The Evolution of Money: From Seashells to Bitcoin — Why We Trust Stuff That Isn’t Real
Hand a stranger a rectangle of printed cotton and he will hand you back a hot meal. Nothing about that rectangle is edible, useful, or scarce in any natural sense. It works because a large number of people you will never meet have agreed, tacitly and permanently, to behave as if it works.
That agreement is the most powerful technology our species has ever built, and almost nobody can explain how it functions.
The Evolution of Money follows the strangest invention in human history from clay tablets recording barley debts to the quiet hum of servers settling claims that exist only as entries in a shared file. Along the way it visits the shells that circulated across three continents, the stone discs that stayed exactly where they were while their ownership traveled, the notched sticks that ran a medieval treasury, the mountain of silver that accidentally globalized the planet, the printing presses that destroyed two revolutions, the meeting rooms where a gold standard was buried, and the anonymous message board post that proposed doing the whole thing without anyone in charge.
Peter Kriger tells this story not as economic history but as the history of a shared belief that has repeatedly been engineered, exploited, over-extended, and rebuilt. Money turns out to be less a substance than a network property: it exists in the space between people rather than inside any object, which is why it can be created by decree and destroyed by rumor. Understanding that is the difference between reading a financial panic as a moral failure and reading it as a system doing exactly what its structure permits.
Witty, skeptical, and full of the kind of detail that changes how you look at your own wallet, this is a book about trust — how it is manufactured, how it is measured, how quickly it evaporates, and what happens next.
Keywords: money, economic history, trust, currency, inflation, banking, cryptocurrency
Contents
Chapter One — The Fiction We All Agree On. 18
Chapter Two — The Barter Myth. 27
Chapter Three — Debt Came First 34
Chapter Four — Cattle, Grain, and the Trouble with Storage. 42
Chapter Five — Shells, Beads, and Feathers. 49
Chapter Six — The Stones of Yap. 55
Chapter Seven — Lydia and the First Coins. 61
Chapter Eight — Athenian Owls and the Silver Mountain. 67
Chapter Nine — Rome and the Slow Poisoning of the Denarius. 73
Chapter Ten — China Invents Paper 79
Chapter Eleven — The Bark of Trees. 85
Chapter Twelve — Sakk, Suftaja, and the Word “Check”. 90
Chapter Thirteen — Templars and the Medieval Wire Transfer 96
Chapter Fourteen — England Ran on Firewood. 103
Chapter Fifteen — Potosí and the First Global Money. 108
Chapter Sixteen — The Anatomy of a Bubble. 113
Chapter Seventeen — John Law’s Machine. 119
Chapter Eighteen — The Bank That Made Debt Permanent 125
Chapter Nineteen — Assignats and Continentals. 131
Chapter Twenty — When Anybody Could Issue. 137
Chapter Twenty-One — The Gold Standard as Religion. 143
Chapter Twenty-Two — The Wheelbarrow Years. 149
Chapter Twenty-Three — 1929 and the Run as Cascade. 155
Chapter Twenty-Four — Bretton Woods. 161
Chapter Twenty-Five — Nixon Closes the Window.. 167
Chapter Twenty-Six — Plastic. 173
Chapter Twenty-Seven — The Money Nobody Governs. 179
Chapter Twenty-Eight — When the Ledger Blinked. 185
Chapter Twenty-Nine — The Speed of Disbelief 192
Chapter Thirty — The Phone Became the Bank. 198
Chapter Thirty-One — Satoshi’s Puzzle. 203
Chapter Thirty-Two — Stablecoins, Central Bank Money, and Programmable Cash. 210
Chapter Thirty-Three — What Money Actually Is. 216
Case Study One — The Bank of Amsterdam and the Secret That Lasted a Century. 233
Case Study Two — Copper Plates, Palmstruch, and Europe’s First Banknote. 239
Case Study Three — The Panic of 1907 and the Private Central Bank of One Man. 244
Case Study Four — Cigarettes, and the Economy That Assembled Itself 249
Case Study Five — Company Scrip and the Store That Owned You. 254
Case Study Six — Six Months Without Banks: Ireland, 1970. 259
Case Study Seven — The Swiss Currency That Only Exists Between Businesses. 264
Case Study Eight — Brazil Cures Hyperinflation With an Imaginary Currency. 269
Case Study Nine — Thailand, 1997: Borrowing in a Currency You Cannot Print 274
Case Study Ten — The Parking Garage Stock Exchange. 278
Case Study Eleven — Cyprus, 2013: The Depositors Pay. 283
Case Study Twelve — Sixty Euros a Day: Greece, 2015. 287
Case Study Thirteen — Iceland Lets the Banks Fail 291
Case Study Fourteen — India Cancels Its Cash. 296
Case Study Fifteen — Nigeria Redesigns Its Money Before an Election 300
Case Study Sixteen — Mackerel, Ramen, and the Prison Economy. 304
Case Study Seventeen — The Largest Currencies Nobody Regulates 308
Case Study Eighteen — The Economy That Runs Inside a Video Game 312
Case Study Nineteen — The Bristol Pound and the Limits of Local Money. 317
Case Study Twenty — Replacing the Money of Three Hundred Million People. 321
Preface
There is a small experiment you can run on a friend, and it costs nothing except the friendship of anyone who dislikes being made to think before lunch. Ask them what money is. Not what it does, not what it buys, not whether they have enough of it, but what it actually is. Most people begin confidently and lose momentum around the fourth word. They say it is a means of exchange, which is a description of a job rather than a description of a thing. They say it is value, which pushes the question one step backward without answering it. They say it is backed by gold, which has not been true for anyone alive under the age of fifty. Eventually, if they are honest, they arrive at a shrug, and the shrug is the correct answer. Money is the thing we all use constantly, depend on absolutely, and understand hardly at all.
This is stranger than it sounds. We are generally suspicious of arrangements we cannot explain. We want to know how the plane stays up, how the vaccine works, how the bridge holds. But we hand over pieces of paper and plastic rectangles and streams of electrons a hundred times a month without a flicker of curiosity about why anyone accepts them. The incuriosity is not laziness. It is a symptom of how well the system works. Infrastructure becomes invisible in proportion to its reliability, and money has been reliable enough, for long enough, in enough places, that it has slipped below the threshold of notice — like plumbing, or grammar, or the assumption that the person driving toward you will stay on their side of the road.
The invisibility ends abruptly. It ended for Germans in 1923, for Americans in 1933, for Argentines with tedious regularity, for Zimbabweans in 2008, for anyone who has stood in a line outside a bank that was not going to open. In those moments the plumbing becomes horribly visible, and everyone discovers at once that they had been trusting something rather than owning something. The paper does not change. What changes is the belief attached to it, and the belief turns out to have been the whole product all along.
So this book takes a position early, and I would rather state it plainly than smuggle it in: money is not an object. It is an agreement that we have found various ways of storing. Sometimes we store it in metal, sometimes in paper, sometimes in clay, sometimes in a chain of cryptographic hashes maintained by strangers with strong opinions about monetary policy. The storage medium changes constantly and dramatically. The agreement underneath changes hardly at all. Everything interesting about the history of money is the history of how that agreement has been made, kept, transferred, mistrusted, restored, and occasionally shredded.
This might sound like a debunking, and debunkings are cheap. It is not one. Saying that money is a shared fiction is only devastating if you believe fictions are weak. They are not. Nations are shared fictions and they field armies. Corporations are shared fictions and they build cities. Marriage is a shared fiction and it reorganizes entire lives. The category of things that exist because we jointly behave as though they exist includes most of what makes human society different from a large troop of clever primates. Money belongs to that category, and it is arguably the most successful member of it, because it is the one fiction that almost every human community has independently arrived at, in wildly different forms, without needing to be persuaded.
What follows is a history of those forms. It runs from the accounting systems of the ancient Near East, which handled credit long before anybody thought to mint a coin, through the shells and beads and metal ingots that moved along trade routes, through the invention of coinage in a small kingdom in Asia Minor and its spread by conquest and convenience, through the Chinese experiments with printed paper that Europeans found impossible to believe when they were described to them, through the medieval instruments that let a merchant move a fortune from one city to another without physically moving anything, through the silver flood that turned a lumpy set of regional economies into something recognizably global, through the age of banks and bubbles and standards, through the twentieth-century decisions that quietly untethered the entire world from metal, and into the present, where money is mostly a number in a database and the newest proposals are attempts to change who is allowed to write in it.
It is a long journey and I have tried to keep it moving. Certain conventions of economic history have been left behind on purpose. There is no mathematics in this book. Not one equation, not one demand curve, not one Greek letter standing in for something that could have been said in English. This is not because the mathematics is wrong or unimportant; it is because everything I want to say can be said without it, and because equations have a way of ending conversations rather than opening them. When a formula appears on the page, most readers experience a small polite blackout and resume attention several paragraphs later. I would rather keep everyone awake.
I have also tried to resist a particular vice of the genre, which is the tendency to present the history of money as a triumphal march from primitive to sophisticated, with each stage improving neatly upon the last. That story is tidy, teachable, and mostly false. The sequence was not a ladder. It was a series of local solutions to local problems, many of them abandoned, several of them better than what replaced them, a few of them so bizarre that they still make excellent dinner conversation. Sophisticated credit instruments existed thousands of years before coins. Paper currency was invented, used successfully, catastrophically over-issued, abandoned, and forgotten for four centuries before anyone tried it again. Societies with no writing ran elaborate systems of obligation with perfect competence. Progress, insofar as it happened, happened sideways and in circles.
The other thing I have tried to do is keep the people in the picture. Monetary history is often written as though currencies moved themselves — as though the denarius simply lost value the way a pond evaporates. Currencies do not evaporate. Someone decides. Someone in a mint decides how much silver goes into the alloy this year. Someone in a treasury decides how many notes to print by Friday. Someone in a committee decides that convertibility is suspended as of Sunday evening. These decisions are usually made by tired people under pressure, with incomplete information, trying to solve a problem that seemed more urgent than the one they were creating. Understanding money means understanding the incentives of the people permitted to make it, which is why so much of this book is about political situations that appear, at first glance, to have nothing to do with economics.
A quiet thread runs beneath the chapters, and it may be worth flagging so that it does not seem to arrive from nowhere in the closing stretch. Monetary systems behave like other systems built out of many mutually dependent parts. They tolerate a great deal of strain and then stop tolerating it very suddenly. Confidence in them is not a dial that turns smoothly downward; it is a state that holds and then flips, because each participant is watching the others and adjusting accordingly. This is why monetary collapses are so consistently described by their victims as arriving overnight, when the underlying deterioration was visible for years to anyone who cared to look. It is also why restoring confidence is so much harder than losing it, and why the restorations, when they work, tend to involve some theatrical act of commitment rather than a gradual improvement in the numbers. Readers who care about such things will recognize this as a family resemblance between monetary panics and other cascading failures. Readers who do not care may safely ignore the whole line of thought and enjoy the shipwrecks.
One more warning about scope. This is not a book about how to make money, keep money, or invest money. It contains no advice, and any reader who takes financial guidance from a historical narrative about tally sticks deserves whatever happens next. Nor is it a partisan book about what monetary policy ought to be. There are strong opinions in circulation about central banks, gold, deficits, and digital currencies, and they are held by serious people on several sides. I have tried to describe what has actually happened when various arrangements were tried, including the parts that embarrass every camp. The gold standard did impose discipline and it did also deepen depressions. Fiat currency has enabled both recovery and ruin. Cryptocurrencies did solve a genuinely hard technical problem and did also reproduce, at speed, most of the frauds of the nineteenth century. History is generous with evidence and stingy with vindication.
A word about vocabulary, since imprecision here causes more confusion than any other single thing. Money, currency, cash, and credit are not synonyms, though newspapers use them interchangeably every day. Currency is the physical or official form issued by an authority. Cash is whatever settles a transaction on the spot with no residual obligation. Credit is an obligation that has been made transferable, which is to say a promise that can be handed to someone else. Money is the broader category that all of these fall into, defined not by what it is made of but by what it does: it lets people compare unlike things, settle debts, and carry purchasing power across time. Almost every argument about whether some new instrument counts as real money is actually an argument about which of those three jobs it performs and how badly. Keeping the jobs separate in your head will make several later chapters much less bewildering, especially the ones where a society is running on credit while insisting it is running on metal.
The book is built to be read straight through, but it is not a single argument stretched over three hundred pages. Each chapter carries its own episode and its own point, and several of them will contradict the impression left by the chapter before, which is intentional rather than careless. Behind the narrative chapters sit a set of case studies drawn from outside the main story — episodes I could not fit into the chronological line without breaking it, ranging across company scrip, occupation currencies, prison economies, ration systems, loyalty points, and several modern failures whose participants would be startled to learn how closely they were following a script written centuries earlier. There is also a glossary, because the field is thick with terms that sound technical and mean something quite simple, and a timeline, because it helps enormously to see how much of this happened simultaneously in places that had never heard of each other.
The last thing to say is about the title, which promises to explain why we trust stuff that isn’t real. The short answer, offered here so that no one feels cheated if they never reach the end, is that we do not trust the stuff. We trust each other, indirectly, through the stuff. A banknote is a portable piece of confidence in a very large number of strangers — that they will accept it tomorrow, that the institution behind it will not do anything catastrophic this quarter, that the state will keep demanding taxes in it, that the shopkeeper down the road will not suddenly decide he prefers chickens. The note itself is merely the physical token of a social fact. When the social fact stays intact, the token can be almost anything: a lump of metal, a slip of paper, a groove cut into wood, a line in a ledger, a cryptographic signature. When the social fact dissolves, no token in the world will save it, and people have tried everything, including gold.
That is the argument. The rest is evidence, and the evidence is more entertaining than the argument, which is the usual arrangement with human affairs. It involves cowrie shells hauled across an ocean, a king who tested purity by boiling metal, an emperor whose paper notes were printed on the inner bark of a tree, a French monarchy destroyed by a Scottish gambler with a brilliant idea, an English treasury that burned down its own building while disposing of its accounting records, and a pizza that cost enough Bitcoin to buy a small island. If nothing else, you will finish this book with a better class of anecdote.
Money is the largest, oldest, most consequential act of collective imagination our species maintains. It deserves better than a shrug. Let us go and look at it properly.
Chapter One — The Fiction We All Agree On
In January 1991 the government of Somalia ceased to exist. The president fled, the ministries emptied, and the central bank in Mogadishu was looted so thoroughly that the building itself was eventually stripped of its fittings. There was no longer any institution issuing the Somali shilling, no authority standing behind it, no reserve of anything, no promise from anyone. By every textbook account the currency should have died on the spot.
It did not. Somalis went on using the old notes for years. They circulated, they were counted, they bought goats and rice and mobile phone minutes. Because no new notes were being printed by any legitimate body, the existing stock became gradually more valuable relative to the goods available, until enterprising businessmen began commissioning fresh printings abroad, using the old plates, and flying them in. These were, by any legal definition, counterfeits — forgeries of the currency of a state that no longer existed. Somali market traders accepted them anyway, examined them for quality, and rejected the ones that were badly made. The economy ran, in a rough and often violent way, on money that had no issuer, no backing, and no legal status, distinguished from worthless paper only by the collective decision to keep treating it as money.
This is the cleanest natural experiment we have. Strip away everything a currency is supposed to require — the state, the bank, the reserves, the law — and if enough people continue to accept the tokens, the tokens continue to work. Add all of those institutions back and remove only the acceptance, and the notes become fire-lighting material within days. Whatever money is, it evidently lives in the acceptance rather than in the paper or the promise.
That claim has an uncomfortable feel to it, because it seems to make money arbitrary, and money does not feel arbitrary. It feels like the hardest and least negotiable fact in most people’s lives. But there is no contradiction. A thing can be entirely dependent on collective belief and still be immovably real for any individual caught inside it. Language works this way. There is nothing about the sound of the word “water” that connects it to the substance; the connection is pure convention, invented and maintained by speakers. And yet no individual English speaker can decide tomorrow that “water” now means “ladder.” The convention is arbitrary in origin and iron in operation. Money is the same species of fact. Your inability to declare your own banknotes worth double is not evidence that their value is objective. It is evidence that you are one participant in a very large agreement and the agreement does not take requests.
It helps to notice what money is being asked to do, because the jobs are separable and are often performed by different things at the same time. It has to serve as a unit of account, so that a horse and a haircut and a year of rent can be compared on a single scale. It has to serve as a medium of exchange, so that the person selling the horse does not have to want a haircut. And it has to serve as a store of value, so that today’s work can be turned into next winter’s food. Societies have repeatedly split these functions across different objects. Prices might be reckoned in one commodity, settled in another, and savings held in a third, with everyone perfectly comfortable about it. The assumption that a single thing should do all three is a modern habit, not a law of nature.
The awkward question is why anyone accepts the token in the first place. There is a neat circular answer: I accept it because I am confident you will accept it, and you accept it because you are confident about the next person, and so on around a loop that has no origin. Circular reasoning is normally a vice, but here it describes the mechanism accurately. Acceptance is self-confirming. This is why currencies can be extremely stable for long periods and then destabilize with alarming speed: the loop holds together while everyone believes it holds together, and there is no gentle intermediate state between believing and not.
Circularity alone would be a shaky foundation, so most successful monetary systems have bolted something onto it. The most reliable bolt has been taxation. If a government demands payment of taxes in a particular token, and enforces that demand with the full unpleasantness at its disposal, then every household in the territory needs to obtain that token whether or not it privately believes in it. Demand for the token becomes as certain as the tax collector, and the tax collector is very certain indeed. This single mechanism explains a startling number of otherwise puzzling episodes in monetary history, including several colonial administrations that introduced taxes for the explicit purpose of forcing people who were perfectly happy without money to start earning some.
