Perpetual motion at Asheville

Two things I did last week. The first was to catch the final week of the M.C. Escher exhibition at Somerset House. It was an extraordinary show. The second was to read the statement issued by the G20 finance ministers and central bank governors, who met in Asheville, North Carolina, on 31 August and 1 September. The Chair’s Statement ran to seventeen paragraphs and was agreed by everyone present except China.

Three of those paragraphs repay being read together, which is not how documents like this are designed to be read.

Paragraph 7 identifies the common impediments to growth. Regulatory and administrative burdens come first on the list, ahead of inefficient tax regimes, inadequate investment and the cost of capital. The remedy is stated without qualification: simplify and reduce red tape and overly burdensome regulations.

Paragraph 15 sets out the financial-sector programme. Ministers commit to modernising regulatory and supervisory frameworks to ensure financial stability, and to promote the resilience and efficiency of their financial systems, in order to support strong and durable economic growth.

Paragraph 9 turns to the standard-setters. It calls on them to support private sector-led innovation and to commit to the principle that standards should be “well-calibrated and fit-for-purpose”.

Read on its own, each paragraph is reasonable. Read together, they describe a structure.

The wheel turns

Escher’s Waterfall is the closest thing I know to the shape of this document.

A stone aqueduct carries water along a zigzag channel. Follow it span by span, and every stretch runs downhill, as water does. At the end of the run, the channel opens over a drop; the water falls, and it drives a mill wheel. And the foot of the fall is the head of the channel. The water has descended continuously and arrived back where it started.

Every joint in that structure is sound. The deception does not live in any single span; it lives in the sequence. A viewer can inspect each stretch of the aqueduct in turn, find nothing whatever wrong with any of them, and be entirely taken in.

That is how these documents are read. Paragraph by paragraph, each within its own frame, each drafted by people who could defend it on its own terms and very likely did. Nobody has to lie for the document to describe something that cannot exist. They only have to write it in sections.

Two further features of the print matter here.

The first is that the structure does work. The wheel turns. Escher has not merely drawn an impossibility; he has drawn a perpetual motion machine, an impossibility that yields output. The Asheville statement makes the same offer — lower burden, undiminished stability, and growth extracted from the difference.

The second is that nobody in the picture is troubled by it. A woman hangs washing on the terrace below. The aqueduct is weathered, in use, unremarkable. The impossibility is not a scandal in the world of the drawing. It is the built environment.

The absence is the point

None of which makes the statement wrong. It makes it incomplete. What is missing is a single sentence identifying where the water is being lifted.

It is worth being clear about what is not being claimed. There is nothing improper about a G20 that wants growth, stability and innovation at once. Wanting all three is the job. Every regulator I have known has lived inside that constraint, and the good ones did not resent it. The impossibility of maximising every objective simultaneously is a permanent feature of the work, not a failure of drafting.

The difficulty arrives one step further on. A trade-off that is named can be argued about, monitored, revisited and — if it turns out badly — owned. A trade-off that is not named still gets made. It simply gets made further down, by people with less authority and no mandate to make it: by the supervisor who drops the third line of enquiry because the resourcing was cut, by the policy team that writes the softer guidance because the signal from above was to reduce burden, by the head of division who does not escalate because escalation is now understood to be friction. None of those people chose the trade-off. Collectively they made it.

Supervisory capacity is finite. Every hour not spent on an examination is an hour of burden removed from a firm and an hour of assurance not obtained. Every prudential requirement relaxed is capital returned to productive use and a buffer that is no longer there. The exchange rate between these things is genuinely uncertain. The direction of the trade is not.

And when the failure comes, as it does, the record will show only that everyone wanted everything.

Four things that moved

The evidence that a trade is being made rather than a hope expressed sits in the financial-sector paragraph, and it becomes visible only by comparison.

The last occasion the G20 finance track agreed on a full text was Durban, in July 2025. Its financial-sector sentence reaffirmed a commitment to addressing vulnerabilities and to promoting an open, resilient and stable financial system, which supports economic growth, and which rests on the consistent, full and timely implementation of all agreed reforms and international standards, Basel III among them.

Set that against paragraph 15 of Asheville, and four things have moved.

Vulnerabilities have become potential vulnerabilities.

An open, resilient and stable system has become one characterised by stability, resilience and efficiency. Openness has quietly departed, having been the post-crisis settlement’s word for cross-border resilience.

The commitment to consistent, full and timely implementation has gone. That formula is not decoration. The G20 has used it since 2012; it runs through the Basel Committee’s implementation reports to ministers and the Buenos Aires communiqué of March 2018, and its persistence across thirteen years and several presidencies is what gave it force. In its place, Asheville records that members “shared updates on our respective proposals and regulations, including implementing the final components of Basel III”. Ministers no longer commit. They report. That is a change in the mood of the verb, and in a document of this kind the mood of the verb is the policy.

