The Alchemy of Finance

The argument

Soros studied at the London School of Economics with the ambition of becoming a philosopher, and Karl Popper’s teaching marked him permanently: knowledge advances by conjecture and refutation, certainty is unavailable, and the open society is the political form of admitted fallibility. He then spent three decades running money, and this book is the settlement between the two careers. Its claim is that the philosophy was never abandoned; it was capitalized. The Quantum Fund, on this account, is a research program, and the book is its lab report.

The foundation is an observation about situations that contain thinking participants. A scientist studying a storm does not change the storm by forecasting it. A market participant forecasting a market is part of the market; his forecast, expressed as a position, moves the thing forecast. Soros names the two functions that collide here: the cognitive function, in which we try to understand the situation, and the participating function, in which our actions change it. Where both operate at once, neither can complete. The facts we are trying to read are partly composed of our readings of them, and understanding becomes structurally, not accidentally, imperfect. He calls this condition reflexivity, and the name is chosen against the grain of the social sciences, which had spent a century pretending the condition away in order to look like physics.

The polemical target is equilibrium economics. Supply and demand curves are treated in the textbooks as independently given, with price the innocent messenger that reconciles them. Soros answers that in financial markets the curves are not independently given, because expectations shape both, and expectations are shaped by prices. The fundamentals that valuation is supposed to anchor to are themselves price-sensitive: a rising stock price lets a company raise capital cheaply, acquire, hire, and thereby produce the earnings that justify the price; a strengthening currency suppresses inflation and attracts the inflows that strengthen it further; a rising collateral value invites the lending that raises it. Prices are never a passive reflection. They are an active ingredient, and so markets cannot converge on truth; they can only oscillate around a moving target that their own oscillation moves. Where the textbook says equilibrium, Soros says process, and the process has a characteristic shape.

That shape is the boom-bust model, the book’s most usable machine. A trend exists, at first unrecognized. A prevailing bias forms around it and begins to feed it: the trend validates the bias, the bias finances the trend. The process survives its early tests, and each survival deepens conviction. Somewhere in the middle the gap opens: perception has moved further than the underlying reality can follow, and the process now requires the bias to strengthen merely to hold position. Eventually the flaw in the perception becomes undeniable, conviction wavers through a twilight period, and then the sequence runs in reverse, compressed, because the same reflexive loop that built the structure over years dismantles it in months. The asymmetry is not decoration; it follows from credit. Booms are built on borrowing against collateral whose value the borrowing inflates, and when the direction turns, lender and collateral fail together. Credit and collateral are one organism, and the book’s account of their reflexive embrace is, to my reading, its permanent contribution.

The case studies do the arguing. The conglomerate boom of the sixties: acquisitive companies bought earnings with overpriced stock, the market priced the acquired growth as if it were organic, and the pricing was precisely what made the next acquisition affordable. The perception created the performance it claimed to be measuring, until the supply of believable acquisitions ran out and the whole structure reversed. The mortgage trusts: Soros published his analysis of their reflexive cycle in advance, boom then bust, and the cycle arrived roughly on schedule, one of the rare cases where the theory was used prospectively in print. Sovereign lending through the seventies: banks lent to countries whose creditworthiness was improved by the lending itself, measured by ratios the loans flattered, until 1982 collected the whole misunderstanding at once. And the strong dollar of the early eighties, which he names the Imperial Circle: high American interest rates and fiscal deficits drew in foreign capital, the inflows strengthened the dollar, the strong dollar suppressed inflation and justified the policy mix, and speculative capital chased the trend it was creating. A freely floating currency, he argues, is not a self-correcting system but a cumulatively destabilizing one, because the flows that dominate it are trend-following by nature. Each case has the same skeleton: a loop mistaken for a line.

Then the book does something no other investment classic attempts. Beginning in August 1985, Soros keeps a real-time diary: the macro hypotheses, the positions expressing them, the revisions as events answer. The Plaza Accord lands a month in, the dollar breaks, and the fund’s results during the experiment were extraordinary; but the profits are not the point of the exercise, and he says so. The point is epistemological hygiene. A thesis written down before the outcome cannot be quietly improved afterward; the diary preserves the misjudgments, the reversals, the positions taken for reasons that dissolved. It is conjecture and refutation with a brokerage account, Popper run at market speed.

The title carries the last claim. In natural science, theories do not alter the phenomena they describe; chemistry works whether or not the molecules have read it. In financial markets, theories are ingredients: a widely believed model changes the prices it models. That is alchemy in the strict sense, operations that transform the substance they handle, and Soros’s inversion is deliberate: it is the economists, with their equilibrium and their borrowed physics, who are practicing the false science, and the speculator, who admits he is inside the experiment, who is being honest about the material.

Working notes

The book completes a shelf. Kindleberger gives the natural history of manias, the taxonomy of a process repeating across four centuries; Soros supplies the mechanism, the loop that makes the taxonomy inevitable; Lowenstein’s autopsy of LTCM shows what happens when brilliant men lever the equilibrium assumption Soros spent this book dismantling. Read in that order, the three form a single argument: bubbles are not anomalies in the system, they are the system, observed at one of its two speeds.