The other traditional bolt is convertibility: the issuer promises to swap the token for something else on demand, usually metal. This looks like the opposite of a fiction — the note is only a claim check on a real object in a real vault. But the appearance is misleading, because the promise only works if it is never widely tested. No issuer in history has held enough metal to redeem all its outstanding claims simultaneously, and none has pretended otherwise in private. Convertibility is not a guarantee that the fiction is unnecessary; it is a device for making the fiction credible. The metal is a prop, and everyone in the theater has agreed not to look behind it.
None of this means money is fake, and here the argument has to be defended against its own most popular misuse. There is a genre of commentary that discovers money is a social construct and concludes triumphantly that it is therefore not real and could be abolished by an act of will. This is like discovering that the rules of chess were invented and concluding that you may therefore move your bishop in a straight line. Constructed things are still binding on everyone inside the construction. Debts constructed out of nothing have removed people’s houses. Currencies constructed out of nothing have decided which countries could import food. If anything, the constructed nature of money makes it more consequential, not less, because it means the arrangement can be altered by decisions rather than being fixed by physics — and decisions have interests behind them.
Two further features of the arrangement deserve notice, because both will recur constantly. The first is that producing the token is nearly always far cheaper than what the token commands. A modern banknote costs a few cents to manufacture and buys a great deal more than a few cents’ worth of anything. That gap is called seigniorage, after the medieval seigneur who ran the mint and kept a slice, and it is the reason issuing money has been a jealously guarded privilege in every society that has had it. Whoever controls the issue receives real goods in exchange for cheap tokens, and the transfer is invisible to almost everyone it affects.
The second is that money is a network good, which is to say its usefulness rises with the number of other people using it and collapses when they stop. This has an odd consequence: monetary systems tend toward winner-takes-most outcomes within any territory. Two currencies rarely coexist comfortably for long; one becomes the thing prices are quoted in and the other becomes a curiosity, unless a law props the loser up. It also explains the phenomenon known as dollarization, in which a population abandons its own perfectly legal national currency in favor of somebody else’s, not out of admiration for the foreign government but because everyone else in the market has already done so. Nobody votes on this. It happens the way a crowd decides which side of the corridor to walk on.
Which brings us to the character of the story ahead. Because money is an agreement, its history is a political history disguised as a technical one. Every change in the form of money has redistributed something: from creditors to debtors or the reverse, from countryside to city, from one empire to its trading partners, from savers to spenders, from those who could obtain the new instrument early to those who received it late. Debasing a coinage was not a mistake made by rulers who failed to understand metallurgy; it was a transfer from wage-earners and lenders to the treasury, performed by people who understood exactly what they were doing and hoped nobody else would. Suspending convertibility was not an accident of paperwork. Devaluing against a neighbor was not an act of God.
The forms will change beyond recognition over the following chapters. Grain in a temple silo, metal weighed on a scale, metal stamped with a face, notched wood, printed paper, engraved plates, telegraph instructions, punched cards, magnetic stripes, database entries, cryptographic signatures. It is tempting to read that list as a march toward abstraction, from solid things to ideas. The truth is closer to the opposite: it was always an idea, and the solid things were interfaces. What has actually changed over five thousand years is not the degree of abstraction but the identity of the people who get to keep the ledger, and the speed at which the ledger can be updated.
So the question worth carrying through this book is not “what was money made of.” It is “who was trusted, by whom, to record what, and what happened when that trust ran out.” Somalia is instructive precisely because it shows how much of the apparatus can be removed while the underlying agreement survives. Other chapters will show the reverse: apparatus in perfect working order, vaults full, laws in place, and the agreement evaporating anyway over the course of a single afternoon.
Chapter Two — The Barter Myth
Nearly every economics textbook published in the last two centuries opens the story of money the same way. Long ago, before money, people traded goods directly. This was inconvenient, because the shoemaker who wanted bread had to find a baker who happened to want shoes. Economists call this the double coincidence of wants, and it is a genuine problem. To solve it, the story goes, communities gradually settled on some commonly desired commodity as an intermediary, and money was born, spontaneously, out of the sheer annoyance of the alternative.
It is a lovely story. Adam Smith told a version of it in 1776, complete with a butcher, a brewer, and a baker. It has the shape of a good origin myth: a problem, an ingenious fix, and a moral about the natural genius of commerce. It has been repeated so often that most educated people believe it is a historical finding rather than a thought experiment.
There is one difficulty, which anthropologists have been pointing out politely for over a century and impolitely for the last few decades. Nobody has ever found it. Not one society, anywhere, at any period, has been documented running its internal economy on direct barter and then inventing money to escape the inconvenience. Caroline Humphrey summarized the state of the evidence in 1985 with the observation that no example of a barter economy, pure and simple, has ever been described, let alone the emergence of money from one; all available ethnography suggests there never has been such a thing. The field has spent the decades since looking harder and not finding it.
What the ethnography finds instead is credit. In small communities where everyone knows everyone, exchange is handled by memory and obligation. You give your neighbor half a pig because he is your neighbor, and both of you know, without any written record, that the balance is now tilted and will be corrected at some future point in some acceptable form — labor, a daughter’s wedding gift, help with a roof, a share of the next hunt. Nobody settles up immediately, because settling up immediately is what you do with strangers. Among people you expect to see for the rest of your life, an unsettled obligation is not a problem to be eliminated; it is the connective tissue of the community. Insisting on paying a friend on the spot is mildly insulting in most human societies, including our own, which is why offering cash to a dinner host is a social catastrophe.
Barter does exist, but it turns up in specific and revealing places. It appears between communities rather than within them — at borders, in ports, between groups with no shared law, no shared enforcement, and no expectation of continued relationship. The famous silent trade described by Herodotus, in which Carthaginians left goods on a beach and withdrew while the locals left gold and withdrew in turn, is barter between parties who did not trust each other enough to meet. Barter is what you do when credit is impossible, which is nearly the reverse of the textbook sequence.
The other place barter appears is in the wreckage of a monetary system. When the ruble’s value became unpredictable in the nineteen-nineties, Russian industry reverted to swapping outputs directly on a spectacular scale: factories paid suppliers in tires, in cement, in fertilizer, in bricks, sometimes in unsold inventories of goods nobody wanted. Estimates put the share of industrial transactions conducted without money above half at the peak. Workers were paid in the products of their own plants and sent to sell them at the roadside, which produced the memorable spectacle of towns where every third person was selling identical crystal glassware. In Argentina after 2001, neighborhood barter clubs enrolled millions of members. Barter is not the ancestor of money. It is what people fall back on when money fails, and everyone involved describes the experience as a nightmare.
There is a third setting worth mentioning, and it embarrasses the myth further. Many societies conducted enormous flows of goods through gift exchange, which is neither barter nor commerce but works better than both at holding a network of islands or clans together. In the kula ring of the Trobriand Islands, men sailed dangerous open water to hand over shell necklaces and armbands, which traveled in fixed opposite directions around a circuit of islands and could not be permanently kept. The valuables had names and histories; possessing one temporarily conferred prestige, and passing it on obliged the recipient to reciprocate later with something of comparable standing. Meanwhile, in the shadow of these ceremonial voyages, a great deal of ordinary practical trade got done. The prestige exchange created and maintained the relationships within which the mundane exchange became safe. That is precisely backwards from the textbook, in which relationships are irrelevant and only the goods matter.
The barter story also quietly assumes that the difficulty being solved is finding a counterparty. But in a village the counterparty is your cousin, and the actual difficulty is arithmetic across time: remembering that he helped with the harvest two years ago, that you lent him a plow, that his mother fed your children during an illness, and arriving at a shared sense of where the balance now stands. Human beings are superb at this within a few hundred relationships and hopeless beyond it. Every monetary technology in this book is, at bottom, a prosthetic for that limit — a way of extending trust past the range at which memory and reputation work unaided.
Why does the myth persist? Partly because it is useful for teaching, since it isolates one real function of money in a clean way. Partly because it is flattering: it presents money as an efficiency innovation arising from voluntary exchange, which sits comfortably with a particular view of how societies work. And partly for the same reason all origin myths persist, which is that a story explaining where something came from is psychologically satisfying whether or not it is true, and nobody enjoys replacing a clear narrative with a shrug.
The consequences of getting it wrong are not merely academic. If money arose from barter, then money is prior to debt: first you have coins, then you can lend them. If money arose from systems of obligation, then debt is prior to money, and coins are a late technology for handling obligations between people who cannot rely on memory or trust. Those two pictures produce completely different intuitions about, for example, whether a government can meaningfully “run out” of its own currency, or whether a debt is a sacred moral fact or an accounting convention that societies have periodically rewritten when the arithmetic became socially intolerable.
There is also a subtler error buried in the barter story: the assumption that the hard problem is exchange. It is not. Exchange is easy. The hard problem is memory — keeping track of who owes what to whom, across time, across distance, and among people who may lie about it. A village of two hundred can hold that ledger in its collective head, gossip serving as the audit function. A city of fifty thousand cannot. What has to be invented is not a lubricant for swapping goods; it is an external memory system reliable enough that strangers will act on it.
Which is exactly what shows up in the earliest evidence, in the place where the writing happens to survive.
Chapter Three — Debt Came First
Around 3200 BCE, in the temple complexes of southern Mesopotamia, administrators had a problem that would be recognizable to anyone who has run a warehouse. Barley was arriving, being stored, being disbursed to workers, being loaned to farmers, being owed by some households and prepaid by others. The quantities were large, the transactions were numerous, and the officials responsible for them were mortal and occasionally dishonest. They needed records.
The solution was to press marks into damp clay with a cut reed, let the tablet dry, and keep it. Tens of thousands of these tablets survive, and the overwhelming majority of the oldest ones are not poetry, law, or prayer. They are receipts. Ration lists. Records of who delivered how much and who is behind. Writing, one of the two or three most consequential inventions in human history, appears to have been invented by accountants, for accounting, and only later borrowed by people with something more interesting to say.
What those tablets record is a fully functioning credit economy operating without coins, which would not be invented for another two and a half millennia. The temple and palace ran an internal unit of account: the shekel, defined as a specific weight of silver, roughly one-hundred-eightieth of a mina, itself corresponding to a fixed measure of barley. This unit was used to price everything — rent, wages, fines, oxen, beer, and the services of a boatman — but the silver itself mostly stayed where it was. Ordinary people did not walk around with it. Beer was bought on a tab at the alehouse, settled at harvest, in grain. The shekel was a yardstick, not a thing that changed hands.
This is the single most useful correction the ancient Near East offers. A society can price everything in silver, keep meticulous accounts in silver, enforce contracts in silver, and conduct almost no transactions in silver. The unit of account and the means of settlement had come apart four thousand years before anyone thought this required a theory.
The tablets also record interest, and record it early. Loans of grain and silver carried rates that were more or less standardized: commonly twenty percent on silver and thirty-three and a third percent on barley, the higher rate on grain reflecting its greater riskiness and the seasonal desperation of the borrowers. The Sumerian word for interest, mas, also meant a young goat or calf, which suggests the original metaphor: a loan of livestock naturally produces offspring, so a loan naturally produces increase. The metaphor was extended to metal, which does not breed, an extension that would trouble moralists from Aristotle to the medieval canon lawyers and remains the crux of the argument in Islamic finance today.
Debt in this world was not an abstraction. Behind the tablet stood real collateral, and the collateral was frequently a person. A farmer whose harvest failed borrowed against the next one; when that failed too, he pledged his sheep, then his field, then his daughter, then his wife, then himself. Debt bondage was the standard mechanism by which free households were converted into dependents, and it was the standard cause of social explosion. Peasants who saw no way out fled to the hills, joined bandits, or simply stopped farming, and a king with an emptying countryside and a growing population of armed debtors in the hills had a serious problem.
The Mesopotamian solution was radical by modern standards and entirely routine by theirs. New kings, and kings in trouble, declared a clean slate. The Akkadian term was andurarum, the Sumerian amargi — the latter written with signs meaning something close to “return to mother,” since freed debt-slaves went home. Consumer debts were annulled, pledged land returned, bonded family members released. Commercial loans between merchants were generally left alone, since the point was not to abolish credit but to prevent the peasantry from being permanently absorbed by it. Some three dozen such declarations are documented across two thousand years. The Hebrew Jubilee described in Leviticus, with its seven-times-seven year cycle, is the same institution given a calendar and a theology.
Two things follow from this that are worth carrying forward. The first is that the sanctity of debt is not a timeless human intuition. It is a specific legal and moral position, adopted at particular times, and the earliest recorded societies with sophisticated credit systems treated periodic cancellation as a normal instrument of statecraft — less a betrayal of contract than routine maintenance, like dredging a canal. The second is that these societies had already worked out something modern finance rediscovered painfully: an economy in which debts grow arithmetically faster than the capacity to pay them will eventually produce either mass default or mass unrest, and it is generally cheaper to choose the moment yourself.
The Code of Hammurabi, promulgated around 1750 BCE, reads in places like a price list precisely because it was trying to hold this system together. It sets the wage of a field laborer, the hire of an ox, the fee for setting a bone, and the penalty for a housebuilder whose work collapses. It caps interest and voids loans on which the creditor has taken more than the legal rate. It states that if a storm or drought destroys a crop, the debtor owes no interest that year — an ancient act of God clause. These are not the concerns of a barter society groping toward the invention of coins. They are the concerns of an administration managing a dense web of financial obligations with tools of real sophistication.
There is even a plausible ancestor of writing itself in the accounting apparatus. For several thousand years before the tablets, Near Eastern sites yield small clay tokens in distinct shapes — cones, spheres, disks — that appear to represent quantities of specific commodities. At some point these tokens began to be sealed inside hollow clay balls, called bullae, presumably so that the parties to an agreement could not alter the count. This created an obvious inconvenience: to check the contents you had to break the container. Someone therefore began impressing the tokens into the outside of the ball before sealing it, so the exterior displayed what the interior contained. And once the marks on the outside were doing the work, somebody eventually noticed that the tokens inside had become ceremonial, and flattened the ball into a tablet. If this reconstruction is right, then the entire tradition of written literature descends from a tamper-proof invoice.
The system also reached well beyond the temple walls. In the early second millennium BCE, merchants from Assur maintained a trading colony at Kanesh in Anatolia, a thousand kilometers away, and their archives have survived in bulk: thousands of letters and contracts on clay, packed in envelopes, complaining about late payments in language any modern supplier would recognize. Tin and textiles went north by donkey caravan, silver came south, and the whole operation ran on partnership contracts in which investors put up capital for a term of years and shared the return. Family firms extended credit across borders they did not control, to counterparties they saw once a season, enforced by nothing stronger than reputation and the threat of exclusion from future business. Long-distance finance did not wait for coins, or banks, or states willing to enforce foreign contracts.
At the other end of the social scale, the same accounting mind produced the bevel-rim bowl, a crude mass-produced vessel found by the tens of thousands across Uruk-period sites. The bowls are roughly uniform in size, ugly, and made so cheaply that they appear to have been discarded after use. The most persuasive interpretation is that they were ration containers: a standardized daily measure of barley issued to workers. Whether or not that reading is exactly right, the impulse behind it is unmistakable. Somebody wanted the portions to be equal, countable, and beyond dispute. Standardization of the unit came before anybody thought to stamp a face on it.
None of this looks like the textbook story. There is no barter phase, no spontaneous emergence of a favored commodity, no coins for two and a half thousand years. What there is instead is a unit of account imposed by an institution, a dense network of obligations recorded in a durable medium, an enforcement apparatus with unpleasant powers, and a political mechanism for wiping the record when the obligations grew socially unbearable. Every element of a modern monetary system is present except the physical currency, which turns out to be the last piece to arrive and the least essential.
The next question, then, is what happened when people did start using actual commodities as money — and why so many of them turned out to be so spectacularly badly suited to the job.
Chapter Four — Cattle, Grain, and the Trouble with Storage
The English word “pecuniary” comes from the Latin pecunia, money, which comes from pecus, cattle. “Capital” comes from caput, a head, as in head of livestock. The medieval “fee,” ancestor of both the fee you pay a lawyer and the feudal fief, descends from a Germanic root meaning cattle. Three of the most abstract words in modern finance are fossilized cows.
This is not a coincidence, and it is also not quite the evidence for cattle-as-currency that it is usually taken to be. Cattle were the principal form of storable wealth across an enormous stretch of the ancient world, from Ireland to India, and wealth measured in cattle was wealth counted in a unit everyone understood. But a cow makes a wretched medium of exchange, and the reasons why are worth spelling out because they define what a monetary commodity actually needs.
A cow cannot be divided. Half a cow is not worth half a cow; it is worth a great deal less, and only for about three days. A cow is not uniform: a healthy six-year-old milker and a scrawny bullock with a limp are both “one cow” and nobody in a bargaining position will accept that they are equivalent. A cow requires maintenance, which means holding wealth in cattle costs you money continuously, in fodder and labor and land. And a cow can die, which is a form of monetary risk not shared by silver. Livestock served brilliantly as a store of value and as a unit for reckoning fines and bride-price — the early Irish law tracts price everything from an insult to a severed finger in cows — and served terribly for buying a jug of beer.
