. Licensing regimes are published in advance, converge on a visible template, and change business models before they change results. That interval is where the information advantage sits, and it requires reading supervisory documents rather than market commentary.

15. What Would Change Our Mind

Each forecast carries its own weakening condition. Three developments would undermine this article's argument as a whole.

—   A major state reclaims direct provision of a function that had moved to licensed private delivery, and the private version loses volume. That would show the licensor role is a transitional arrangement rather than the destination.

—   Unbundled provision reaches lower-income and less-documented populations at terms comparable to the capable, which would falsify the two-tier reading directly and would be the most welcome result in this article.

—   Enforcement cost falls materially — cross-border seizure, judgment enforcement or compulsion becoming genuinely portable. That is the third constraint, and if it moves, the split line moves with it and this article's central prediction about where the bundle divides is wrong.

The running tally, and a correction taking effect. This article's case material spans the European Union, the United States, Hong Kong, Brazil, Japan, Singapore, the Gulf, Argentina, Turkey and Nigeria, and its four forecasts resolve against international and multi-jurisdictional sources rather than European ones. Across Series I the count of forecasts resolving against European institutions now stands at eleven of twenty-four. The concentration recorded in Article 4 is being corrected by finding better-documented material elsewhere rather than by adding weaker examples for balance, which was the stated remedy, and it is working.

Founder's Lens

[ EDITORIAL GATE — WRITTEN BY HAND BEFORE PUBLICATION. Never generated. Replace this marker with the founder's text, or record a suspension. ]

16. Bottom Line

The state arrives as a package, and the package is not a natural kind. Those functions were delivered together because coordination, information and enforcement were expensive, and bundling them was the only affordable arrangement. Two of those three costs have collapsed and one has not, which predicts where the bundle splits: functions that work at distance leave, and functions that require the capacity to compel stay.

Money went first and went furthest. A payment rail carrying more annual volume than a global card network, denominated mostly in dollars, issued by parties that are not central banks.

And the state did not lose the function. It licensed it. Authority moved from provision to permission, and the states able to make that move are the ones with the capacity Article 4 described. The rest are territories where a function is performed under conditions set elsewhere.

For the individual, the trade is real in both directions and only one side is usually counted. Concentration falls, which is a genuine gain. What falls with it is a set of things that were never itemised because nobody chose them: statutory obligations that could not be revised, a defined route to contest a failure, and a provider that could not exit. Trading those for optionality can be a good trade. It is a trade.

And the diversification is often nominal. Two banks on one rail, two currencies on one reserve, three services under one supervisor. The correction this series keeps making applies to your own life as directly as to any supply chain: count the upstream origins, not the providers.

The bundled state was the most concentrated dependency most people will ever hold, and almost nobody assessed it as one. Unbundling is the first opportunity to — and the map is only useful if it is made before it is needed.

Forecast record

Four forecasts, one per horizon, each with a threshold, a named verifier and a resolution date, recorded before the outcome is known.

Correlation and jurisdiction. All four share a parent cause in the trajectory of stablecoin regulation, so this is a single-family set rather than four independent observations. It is, however, the first set in Series I that does not concentrate on one jurisdiction: the resolution sources span supervisory records, legislation and market data across several regions. Correcting a jurisdictional concentration by increasing a mechanism concentration is a real trade and we are recording it rather than presenting the improvement alone.

Directional statements elsewhere in this article carry no threshold and are deliberately excluded from the record.

Sources

Two entries here are weaker than the rest and are marked accordingly. The Treasury exposure figure is an issuer attestation rather than an audited statement, and the comparison to sovereign holdings is our own. The emerging-market adoption point is a reported observation with no official statistics behind it — we use it to establish that the pattern exists and make no claim about its size.

In this series

—   Previous: Article 6, Adaptive vs Rigid — why the useful question about a system is not the one usually asked.

—   Next: Article 8, After the Package — what holds a system together once the functions have separated, and where Series II begins.

—   The method behind the Chaos Index and this series: /methodology

THRIVE IN CHAOS

Decision Intelligence for an Uncertain World

Analysis → Forecast → Recommendations  ·  Signal → Meaning → Action → Stability

Signal Over Noise  ·  thriveinchaos.ai

AI intelligence system with human editorial oversight.

Forecasts are probability-based analytical assessments, not certainties. This material supports independent judgment and does not constitute financial, legal or investment advice. Nothing here is a recommendation to acquire, hold or dispose of any instrument.