And the fourth. In Durban the financial system was the object of the commitment, and economic growth was what such a system supports. In Asheville the frameworks are modernised in order to support growth. Stability has moved from end to means.

That last change is not a softening of tone. It is a reversal of what the enterprise is for.

Two objections can be met quickly. There is no intervening text in which the older language might be watched eroding: Durban was the last full statement the finance track agreed, the October 2025 meeting in Washington managed only a Chair’s Summary, and the first meeting of the United States presidency, in April, produced no agreed statement at all. And Durban was a communiqué where Asheville is a Chair’s Statement, which is not quite like-for-like — but the Asheville text is written throughout in the collective first person, records no divergence in its body, and confines dissent to a footnote. It is a communiqué in all but the label.

Paragraph 3 offers its own account of how the text became shorter, welcoming the “successful streamlining of the G20 Finance track’s structure and agenda”. The commitment was not argued away. It was tidied.

The surrounding language points the same way. The FSB’s forthcoming stocktake is described as one that will “identify widespread regulatory and supervisory modernization efforts across all jurisdictions” — a stocktake framed by its expected finding rather than by its question. And paragraph 9’s instruction to the standard-setters is a quiet repositioning of what standards are for. Bodies established to raise a floor are being asked to calibrate against private-sector innovation. Both propositions may be defensible. Neither is presented as a choice.

Paragraph 16, on digital assets, reproduces the shape in miniature: responsible frameworks that preserve financial stability, support economic growth, and establish clear pathways for innovation. The same circuit, drawn smaller, and again no account of what drives it.

The old words were not kept either

There is a fair objection to all of this. The formula that has gone was not, on any reasonable reading, being met.

In October 2025, the FSB Chair wrote to these same ministers, reporting that full, timely, and consistent implementation of the agreed reforms had not been achieved, and that the financial system remained vulnerable in consequence. The sentence was reaffirmed annually and unmet annually. To mourn its passing looks sentimental.

But the value of a standard that nobody meets is not compliance. It is that the shortfall can be named. The FSB Chair could write that letter because the sentence existed — it gave him a yardstick, a published measure, and a deficit he could quantify and put in front of ministers. Remove the sentence and the shortfall does not close. It stops being a shortfall. Incomplete implementation ceases to be a gap against a commitment and becomes, simply, the position.

There is a second consequence. If implementation was already incomplete, then part of the burden now to be reduced is burden that was never fully imposed. The modernisation agenda is not trimming an over-built structure. It may be ratifying an under-built one.

And there is a respectable case on the other side, which is that a promise nobody keeps is worse than no promise, and better dropped openly than repeated as ceremony. That case can be made. Making it would have taken one sentence. Nobody wrote it.

Not long to wait to see if this is permanent

The reading offered here is testable, and the test arrives shortly. The FSB’s stocktake and its consultation report on modernisation principles are both due, and between them they will show what modernisation is taken to mean.

If it means reporting burden, data duplication and process — work that is useful and long overdue in most jurisdictions — then a drafting change has been overread here, and the older commitment can be expected to resurface in Bangkok. If it reaches capital, liquidity or the resolution architecture, then paragraph 15 was the announcement, and the omission was the point. The Thematic Peer Review on Public Sector Backstop Funding Mechanisms, flagged in the same paragraph, is worth watching alongside it: reviewing the backstops while easing the requirements is a recognisable pattern, and not an encouraging one.

The paragraph that is not there

A complete version would not have been difficult to draft. Something to the effect that members accept a somewhat higher probability of individual firm failure in exchange for a materially lower compliance burden, that this is a considered judgement about where the balance had drifted, and that the FSB will report annually on whether the trade is delivering what was expected of it.

That paragraph would have been contentious. It would also have been true, defensible, and capable of being tested against events in five years’ time. Its absence is not a drafting oversight. It is the choice.

One further detail is worth recording. China objected to paragraphs 4, 10, 11 and 13 — the geopolitical language, the two paragraphs on global imbalances, and the paragraph on sovereign debt. It did not object to paragraphs 7, 9 or 15.

The device is worth a moment. The G20 more often records dissent without attributing it, in the familiar formula about most members holding one view and others another. Here the chair named both the dissenter and the paragraphs. What results is an unusually precise map of where the group divided, and financial regulation is not on it.

On the international politics of trade and debt, the G20 divided. On the direction of financial regulation, it did not. The consensus, such as it is, is the deregulatory one. That is the more significant fact about Asheville. The wheel turns, and nobody in the document says what is turning it.

G20 Chair’s Statement, Second Meeting of G20 Finance Ministers and Central Bank Governors. Asheville, USA. 1 September 2026. Retrieved from https://home.treasury.gov/news/press-releases/sb0620.