Grain has the opposite profile. It divides perfectly, it is uniform enough to be measured by volume rather than inspected individually, and it is wanted by everyone, since everyone eats. Ancient Egypt built an entire monetary architecture on this. Under the Ptolemies the country ran what amounted to a national grain bank: harvests were deposited in royal granaries, depositors held credit against their stores, and transfers between accounts could be made at a distance by written order, with a central office in Alexandria clearing the balances. Grain never moved. Only the entries moved. This was a functioning giro system, using wheat as the reserve asset and papyrus as the transaction layer, and it was doing in the third century BCE approximately what a modern clearing bank does, with better weather.
Grain’s defect is that it rots, and it takes up an appalling amount of room. A year’s wages in wheat requires a shed. Damp destroys it, rats eat it, and its value swings wildly with the harvest, so a debt fixed in grain can double or halve in real terms depending on the rain. Every commodity money carries this carrying cost, and it produces a curious behavioral effect: when holding your wealth is expensive, you spend it. Economists call the phenomenon demurrage, and monetary reformers have periodically tried to build it into currencies deliberately, on the theory that money which slowly evaporates will circulate faster and stimulate trade. The most famous experiment ran in the Austrian town of Wörgl in 1932, where locally issued notes required a monthly stamp to retain their value; the local economy revived so conspicuously that the Austrian central bank shut it down within a year, on the reasonable grounds that it had not authorized anyone else to print money.
Between the cow and the grain sack, societies tried nearly everything that could be counted. Iron spits in early Greece, six of which made a handful — the drachma, from the verb meaning to grasp, later became the name of a silver coin, so that classical Athenians were buying olives with a unit named after a fistful of skewers. Bricks of compressed tea in Central Asia and Siberia, which had the notable advantage of being drinkable when the market failed, and which circulated into the twentieth century. Salt bars in Ethiopia, called amole, cut to standard size in the Danakil depression and carried inland, where they served as small change well within living memory. Cocoa beans in Mesoamerica, valuable enough that the Aztecs recorded a case of a counterfeiter who emptied the husks and refilled them with mud, thereby inventing the debased coin without needing metallurgy.
Japan’s version is instructive because it lasted so long alongside coins. Under the Tokugawa, the wealth of a domain and the stipend of a samurai were reckoned in koku, a volume of rice notionally sufficient to feed one adult for a year. A lord was not described as rich; he was described as a hundred-thousand-koku lord. Samurai received rice, and since armed gentlemen were forbidden to engage in trade, they sold their rice through brokers in Osaka, who advanced them cash against future stipends and eventually built the world’s first organized futures market on the resulting paper. A warrior class that despised commerce ended up financed by a derivatives exchange, which is the kind of outcome history arranges when nobody is looking.
The pattern across all of these is the same. A commodity gets chosen because it is already wanted, already produced, and already trusted. Then the monetary role begins to distort it. Producers make more of it than any consumer would want, because the surplus is not food or clothing but purchasing power. The Aztec counterfeiter, the tea brick adulterated with twigs, the salt bar cut a fraction short: each is the same story of a good whose value has floated free of its usefulness, at which point the incentive to game it becomes overwhelming.
Which explains the long migration toward metal. Metal cannot be eaten, which sounds like a disadvantage and is in fact the point. Its value is almost entirely monetary and ornamental, so it does not fluctuate with the harvest. It does not rot, does not need feeding, divides without loss, and can be melted and recombined. It is heavy relative to its bulk, which makes it awkward to carry and very awkward to steal in quantity. Above all it is hard to fake convincingly, because density and color and behavior under heat are difficult to counterfeit without the actual material.
But metal by weight is not yet coinage, and the gap between them lasted millennia. Silver circulated in the ancient Near East as bits, rings, and spirals, snipped off and weighed on a balance for every transaction. This works, but it requires a scale, a set of standard weights, and mutual confidence that the weights have not been shaved. Every serious market had disputes about weights, which is why so many ancient legal codes and prophetic denunciations concern themselves with honest measures. And weighing does nothing about purity: silver alloyed with lead looks like silver and weighs slightly more, and telling the difference required an assay that few merchants could perform in a doorway.
So the outstanding problem, by the middle of the first millennium BCE, was not the invention of money. Money had existed for three thousand years. The problem was verification at the point of sale — how to let two strangers agree instantly on how much value had just changed hands without a laboratory and a court. Before anyone solved it with a stamp, several societies solved it with objects that were valuable precisely because they were impossible to produce locally.
Chapter Five — Shells, Beads, and Feathers
If you had asked a merchant in eighteenth-century Bengal, a slave trader on the Bight of Benin, a Chinese antiquarian, and a fisherman in the Maldives what a cowrie shell was, you would have received four answers: currency, currency, an ancient character in the writing system meaning wealth, and dinner. The cowrie is the closest thing to a global currency the pre-modern world produced, and it achieved that position without any state, any treaty, or any central authority whatsoever.
Cypraea moneta is a small, glossy, porcelain-hard shell with a distinctive slit opening. It is durable enough to survive decades of handling, small enough to string in the hundreds, uniform enough that counting is unambiguous, and effectively impossible to counterfeit, since it is a biological product with a texture no craftsman could imitate. Critically, it comes from a limited set of warm shallow waters, above all the Maldives, where the harvest was managed as a royal monopoly. Ibn Battuta, visiting in the fourteenth century, described the collection method — palm fronds sunk in the lagoon, shells clinging to them, the animals left to rot away in pits — and noted that the local rulers exported them by the shipload in exchange for rice.
From the Maldives the shells moved to Bengal, where they served as small change for a market economy of enormous size, then overland and by sea to West Africa, where European traders discovered a lucrative and morally squalid arbitrage: shells could be bought for almost nothing in the Indian Ocean and were money in Africa. Dutch and English ships carried cowries as ballast on the outward leg, and they were exchanged for enslaved people. The mathematics of that trade are not something one can present neutrally, and the fact that a currency worked well is the least important thing about it.
What the cowrie demonstrates is that a monetary system can be extremely stable for centuries and then be destroyed from outside in a single generation. As long as supply depended on Maldivian labor and Indian Ocean shipping, the quantity in circulation grew slowly and predictably. Then in the nineteenth century European traders began importing a second species from the East African coast, cheaper and available in far greater volume, and shipping it in industrial quantities using steam vessels. The West African cowrie economy inflated catastrophically. Prices in shells multiplied several times over within decades, savings held in shells dissolved, and the currency that had run the region’s trade for four hundred years was finished — not by a policy error, not by war, but because somebody found a bigger supply and a faster boat.
The North American parallel is wampum, and it followed the same arc at higher speed. Wampum was not a found object but a manufactured one: cylindrical beads laboriously drilled from the shells of whelk and quahog, white and dark purple, strung and woven into belts. In Iroquois and Algonquian societies the belts were not money in the market sense. They were records: mnemonic devices whose patterns encoded treaties, condolence rituals, and the terms of alliances, presented formally and read aloud by those trained to interpret them. Value lay in the labor, the meaning, and the ceremony.
Dutch and English colonists, arriving with metal tools and a keen interest in the fur trade, noticed that wampum was accepted by inland peoples who had no use for European goods. Coin was chronically short in the colonies, so wampum was pressed into service as ordinary currency: Massachusetts made it legal tender in 1637, with a fixed rate against the shilling, and it was accepted for taxes, tuition at Harvard, and ferry fares. Then the colonists applied steel drills and, eventually, water-powered machinery. A workshop in New Jersey run by the Campbell family produced wampum industrially into the late nineteenth century. Beads that had taken a skilled worker days to make could now be turned out by the thousand. The colonies dropped wampum as legal tender within a few decades, and its value in the interior fell steadily thereafter. A currency whose scarcity depended on the difficulty of drilling a hole did not survive the arrival of a better drill.
Other societies chose objects so specific that no outside supply could exist at all. On Santa Cruz in the Solomon Islands, the high-value currency was coils of red feathers, thousands of them from the scarlet honeyeater, glued in overlapping rows onto a fiber band and rolled into a hoop. Making one required a specialist, a season of trapping, and knowledge held by a small number of families. Coils were graded by brightness, since the color faded with age and handling, so the currency had a built-in depreciation schedule and a quality hierarchy running through ten distinct classes. They were used above all for bride-price, the transaction that mattered most and needed the most reliable valuation.
On Rossel Island, off New Guinea, the anthropologist W. E. Armstrong described a system so elaborate that he initially believed the islanders were charging compound interest. Two separate currencies, one of polished shell disks used by men for major transactions and one of shell strings used by women for lesser ones, each divided into a ranked series of named values, with strict rules about which rank could be substituted for which, and time-dependent obligations attached to loans. Later scholars argued Armstrong had over-mathematized what he saw. Either way, a community of a few thousand people with no writing maintained a two-tier currency with ordinal denominations, which should end any lingering assumption that monetary sophistication tracks technology.
The common thread is that every one of these systems worked precisely as well as its scarcity held. The shell, the bead, and the feather were not valuable because of beauty, though they were beautiful, and not because of usefulness, since none of them did anything. They were valuable because getting more of them was hard in a way everyone could verify. When the difficulty was removed — by a steamship, a steel drill, a machine — the value went with it, quickly and without mercy, and the people holding savings in the old medium simply lost them.
That leaves one more object, on one more island, which took the logic to its limit by making the money so difficult to move that it was eventually not moved at all.
Chapter Six — The Stones of Yap
Somewhere in the Pacific, off the coast of the island of Yap, there is a large limestone disk lying on the seabed. It has been there since some point in the nineteenth century, when the canoe raft towing it was caught in a storm and the crew cut it loose to save themselves. Everyone agreed the stone had been quarried, carved, and successfully transported up to the moment of the accident. Everyone therefore agreed that its owner still owned it. It remained part of his wealth, was included in his transactions, and passed to others in exchange for goods and services — all without any prospect of ever being seen again.
The rai stones of Yap have become the standard illustration in this subject, and there is a risk of treating them as a charming curiosity. They are not a curiosity. They are the clearest physical demonstration ever produced that money is a record rather than an object, and the islanders understood this considerably better than the visiting economists did.
The stones themselves are disks of crystalline calcite, pierced through the center so a pole could be run through them, ranging from the size of a saucer to nearly four meters across. Yap has no limestone. The stone came from Palau, some four hundred kilometers away across open ocean, and later from Guam. Expeditions had to negotiate with the local landowners for quarrying rights, cut the disk with shell and stone tools, and raft it home on bamboo. Men died on these voyages, and this mattered: a stone whose transport had cost lives was worth more than an identical stone that had come home easily. The value was not in the calcite. It was in the story attached to it — who quarried it, when, at what cost, and how many people could vouch for the account.
Because the largest stones weighed several tons, they were not moved when ownership changed. They stood outside a village meeting house or beside a path, and everyone simply knew whose they now were. There was no register, no title deed, no clerk. The community held the ledger in shared memory, updated it publicly when a transfer was agreed, and enforced it by the simple fact that lying about the ownership of a four-meter stone in front of the neighbors who witnessed the transfer is not a promising strategy. Ownership was information, and the stone was merely the object the information pointed at, which is why the stone at the bottom of the sea worked perfectly well.
The account reached economists through a 1910 article by the anthropologist William Henry Furness, which was read with delight by John Maynard Keynes and later used by Milton Friedman, two men who agreed on almost nothing else. Friedman drew a comparison that deserves to be better known. In 1932, alarmed by rumors that the United States would abandon the gold standard, the Bank of France asked the Federal Reserve to convert its dollar holdings into gold. The Federal Reserve did not ship any gold to Paris. Staff went into the vault under Liberty Street, moved bullion from one compartment to another, and labeled the second compartment as French property. Headlines announced a drain on American gold. Nothing had gone anywhere. A quantity of metal in a basement in Manhattan had been relabeled, and this was understood by everyone concerned to be a real economic event with real consequences for interest rates and confidence.
There is no meaningful difference between the drowned stone and the relabeled gold except that one arrangement is administered by villagers with excellent memories and the other by clerks with excellent filing systems. In both cases the physical object is inert and stationary, and the thing that moves is a claim recorded in a trusted place. It is worth holding on to this the next time somebody insists that digital money is a dangerous novelty because nothing tangible changes hands. Nothing tangible has changed hands for most of monetary history. The tangibility was always for reassurance.
Yap also supplies an unusually neat demonstration of what happens when the difficulty of production is removed. In the late nineteenth century an Irish-American adventurer named David O’Keefe was shipwrecked on the island, recovered, and worked out that Yapese labor could be hired for copra harvesting if he paid in the one thing the islanders actually wanted. He acquired a ship, sailed to Palau, and began quarrying rai with iron tools and hauling them home in a hold rather than on a raft. The stones arrived by the dozen. They were also, in the Yapese assessment, worth less: they had cost no lives and no hardship, and the islanders discounted them accordingly, maintaining a two-tier valuation between old stones and O’Keefe stones. A community with no economics faculty spontaneously invented a distinction between hard money and easy money and applied it to the same physical substance.
The German administration, which acquired the islands in 1899, provided the final lesson. Wanting roads built and finding the islanders unenthusiastic, officials sent men out to paint black crosses on selected rai stones, declaring them the property of the government until the work was done. No stone moved. No force was applied. The marks were removed when the paths were finished, and the roads got built. This is taxation and property seizure conducted entirely in the symbolic layer, and it worked because the symbolic layer was where the money had always lived.
The uncomfortable implication for anyone attached to physical currency is that Yap did not represent a primitive stage on the way to something better. It represented a ledger system with an unusually theatrical set of markers, closer in its logic to a modern payment network than to a purse full of coins. The islanders had solved the memory problem with public witness and shared memory, which works up to a few thousand people who all know each other. The rest of the world had to solve the same problem for millions of strangers, and the solution it eventually hit upon was to shrink the object until it fit in a hand and stamp it with a face that could be recognized instantly by someone who knew nothing about the person offering it.
Chapter Seven — Lydia and the First Coins
The river Pactolus ran down from Mount Tmolus through the Lydian capital of Sardis, in what is now western Turkey, and it carried gold. Not nuggets — fine grains, washed out of the mountain and settled in the gravel, recoverable by anyone patient enough to work a sheepskin through the shallows. The Greeks explained the deposit by saying that King Midas had bathed in the river to rid himself of his catastrophic gift. The Lydians, who had the deposit rather than the myth, simply became rich.
The gold was not pure. It came out of the ground alloyed with silver in a natural mixture the Greeks called electrum, after the word for amber, which it resembles in color. And electrum has a nasty property for anyone trying to trade with it: the ratio of gold to silver varies from one deposit to another and one grain to the next, and it cannot be judged by eye. A lump might be seventy percent gold or forty. Since gold was worth roughly ten times silver, that difference was the difference between a fortune and a swindle, and no merchant standing in a market could tell.
The available test was the touchstone, a piece of dark slate against which the metal was rubbed to leave a streak, compared against streaks from needles of known composition. It worked, in the hands of an expert with a good set of needles and good light, and it was hopeless for the ordinary business of buying bread. What Lydia did, somewhere around 600 BCE, was to solve the problem administratively rather than technically. Lumps of electrum were made to a standard weight, and then struck with a punch bearing a mark — a lion’s head, in the royal issues. The mark did not describe the metal. It said that an authority stood behind the piece and would be embarrassed, or worse, if the piece turned out to be short.
This is the invention. Not metal money, which was three thousand years old, and not standard weights, which were older. The invention was outsourced verification: a small object carrying a statement of guarantee that a stranger could evaluate in a second by recognizing a picture. Coinage converted a metallurgical question into a political one, which is a spectacular trick, and it has been the basis of currency ever since. The reason the design on money is always a face, an animal, a monument, or a national emblem is that the design is not decoration. It is the signature.
The earliest pieces were large — a full stater would buy something substantial, perhaps a month of a laborer’s wages — and they were fractioned down to tiny divisions weighing a fraction of a gram, small enough to lose in the folds of a garment. Nearly a hundred of these were found under the foundations of the temple of Artemis at Ephesus, sealed there when the building went up in the sixth century BCE, which is how the dating of the whole business became reasonably firm.
There was a catch that took a generation to notice. If the state guarantees a coin of electrum, and electrum’s composition varies, then the state is guaranteeing something it has not itself verified — or, if it has verified it, the state now knows the true gold content and everyone else does not. Modern analysis of early electrum coins finds their gold content notably lower than the natural river alloy, which suggests the Lydian mint was quietly adding silver and pocketing the difference. Seigniorage arrived with coinage itself, in its first decade, before anyone had a word for it.
Croesus, the last Lydian king and the one whose wealth became proverbial, closed the loophole. Excavations at Sardis have uncovered an industrial refinery: workshops where electrum was parted into its components by cementation, packing thin metal sheets with salt and brick dust and heating them for days so the silver was drawn off as chloride, leaving gold behind. With refined metal available, Croesus issued separate coins of pure gold and pure silver at a fixed ratio between them. This is the first bimetallic currency, and it eliminated the ambiguity that had made the royal stamp necessary in the first place — while, of course, leaving the stamp in place, because by then everyone had learned to trust the picture.
The idea traveled fast in the Greek world, where several hundred independent cities discovered they now had a technology that was simultaneously a payment system and an advertisement. Aegina struck turtles, Corinth struck the winged horse, Ephesus struck the bee. A city’s coinage announced its existence to anyone who handled it, which is why so many small and otherwise negligible communities issued silver they barely needed. The word “money” itself descends from this habit: Roman coins were struck at the temple of Juno Moneta, Juno the Warner, and the name of the building became the name of the product.
