When Central Banks Survey What They Cannot Aggregate
On 26 June 2026 the European Central Bank published the May 2026 wave of its Consumer Expectations Survey. Across roughly 19,000 respondents in eleven euro-area countries, the median perception of inflation over the past twelve months sat at 4.0%, unchanged from April. The median expectation for inflation over the next twelve months fell to 3.5%, from 4.0% the month before. Three-year expectations held at 2.9%; five-year expectations held at 2.4%. Expected unemployment twelve months ahead ticked up to 11.3%.
Read the press release the way the ECB insists — as the best available aggregation of dispersed household beliefs — and the same numbers are an admission of failure. Lower-income quintiles continued to report higher inflation perceptions than higher-income quintiles. Respondents aged 18–34 reported lower inflation perceptions than respondents aged 35–70. The ECB notes, in plain language, that uncertainty about twelve-month inflation expectations "remained at a higher level than before the start of the war in the Middle East." A survey of nineteen thousand people, in eleven countries, still cannot agree on what inflation is. The dispersion is the data.
What the Story Claims
The official reading is sober. The CES is "used for policy analysis" and "complements other data sources." The headline takeaway is that short-term inflation expectations have fallen modestly while medium-term expectations have held steady, suggesting the public's belief in the ECB's 2% target is intact at the five-year horizon. The dispersion by income and age is treated as a feature. The methodology drops in a revealing detail: the headline figures are 2% winsorised means — a statistical adjustment that caps the influence of the most extreme five percent of responses at each tail. The ECB is not measuring a single belief. It is measuring the shape of a disagreement and presenting the middle as the result.
The Austrian Diagnosis
Friedrich Hayek (1899–1992), a Nobel-prize-winning economist who argued that no central authority can possess the dispersed knowledge held by millions of individuals, dedicated his most influential essay — "The Use of Knowledge in Society," published in September 1945 — to exactly this problem. The knowledge problem (the claim that useful economic information is not held in any single office but is dispersed, tacit, and locally known across millions of people) was, for Hayek, not an obstacle to be overcome by better statistics. It was a structural feature of any economy more complex than a village. The right response was to design institutions that let dispersed knowledge act on prices without ever being collected in one place.
Read the May 2026 CES through that lens and every number in the release is a confirmation rather than a surprise. Why do lower-income quintiles report higher inflation perceptions? Because they spend a higher share of their budget on food, energy, and housing — the categories whose prices are most volatile. Why do 18–34 year-olds report lower inflation than 35–70 year-olds? Because they entered the workforce during a period of stable goods-price inflation. Why is the uncertainty around the twelve-month figure higher than before a regional war began in 2024? Because the relevant knowledge — about energy supply, shipping rates, second-round wage effects — is held by millions of separate actors and no survey instrument can collect it.
Ludwig von Mises (1881–1973), an Austrian economist who spent his career showing that centrally planned economies cannot rationally allocate resources without market prices, made the deeper point in 1920. The calculation problem (the claim that, without market prices set by voluntary exchange, no one can tell whether a given allocation of resources is good or bad) does not only apply to steel and wheat. It applies to beliefs. The ECB cannot "calculate" the right household inflation expectation any more than Gosplan could calculate the right price of steel. It can collect answers. It cannot aggregate them honestly, because the honest answer is that they disagree.
The Price That Already Aggregates
The price system, the institution Hayek spent the second half of his career defending, exists precisely to do what the CES cannot. When a euro is spent on a litre of milk, on a tank of diesel, on a kilowatt-hour of electricity, the buyer reveals local knowledge — what they need, what they can substitute, what they will pay rather than go without. The seller reveals what their cost is and what their competitor is charging. The price that emerges is not the average of these knowledges. It is the only point at which the buyer's willingness to pay and the seller's willingness to accept meet. Every price in a market economy is, in this sense, an aggregated belief about the future, computed without anyone surveying anyone.
This is where prediction markets enter the picture. A prediction market is a financial market whose payoffs depend on real-world outcomes — will euro-area inflation exceed 3.0% in December? Will the ECB cut rates before September? The price of a YES contract on any of these questions is the market's estimate of the probability that the event will occur. The estimate updates continuously. It is not a median. It is not a winsorised mean. It is the price at which the marginal buyer is willing to take the other side of the bet — and the marginal buyer, in a well-functioning market, has the strongest information or the most skin in the game.
Why This Matters for Sound Money
Rails to Freedom makes the case in Part 3 (Ethereum as the Solution) that the value of programmable money is not that it is faster or cheaper than the existing financial system. The value is that it lets the price system extend to domains that legacy rails cannot reach. The book returns in its chapter on decentralised knowledge — itself a paraphrase of Hayek's "spontaneous order" (the insight that complex beneficial coordination can emerge from millions of independent decisions without anyone designing it) — to the argument that beliefs about the future can be aggregated through prediction markets without a central counterparty and without a periodic survey instrument.
Part 1 of the book makes the Austrian diagnosis explicit: a central planner cannot aggregate dispersed beliefs any more than it can aggregate dispersed knowledge of resources. The May 2026 CES is a clean illustration. Nineteen thousand respondents, eleven countries, and the headline output is a median that the ECB itself admits is surrounded by persistent, structured disagreement. Sampling produces a statistic. Aggregation produces a price. The ECB's instrument cannot produce the latter, because no survey instrument can. The market can.
What Markets Are Already Doing
The on-chain version of the ECB's CES is already running. Polymarket, the world's largest prediction market, is built on Polygon — an Ethereum Layer 2 (a secondary network that settles transactions back to Ethereum mainnet for security, scaling the same base protocol rather than running a separate chain) — and uses UMA's Optimistic Oracle (a smart-contract system that lets anyone propose a real-world answer, with a bond forfeited if a challenger disputes it within a challenge period) to resolve outcomes. Polymarket currently hosts active contracts on whether euro-area inflation will exceed specific thresholds, on whether the ECB will hold or cut at upcoming meetings, on whether national unemployment rates will move in specific directions. Each contract is, in effect, a continuous, priced, decentralised CES question.
The contrast with the ECB's monthly survey is precise. The CES asks nineteen thousand people what they think. The median of their answers takes three weeks to publish, comes out once a month, and is presented with a wide uncertainty band that the ECB itself notes is elevated. A Polymarket contract asks anyone with money and an opinion to put both on the table, twenty-four hours a day. The price updates every trade. The outcome is verified by UMA's oracle against an agreed-upon public source. The price is the aggregation — the only honest aggregation of dispersed beliefs that exists, because it cannot be sampled away.
Looking Ahead
The ECB will publish the June 2026 CES results on 24 July 2026. They will show another median, another uncertainty band, another set of income- and age-specific breakdowns. The Eurosystem's broader inflation outlook will continue to draw on the CES alongside market-based measures of inflation expectations — the very measures whose logic the CES is trying to replicate.
The Austrian critique is structural. The knowledge the survey tries to collect is, in Hayek's terms, knowledge of time and place — the kind that cannot be transmitted to a central authority in a survey response. It can only be revealed through action, and the most efficient action a household can take to reveal what it believes about inflation is to buy or sell at a price. The market does the aggregation. The ECB documents the impossibility of doing it any other way. That is what the May 2026 CES actually says, once you read past the median.