There is a comforting idea that when a central bank adds money, everything simply gets more expensive by the same amount and nobody is really better or worse off. That has never once been true. Money has to enter somewhere, it reaches people in an order, and the order is the whole story.
Whoever already owns the assets the central bank is bidding up, and whoever sits closest to the point where the new money enters. QE does not post cheques to households: it creates reserves and buys bonds, which lifts bond prices, then the price of everything investors buy instead of bonds. The Bank of England's own July 2012 paper, written at the Treasury Committee's request, said asset purchases raised the financial wealth households hold outside pension funds, and that those holdings were heavily skewed, with the top 5% of households holding 40% of them. Section 04 below lets you run the mechanism yourself, section 05 measures it against 25 years of real US wage and house-price data, and section 07 covers what economists still genuinely dispute about it.
If a central bank creates a trillion dollars, who is holding it thirty seconds later?
Almost nobody can answer this, which is strange, because it is the only question that matters. We talk about money supply as though it were rainfall: it arrives everywhere, evenly, and the ground gets uniformly wet.
It is not rainfall. It is plumbing. The money enters at a specific valve, and the people standing at that valve are holding it before anyone else knows it exists.
Pick a route below. Each one puts the same $100 billion into the economy, and each one puts it in someone else's hands first.
Notice that none of these is a moral judgement. Each route can be the right call at the time. But they are not interchangeable, and pretending they are is how a policy gets described as neutral when its first effect is to hand money to a specific list of counterparties.
If the Fed created trillions, why did none of it show up in your bank account?
Because that is not what QE does. This is the single most common misunderstanding of the policy, and getting it right actually makes the distributional story sharper, not softer.
QE is a swap. The central bank creates reserves, which are a special kind of money only banks can hold, and uses them to buy bonds. Somebody who was holding a bond is now holding money instead. The total amount of financial assets in the world has not changed. Their composition has.
Step through one operation. Watch which line items appear, and which do not.
The key line is the last one. When the central bank buys from a bank, it creates reserves, and reserves are trapped: they can settle payments between banks but they cannot be spent in a shop. When it buys from a pension fund or an insurer, which is where most of the buying actually happened, the fund's commercial bank credits the fund's deposit account. That deposit is ordinary money. Broad money genuinely rises.
So the honest answer is: QE is not printing money and handing it out, but it is not a harmless accounting shuffle either. It puts money into the hands of institutions whose entire job is to go and buy another asset with it.
The central bank bought government bonds. So why did shares, corporate debt and houses all go up?
Because the pension fund that just sold its bonds does not want to sit on cash. It has liabilities to meet and a mandate to be invested. So it buys the next thing along the risk ladder: corporate bonds. The seller of those corporate bonds does the same, and buys equities. And so on, outward.
Economists call this the portfolio rebalancing channel. It is not a fringe theory. It is the mechanism the central banks themselves described, in public, as the intended effect.
Read that again with the distributional question in mind. The policy's stated success condition was that the price of shares would rise. Rising share prices are good for people who own shares. That is not a conspiracy or a side effect. It is the transmission mechanism working exactly as designed.
The awkward part is what comes next, and the Bank published that too, in the same paper: those gains land on a population that does not own assets evenly.
If prices rise 20% and your money rose 20%, you broke even. Didn't you?
Only if both happened at the same instant. They never do.
Prices do not update the moment money is created. They update as the money is spent, one transaction at a time, rippling outward from wherever it entered. Whoever spends it early buys at yesterday's prices. Whoever receives it last buys at today's.
This is not a modern insight. Richard Cantillon was an Irish-French banker who speculated in John Law's Mississippi Company around 1720 and walked away a multimillionaire while the scheme collapsed on everyone else. He then wrote down exactly why it had worked. His Essai sur la Nature du Commerce en Général was written around 1730 and published in 1755, more than twenty years after his death.
The man who named the effect had already used it.
Ten positions, ordered by distance from where the money enters. Everyone starts with the same savings. New money is created and spent down the line. Prices rise as it circulates. Nothing here is animated for effect: every figure is computed from the settings below.