Two other civilizations arrived at coinage on their own, and it is worth noticing how differently. In northern India, from perhaps the sixth century BCE, silver was cut to weight and struck with multiple small punches — sun symbols, animals, geometric marks — apparently applied by successive authorities or merchant guilds, so that a well-traveled karshapana accumulated stamps like a passport collecting visas. In China, coinage grew out of tools: bronze spades and knives, first traded as useful objects, then miniaturized into non-functional token versions, then simplified into the round coin with a square hole that would remain the standard for two thousand years. Chinese coins were cast in molds rather than struck, made of base metal rather than precious, and strung by the thousand through the central hole, which meant Chinese money was designed from the start for the small transactions of ordinary people rather than the large payments of soldiers and states.
That difference in design reflects a difference in purpose that runs through this entire subject. A silver stater is a device for paying an army, settling a fine, or storing wealth. A string of bronze cash is a device for buying vegetables. The first invites hoarding and travels far; the second circulates furiously and stays home. Nearly every dispute about monetary policy in the following twenty-five centuries is a disagreement about which of those two jobs the money is chiefly for.
The Greek world would take the stater and build the first international currency out of it, using an accident of geology quite as convenient as the Pactolus and rather more consequential.
Chapter Eight — Athenian Owls and the Silver Mountain
In 483 BCE, miners working the state-owned silver deposits at Laurion, in the rocky southeastern corner of Attica, struck a rich new vein. The revenue belonged to the city, and the Athenian assembly faced a pleasant question: what to do with a windfall. The conventional proposal was to distribute it, ten drachmas to every citizen, which was real money and would have been extremely popular.
Themistocles argued instead for spending the entire sum on warships. He is reported to have made the case not against Persia, which would have sounded alarmist, but against Aegina, the trading rival across the water, which was a grievance everyone already had. The assembly voted for the ships. Two hundred triremes were built. Three years later the Persian fleet was destroyed at Salamis by those ships, and the political history of the Mediterranean turned on a mining accident and a piece of misdirection in a public debate.
Laurion was worked on a scale that is easy to underestimate. Thousands of shafts and galleries, some over a hundred meters deep, cut through the limestone; ore was crushed, washed in carefully engineered cisterns that recycled scarce water, and smelted in furnaces whose slag heaps are still visible. The labor was performed by enslaved people in numbers that contemporaries put in the tens of thousands, under conditions that everyone at the time understood to be lethal. The silver that made Athenian coinage the most trusted money in the ancient world was extracted by people who had no share in any of it, and no account of the elegance of the coin should be allowed to obscure that.
The coin itself was the tetradrachm, bearing the helmeted head of Athena on one side and her owl on the other, with a sprig of olive and the first three letters of the city’s name. It was struck in enormous quantity and, crucially, its weight and purity did not change. For something like two centuries the Athenian owl was minted to the same standard, through wars, plagues, and political revolutions. Merchants from Egypt to the Black Sea learned that an owl was an owl, and it became the reference currency of the eastern Mediterranean — imitated in Egypt, in Arabia, in Bactria, by authorities who found it easier to copy a trusted design than to persuade anyone to trust their own.
The Athenians also declined to update the design. The Athena portrait remained deliberately archaic, in a style that had been out of fashion for generations, because a familiar face was worth more than an elegant one. This is the earliest recorded instance of a monetary authority understanding that changing the appearance of the currency costs credibility, a lesson that central banks with redesign committees continue to relearn.
Because the coinage was trusted, it became an instrument of empire. At some point in the fifth century Athens issued a decree — known from fragmentary inscriptions found across the Aegean — requiring the cities of its alliance to use Athenian coins, weights, and measures, to close their own mints, and to hand in their local silver for restriking, with a commission retained. Enforcement was assigned to officials in each city, and the text was to be displayed publicly in every marketplace. Whatever the precise date, the intention is plain enough: monetary union imposed from the center, generating revenue for the imposing power and eliminating the awkwardness of allies with independent currencies. The allies were not consulted about it, and several of them mention the matter, in other contexts, with feeling.
Then came the long war with Sparta, and Athens learned the other half of the lesson. By 407 BCE the treasury was empty, the Laurion mines had been abandoned after a Spartan garrison at Decelea gave the enslaved miners the opportunity to escape — which twenty thousand of them promptly took — and the silver was gone. The city melted down the gold ornaments of the statues on the Acropolis and struck emergency gold coins. When that ran out, it issued bronze pieces plated with silver and declared them equivalent to the real thing.
The result was described the following year by Aristophanes in a chorus of The Frogs, comparing the city’s treatment of its politicians to its treatment of its coinage: the old, honest, full-weight pieces have vanished from circulation, and everyone now passes around the wretched new bronze. Nobody spends a good coin when a bad one will settle the same debt, so the good ones go into jars under the floor and the bad ones do all the work. The observation would be attributed two thousand years later to an English financier, but a comic playwright had it first, and put it in a song.
What happened next is the part usually left out. Athens lost the war, was occupied, was stripped of its fleet and its empire, and within a few years had withdrawn the plated bronze, resumed silver coinage at the old standard, and passed a law — the text of it survives, inscribed on stone from 375 BCE — appointing a public tester in the marketplace, a slave owned by the city, whose job was to examine coins on demand, confiscate counterfeits, and confirm that good foreign imitations of Athenian silver were acceptable. Anyone refusing a coin the tester had approved could be prosecuted. A defeated, impoverished city rebuilt its monetary credibility by hiring an official whose entire function was to make the guarantee visible again.
The Athenian experience contains, in miniature, most of what follows in this book. A currency became international because it was reliable rather than because it was ordered to be. It was then used as a tool of political control, which worked until the power behind it weakened. It was debased under military emergency, with entirely predictable consequences that a comedian could see and the government could not admit. And it was restored not by finding more silver but by re-establishing, publicly and institutionally, the mechanism by which strangers could verify what they were being handed.
Rome would repeat the entire sequence at greater length, with a much larger empire, and would fail to arrive at the last step for a very long time.
Chapter Nine — Rome and the Slow Poisoning of the Denarius
Early Romans were, by Mediterranean standards, monetary latecomers and rather embarrassed about it. While Greek cities struck elegant silver, Romans paid in aes rude — shapeless lumps of bronze, weighed on scales. The formal legal ritual for transferring property in Roman law preserved the memory of this for centuries: the parties met with a witness holding a balance, and a piece of bronze was struck against the pan, long after nobody actually weighed anything. Legal procedure keeps the fossils of dead technologies the way language does.
The denarius arrived around 211 BCE, in the middle of the war against Hannibal, when Rome needed to pay large numbers of soldiers over long distances. It was a small silver coin, and its name meant “containing ten,” being worth ten bronze asses. It kept its weight and purity substantially unchanged for two hundred and fifty years, which is a remarkable record for a state at war for most of that period, and it financed the conquest of the Mediterranean world.
The first significant reduction came under Nero in 64 CE: the coin was made slightly lighter and its silver content dropped from near-pure to about ninety percent. This was not a crisis measure. It was arithmetic. An emperor who has just rebuilt a burned capital, is funding grain distributions, and has an army to pay can either raise taxes, which is visible and unpopular, or make each pound of silver into more coins, which is invisible for a year or two. Every ruler who has ever had a mint has faced this choice, and the historical record is close to unanimous about which way they go.
What makes the Roman case instructive is the shape of the decline. It was not a collapse but a two-century slide, each step small enough to be deniable. Ninety percent under Nero, roughly eighty under Marcus Aurelius, sixty under Septimius Severus, forty under his successors. Then, in 215 CE, Caracalla introduced a new coin, later called the antoninianus, which was tariffed at two denarii but contained only about one and a half denarii’s worth of silver. This is the first clearly documented instance of a state creating an explicitly overvalued coin — announcing, in effect, that the difference between the metal and the value was now official policy. The public responded exactly as anyone would expect: the old good denarii disappeared from circulation, and within a generation the antoninianus was the only silver coin anyone actually saw.
By the reign of Gallienus, in the 260s, the “silver” coin was a copper disk with a wash of silver on the surface, perhaps five percent of the whole, applied so that it looked right when new and wore through to red within months. The empire had reached the point where its money was, quite literally, a picture of money. Prices in Egypt, where the papyrus record is best, rose by something on the order of a hundredfold across the third century, and rose fastest at exactly the moments when the silver content dropped fastest.
The mint workers themselves understood the situation with some precision. In 271 CE, when the emperor Aurelian moved to reform the coinage and audit the mint at Rome, the workers revolted. The rebellion had to be put down by the army in street fighting on the Caelian Hill, with casualties reported in the thousands — the historical sources are unreliable about the number and entirely consistent about the fact. The mint staff, it appears, had been skimming metal and issuing underweight coins on their own account, and an audit threatened a business they had been running for some time. The forgery of the currency had reached the point where the forgers were the official mint.
The state’s own behavior gives the game away most clearly of all. Increasingly, the Roman government declined to accept its own coins for its own taxes, demanding payment in goods or in bullion by weight. Soldiers were paid in rations and equipment through the annona rather than in cash. When a government stops trusting its own money, it has effectively published an internal memorandum admitting the whole arrangement, and every merchant in the empire reads it.
Diocletian, restoring order at the end of the century, tried to fix prices instead. The Edict on Maximum Prices of 301 CE set legal ceilings for over a thousand goods and services — wheat, wine, wages for a carpenter, the fee for a lawyer’s services, the price of a lion — with death prescribed for violation. Fragments of the edict have been found across the eastern empire, carved in stone, which tells us it was taken seriously. It failed immediately. Goods vanished from legal markets, since selling below cost is not an improvement on not selling at all, and a black market absorbed the trade. A contemporary Christian writer, no friend of Diocletian, noted with satisfaction that much blood was shed over small and cheap items until the law was allowed to lapse. Price controls address the symptom by making it illegal to report it.
The actual repair came from Constantine, and it came in gold. The solidus, introduced around 309 CE at seventy-two to the Roman pound, was struck to a consistent standard and kept there — not for a reign, but for something close to seven hundred years, through the entire Byzantine period, which makes it the most durable coin in recorded history. It was accepted from Sri Lanka to Scandinavia. A sixth-century Egyptian merchant wrote that all nations conducted their trade in it, and treated this as evidence of divine favor toward the empire.
The solidus worked for a reason that has nothing to do with metallurgy. It was reserved for the transactions the state cared about — taxes, senior salaries, large payments — and the state made itself the guarantor of a single high-value instrument while allowing the small change beneath it to fluctuate. Ordinary people, who dealt in copper, went on suffering. The empire had learned that credibility could be rebuilt, and had also learned that it was cheaper to rebuild it for the wealthy than for everyone.
Meanwhile, at the other end of the continent, a civilization with a very different attitude to bureaucracy was preparing to skip the metal question altogether.
Chapter Ten — China Invents Paper
The province of Sichuan, in the tenth century, had a currency problem of an unusually literal kind. The region was short of copper, so its coinage was cast in iron, and iron is worth far less by weight. To buy a bolt of silk a merchant needed something like sixty kilograms of coins. Contemporary accounts describe shoppers arriving with carts. Whatever else this arrangement discouraged, it discouraged shopping.
The workaround was invented by private businesses, not by the state. Merchants in Chengdu began accepting deposits of iron cash and issuing printed receipts — jiaozi, “exchange notes” — which could be handed on to a third party who could then redeem them at the shop. This is a deposit receipt evolving into a bearer instrument, and it happened for the mundane reason that carrying paper is easier than carrying a cart. Sixteen of the largest merchant houses eventually formed an association to standardize the notes, printed on paper with distinctive seals and designs to frustrate forgery, and by the early eleventh century they circulated widely enough that some of the issuers began quietly issuing more notes than they had iron.
The Song government did what governments do when a private sector invents money: in 1024 it shut the merchants down and took the business over, establishing a state bureau in Chengdu to issue notes with a three-year term and a fixed reserve. For a while the arrangement was disciplined. Notes were retired and reissued on schedule, backed by a defined fraction of metal, and confined to particular regions.
Discipline lasted as long as the military situation did. Confronted with the Jurchen invasions and the loss of the north, the Song and their successors did the arithmetic every besieged treasury does, and printed. The huizi notes of the Southern Song were issued in accelerating volume, their reserve ratio quietly abandoned, their term extended and then ignored. By the middle of the thirteenth century they had lost most of their value, and the government was attempting to prop them up by requiring their use in tax payments — which is the correct policy, applied far too late and in too small a dose.
It was the Mongols who took the idea to its logical extreme. Under Kublai Khan the Yuan dynasty issued the chao, a paper currency with no convertibility whatsoever. There was no promise to redeem it in metal, because the state had confiscated the metal: gold and silver were to be handed in at government offices in exchange for notes, and private transactions in bullion were prohibited. The chao was legal tender across the largest contiguous empire in history, from Korea to the edge of Persia, and for roughly a generation it worked beautifully. Trade across Eurasia was settled in paper issued by a Mongol administration, which is one of the least likely sentences in monetary history and also true.
It ended the way these things end. Military campaigns, an extravagant court, and a series of failed expeditions produced deficits, and deficits produced issue. By the 1350s the notes were being printed in quantities that made them nearly worthless, and popular songs of the period joked about using them as wallpaper. The dynasty fell in 1368 amid rebellion, famine, and flooding, and the paper currency was one grievance among many, though a conspicuous one.
The founder of the succeeding Ming dynasty, having watched all this at close range as a peasant and then as a rebel, drew the wrong conclusion. He restarted paper currency in 1375 with the Great Ming Precious Note — and made two decisions that guaranteed failure. There was no redemption in metal, which was defensible. But the government also declined to accept the notes in payment of most taxes, and it issued no mechanism for withdrawing old notes from circulation. Money that the issuer will not take back has lost the one anchor that keeps a fiat currency afloat. The notes fell steadily, and by the 1420s a note with a face value of one string of cash was trading for a few percent of that. The state kept printing them anyway, chiefly to pay officials, who received them at face value and disposed of them as fast as they could.
What followed is one of the great reversals. China, which had invented paper money four centuries before anyone else and run it at continental scale, abandoned it almost entirely and did not seriously return to it for hundreds of years. The economy moved onto silver — uncoined silver, weighed in ingots and fragments and tested by assayers, which is to say it moved back to the Mesopotamian system. The great Ming tax reform of the sixteenth century, the Single Whip, consolidated dozens of levies and labor obligations into a single payment in silver, which made the fiscal system enormously simpler and also made the largest economy in the world structurally dependent on silver it did not produce.
That dependency will matter a great deal in a later chapter, because the silver had to come from somewhere, and the place it came from was on the other side of the planet.
The Chinese experience settles a question that gets argued about as though it were open. Fiat money is not a modern invention, not a twentieth-century aberration, and not a consequence of central banking. It was invented in the eleventh century, matured in the thirteenth, was destroyed by war finance in the fourteenth, and was abandoned as a bad idea by the people who invented it. The variable in all four cases was not the material. It was whether the issuer could resist the temptation that possession of a printing press creates, and whether the issuer took its own notes back. Those two conditions have determined the fate of every paper currency since, and neither of them has anything to do with gold.
Chapter Eleven — The Bark of Trees
Somewhere around 1298, in a Genoese prison, a Venetian merchant dictated an account of his travels to a cellmate who wrote romances for a living. The resulting book described the eastern world in detail that no European could check, and it contained a chapter that readers found harder to believe than the descriptions of unicorns.
Marco Polo reported that the Great Khan had a mint in the city of Khanbaliq, and that in this mint he made money out of the bark of a tree. The inner layer of the mulberry, between the wood and the outer bark, was soaked, pounded into a pulp, and pressed into sheets, which were cut into rectangles of graded sizes, stamped with officials’ seals, and finished with a vermilion imprint of the Khan’s own seal. With these pieces of tree, Polo wrote, one could buy anything in the empire, and everyone accepted them as readily as gold, and merchants arriving from abroad were required to hand over their gold and silver and take paper in exchange — which they did willingly, since with the paper they could buy the goods they had come for.
European readers concluded that this part, at least, was invented. The manuscript acquired the nickname Il Milione, either from a family name or from a widespread suspicion that it contained a million lies. The problem was not that paper money was implausible in principle. The problem was that in Europe money was metal, metal was value, and the notion that a ruler could declare a strip of processed vegetation to be worth a horse, and have the population agree, sounded like sorcery. Polo, who had lived under the system for years, does describe it as a kind of alchemy, and remarks that the Khan had thereby mastered the art of making gold, which the philosophers had been seeking without success.
The friar William of Rubruck, who had traveled to the Mongol court decades earlier, had mentioned the paper in passing without anyone paying attention. Later visitors confirmed it too. Confirmation did not help; the fact remained outside what a European could imagine doing at home, and it stayed a curiosity rather than a proposal for the better part of four hundred years.
Except once, and the exception is the most useful experiment in this entire book. In 1294 the Ilkhanate — the Mongol regime ruling Persia, cousins of the dynasty in China — found itself with an empty treasury after heavy military spending and a collapse in agricultural revenue. A senior minister who had traveled east proposed the obvious remedy: adopt the chao. Printing blocks were prepared in Tabriz, notes were produced with Chinese characters on them alongside the Islamic profession of faith, and a decree announced that from a fixed date all transactions were to be conducted in paper. Refusal was a capital offense. Gold and silver were to be surrendered.
The economy of Tabriz stopped. Not gradually — within days. Merchants closed their shops rather than sell goods for paper. Those who had goods hid them; those who needed goods could not get them. The bazaars emptied, food became unobtainable in the city, and the population began leaving. There were riots. Within about two months the decree was withdrawn, the experiment abandoned, and the minister responsible was executed, though he had other enemies and other charges against him.
