Somewhere in McKinsey's 2025 research cycle, a sentence was published that should have stopped more readers than it did: wealth grew faster than the real economy. Not alongside it. Faster. In a year when global output expanded at a rate most macro desks would describe as unremarkable, aggregate net worth expanded at a multiple of it.
Crypto Briefing compressed the finding into a news brief. Three claims survived the compression: wealth has decoupled from production, the decoupling is driven by asset inflation rather than value creation, and the consequences are instability and inequality. No numbers. No sample. No definition of whether "wealth" means household net worth, total social assets, or something including leverage. No statement of which price index was used to deflate anything.
I have read enough tokenomics sections to recognise a claim that cannot be falsified. This is one. But it is also testable, just not with the instrument McKinsey chose. There is a ledger that records what people actually paid, at the moment they paid it. It does not update when the headline changes.
The source degradation matters here, and most readers will skip past it. What circulated was not the report. It was a news brief about the report — a second-order compression of a document that itself contained methodology, brackets, and footnotes. By the time a claim reaches a three-paragraph brief, the falsifiable parts have been dropped and the conclusion has been retained. That is the exact failure mode I have spent my career documenting in whitepapers, and it is worth naming before engaging with the substance.
Context
Asset inflation is not a new diagnosis. It is the oldest open wound in financial economics: the observation that the prices of assets which cannot be consumed — equities, land, bonds, digital tokens — can rise indefinitely while the prices of things people actually consume stay flat. The mechanism is not mysterious. Liquidity arrives. Assets have inelastic supply in the short run, so the marginal bid reprices the entire float. Consumer demand is constrained by wages, so the repricing stops at the checkout counter. The result is calm CPI and violent asset charts.
The McKinsey framing adds a specific claim about magnitude. It is not that assets rose — that is trivial and observable. It is that asset appreciation has become the dominant source of wealth formation, eclipsing the surplus generated by actual production. When that happens, the accounting identity between wealth and output breaks.
Wealth is a stock. Output is a flow. Stocks are valued at the discounted present value of expected future flows. If the stock grows faster than the flow, the extra growth must be hiding in the discount rate. There is nowhere else for it to live. A price-to-earnings multiple expanding from 15 to 30 is not a company becoming twice as productive. It is the market deciding that the same stream of future cash is worth twice as much today. Nothing was built. The multiple moved.
In 2017 I ran that arithmetic on 45 ICO whitepapers. The "OmniChain" presale had an emission schedule that guaranteed structural sell pressure from month nine onward — not because the team was malicious, but because the vesting cliff was set against a demand model that assumed continuous new buyers. The math was visible to anyone who plotted two curves on the same axis. Fifteen thousand people read the breakdown. Almost none of them acted on it. That is the lesson I keep relearning: the ledger never lies, only the narrative obscures.
The same arithmetic applies here, one level up. If wealth is being created by discount-rate compression rather than earnings, then a change in the discount rate destroys wealth that was never produced. That is not a prediction. It is a definition, and it is the part of the McKinsey brief that the brief itself did not quantify.
Before moving to evidence, it is worth being explicit about which instruments can and cannot answer the question.
| Metric | What it measures | What it cannot measure | |--------|------------------|------------------------| | Market capitalisation | Marginal price × total supply | Capital actually placed | | Realised capitalisation | Aggregate cost basis of the float | Forward earning power | | Stablecoin supply | Hard collateral in the settlement layer | Velocity, embedded leverage | | ETF net flow | Marginal institutional demand | Holder intent, duration |
Four columns, four failure modes. Any argument built on one of these numbers alone is an argument waiting to be broken.
Core
Crypto gives us something macro data does not: a continuous, public, timestamped record of who paid what, and when. The relevant metric is not market capitalisation. Market cap multiplies a single marginal trade by the entire supply. It is a mark, not a measurement. The relevant metric is realised capitalisation — the sum of every unspent output valued at the price it last moved on-chain.
For Bitcoin, realised cap is the aggregate cost basis of the whole float. It moves only when coins move between wallets, at the price prevailing in that block. It cannot be inflated by a thin order book on a weekend. It is the closest thing this industry has to a weighted-average purchase price for the entire supply, and it is the only wealth number in crypto that requires an actual counterparty.
Market cap minus realised cap is net unrealised profit across every holder. That spread is the on-chain translation of McKinsey's divergence. It is wealth that exists only as a mark.
Look at the 2024–2025 cycle. Following the spot ETF approvals, market capitalisation expanded from roughly $0.8 trillion to a peak above $2 trillion. Realised capitalisation over the same window moved from roughly $450 billion to somewhere between $650 and $900 billion, depending on the measurement window and how aggressively you filter for exchange-internal shuffling. The gap between the two — call it $1.1 to $1.2 trillion at the peak — was not capital. It was repricing. No new dollar entered the settlement layer to create it. The float changed hands at higher prints, and the accounting wrote up the remainder of the supply to match.