Each tile: what that position receives, the price level they face when they spend it, and their net position in real goods once the erosion of their existing savings is counted.
Drag the third slider up. Stickiness is the realistic case: prices in the real world do not reprice in an afternoon, wages least of all. The stickier prices are, the longer the early positions get to shop at the old ones, and the wider the gap becomes.
Now set the second slider to 10, so that everybody gets an equal share. The gap narrows but it does not close, because the person at position 1 still spends before the person at position 10. Equal distribution is not the same as simultaneous distribution.
Over the last quarter century, what happened to the price of a house compared to the wage that has to buy it?
The model above is a model. This is not. Below are three real series, from FRED, running from 2000 to 2025: US house prices, the median weekly wage of a full-time prime-age worker, and consumer prices. Pick a starting year and watch what separates.
Two things in that chart deserve to be said plainly, because only one of them fits the usual narrative.
The first is that wages did not collapse. From 2000 to 2025 the median full-time prime-age wage rose 108% while consumer prices rose 86%, so the typical worker's real wage rose by around 12%. Anyone telling you the median worker got poorer in nominal-versus-CPI terms is not reading the same data.
The second is that the asset ran away anyway. Over that same span house prices rose 224%. Measured in weeks of median wages, a house costs 55% more than it did in 2000. Both facts are true at once. You can be earning more in real terms every year and still be falling further behind the thing you are trying to buy, because your wage is indexed to the consumer basket and the house is indexed to the asset market.
That gap is the Cantillon Effect with the model taken off.
Move the start year to 2008 and the same measure shows only 13%. That is not a bug and it is not the chart being coy: January 2008 was near the top of a housing bubble, so starting there flatters the present. Start at the 2012 trough instead and it is 53%. Any single number here is a choice of start date, and a page that shows you only its best one is selling you something.
What happens when the institution running the policy is ordered to show its distributional workings?
In 2012 the UK Treasury Committee asked the Bank of England to explain the costs and benefits of QE, specifically for the groups who felt they had been hurt by it. The Bank did not deflect. On 12 July 2012 it published The Distributional Effects of Asset Purchases, and the findings are more candid than most critics of QE have ever been.
By that point the Bank had bought £375 billion of assets, almost entirely gilts. It reported that this raised the price of gilts, which raised demand for corporate bonds and equities, which raised those prices too. It then reported who owned them.
A policy whose stated transmission mechanism is raising the price of these assets delivers its first-round gains in proportion to who holds them.
To the Bank's credit, it put the counter-argument in the same document, and it is a serious one: without the purchases, it argued, most people in the UK would have been worse off, with lower growth, higher unemployment and more companies failing. That case deserves to be weighed rather than waved away. A recession prevented is a real benefit, and it is diffuse and invisible in a way that a rising share price is not.
But both things can be true. The policy can have been better than the alternative and have delivered its gains in a specific order to a specific set of people. Cantillon would not have found that confusing at all.
If the mechanism is real, why don't economists agree on the size of it?
Because attribution is hard, and the honest answer is that the mechanism is settled while the magnitude is not. This section exists because a page that only gave you the strong version would be less useful to you.
The evidence that the divergence is real. Moody's Analytics estimates that the top 20% of US earners, those making more than $175,000, account for close to 60% of personal outlays as of the first quarter of 2026, and that their outlays grew 6.5% over the prior year while the bottom 80% grew 2.6%, below inflation. That is the K, and it is wide.
The evidence that it is not all QE. Three serious objections, all worth knowing:
| Objection | Substance | Weight |
|---|---|---|
| Relative vs absolute | Bank of England research has found monetary policy's effect on households was similar in relative percentage terms across the wealth distribution, even where the absolute cash amounts differed enormously. | Strong |
| The headline stat is soft | Moody's revised the top-10% spending share down to 45.8% after a methodology change, from an earlier 49.2%. The Minneapolis Fed has published a review of the underlying data. The K is still visible; the exact number is not gospel. | Strong |
| Confounders | Technology, globalisation, housing supply constraints, tax policy and pandemic fiscal transfers all push the same direction over the same period. Isolating the monetary contribution is genuinely unresolved. | Strong |
So the defensible claim is narrower than the popular one, and still substantial: QE is not the sole cause of asset-price divergence, but it is a policy whose intended mechanism was to raise asset prices, executed at a scale of trillions, on a population that holds those assets very unevenly. You do not need it to be the only cause for the ordering to matter.