The comparison is nearly a controlled trial. The same technology, the same idea, the same imperial family, applied in two places within a few decades. In China it ran for generations. In Persia it lasted a season. The difference was not metallurgy, not education, not any quality of the paper. It was that the Yuan inherited a thousand-year-old bureaucracy with tax registers, licensed mints, a tradition of state-issued instruments, and a population accustomed to official documents having force — and the Ilkhanate had a conquest regime with none of that, ruling a merchant society that traded across borders into places where the notes meant nothing at all. Paper money is not a technology. It is a relationship with an institution, and you cannot import the relationship in the same crate as the printing blocks.
There is a second lesson buried in the Tabriz episode, which is about speed. The Roman debasement took two centuries to work through. The Persian paper failure took eight weeks. The difference is that a debased coin still contains something, so its value falls in proportion to what has been removed, gradually and negotiably. A pure fiat instrument has no floor beneath it. Its value is entirely a state of collective belief, and states of collective belief do not decline smoothly; they hold and then flip. Everyone in the Tabriz bazaar was watching everyone else, and the moment enough shopkeepers closed, staying open became irrational for the rest. Nothing about the notes changed between the Tuesday when they were accepted and the Thursday when they were not.
Europe, meanwhile, went on with metal, and would go on with it for centuries — which raises an obvious question, since European commerce in this period was plainly not conducted by merchants dragging chests of silver across the Alps. Something else was moving the value. It was, and had been for some time, moving on paper, but paper of a completely different kind: not currency issued by a ruler, but private promises between named parties, which is a technology with an entirely separate ancestry running back through the trading cities of the Islamic world.
Chapter Twelve — Sakk, Suftaja, and the Word “Check”
Every time an English speaker writes a check, they use a word that has traveled from Persian through Arabic into medieval Latin and out the other side, and which originally meant a written document, an authorization to pay: sakk. Its plural is sukuk, a term that has come back into English in the last few decades attached to Islamic bond instruments, so the word has made the round trip.
The financial world it comes from was shaped by a prohibition. The Quran forbids riba, an increase stipulated on a loan, and the classical jurists interpreted this strictly: lending a hundred dirhams and requiring a hundred and ten back is forbidden, regardless of who benefits and how the money was used. This is not an unusual position historically — Christian canon law took the same view, as did Aristotle, on the ground that metal is barren and cannot reproduce — but the Islamic world enforced it across an enormous commercial zone during a period of intense long-distance trade, and the result was several centuries of extremely inventive workarounds and, more importantly, several genuine alternatives to interest-bearing debt.
The main alternative was partnership. In the mudaraba, one party supplies capital and the other supplies labor and expertise; profits are divided according to an agreed ratio, and losses fall on the capital, since the working partner has already lost their time. This is equity finance, and it is a considerably more sophisticated instrument than a loan, because it aligns the interests of the parties instead of setting them against each other. Italian merchants adopted the identical structure under the name commenda, and it financed the medieval Mediterranean trade that made Venice and Genoa rich. Whether the Italians borrowed the form directly or arrived at it independently is argued about; the resemblance is close, and the trade routes were shared.
For moving value, the instruments were the suftaja and the hawala. A suftaja was a document obtained from a banker in one city, payable by an associate in another, so that a merchant traveling from Baghdad to Aleppo carried a letter rather than a purse. This solved the problem that bandits do not accept correspondence. It also came uncomfortably close to a fee for the use of money, which is why the jurists debated it endlessly, some permitting it on the grounds that the benefit was the avoidance of risk rather than an increase on a loan, others forbidding it precisely because a benefit accrued. The debate never fully resolved, and the instrument was used anyway, which is a pattern anyone who has read a tax code will recognize.
Hawala worked without any document at all, on reputation. A customer pays an agent in one city; the agent contacts a counterpart elsewhere, who pays out the equivalent to the named recipient, identified by an agreed password or a token detail. No funds cross the distance. The two agents settle their mutual balance later, in whatever direction it happens to run, often by offsetting against trades going the other way. The system is fast, cheap, leaves minimal records, and works wherever the two agents trust each other, which in practice means wherever there are family or communal ties. It still moves very large sums today, chiefly in remittances, and it still works on exactly the same principle, which is that the money does not travel — only the information does, and the accounts are reconciled at leisure. Yap, again, with a bigger network.
Abbasid Baghdad had professional bankers, the jahbadh, who held deposits, managed the accounts of the caliphal treasury, advanced funds against future tax revenue, and issued the written payment orders that gave us the word. Tenth-century sources describe the arrangement in which the government contracted with banking houses to smooth its cash flow, receiving advances in return for the right to collect. Every large market town had money changers, the sayrafi, who dealt in the multiple coinages circulating simultaneously and who shaded gradually into deposit banking. A merchant could keep a running account, draw on it by written order, and settle with other merchants by transfer between accounts at the same house, with no coin moving anywhere.
The prohibition on interest also generated a fine tradition of formal evasion, of which the double sale is the most elegant. I sell you a bolt of cloth for a hundred and twenty dirhams payable in a year, and then immediately buy it back from you for a hundred in cash. You now have a hundred, you owe a hundred and twenty in a year, and no interest has been charged, because two entirely licit sales have occurred and the fact that the cloth never left the shelf is a matter for philosophers. Christian Europe developed its own versions of the same maneuver, generally by disguising interest as a penalty for late payment or as a difference in exchange rates between currencies. When a rule blocks a widely desired transaction, the transaction survives and the paperwork becomes baroque.
None of this should be read as pious hypocrisy or as clever cheating. The instruments were real and the constraints shaped them productively. A financial culture forbidden to lend at interest developed partnership finance, risk-sharing, transferable payment orders, deposit accounts, clearing between merchants, and remittance networks spanning from Spain to India — by the tenth century, several hundred years before comparable arrangements were normal in northern Europe. The Italians, sitting at the interface, learned the techniques and carried them north, where a small number of extremely well-armed monks were about to apply the same idea at scale.
Chapter Thirteen — Templars and the Medieval Wire Transfer
A pilgrim setting out from London for Jerusalem in the year 1200 faced a journey of several thousand miles through territories with varying degrees of law, and he needed to carry enough money to eat for a year. This is an invitation printed in large letters and addressed to every bandit between the Channel and the Levant. The problem was structural and everybody knew it, and the institution that solved it was, improbably, a monastic order of armed knights.
The Poor Fellow-Soldiers of Christ and of the Temple of Solomon were founded to protect pilgrims on the roads of the Holy Land, and they discovered fairly quickly that the most effective protection was to ensure the pilgrim had nothing worth stealing. A traveler deposited his funds at the Temple in London or Paris and received a document. At his destination he presented the document to another Templar house and drew the equivalent in local coin, minus a charge for the service, which was described as a fee for administration and safekeeping rather than as interest, because the alternative was a conversation with the Church about usury that nobody wished to have.
This required exactly the infrastructure the Order happened to possess: a network of fortified houses across Europe and the Near East, staffed by men bound by vows, communicating by regular courier, keeping written accounts, and answerable to a single command. It was also, incidentally, immune to most of the ordinary risks of medieval banking, since robbing a Templar preceptory meant fighting the garrison and then being excommunicated. Within a few generations the Order was managing the treasuries of kings. The Paris Temple effectively functioned as the treasury of France for a century; the Templar house in London held the crown jewels of England as security against a royal debt, and Henry III borrowed from the Order routinely.
The arrangement ended badly, and the ending is itself a monetary event. Philip IV of France, having debased his own coinage repeatedly, expelled his Jewish and Lombard creditors and seized their assets, and still faced a deficit, turned in 1307 on the largest creditor remaining. The Templars were arrested across France in a coordinated dawn operation, charged with an assortment of blasphemies, tortured into confessions, and the Order was dissolved. Philip did not, in the end, recover as much treasure as he had hoped, which is the usual result of destroying a financial institution in order to improve one’s balance sheet.
By then the technique had been thoroughly learned by people with no vows at all. The Italian merchant banks of Florence, Siena, Lucca, and Genoa had developed a document that would remain the backbone of international commerce for six hundred years: the bill of exchange. Its structure is worth understanding, because it is a small masterpiece of legal engineering built to satisfy two authorities at once.
Four parties are involved. A merchant in Florence pays a banker there in florins. The banker writes a bill instructing his correspondent in Bruges to pay a specified sum in Flemish currency to a named person at a stated future date. The bill travels north by courier; the money does not travel at all. Because the payment is in a different currency, in a different place, at a later time, the transaction is legally an exchange of currencies rather than a loan, and the profit is embedded in the exchange rate rather than stated as a rate of interest. The Church could hardly forbid the exchange of currencies, since exchange rates genuinely fluctuate and genuine risk is being taken. That the rates quoted between the great banking houses reliably produced a return of ten to fifteen percent per annum, and reliably in the same direction, was a coincidence upon which the theologians did not care to dwell.
Bankers eventually stopped pretending. In the practice known as dry exchange, a bill was drawn on a foreign city and then a return bill was drawn straight back, so that the funds made a round trip on paper and ended precisely where they started, with nothing having been transferred except a sum of interest disguised as two exchange operations. Everyone involved understood this. It appears in the manuals.
Settlement happened at the great fairs, above all those of Champagne, where merchants from across Europe met on a fixed calendar. The final days of each fair were reserved not for trade but for clearing: bills were presented, obligations offset against one another, and only the residual balance settled in coin. A merchant might arrive owing thousands and leave having paid a few dozen, because most of what he owed cancelled against what he was owed. Multilateral netting — the process that sits at the heart of every modern payment system — was being conducted in a muddy field in the twelfth century by men with abacuses.
The great banking families that grew from this trade had one structural weakness, which is the same weakness their successors have never eliminated: they lent to kings. The Bardi and Peruzzi houses of Florence financed Edward III of England at the start of the Hundred Years’ War. Edward, having spent the money on an unsuccessful campaign, declined to repay. Both houses failed in the 1340s, taking a good part of the Florentine economy with them, and a chronicler of the period recorded the disaster with the bitterness of a man who had lost his deposits.
The Medici, arriving a century later, drew the appropriate conclusion and organized their bank as a set of legally separate partnerships — Florence, Venice, Rome, Geneva, Bruges, London — each with local partners holding a stake, each capable of failing without dragging the others down. This is the holding company structure, invented for the excellent reason that the branch manager in London might do something irreversible. When the London branch did eventually do something irreversible, by lending heavily to Edward IV, the damage was contained — though the family’s general decline in the late fifteenth century owed a good deal to the same disease that killed the Bardi.
The final ingredient arrived in print in 1494, when a Franciscan friar named Luca Pacioli published a mathematics textbook containing a section describing the bookkeeping method already in use among Venetian merchants: every transaction recorded twice, as a debit in one account and a credit in another, so that the books balance and an error announces itself. Double-entry does not merely record business; it makes a business legible as a single object with a capital, a profit, and a position, distinct from the person who owns it. Once you can see an enterprise that way, you can sell shares in it, audit it, tax it, and lend against it. Nearly everything in the following four hundred years depends on a bookkeeping convention popularized by a mathematics teacher who also wrote a treatise on card tricks.
Chapter Fourteen — England Ran on Firewood
For roughly seven hundred years, the accounts of the English Crown were kept on sticks. Not as a quaint survival or a ceremonial gesture — as the actual, legally binding, primary record of what had been paid to the Exchequer and by whom. The sticks were hazel or willow, squared off, about the length of a forearm, and they worked well enough that they outlasted the Plantagenets, the Tudors, the Stuarts, and the Napoleonic Wars.
The system was ingenious in its simplicity. A sum received was recorded by cutting notches across the width of the stick, with the size of the cut indicating the denomination: a notch the width of a palm for a thousand pounds, a thumb for a hundred, a barleycorn for a penny, with intermediate widths in between. The name of the payer and the date were written on the flat sides in ink. Then the stick was split lengthwise through the notches, producing two pieces that each carried half of every cut. The Exchequer kept the shorter piece, the foil; the payer took the longer piece with its handle, called the stock — which is why a person who holds an interest in an enterprise is still called a stockholder, and why the word stock came to mean a financial claim rather than a length of wood.
The security of the arrangement is the elegant part. Wood grain is irregular, and a stick split along its length produces two halves whose surfaces are unique to each other. No forger could produce a matching piece, because matching required the original tree. When the two halves were brought together and the notches aligned, the tally was proved. Anyone who has worried that ancient bureaucracies must have been easy to defraud should sit for a moment with the realization that medieval England had a tamper-evident, cryptographically unique receipt system running on coppiced woodland.
What turned this from an accounting device into money was the practice of tallies of assignment. If the Crown needed to pay a supplier but had no cash in hand, it could issue a tally against a specific future revenue — the customs of a particular port, the taxes of a particular county — and hand it to the creditor, who would present it to the sheriff or the customs officer when the revenue came in. This is a bearer instrument representing government debt, payable by a designated collector, and it could be sold. A supplier who needed cash now rather than at Michaelmas sold his tally to somebody with patience, at a discount. The size of the discount reflected how likely the revenue was to materialize and how badly the seller needed money, which is to say that fourteenth-century London had a functioning secondary market in government paper, priced by risk, denominated in wood.
The system had a limit, and Charles II found it. Having issued tallies against future revenues far in excess of what those revenues would produce, the Crown announced in 1672 what became known as the Stop of the Exchequer: repayment on a large class of obligations was suspended. The goldsmith bankers of London, who had been buying government tallies with money deposited by the public, discovered that their principal asset had stopped paying. Several failed. Their depositors, who had merely wanted somewhere safe for their coin, lost it. This is a sovereign default with a banking crisis attached, staged in 1672, containing every element that would be presented as unprecedented in later centuries.
Tallies were finally abolished in 1826, by which time they were an embarrassment maintained by clerks whose posts depended on them. That left the Exchequer with an enormous accumulation of obsolete sticks, and in October 1834 somebody decided the simplest disposal method was to burn them in the two furnaces that heated the House of Lords. The stoves were overfilled, the flues overheated, the panelling caught, and by nightfall both Houses of Parliament were ablaze. The old Palace of Westminster burned to the ground before crowds who lined the bridges to watch, among them the painter Turner, who produced two canvases from the spectacle. The building that replaced it is the one with the clock tower, and it exists because the government of the day mishandled the disposal of its own accounting records.
Charles Dickens, describing the affair years later in a speech on administrative reform, drew the moral with visible enjoyment: officials clung to the sticks long past all reason, and when finally forced to dispose of them, contrived to destroy the seat of government rather than simply give the firewood to the poor who were freezing in the neighborhood. It remains the most literal case on record of a country being burned down by its own back office.
Behind the comedy sits a serious point that the following chapters will keep returning to. The tally was money in every sense that mattered: it was accepted in payment, it circulated between parties, it had a market price, and it represented a claim on a future flow of value. It was also a stick. The English state, in the same centuries, was minting fine silver and worrying about its purity, while the actual working capital of the realm moved around in the form of notched hazel. The metal got the attention and the theory. The wood did the job.
Chapter Fifteen — Potosí and the First Global Money
In 1545, according to the story the Spanish told, an Andean herder named Diego Gualpa was climbing a mountain in what is now Bolivia when the wind knocked him down, and grabbing at a shrub to stop his fall he pulled up a root with silver ore clinging to it. The story is almost certainly a legend, and local people had known about the mountain for some time. What is not legendary is the mountain, which the Spanish called Cerro Rico, the Rich Hill, and which turned out to contain the largest concentration of silver ever found.
Within thirty years the town at its foot had become one of the largest cities in the world, at an altitude of four thousand meters, in a landscape that grows nothing. Everything — food, timber, cloth, wine — had to be hauled up from the lowlands, and it was, because the mountain paid. Estimates of the total production over three centuries run to tens of thousands of tons of silver.
The labor came from the mita, a rotational draft adapted from an Inca institution and repurposed at a scale the Inca had never contemplated. Communities across a vast catchment area were required to send a proportion of their adult men to Potosí for a year’s service. They walked for weeks to get there, often with their families, and a substantial fraction did not walk back. Underground conditions were lethal in themselves, and above ground the refining process used mercury — shipped from the mines at Huancavelica, where conditions were, if anything, worse — amalgamated with crushed ore in open patios that workers trod barefoot. Mercury poisoning was routine. The demographic damage to the Andean highlands was severe and permanent, and modern studies still find measurably worse outcomes in the districts that were subject to the draft.
The coin struck from all this was the eight-real piece, the peso de a ocho, the piece of eight. It was large, heavy, of dependable purity, and produced in such quantity that it became the first genuinely global currency. It circulated in the Caribbean, in the Mediterranean, in West Africa, in India, in China, and in the British North American colonies, where it remained the principal coin in daily use through the Revolution and beyond — the reason the United States adopted a decimal dollar rather than a pound is that the Spanish dollar was already the money people had in their pockets. It was frequently cut into segments for small change, eight of them, which is where the American slang for a quarter as two bits comes from.
The greater part of it went east. China, having abandoned paper and shifted its entire fiscal system onto silver, was the deepest sink for the metal in the world, and gold was relatively cheaper there than in Europe, which meant silver bought more in Canton than in Seville. Ships crossed from Acapulco to Manila once a year, the galleon carrying silver west and returning with Chinese silk and porcelain. A recognizable global trading system — American mines, Asian manufactures, European consumers, African labor stolen into the plantations that fed the whole arrangement — assembled itself around the flow of one metal, and the connection was direct enough that Chinese fiscal policy and Andean mining output moved the price of bread in Castile.
