Why Measurement Reform Fails

The 60-Second Version

1. The Measurement Problem

GDP was designed by Simon Kuznets in 1934 as a wartime production measure. He explicitly warned against using it as a measure of welfare. The warning was ignored at Bretton Woods in 1944, and GDP became the world's dominant economic indicator — not because it measured the right thing, but because it measured one thing with institutional legibility.

The core error is not that GDP is a bad measure — everyone from Kuznets onward has said that, and saying it again changes nothing. The error is treating the problem as intellectual when it is structural. Bretton Woods did not make GDP persuasive; it made GDP required. Debt-to-GDP and deficit-to-GDP are the terms in which sovereign borrowing is priced, so every government faces a mechanical compulsion to grow the denominator, whatever that does for anyone. A measure no one defends and everyone requires is not displaced by a better argument.

2. Lock-In, Not Suppression

This lineage is often described as suppressed. It was not, and the weaker word is the stronger claim. Ostrom won the Nobel Memorial Prize; Keynes reshaped the institutions of the post-war order; Graeber was a bestseller. None of it was hidden. What happened is that none of it was ever written into a lending covenant, a central bank mandate, or a sovereign risk model — and a measure that is not load-bearing anywhere can be admired indefinitely without displacing anything. That requires no conspiracy, only the ordinary mechanics of disciplinary and institutional formation, which is what makes it hard to undo.

Five thinkers each saw a piece of the same account:

The Das Adam Smith Problem — the supposed contradiction between Wealth of Nations and Theory of Moral Sentiments — dissolves under Circulatory Economics. Both books describe the same person seen whole: self-interested and sympathetic, transactional and relational.

3. Velocity, Honestly Stated

The natural candidate for a circulatory measure is velocity — the rate at which value changes hands. Measured money velocity has fallen through a long expansion in advanced economies, which is at least suggestive: output rising while each unit of money turns over less often.

Falling velocity is a symptom worth explaining, not a mechanism that explains. The better evidence measures circulation directly: Egger et al. (2022, Econometrica) estimate a general-equilibrium multiplier from unconditional cash transfers in Kenya, in which the transfer’s effect on non-recipients is observed rather than inferred from an accounting identity. The 2008 response is the concrete case in the other direction: a very large quantity injected, output stabilised, and the money moved into asset prices — which are pooling mechanisms rather than circulating ones.

Gift vs Transaction: 8 Outputs vs 1

4. The Unit of Analysis

Orthodox economics begins with the individual rational actor — homo economicus. Circulatory Economics begins with the relational unit: the relationship and the flow between actors.

Behavioural economics made a conservative move: keeping the individual unit but noting it is “irrational.” Circulatory Economics makes the radical move: changing the unit entirely. The individual isn't irrational — the unit of analysis is wrong.

5. The Three Instruments

A paradigm without measurement is philosophy. CE provides three instruments — each measuring a different dimension of circulatory health — that together constitute the first formal apparatus for the paradigm.

6. The Architectural Intervention

Architectures of Ease achieves cooperative circulation through three mechanisms — without enforcement, monitoring, or punishment:

Constitutional governance forms the substrate — the foundational architecture within which these mechanisms operate. Constellation provides this: commitments, constraints, and contestation that constitute the operating environment rather than oversight imposed after the fact.

7. Implications

Golden Age vs Financialisation

Background context, not the paper’s evidence. These are widely cited US figures assembled here for orientation; the paper makes no era comparison and rests none of its argument on one.

8. The One Pool Where This Is Testable

The design constraint says the test must not need anyone’s permission. One pool qualifies: capital already donated, already outside private hands, held in vehicles whose declared purpose is public benefit, and distributing at a mandated minimum. The vehicle exists; the engine does not.

Whether that pool is static is where the measurement question turns on itself. The industry aggregate says it moves briskly: US donor-advised funds granted $64.89bn in FY2024, a 25.3% payout against prior year-end assets — three times the 8.1% managed by private foundations, which are mandated at 5%. Read straight, that refutes the premise of this section.

It does not survive disaggregation. Recomputing from the same Form 990 data on the denominator IRS statisticians prefer — year-end assets plus grants made during the year — gives a median sponsor payout of 9.7% for 2023, stable between 9 and 10% across four years. The gap is not a rounding dispute, and each of its three sources is a measurement artefact rather than a fact about philanthropy:

• Donation processors counted as endowed pools. Workplace-giving and crowdfunding rails grant out roughly 82% in a year and hold a median account of $305. They are averaged with national sponsors holding a median account of $390,541. Pass-through payment infrastructure lifts the composite.
• Transfers between funds count as grants. DAF-to-DAF giving reached an estimated $7.5bn in 2024, about 12% of all DAF grant dollars. None of it reaches an operating charity. In 2023 the largest single grant made by the largest national sponsor went to another sponsor.
• The denominator is chosen. When the series’ previous publisher moved from year-end to beginning-of-year assets in 2014, reported distributions rose from about 15% to about 20% with nothing changing in the world. Two of the three largest sponsors use a five-year average of year-end assets, lowering the denominator further.

At the level where the question is actually posed — the individual account — the pool is close to immobile: a 2021 Council of Michigan Foundations study found a median payout of 3.1% with over a quarter of accounts granting nothing in a year, and a 2022 California Attorney General audit found a fifth of sampled accounts paying out under 5%. ⭐ Not capital that is scarce, and not capital anyone is withholding improperly, but capital that has already left private hands, already received its treatment, is legally dedicated to public benefit — and is not moving. A clot is not a shortage.

⚠️ Sources disagree and both have interests: the 25.3% figure is published by the industry’s own collaborative; the 9.7% recomputation is by an independent institute that contests it. The figures are US and FY2023–24. The claim they support survives only at the account level — the aggregate refutes it, and that inversion is the point rather than an inconvenience.