It is tempting, on a page published by a Bitcoin infrastructure company, to end with the obvious. So here is the honest version.
Bitcoin has no discretionary issuance. There is no committee that can decide to buy assets, no counterparty list, no first recipient chosen by policy. The issuance schedule was fixed in advance and is known to everyone equally. In that specific sense, the Cantillon Effect in issuance is absent, and that is a real structural difference rather than a marketing claim.
What it does not fix: early holders of a scarce asset capture enormous gains as adoption spreads, which is its own distributional story, just one driven by timing of belief rather than proximity to a policy desk. Anyone claiming a fixed supply makes distribution equal is selling you something. It makes distribution predictable. Those are different words and only one of them is a promise.
On 18 August 2026 the US Treasury announced it would at least double the size of its liquidity-support buybacks for long-dated securities, from $2 billion to at least $4 billion per operation, effective 9 September and running through 4 November. The 30-year yield fell within the hour.
Buybacks are not QE. They are debt management, funded by issuance, not money creation, and it matters to say so. But look at the shape: an announcement is made, and the people holding those specific long-dated bonds are repriced immediately. The person holding a mortgage rate set last month, or a savings account, finds out later and receives nothing. That is the ordering, visible in an afternoon, in an instrument that is not even monetary policy.
True or false. Each one corrects a claim that circulates widely, in both directions.
Ask where the money enters, not how much there is. The quantity tells you almost nothing about who gains. The entry point tells you nearly everything. Any analysis that discusses only the size of a programme has skipped the question that determines the outcome.
Neutral money is a modelling assumption, not a property of the world. It is a useful simplification for some questions and a serious error for distributional ones. Prices adjust sequentially because spending happens sequentially. There is no version of this where everyone repriced at once.
Real wages rising and affordability falling are not contradictory. Your wage is measured against a basket of consumer goods. The asset you are saving for is not in that basket. Between 2000 and 2025 US real median wages rose about 12% while a house cost 55% more in wage-weeks. Both numbers are correct.
Read the central banks directly. The most damaging single statistic about QE's distribution was published by the Bank of England, about itself, in 2012, at Parliament's request. The primary sources are more candid than the commentary in both directions, and they are free.
Whoever holds the assets the central bank is bidding up, and whoever is closest to the point where the new money enters. The Bank of England's own July 2012 paper, written at the Treasury Committee's request, found that asset purchases raised the value of households' financial wealth held outside pension funds, but that those holdings were heavily skewed, with the top 5% of households holding 40% of them.
The observation that new money does not raise all prices at once, so whoever spends it first buys at the old prices and whoever receives it last buys at the new ones. It is named for Richard Cantillon, an Irish-French banker who made a fortune in John Law's Mississippi Company around 1720 and then described the mechanism in his Essai, written around 1730 and published in 1755.
No. QE is an asset swap: the central bank creates reserves and uses them to buy bonds, so a pension fund or bank ends up holding money where it previously held a bond. Reserves themselves cannot leave the banking system and be spent in shops. When the seller is a non-bank, however, a new commercial bank deposit is created, so the quantity of broad money does rise.
It is one contributing factor among several, and the size of its contribution is genuinely contested. Moody's Analytics estimates the top 20% of US earners account for close to 60% of personal outlays as of the first quarter of 2026, but Bank of England research has also found that monetary policy's effect on households was similar in relative percentage terms across the wealth distribution even where the absolute cash amounts differed enormously.
From roughly $0.87 trillion before the 2008 crisis to a peak near $8.97 trillion in April 2022, then down by about $2.2 trillion of quantitative tightening between June 2022 and October 2025. It stood at $6.7 trillion as of March 2026, around 21% of nominal GDP, and the FOMC judged on 10 December 2025 that reserves had reached ample levels and began buying shorter-term Treasuries again.