Europe, receiving all this silver, discovered inflation. Prices across western Europe rose perhaps five- or sixfold over the sixteenth century, a rate that would seem trivial today and was unprecedented then, in societies whose price expectations were essentially flat across generations. Wages lagged badly, since employers adjust them last, so the standard of living of ordinary workers fell while the incomes of landlords and merchants rose. A French jurist named Jean Bodin, arguing in 1568 with an official who blamed the phenomenon on debasement, made the case that the principal cause was the abundance of gold and silver arriving from the Indies. That is the quantity theory of money, stated plainly, in the middle of the sixteenth century, by a man better known for writing about sovereignty and, less creditably, witchcraft.
The final irony belongs to Spain. A state with unlimited access to the world’s silver supply managed to default on its debts at least six times between 1557 and 1647. The bullion arrived at Seville, was counted, and passed almost immediately to Genoese and German bankers to service loans taken out against next year’s fleet. It financed wars in the Netherlands and Germany that Spain did not win. Domestic industry withered because imports were cheap and the incentive to produce anything was weak. By the seventeenth century Spain was issuing debased copper coinage at home while shipping fine silver abroad, and the country that had found the mountain was poorer, relative to its neighbors, than before it found it.
This is the first appearance of a pattern that would be given a name only in the twentieth century, when several oil-rich states discovered the same thing: a windfall of monetary metal is not wealth. It is a claim on other people’s production, and a state that mistakes the claim for the production will spend the windfall on soldiers, import everything else, and be left, when the mountain runs low, with neither the silver nor the industries it might have built. Potosí made the world’s first global currency, killed an enormous number of people making it, and impoverished the empire that owned it. On the modern Bolivian coat of arms the mountain still appears, which is either irony or memory, depending on how you look at it.
Chapter Sixteen — The Anatomy of a Bubble
The Dutch Republic in the 1630s was the wealthiest society in Europe per head, and it had a fashionable problem: what to do with the money. Land was scarce, the great trading companies were closely held, and a merchant with a surplus had limited outlets. Into this situation arrived a flower.
The tulip had come from Ottoman gardens by way of Vienna in the previous century, and it did something in Dutch soil that made it commercially interesting. Certain bulbs produced blooms with dramatic flames and feathers of contrasting color streaking the petals — the effect prized above all others, and utterly unpredictable. Nobody at the time knew why it happened. It is caused by a virus carried by aphids, which weakens the plant and makes the striped varieties propagate slowly, which is precisely why they were rare and expensive. The most sought-after cultivars had names like admirals and generals, and the grandest of them, Semper Augustus, was owned in a handful of bulbs by a man who declined to sell.
Because bulbs can only be lifted from the ground in summer, trading during the rest of the year had to be conducted in bulbs still buried, which meant contracts for future delivery. The Dutch called it windhandel, the wind trade, since nothing was being handed over but air. The trade moved out of the hands of connoisseurs and into taverns, where it was conducted through elaborate ritualized auctions among people who had never grown a flower, with payment due at lifting in the summer, and with the contracts themselves changing hands repeatedly in the interim. During the winter of 1636 to 1637 the prices in these contracts multiplied enormously, and then, in the first week of February 1637, at an auction in Haarlem, there were no bidders. Within days the market ceased to exist.
That is the standard account, and it is largely true. What is not true is nearly everything that has been built on top of it. The popular version — that the Dutch economy was ruined, that merchants leapt into canals, that maids were bankrupted, that the nation went mad — comes from a nineteenth-century compendium of popular delusions, which drew in turn on satirical pamphlets published by moralists at the time for the purpose of ridiculing speculators. The archival work of recent decades, above all by Anne Goldgar, found something more sober: the number of people involved was in the hundreds rather than the tens of thousands, they were concentrated among wealthy merchants and skilled artisans who could afford the loss, almost no contracts were ever fully paid, the courts declined to enforce them, and no bankruptcies attributable to tulips can be identified. There was no macroeconomic damage whatsoever. There was, however, a great deal of anxiety about honor and trust, because a society in which a man’s written promise could be repudiated by a court had discovered something uncomfortable about itself.
It is worth being precise about this, because tulip mania has become the lazy analogy for every price rise anyone dislikes, and lazy analogies obscure the mechanism. So here is the mechanism, which is genuinely general.
A bubble begins with something real. There must be a plausible story about why the new thing is valuable, and the story is usually correct: the striped tulips genuinely were rare and beautiful, the railways genuinely did transform transport, the internet genuinely did change commerce. Displacement, the economic historian Charles Kindleberger called it — an event that changes the outlook and creates new profit opportunities that nobody has priced yet.
Next comes credit, in one form or another. Prices cannot rise far on savings alone; there has to be a way of buying more than one can afford, whether through forward contracts settled later, margin loans, mortgages, or an obliging bank. Every serious bubble in history has been financed rather than saved into, which is why the ones that end in ruin are the ones the banking system participated in, and why tulips — financed by promises between individuals that the courts then voided — did so little damage.
Then the character of the buyer changes. Early purchasers buy for the thing itself: they want the flower, the house, the share of the railway. Later purchasers buy because the price is rising, which is a completely different reason and requires only the belief that somebody else will pay more. At this stage the asset’s actual properties stop mattering, and the market becomes a pure exercise in guessing what other people will guess. Newcomers arrive who know nothing about the underlying business and do not need to. Anecdotes circulate about ordinary people making extraordinary sums, which is the most powerful recruiting device ever devised, because it converts caution into the fear of being left out.
And then the supply of new entrants runs out, and this is the part that surprises people every time. Nothing needs to happen. No bad news is required. The Haarlem auction did not fail because of a disease or a war or a decree; it failed because on that particular day the room contained nobody willing to bid, and the news of that traveled. The price was never a property of the tulip. It was a property of the crowd, and crowds are not a smoothly declining function — they hold their configuration and then reorganize suddenly, because each member is watching the others.
The final phase is the moral one, and it is the least useful. Afterward the participants are described as having been insane, greedy, or stupid, and pamphlets are written. This is comforting because it implies the rest of us, being sensible, are immune. But the individuals in a bubble are behaving rationally given what they believe about everyone else, which is the same reason a bank run is rational for each depositor and catastrophic for all of them. Calling it madness is a way of avoiding the more disturbing conclusion, which is that the structure produces the outcome and the personalities are interchangeable.
Two of the three great bubbles of the following century would test that proposition at a scale the flower market never approached, and they would do it in the same year, in two countries, with two men who had known each other.
Chapter Seventeen — John Law’s Machine
John Law was a Scottish banker’s son who killed a man in a duel in London in 1694, was sentenced to hang, escaped from prison under circumstances that suggest bribery, and spent the following two decades on the continent supporting himself by gambling. He was extremely good at gambling, in the specific sense that he understood probability better than the aristocrats he played against and calculated the odds while they trusted their luck. He was also, in his spare time, developing a monetary theory, and it was a good one.
Law’s argument was that a country’s prosperity is limited by the quantity of money in circulation, that gold and silver are an arbitrary and inconveniently rigid basis for that quantity, and that a well-run bank could issue paper backed by something more expandable — land, or the revenue of a going concern — thereby putting idle resources to work. He was two centuries ahead of the consensus and, as it turned out, precisely as dangerous as such a person usually is when handed an entire country.
The country was France, and it was available because Louis XIV had died in 1715 leaving debts of an almost unmeasurable size, a five-year-old great-grandson on the throne, and a regent, the Duke of Orléans, who was open to suggestions. Law had a suggestion. He was permitted to found a private bank in 1716, which issued notes; the notes were good, they were accepted for taxes, and they worked well enough that the bank became the Royal Bank two years later, with the state guaranteeing the notes.
The second half of the machine was the Mississippi Company, which held the trading monopoly for the French territory of Louisiana — an enormous region about which almost nothing was known, described in promotional material as fertile, temperate, and rich in gold. It was in fact swamp and forest, the colony was tiny, and the settlers sent there died in impressive numbers. This did not matter, because the shares were not really about Louisiana.
What Law actually built was an engine for converting the national debt into equity. Holders of government bonds — those depreciated, doubtful, unsellable claims on a bankrupt crown — were invited to exchange them for shares in the Company. The Company absorbed the debt and, in return, acquired further monopolies: tobacco, the mints, the collection of taxes across the whole of France. Each new acquisition made the shares more attractive, which raised the price, which allowed a further issue, which allowed the absorption of more debt. The bank, meanwhile, lent notes to people so that they could buy the shares. The two halves fed each other.
The price of a share went from around five hundred livres to ten thousand within two years. The narrow street of Rue Quincampôix, where the trading took place, became so crowded that a hunchback is said to have made a living renting out his back as a writing desk. Fortunes were made by servants and coachmen and, in the case of one bookseller’s widow, by holding on longer than her betters. The word millionaire was coined in Paris at this moment, because the sums required a new noun. Law himself was made Controller General of Finances, converted to Catholicism to qualify for the post, and became for a season the most powerful man in Europe.
The unwinding began when a few large holders decided to take their profits in metal. The Prince de Conti is supposed to have sent wagons to the bank to convert his notes into coin. Others followed quietly. Law’s response was to make the exit illegal: private holdings of gold and silver above a small limit were prohibited, jewelry manufacture was restricted, informers were rewarded, and searches were authorized. This is the point at which a monetary experiment becomes a police matter, and it is a reliable signal that the arithmetic has already lost.
The shares fell through 1720. Law attempted a controlled devaluation of the share price and the notes together, published as a decree in May, which succeeded in destroying the last of the confidence at a stroke; the decree was reversed within days, which destroyed confidence in the reversal. By the end of the year the system was over. Law fled France with very little, having refused, to his credit, to steal on the way out, and died in Venice nine years later, gambling.
Across the Channel the same year, the South Sea Company was running a structurally identical scheme with better manners: government debt exchanged for company shares, a trading monopoly of no real value — in this case the right to supply enslaved people to the Spanish colonies, which Spain had no intention of honoring — and a share price that rose eightfold before collapsing. Isaac Newton, then Master of the Mint, sold early at a profit, bought back in at the top, and lost a sum equal to many years of his salary. The remark attributed to him afterward, that he could calculate the motions of the heavenly bodies but not the madness of people, is probably apocryphal and is certainly earned.
The consequences diverged sharply. Britain absorbed the damage, passed legislation restricting the formation of joint-stock companies, prosecuted a few directors, and carried on developing its financial system, which within a century would be the most sophisticated in the world. France drew the opposite lesson. Paper money and banks had been discredited so thoroughly that the country had no significant note-issuing bank for the better part of a century, and public finance reverted to the tax farms, the venal offices, and the ruinously expensive borrowing that would still be unresolved in 1789. The most enduring product of Law’s system was a national allergy, and the fiscal crisis that allergy helped preserve would eventually cost the monarchy considerably more than ten thousand livres a share.
Chapter Eighteen — The Bank That Made Debt Permanent
In 1694 the English government was losing a war to France and could not pay for it. A Scottish projector named William Paterson proposed an arrangement of the kind that usually goes nowhere: a group of subscribers would lend the Crown one million two hundred thousand pounds at eight percent, and in exchange would receive a royal charter incorporating them as the Governor and Company of the Bank of England, with the right to issue notes and conduct banking business. The subscription filled in twelve days.
The important feature was not the loan but its permanence. The principal was never to be repaid. The government would service the interest out of specified taxes, and the debt would sit there indefinitely, transferable between holders. This sounds like a defect and was the entire innovation. A debt that never falls due cannot precipitate a crisis at maturity; it becomes an asset the holder can sell to someone else at a market price, and the government’s obligation is reduced to a manageable annual payment rather than a terrifying lump. Around this arrangement grew a market in transferable government securities, and with it the ability of the English state to borrow at rates its rivals could not approach.
The consequences were military before they were economic. France was the larger, richer, more populous country throughout the eighteenth century, and it lost most of the wars, because Britain could raise money faster and more cheaply and could therefore keep fighting after its opponent had run out of options. A financial technique invented to cover one year’s deficit became the structural advantage that decided a century of conflict. Historians call the whole complex — the funded debt, the Bank, the stock market, the insurance industry that grew alongside them — the Financial Revolution, and it is at least as consequential as the industrial one that followed.
The Bank’s first serious crisis arrived within two years and had nothing to do with paper. England’s silver coinage had been clipped so persistently that the circulating coins contained roughly half their nominal metal, and in 1696 the government undertook a great recoinage, calling in the old money. The question was whether to reissue at the old standard, which meant that anyone holding clipped coin lost the difference, or to devalue formally and recognize what the coins actually contained.
The debate was conducted at an unusually high level. William Lowndes, the Treasury secretary, argued for devaluation on practical grounds: the coins were what they were, and pretending otherwise would cause deflation and hardship. John Locke argued that silver content was the definition of the pound, that a pound was a weight and not a decision, and that any devaluation would be a fraud on creditors and a breach of public faith. Locke won, largely because his position had the moral high ground and the creditors had the political power. The recoinage went ahead at the old standard, the money supply contracted violently, credit froze, the Bank suspended payments briefly, and the country endured two years of severe deflation and unrest. Isaac Newton, appointed Warden and then Master of the Mint, spent the period getting the mints to work at unprecedented speed and pursuing counterfeiters through the taverns of London with a zeal that surprised everyone who thought of him as a mathematician; he sent several to the gallows.
The episode is the first full-dress performance of an argument that recurs in every chapter from here on. Is money a natural fact that governments merely certify, or an instrument that governments manage? Locke’s position is intellectually clean and has the enormous advantage that it removes discretion, and therefore removes the temptation every treasury faces. Lowndes’s position is messy and admits that the definition of value is a policy choice, which invites abuse but at least allows the abuse to be discussed openly. Both men were right about what the other’s solution would cost.
A century later the Bank abandoned Locke’s principle without much ceremony. In February 1797, with a French invasion scare and rural banks failing, gold reserves fell to a level that could not meet demand, and the government ordered the Bank to stop paying out gold. The Restriction, intended as a temporary emergency measure, lasted twenty-four years. During it, the pound was pure paper, the country fought and won the Napoleonic Wars, and prices rose substantially and then fell again — an entire generation of monetary experience conducted off the metal standard by a state that continued to insist it believed in metal.
It was during the Restriction that the Bank acquired its nickname. A cartoon by James Gillray in 1797 depicted the institution as an elderly woman in a dress made of banknotes, seated on a strongbox, being assaulted by the Prime Minister after her gold — the Old Lady of Threadneedle Street, complaining loudly about her virtue while her chest is being emptied. Two decades later George Cruikshank produced a more savage image: a mock banknote, issued in response to the hanging of people convicted of passing forged notes during a period when forgery was rampant and the currency poorly made. His note is decorated with corpses on a gibbet, signed by the hangman, and payable in death. It is said to have contributed to the ending of capital punishment for the offense.
Which is a useful reminder that behind the elegant machinery of funded debt and note issue there were people being hanged for holding the wrong piece of paper. The Bank of England had made the English state solvent, credible, and militarily formidable. It had also made the question of who is permitted to create money a matter of life and death, quite literally, and the next century would see several countries answer that question in ways that made the English arrangement look positively restrained.
Chapter Nineteen — Assignats and Continentals
Revolutions are expensive, and they are launched by people who have just repudiated the authority that used to collect the taxes. This puts a new government in an awkward position roughly six months in, and there is one solution available to anyone who controls a printing press.
The Continental Congress that declared American independence had no power to tax at all. It could requisition funds from the states, which mostly ignored it, and it could borrow abroad, which it did with some success from France. In the meantime it authorized the issue of Continental currency: paper dollars, promised to be redeemable in Spanish milled dollars at some unspecified future point, in quantities that grew from two million in 1775 to something over two hundred million by 1779. The states issued their own paper alongside it. Nobody was withdrawing any of it, because withdrawal requires taxation.
The currency depreciated from near parity to something on the order of a hundred to one by 1780, and eventually to a thousand to one. It became worthless in the specific sense that shopkeepers papered walls with it and papers ran cartoons about it, and it left the language a phrase for something entirely valueless that survived a century after the notes themselves were forgotten.
The depreciation had assistance. The British, correctly identifying the currency as a war target, ran a counterfeiting operation from a ship moored off New York, advertising in loyalist newspapers for travelers willing to carry quantities of the imitation notes into the interior, supplied free of charge. The counterfeits were reportedly better made than the originals, since the Continental Congress was printing on poor paper with worn plates. This is the earliest example of monetary warfare conducted as a deliberate operation, and it would not be the last in this chapter.
Congress eventually revalued the notes at forty to one and let them expire, funding the remainder of the war on French loans and the personal credit of a few merchants. The experience left a mark on the American constitutional settlement that is still visible: the Constitution gives Congress power to coin money and forbids the states from issuing bills of credit, and the founders’ hostility to paper money runs through the debates of the period like a cold draft. It also, more practically, produced a generation of American politicians who understood that a government which cannot tax cannot borrow either, since a lender is really buying a claim on future revenue and will notice if there is none.
The French case, a decade later, was more sophisticated and ended worse. The revolutionary Assembly of 1789 faced the same problem — a bankrupt state, a collapsed tax system, and urgent obligations — but it had an asset. It had confiscated the lands of the Church, which amounted to a substantial fraction of the cultivable soil of France. The assignat was originally a bond assigned against those lands, bearing interest, to be used to purchase them, and cancelled when it did so. This is a genuinely clever design: the notes were backed by real property, and the mechanism for retiring them was built in.