Now compare that gap to the money that actually arrived. Net stablecoin issuance across 2024 and 2025 — dollar-denominated claims minted into the system, the closest thing to hard collateral — grew by tens of billions, not trillions. In 2020 I built a script to track APY sustainability across 12,000 Uniswap and SushiSwap pool transactions during DeFi Summer. The same pattern emerged then: headline yields were dominated by token emissions, while the hard capital entering the pools was an order of magnitude smaller. The ratio between marked wealth and placed capital has always been the tell. It was the tell in the yield farms. It is the tell here.
The 2025 ETF pipeline sharpens the picture. My dashboard tracked real-time institutional inflows against retail demand, processing roughly ten million daily transactions to construct a Smart Money Index. What the data showed was not that institutions bought a lot. It was that institutions became the marginal buyer. The marginal buyer sets the price. When the marginal buyer is a flow — a monthly allocation, a rebalancing mandate, a rate-sensitive advisory model — then the mark and the flow are the same variable. Price stops being an estimate of value. Price becomes a function of whether the flow continues next month.
Where this becomes mechanical rather than philosophical is in lending markets. A marked-up collateral base does not stay marked. It gets accepted as security against borrowings at a loan-to-value ratio. A $1 valuation on a token or a wrapped asset supports roughly $0.6 to $0.8 of new credit. That credit is spent, re-deposited, and re-hypothecated. Asset inflation that circulates through collateralised lending is not a wealth effect. It is a credit expansion wearing a wealth effect's clothes. When the mark compresses, the LTVs do not politely follow. They trigger. Liquidations are the mechanism by which a valuation-driven expansion converts itself into a realised contraction, and they execute without discretion.
This is where the Terra comparison stops being academic. In early 2022 I spent three weeks reconstructing Anchor Protocol deposit flows before the depeg. The withdrawal pattern was visible weeks ahead of the collapse: the marginal depositor had shifted from sticky capital chasing a yield to hot capital chasing an exit. Nobody needed a forecast. They needed a flow chart and the willingness to look at it.
The structure here is different in kind but identical in logic. When the marginal price is set by a recurring flow, the asset's valuation is a derivative of that flow's continuation, not of any productive capacity underneath it. Minsky described this as stability breeding instability. On-chain, it looks like a gap between realised and market capitalisation that nobody can close without selling — and selling is what closes it.
One more note on crypto's peculiar exposure. Equity holders can, in principle, be paid from earnings. Land can be farmed. A digital token has no cash-flow leg at all — its only fundamental is the marginal buyer's willingness to hold. That makes it the purest test instrument for the McKinsey thesis rather than a counterexample to it. If asset inflation is real, crypto is where it runs fastest and where it snaps hardest. An algorithm does not sleep, nor does it feel fear — it just updates the mark.
Contrarian
The consensus reading of the McKinsey brief is already forming, and it is wrong in a specific, expensive way. The reading goes: asset inflation is structural, therefore hold assets, therefore crypto is the hedge.
Correlation is a suggestion; causality is a truth. Crypto is not a hedge against asset inflation. It is the highest-beta expression of it. Both are repriced by the same variable — the availability and cost of liquidity. A portfolio holding equities and bitcoin as a defence against currency debasement is not diversified. It is levered to a single factor and has told itself a story about diversification. When that factor reverses, the two positions fail on the same day, in the same direction, for the same reason.
The second blind spot is subtler. The McKinsey framing treats wealth as a stock to be defended. That framing implies the wealth is sitting in accounts, waiting to be spent or taxed or inherited. But a stock valued at the margin is not a stock that can be liquidated at that value. If every holder of the marked-up supply attempted to convert, the mark would collapse before the conversion completed. The aggregate wealth figure is not a reservoir. It is a balance that exists only in equilibrium, and equilibrium is a description of a state, not a guarantee of one.
The third blind spot belongs to my own industry. The number that actually represents placed capital — stablecoin supply — is the one nobody treats as a macro variable. Yet it is the collateral layer beneath the entire venue system. Collateral is not defensive. Collateral is the thing that gets sold first when margin calls arrive. Trust the hash, not the headline. The hash says the hard capital is small. The headline says the wealth is large. Both are true, and only one of them can be redeemed.
Takeaway
Two numbers will resolve this before any narrative does. First, the ratio between realised and market capitalisation: as it compresses toward one, either the mark is being converted into genuine transactions, or the mark is being written down. There is no third path. Second, net stablecoin issuance: if hard collateral keeps growing while the mark compresses, the system is maturing. If both fall together, the flow was the floor.
The question for 2026 is not whether wealth outpaced the real economy. It already did. The question is whether the marginal buyer is still a dollar flow or has quietly become a credit flow — because those two fail in very different ways, and only one of them appears on a chart before it happens.