What a second reading made clear is that the diary is the best technology in the book, better than the theory it was built to test. Every thesis I have watched fail, mine and others’, failed with the assistance of edited memory: the entry reasons quietly revised to fit the exit, conviction backdated, doubt erased. Ink forbids this. Soros publishing his hesitations and reversals in real time, at the height of his powers, is a more severe act of discipline than any position he describes, and it is the one part of the book that transfers to any practitioner whole, without translation.

Reflexivity also named something I had seen and lacked a word for: the moment when analysis becomes participation. A large fund publishing research on an illiquid asset is not describing the market; it is operating on it. A valuation framework, adopted widely enough, becomes a price driver and stops being a measurement. The observer effect is not a philosophical curiosity here; it is a description of what liquidity does to belief. Lefèvre’s operators knew this with their hands, running pools that manufactured the appearance of demand. Soros knew it with his whole epistemology, and wrote the physics for what the old tape readers practiced as craft.

The fallibility doctrine runs deeper than markets, and the book knows it, though it says so awkwardly. A mind that takes its own error as the base case, that holds every belief as a position awaiting refutation, is practicing something older than portfolio management. The serious contemplative traditions begin exactly there: the knower is inside the known, the instrument is implicated in the reading, and certainty is the one state that guarantees blindness. I hold that this is why the book’s stance felt familiar before it felt true. Soros arrived at epistemic humility through Popper and P&L; the inquiry I keep on the other side of my life arrives at it through stiller means; the doctrine is recognizably one doctrine, and the market is simply the place where holding it is paid and violating it is fined.

A note on the prose, since every reader collides with it: the book circles, repeats, qualifies, and defines its terms three times in three chapters, each slightly differently. I no longer read this as bad editing. It is a man performing his own thesis, unable to stand outside his subject long enough to describe it cleanly, and saying so. The style is the argument, which does not make it pleasant.

Where I push back

The irony at the center of the book is that a devoted Popperian built an unfalsifiable theory. Reflexivity explains every boom after it busts and issues no prediction that could embarrass it in advance; a trend that continues confirms the self-reinforcing phase, a trend that reverses confirms the crossover, and no observation is permitted to count against the framework. Popper’s whole demand was that a theory forbid something. Reflexivity forbids nothing. It is a lens, and a genuinely superior one, but the book insists on calling it a theory, and on that word Soros fails his own master’s test.

Second, the general claim overreaches. Soros wants reflexivity as a foundation for social science, a theory of history, ethics, and politics; what he demonstrably has is a powerful special case, credit-fed perception loops in leveraged financial markets. Inside that domain the model earns its keep on every page. Outside it, the argument thins into assertion, and the later chapters, where markets recede and world order advances, read like a rich man being humored by his editor. The book is strongest exactly where its author is an expert and weakest exactly where he is a philosopher, which is the reverse of what he wanted.

Third, the record does the persuading, and it should not. Readers credit the theory with returns that the diary itself shows were produced by feel: the timing, the sizing, the willingness to reverse in a day, none of which reflexivity supplies. The theory told him where to look. Instinct told him when to act, and instinct is not in the book because it cannot be. The dangerous reader takes the framework, lacks the instinct, and discovers that a map of loops is not a schedule of turns; he shorts the boom early, at size, with Soros’s vocabulary in his mouth, and the boom buries him in the self-reinforcing phase, exactly as the model, correctly read, said it might.

How it enters the work

BlockHedge trades the most reflexive market that has ever existed, and I mean that as a technical description. In crypto the price is not a signal about the fundamentals; over whole regimes it is the fundamentals. A rising token funds the treasury, attracts the developers, deepens the liquidity, secures the network whose security justifies the token: the loop Soros diagrammed for conglomerate earnings, run without the friction of quarterly reporting. DeFi rebuilt his credit-collateral organism natively: lending against collateral whose price the lending inflates, liquidation cascades as the same loop in reverse, compressed from years into hours. Stablecoin runs, funding-rate spirals, reflexive treasuries: the book’s case studies, refactored into software. Nothing else on this shelf describes the asset class this exactly, which is why the firm treats reflexivity not as theory but as terrain.

The operational translation is the gap. For every core position the desk maintains two ledgers: what the price implies is being believed, and what the chain and the cash flows can actually show. The thesis rides the self-reinforcing phase while the gap is narrow or funded, and reduces as the story must strengthen just to hold the numbers still, which is Soros’s twilight, made measurable. The model never calls the turn; we do not ask it to. It tells us which exposures are loops rather than lines, and loop exposure is sized so that the reversal, arriving on its own schedule and at reflexive speed, is an event and not an ending.

The diary became firm process without modification. Every thesis is written before the position exists: the loop claimed, the gap measured, the invalidation named. The document is dated and never edited, only appended, and the quarterly review reads entries against outcomes with the entries’ own words. It is the real-time experiment institutionalized, and it has the effect Soros reports: it converts fallibility from a mood into a procedure. The desk does not aspire to be right. It aspires to notice being wrong at a speed the loop cannot outrun.

And the doctrine crossed the last border on its own. Holding my own beliefs as positions, base case fallible, open to refutation, revised in ink, has become the discipline I bring to inquiry that has nothing to do with capital; the same posture, held in stillness instead of size. Soros would resist the company his book keeps there. The book keeps it anyway.