Every safeguard was then removed in sequence, each time for a good reason. The interest was dropped, to make the notes circulate as currency. The denominations were reduced from a thousand livres to five, to put them in everyday hands. Retirement by land purchase slowed, because the government preferred to sell land for cash and keep the notes outstanding. And the volume was increased, because there was a war on against most of Europe, and the alternative was defeat. Issue reached tens of billions of livres against a land stock that had been worth a fraction of that.
The measures taken to defend the currency escalated in the familiar direction. Refusing to accept assignats at face value became a criminal offense. Dealing in gold and silver was prohibited, then made a capital offense. Quoting different prices for cash and paper was forbidden. The Law of the Maximum imposed price ceilings on grain and then on a long list of goods, enforced by revolutionary committees, with the guillotine available at the far end of the enforcement chain. Peasants stopped bringing food to market, cities went hungry, and requisitioning columns went into the countryside to take it by force. Every one of these steps was a response to a real problem and every one made the underlying problem worse.
The British contribution this time was industrial. A counterfeiting operation was run from London with the involvement of émigré royalists and the tacit support of the government; forged assignats were produced in bulk and shipped into France, and while the scale is disputed, contemporaries on both sides believed it substantial. When your enemy’s currency is already collapsing, adding to the supply is cheap sabotage.
By 1796 the assignat was worth well under one percent of its face value. The Directory conducted a public ceremony in the Place Vendôme, breaking the printing plates and burning them before a crowd, which was theater but effective theater. A replacement instrument, the mandat territorial, was issued with a direct claim on specific land, and depreciated to nothing within months, because by then nobody would touch anything printed by that government. France returned to metal, and Napoleon’s franc, defined by law in 1803 as a fixed weight of silver, held its definition for over a century — an unusually long monetary settlement, purchased by an unusually thorough national trauma.
What the two episodes share is not the printing. It is the sequence: a government without a tax base issues claims, defends them with law rather than with revenue, criminalizes the alternatives, and ends by destroying its own instrument. What separates them is instructive too. The Americans stopped when the notes stopped working and moved on. The French kept going for years, adding coercion at each stage, because they were fighting for survival and had persuaded themselves that acceptance was a matter of patriotism rather than arithmetic. Money is not much affected by patriotism. It is affected by whether the issuer takes it back.
Chapter Twenty — When Anybody Could Issue
A customer walks into a store in Cincinnati in 1855 and offers a five-dollar note. The shopkeeper takes it, and then reaches under the counter for a printed book. He is not checking whether the note is genuine. He is checking what it is worth. The issuing bank turns out to be in Maine; the current discount is fourteen percent; the customer is informed that his five dollars will settle four dollars and thirty cents of the bill. Nobody in the shop finds any of this remarkable, because this was ordinary retail life across much of the United States for several decades before the Civil War.
The situation arose from a national argument. Two attempts at a central bank had been made and abandoned, the second killed by Andrew Jackson, who regarded the Bank of the United States as an aristocratic monster and vetoed its recharter in 1832. What replaced it was nothing at the federal level and, in state after state, a system called free banking: rather than requiring a special legislative charter for each bank, states passed general laws under which anybody meeting stated conditions could open one and issue notes. The conditions usually required depositing state or federal bonds with the authorities as security for the notes.
The result was thousands of note-issuing banks, each with its own designs, and something on the order of ten thousand distinct varieties of banknote circulating simultaneously. Since a note was a promise to pay coin at the issuing bank’s counter, its value depended on how likely the bank was to honor it and how much trouble it would be to get there. Notes therefore traded at a discount that increased with distance and with doubt. An entire publishing industry sprang up to serve the problem: bank note reporters, issued weekly, listing current discounts by institution, alongside counterfeit detectors describing the distinguishing features of known forgeries. Bank tellers and merchants subscribed the way one subscribes to a trade journal, and the information was out of date the moment it was printed.
Counterfeiting flourished, naturally, because no shopkeeper in Ohio could possibly know what a genuine note from a bank in Maine looked like. Estimates of the share of circulating paper that was fraudulent in some way — outright forgeries, altered denominations, notes of banks that had failed, notes of banks that had never existed — run into the tens of percent. There is a category of nineteenth-century American crime that consists entirely of printing plausible money and taking it somewhere far away.
The wildcat bank belongs to this world, and the name is worth explaining because it is usually misused. A wildcat bank was one that met the letter of the law — depositing the required bonds, printing handsome notes — while establishing its redemption office somewhere so remote that presenting a note for payment was impractical: out among the wildcats, as the phrase had it. Michigan’s free banking law of 1837 produced a famous crop of these, along with the equally famous practice of moving the same chest of specie from one bank to the next ahead of the inspector, so that each institution could be found in possession of adequate reserves on the day it was examined.
And yet the picture is not the simple morality tale it became. Careful work on the period has found that most free banks were ordinary businesses that failed for ordinary reasons, chiefly because the state bonds they had deposited as security lost value — which is to say the banking failures of the era were frequently caused by the collapse of state government credit rather than by fraud. Where states required diversified, high-quality collateral, banks failed rarely. Where states permitted banks to post their own dubious paper, they failed constantly. The variable was the regulation, not the freedom.
The private sector also solved a good part of the problem itself, in New England. The Suffolk Bank of Boston began accepting country bank notes at par on the condition that the issuing bank keep a deposit with it, and returned notes for redemption promptly and relentlessly. Since a bank that had over-issued would find Suffolk on its doorstep with a sack of its own paper, over-issue became painful. New England notes consequently circulated at or near face value throughout the region — a private clearinghouse imposing monetary discipline for profit, without any legislation, several decades before a central bank existed.
Across the Atlantic, Scotland had run a system with even fewer rules and considerably better results for over a century. Scottish banks issued their own notes, competed vigorously, established branch networks, cleared each other’s paper twice a week in Edinburgh, and — the crucial detail — operated under unlimited liability, meaning that shareholders’ personal fortunes stood behind the notes. Failures happened, and when they did the shareholders were ruined and the noteholders were paid. Scottish banks also included an option clause in their notes permitting a temporary delay in redemption at a penalty rate, which is a private version of a suspension of convertibility and rather more honest than the government kind. Advocates of free banking cite Scotland; critics point out that it operated in the shadow of London and the Bank of England, and both are correct.
The American arrangement ended for fiscal reasons, not intellectual ones. The Union needed to finance a war, and the National Banking Acts of 1863 and 1864 created federally chartered banks issuing uniform national notes backed by federal government bonds — an arrangement that simultaneously standardized the currency and created a captive market for war debt. To finish the job, Congress imposed a ten percent tax on state bank notes in 1865, which ended their issue immediately. The chaos of ten thousand currencies was resolved not by argument but by taxation, which is how most monetary arguments are resolved.
One thing survived from the free banking era and deserves a mention, since it is the reason the state banks did not simply disappear. Deprived of the right to issue notes, they turned to a product that was not taxed: the deposit account, transferable by check. Within a generation checkable deposits were a larger part of the money supply than currency, and remain so. A tax designed to eliminate private money creation succeeded in eliminating the paper and left the banks creating money in a form nobody had thought to legislate about.
Chapter Twenty-One — The Gold Standard as Religion
Britain adopted the gold standard by accident, and the accident was committed by the greatest scientist of the age. As Master of the Mint, Isaac Newton was responsible in 1717 for setting the official price at which the mint would exchange gold and silver. He set the value of the gold guinea slightly too high relative to silver, judged against the ratios prevailing in continental markets. The consequence was mechanical: it became profitable to bring gold to England and to take silver out. Silver coin drained away to the Continent and to Asia, gold accumulated, and within a few decades Britain was on gold in practice while remaining on both metals in law. Formal recognition came in 1821. An empire’s monetary constitution was decided by a rounding error in a table.
The system that grew from this was never designed. It assembled itself in the 1870s when Germany, flush with the indemnity extracted from France after the Franco-Prussian War, converted to gold and dumped its silver on the market; the fall in the silver price pushed most other trading nations to follow, since nobody wished to be left holding the depreciating metal. The United States dropped silver from its coinage in 1873 in a piece of legislation that passed with little debate and was denounced for the following quarter-century as the Crime of ’73. By the 1880s most of the trading world was on gold, and the arrangement lasted until the guns of August 1914, which is why it is remembered with a nostalgia that its actual operation does not entirely justify.
The mechanism was described by David Hume more than a century before it existed. A country importing more than it exports must settle the difference in gold; gold leaving the country reduces the money in circulation; less money means lower prices; lower prices make its exports attractive and imports expensive; the flow reverses. The system self-corrects, without any official deciding anything. It is a beautiful piece of reasoning and it works, on one condition: that wages and prices are free to fall. That condition is where the trouble lives, because prices fall by putting people out of work, and the elegance of the adjustment mechanism is not apparent to the person being adjusted.
The classical gold standard did deliver something valuable. Exchange rates between the major economies were fixed and stable for four decades, which made long-term international investment possible on a scale not seen again until the late twentieth century. Capital flowed from Britain into railways in Argentina and mines in South Africa on the strength of a contract that would be repaid in a currency whose value could be predicted. Adherence functioned as a signal of respectability — economic historians have called it a good housekeeping seal of approval — and countries that joined it borrowed more cheaply than those that did not. That, rather than any monetary theory, is why so many governments were desperate to be on it.
The cost was borne unevenly. Between 1873 and 1896 the world’s gold supply grew more slowly than its output of goods, and prices fell more or less continuously for a generation. Falling prices are pleasant for anyone with a fixed income and ruinous for anyone with a fixed debt, because the debt stays the same while the money to pay it becomes harder to get. American farmers, who had borrowed to buy land and machinery and who sold crops at prices that fell year after year, were being slowly crushed by an arrangement they had never voted for and which was defended in the language of moral obligation.
Out of this came the most extraordinary monetary speech ever delivered. At the Democratic convention in Chicago in 1896, William Jennings Bryan, aged thirty-six, argued for the free coinage of silver — that is, for monetary expansion — and finished by telling the assembled financial interests that they should not press down upon the brow of labor a crown of thorns, nor crucify mankind upon a cross of gold. He won the nomination on the strength of it. He lost the election, and lost twice more afterward, and the gold standard survived, whereupon new discoveries in South Africa and the Yukon plus the cyanide extraction process expanded the gold supply and quietly delivered the inflation Bryan had demanded. There is a persistent reading of The Wonderful Wizard of Oz, published four years later, as an allegory of all this — the yellow brick road, the silver slippers that Hollywood turned ruby, the humbug wizard behind the curtain. The author never said any such thing, and the theory was invented by a schoolteacher in 1964. It is too good to abandon and too unsupported to assert.
The war ended the system in a fortnight. Belligerents suspended convertibility, printed, and spent. What followed was the great mistake, and it was made by intelligent people acting on principle. Through the 1920s country after country returned to gold, because the pre-war order was remembered as normality and returning to it seemed like a return to health. Britain went back in 1925, at the pre-war parity, which meant declaring the pound to be worth substantially more than the market thought — and therefore requiring British wages and prices to fall by roughly the same margin to keep exports competitive.
Keynes published a pamphlet titled The Economic Consequences of Mr. Churchill, which argued that the Chancellor had committed the country to deliberate deflation, that deflation meant unemployment, and that the process would fall hardest on the coal industry, where wages would have to be cut first. The following year the coal owners cut wages, the miners struck, and the General Strike of 1926 shut the country down for nine days. Churchill later described the return to gold as the worst decision of his life, which, given the competition, is a substantial admission.
The deeper problem was that the restored standard was not the old one. Before the war it had been managed, informally and skillfully, by a small group of central banks led by London, which cooperated in crises and understood that the rules could be bent quietly. After the war the financial center of gravity had moved to New York, France had accumulated gold and was determined to keep it, Germany was paying reparations with borrowed money, and nobody was in charge. Countries held each other’s currencies as reserves, so a loss of confidence anywhere could pull gold from everywhere. The machine had been rebuilt without the mechanic, and it was now wired so that a failure in one part would propagate through the rest.
It would be tested in 1931, and it would fail catastrophically. But the first great monetary disaster of the interwar period had already happened, in a country that had left gold and could not find its way back, and where the collapse ran in the opposite direction entirely.
Chapter Twenty-Two — The Wheelbarrow Years
The photograph everyone has seen shows children building a tower out of banknotes, or a woman feeding a stove with them, or a man wheeling a barrow of paper to buy bread. The images are real, they date from the autumn of 1923, and they have become such an efficient shorthand that the actual mechanics of the German inflation have been almost entirely replaced by them. What the pictures cannot show is that the process took nine years, that it was survivable and even profitable for a substantial part of the population until quite late, and that it was ended in a matter of weeks by an act of theater.
The inflation began with the war, as all of them do. Germany financed the conflict overwhelmingly by borrowing rather than taxing, on the assumption — stated openly by officials — that the defeated enemy would pay the bill afterward. The mark had already lost most of its pre-war value by the armistice. What followed was a defeated state with a collapsed tax system, revolutionary unrest, an army to demobilize, a population to feed, and, from 1921, a reparations schedule denominated in gold marks that the government regarded as impossible and its creditors regarded as agreed.
The decisive escalation came in January 1923, when France and Belgium, exasperated by shortfalls in coal and timber deliveries, occupied the Ruhr. The German government responded with passive resistance: workers in the occupied zone were told to strike, and the state undertook to pay their wages. This is a government committing to pay the wages of an entire industrial region while receiving no output and no tax revenue from it, and there was exactly one way to fund it. The Reichsbank printed. By the summer, more than a hundred private printing works with nearly two thousand presses were producing currency around the clock, and the Reichsbank publicly congratulated itself on the achievement of meeting demand.
The velocity is what turns inflation into hyperinflation, and velocity is a behavior rather than a policy. Once people expect prices to rise, holding money becomes irrational, so everyone spends immediately, which raises prices, which strengthens the expectation. Workers were paid twice a day and wives waited at the factory gate to run to the shops. Restaurants stopped printing prices. Firms issued their own emergency money — Notgeld — in enormous variety, some of it charming enough to be collected as souvenirs at the time. At the peak in October and November 1923 prices were doubling roughly every three to four days, and the exchange rate reached over four trillion marks to the dollar, which is a figure that conveys nothing at all, which is precisely the problem with the whole subject.
The distribution of pain was uneven and politically fateful. Anyone holding a mortgage or a business loan watched the debt evaporate, and the great industrial fortunes of the period were substantially built on repaying in worthless currency loans taken out in sound money. Anyone holding savings, insurance policies, war bonds, or a pension — the thrifty, cautious, patriotic middle class who had lent to the state out of duty — lost everything, and lost it in a way that felt like a swindle rather than a misfortune. Farmers with produce did well; city dwellers with salaries did badly. A social order in which prudence had been rewarded was inverted in view of everyone, and the memory of it did not fade when the prices stopped moving.
The ending is the genuinely instructive part. In November 1923 the government introduced the Rentenmark, exchanged at one to a trillion old marks, and it was backed by — nothing that would survive scrutiny. The nominal security was a mortgage on German agricultural and industrial land, which could not have been foreclosed on in any practical sense and which nobody expected to claim. What accompanied the new note mattered far more: the issue was strictly limited by law, the Reichsbank was forbidden to discount further government paper, the budget was slashed with hundreds of thousands of public employees dismissed, and passive resistance in the Ruhr was called off. The commissioner in charge, Hjalmar Schacht, ran the operation from a converted cleaning cupboard with one secretary, a detail he told with relish for the rest of his life.
It worked immediately. Prices stabilized within days. The country that had been doubling its price level twice a week was, within a month, transacting normally in a currency with imaginary backing. This is the fact that ought to be quoted whenever anyone insists that money must be backed by something. What the Rentenmark was backed by was the visible, credible, verifiable end of the printing — and the population, watching the government cut its own budget and abandon the Ruhr policy, concluded that this time the issuer meant it.
The pattern generalizes with unusual reliability. Hyperinflations do not fade out gradually as the printing slows; they stop, abruptly, when the fiscal regime visibly changes and the change is believed. Austria and Hungary ended theirs in the same period by the same method, under League of Nations supervision that made the commitment credible by making it external. The reverse is equally consistent: every attempt to end one by decree, price control, or currency reform without fixing the deficit has failed within months, because the public correctly identified the announcement as cosmetic.
Germany does not hold the record, incidentally, and it is worth knowing who does, if only to keep the ranking straight. Hungary in 1946, recovering from occupation, ended with prices doubling roughly every fifteen hours and issued a note denominated in a quintillion pengő, the highest denomination ever printed anywhere. Zimbabwe in 2008 and Yugoslavia in 1994 sit between the two. In every case the sequence was the same: a state facing an obligation it could not fund, choosing the invisible tax over the visible one, and losing control of the expectations that made its money work.
The German lesson was learned, thoroughly and perhaps too well. The horror of 1923 became the central monetary memory of the country, embedded in the mandate of its post-war central bank and, later, in the design of the European one. That is an entirely defensible response, with one complication that historians have argued about ever since: the events that brought the National Socialists to power occurred a decade after the inflation, during a savage deflation caused by the opposite policy, and it was mass unemployment rather than wheelbarrows that filled their meetings. Both catastrophes were monetary. Only one of them is on the postcards.
Chapter Twenty-Three — 1929 and the Run as Cascade
The stock market crash of October 1929 is the most famous financial event in history and one of the less important ones. Share prices fell hard over a few days, wiping out a good deal of paper wealth belonging to a small minority of Americans — fewer than one household in ten owned shares at all. Markets had crashed before without producing a decade of misery. What turned a crash into the Great Depression was not the stock market but the banking system, and it took three and a half years to do it.
The mechanics of a bank are worth stating plainly because they explain everything that follows. A bank takes deposits, promises to return them on demand, and lends most of them out for years at a time. It cannot possibly honor all its promises at once, and it does not need to, because on any normal day withdrawals and deposits roughly offset. The arrangement is not a fraud; it is what converts idle savings into houses, factories, and inventory. It is also, structurally, a bet that everyone will not ask at the same time — and the moment a depositor suspects that others may ask, the rational move is to ask first. Bank runs are not panics in the sense of irrationality. They are panics in the sense that each individual is behaving sensibly and the aggregate is a disaster.
The American banking system of 1930 was uniquely vulnerable, because political hostility to large banks had produced a country with twenty-five thousand of them, most tiny, most confined by law to a single state or even a single office, and therefore undiversified in every direction: one town, one crop, one industry. A bad harvest in a county could kill its bank. Canada, next door, with nationwide branch banking and a tenth as many institutions, went through the same depression without a single bank failure. This is as close to a controlled comparison as monetary history offers, and the variable was the structure of the network rather than the severity of the shock.
The failures came in waves rather than a single collapse. An agricultural wave in late 1930, culminating in the failure of a large New York institution unfortunately named the Bank of United States — a private bank with no official status whose name persuaded many depositors abroad that the American government had gone under. A second wave in the spring of 1931. Then the shock crossed the Atlantic in reverse when Austria’s largest bank, the Creditanstalt, failed in May 1931, pulling down German banks that were entangled with it, forcing Germany into capital controls, and finally driving Britain off the gold standard in September. The international monetary machine that had been rebuilt without a mechanic was now transmitting failure from each part to the next.
At each stage a central bank could have flooded the system with liquidity, and at each stage this was not done adequately. The rule had been written down sixty years earlier by Walter Bagehot, editor of The Economist, in a book about the London money market: in a crisis, lend freely, at a penalty rate, against good collateral. Lend freely, because a run is a shortage of confidence and can be met with a demonstration that the money is available. The Federal Reserve, a decentralized institution barely fifteen years old, with regional banks that disagreed with each other and a leadership convinced that failing banks deserved to fail, did approximately the opposite. It raised rates in October 1931 to defend the gold parity, in the depths of a banking collapse, and the money supply of the United States contracted by roughly a third between 1929 and 1933.
That contraction is the mechanism by which a financial event became a human one. When a bank fails, its depositors’ money does not go somewhere else; it ceases to exist, because deposits are created by lending and destroyed by its reversal. Less money chasing the same goods means falling prices, and falling prices raise the real weight of every existing debt, which causes more defaults, which kills more banks. The economist Irving Fisher, who had lost his own fortune in the crash and had famously declared shortly beforehand that stocks had reached a permanently high plateau, worked out the process and named it debt deflation: the more the debtors pay, the more they owe.
It ended in March 1933 with a piece of stagecraft as important as the Rentenmark. Roosevelt, inaugurated on the fourth, declared a national bank holiday on the sixth, closing every bank in the country. Examiners went through the books; sound institutions were licensed to reopen within days, doubtful ones held back. On the evening of the twelfth he broadcast to the nation the first of his fireside chats, explaining in plain language what a bank does with deposits and why it is safer to keep money in a reopened bank than under a mattress. When the doors opened the following morning, deposits flowed back in. A government had ended a national panic by explaining fractional reserve banking on the radio, which remains the single most successful piece of public financial education ever attempted.
The permanent fix arrived in the Banking Act of the same year: federal deposit insurance, guaranteeing accounts up to a limit. The idea had been proposed and rejected repeatedly for fifty years, opposed by the banking industry and by Roosevelt himself, on the reasonable ground that insuring deposits encourages banks to take risks with them. It also, however, removes the reason for the depositor to run, and by removing the reason it removes the run. Bank failures in the United States dropped from thousands a year to a handful. It is the clearest case in this book of a purely psychological intervention producing a structural result: nothing about the fragility of the underlying arrangement changed, and the fragility stopped mattering because nobody had any reason to test it.
There was a less edifying measure alongside it. In April 1933 an executive order required Americans to surrender their gold coin and bullion to the Federal Reserve at twenty dollars sixty-seven cents an ounce, with criminal penalties for retention. The following January the government revalued gold at thirty-five dollars an ounce. Citizens had been required to sell at one price and the state had then marked the same metal up by nearly seventy percent, booking the difference as a profit. Whatever one thinks of the policy — and it did break the deflation, and recovery began that year — it settled a question that had been open since Locke. The gold content of the dollar was not a fact of nature. It was a number, and the government could change it on a Tuesday.
Chapter Twenty-Four — Bretton Woods
In July 1944, while the fighting in Normandy was still going on, seven hundred delegates from forty-four countries assembled at a resort hotel in the White Mountains of New Hampshire to design the monetary system of a world that did not yet exist. The hotel had been shut for years and reopened hastily; the plumbing failed, the staff were overwhelmed, delegates slept in corridors, and the Soviet delegation communicated with Moscow by telegram about every clause. Out of three weeks in that building came an arrangement that governed international finance for a quarter of a century.
Everyone present agreed on what they were trying to prevent. The interwar years had produced competitive devaluations, in which countries cheapened their currencies to steal export markets from each other; exchange controls; bilateral clearing arrangements that fragmented trade into a web of barter deals between states; and, in the general view of the delegates, a monetary disorder that had contributed materially to the rise of the regimes they were currently at war with. The goal was fixed exchange rates, to make trade predictable, without the rigidity of the gold standard that had forced countries to choose between their currency and their employment.
Two plans were on the table, and they reflected the positions of a creditor and a debtor. John Maynard Keynes, representing a Britain that had bankrupted itself fighting since 1939, proposed an international clearing union with its own unit of account, which he called bancor, in which both deficit and surplus countries would face pressure to adjust — the surplus countries penalized for accumulating balances, on the argument that a persistent surplus is as much a cause of imbalance as a deficit. Harry Dexter White, representing a United States that held most of the world’s gold and expected to run surpluses indefinitely, proposed a fund of subscribed national currencies with rules for borrowing from it, and no obligation on creditors whatsoever. White’s plan won, for the reason that such contests are usually decided: he had the gold.
The system that emerged was a gold standard with one participant. The dollar was fixed to gold at thirty-five dollars an ounce, and every other currency was fixed to the dollar, with a narrow band of permitted movement and the option to adjust the peg in case of what was elegantly termed fundamental disequilibrium. Convertibility into gold was available only to foreign governments and central banks, not to citizens — Americans remained forbidden to own bullion until 1974. Two institutions were created to manage the arrangement: a Fund to lend to countries in temporary difficulty, and a Bank to finance reconstruction and later development.
The Soviet delegation signed the agreements and then declined to ratify them, which surprised the Americans and should not have. White, the principal architect, was later accused of having passed information to Soviet intelligence; he denied it before a congressional committee in 1948 and died of a heart attack three days afterward, and the evidence that has emerged since suggests the accusations had substance. The founding document of the post-war capitalist monetary order was drafted, in part, by a man who was probably a Soviet source, which is the kind of detail that would be rejected by an editor as implausible.
The system did not begin working for over a decade. European currencies were not convertible until 1958, and the intervening years were dominated by a dollar shortage: everyone needed dollars to buy American goods, and nobody had any, which was resolved by the Marshall Plan pushing dollars out into Europe in quantities large enough to restart trade. Once it did work, it worked spectacularly. The two and a half decades that followed produced the fastest growth in the history of the industrial world, with stable exchange rates, expanding trade, and — a point rarely made — tight controls on the movement of capital across borders, which nearly every government maintained and which are usually omitted from nostalgic accounts of the period.
The flaw was identified early, and by name. In 1959 a Belgian economist named Robert Triffin explained to the United States Congress that the arrangement contained a contradiction it could not survive. The world needed dollars for reserves and for trade, and the supply of dollars could only grow if the United States ran deficits. But the more dollars accumulated abroad, the less credible the promise to convert them all into a fixed and slowly growing stock of gold. Confidence in the dollar therefore depended on the United States doing something that steadily undermined confidence in the dollar. Either the world would be starved of liquidity or the anchor would eventually break, and there was no third option.
By the 1960s the arithmetic was visible to anyone who cared to look. Foreign dollar claims exceeded American gold reserves and the gap widened every year, driven by military spending abroad, domestic programs, and the growing appetite of European central banks. Eight countries formed a London Gold Pool in 1961, coordinating sales to hold the market price at thirty-five dollars, which is a cartel of central banks defending a price they knew to be wrong. France withdrew from it in 1967 and began systematically converting its dollar holdings into metal, shipping it home, in a policy its finance minister had described as a protest against the exorbitant privilege of the country that could settle its foreign debts by printing its own currency. The Pool collapsed in March 1968 after a run so heavy that the London market was closed for two weeks.
What remained after that was a two-tier arrangement: an official price of thirty-five dollars for transactions between central banks, and a free market price that was higher and rising. A fixed price that only applies to insiders and only holds while the insiders agree not to test it is not a monetary standard. It is a gentleman’s agreement with a shrinking number of gentlemen, and by the summer of 1971 the British government would ask to convert three billion dollars, and the last of them would run out of patience.
Chapter Twenty-Five — Nixon Closes the Window
On Friday the thirteenth of August 1971, fifteen men flew by helicopter to Camp David under instructions to tell nobody where they were going. Some had not been told themselves. Over the weekend they designed a new international monetary order, and on Sunday evening the President went on television to announce it, preempting the most popular program in America and irritating a nation of viewers who had been expecting a western.
The immediate trigger was a British request. London had asked the United States Treasury to guarantee three billion dollars of its holdings against devaluation, or convert them, and the sum was a substantial fraction of what remained in the vaults. It was the last of a long series. Foreign claims on American gold had exceeded the American gold stock for years, and everybody in the room understood that the window was going to be closed; the only question was whether it would be closed by decision or by a stampede.
The announcement had three parts, and the packaging was as deliberate as the substance. Convertibility of the dollar into gold was suspended — temporarily, the President said, and it has been temporary for over fifty years. A ten percent surcharge was imposed on imports, to force trading partners to the negotiating table. And a ninety-day freeze was placed on wages and prices, which had nothing to do with the gold decision and everything to do with the fact that suspending convertibility sounds like a defeat, whereas a bold three-part program against inflation sounds like leadership. The freeze was wildly popular, the stock market rose sharply the next day, and the fact that the United States had just defaulted on its central obligation to the rest of the world went down as a triumph.
John Connally, the Treasury Secretary who ran the meeting and the messaging, had summarized his negotiating philosophy earlier that year in a remark to European finance ministers that has been quoted ever since: the dollar is our currency, but it is your problem. It was not diplomacy. It was an accurate description of the balance of power, and the Europeans, who had been complaining for a decade about the privilege of a country that could settle its debts in money it printed itself, discovered that complaining was the full extent of their options.
An attempt was made to rebuild the fixed-rate system. In December the major economies met at the Smithsonian Institution in Washington and agreed a new set of parities, with the dollar devalued against gold to thirty-eight dollars an ounce and other currencies revalued upward. Nixon called it the most significant monetary agreement in the history of the world, which was an unusual claim for an arrangement that lasted fourteen months. Speculation resumed almost immediately, the dollar was devalued again in February 1973, and by March the major currencies were floating against each other. They have been floating ever since.
It is worth pausing on what that means, because it is the largest single change in the subject of this book and it happened without a vote anywhere. From that point on, no major currency has had any defined value in terms of anything at all. A dollar is not a claim on gold, silver, land, grain, or labor. It is a claim on nothing whatsoever. Its value is determined entirely by what it will buy, which is determined by how much of it exists relative to the goods available, which is determined by the decisions of a central bank and the lending behavior of commercial banks. The entire planet had, in less than two years, moved onto a monetary basis that every previous generation would have regarded as an emergency measure of the sort that precedes disaster.
Disaster duly appeared, though its causes were tangled. Through the 1970s inflation in the industrial world reached levels not seen in peacetime. The oil embargo of 1973 quadrupled crude prices, and it is worth noting that the producing states had a monetary grievance as well as a political one: their oil was priced in dollars, and the dollars had been losing value for years, so a nominal price that looked stable had been falling in real terms. Wage and price controls in America were extended, produced shortages and a farcical bureaucracy, and were abandoned. By 1979 American inflation was in double digits and the belief that governments could control it had substantially collapsed.
The repair came from a familiar direction: a public, painful, credible commitment. Paul Volcker, appointed to the Federal Reserve in 1979, announced that the central bank would target the quantity of money and accept whatever happened to interest rates as a result. What happened was that rates went above twenty percent, the economy went into the deepest recession since the 1930s, unemployment reached nearly eleven percent, farmers drove tractors to Washington and blockaded the building, and construction workers mailed him sawn pieces of two-by-four to represent the houses they were not building. He did not reverse course, and inflation fell from over thirteen percent to under four within three years.
The lesson institutionalized from that episode governs monetary policy to this day. If money has no anchor in metal, it needs an anchor in expectations, and the only way to produce one is an institution that will impose real costs to defend its word. Hence the wave of central bank independence in the 1980s and 1990s, hence formal inflation targets — New Zealand first, in 1990, with a target written into the governor’s contract — and hence the enormous attention paid to the exact wording of central bank statements, which would be bizarre if the words were not themselves the anchor.
So the modern arrangement is this: money is a promise, issued by a state, managed by an institution that is deliberately insulated from the politicians who appoint it, and worth what it is worth because that institution has established a reputation for making its word expensive to doubt. This is a considerably stranger arrangement than gold, and it has performed considerably better than the interwar gold standard did. It also depends entirely on a form of trust that took a recession and several thousand pieces of lumber to establish, and which nobody has found a way to establish more cheaply.
Chapter Twenty-Six — Plastic
The founding legend of the credit card industry involves a businessman named Frank McNamara who, dining at a New York restaurant in 1949, discovered he had left his wallet at home and had to be rescued by his wife. Humiliated, he conceived of a card that would let a member charge meals at any participating establishment. The company later admitted the incident had been improved for promotional purposes, which is fitting, since the entire industry runs on the difference between what people have and what they can be persuaded they can afford.
Diners Club launched in 1950 with a few hundred cardholders and fourteen New York restaurants. The card was cardboard. The business model was the important part: the club charged the restaurant a percentage of the bill and the member an annual fee, and it settled with both. This is a three-party system in which the card company sits between buyer and seller and takes a cut for solving a problem neither of them can solve alone — the seller cannot assess the creditworthiness of a stranger, and the buyer cannot prove it.
Banks entered a decade later and immediately encountered the fundamental obstacle of this business, which is that no merchant wants to accept a card nobody carries, and no customer wants to carry a card nobody accepts. Bank of America solved it with an act of institutional recklessness that would be illegal today. In September 1958 it mailed sixty thousand live, activated, unsolicited credit cards to residents of Fresno, California, followed by hundreds of thousands more across the state. This was known internally as the drop. Merchants signed up because the cards were already in customers’ hands; customers used them because they had arrived free in the post. It also produced fraud and delinquency on a scale that cost the bank something near twenty million dollars, a public scandal, and eventually a federal law banning the practice. The program survived, and by the mid-1960s was profitable.
The technical mess was worse than the financial one. Licensee banks across the country were running incompatible operations, transactions took weeks to clear, and paper drafts moved physically between institutions in sacks. A licensee executive named Dee Hock proposed, in 1970, that the whole thing be reorganized as a cooperative owned by the member banks, with no shares traded, no central owner, and a set of rules that any member could follow while competing fiercely with the others. The result eventually took the name Visa, chosen because it was pronounceable in most languages and meant something reassuring in all of them. Hock later wrote books about self-organizing systems and coined a word for what he thought he had built, which is what happens when a payments executive reads biology.
The physical technology arrived in parallel and has a domestic origin story that, unusually, appears to be true. An IBM engineer named Forrest Parry was trying to attach a strip of magnetic tape to a plastic identity card and failing, because every adhesive he tried distorted the tape. His wife, who was ironing, suggested he try the iron. The heat bonded the strip cleanly. The magnetic stripe made automated authorization possible, and automated authorization made the whole system scale from a network of restaurants to a network of everything.
What the card actually does is worth spelling out, since almost nobody who uses one could describe it. When a card is presented, no money moves. A message goes to the merchant’s bank, which routes it through the network to the card issuer, which checks the account and returns an authorization. The merchant is credited by its own bank, the issuer debits the cardholder, and the two banks settle their net positions with each other later, through the network, on a schedule. The purchase is a set of promises to adjust ledgers, made in about a second, and the goods leave the shop before any of the adjustments happen. Multilateral netting again, exactly as at the fairs of Champagne, with the addition of fiber optic cable.
The money in this system is made from a fee structure that most people never see. The merchant pays a discount on each sale, the larger part of which — the interchange fee — goes to the bank that issued the card. This creates a market in which issuers compete to give cardholders benefits, since a cardholder is a revenue stream, and merchants have almost no ability to refuse the resulting costs, since refusing a major card network means refusing a large share of customers. Airline miles, cash back, and airport lounges are funded by a charge embedded in the price of everything, including for the people paying cash, who receive none of the benefits. It is a quiet transfer from the poor to the comfortable, running continuously, at a scale most people would find startling if it were itemized on the receipt.
The genuinely transformative product was not the charge card, which had to be paid in full each month, but the revolving balance, which did not. Once a cardholder could carry debt indefinitely at a high rate, the card stopped being a payment convenience and became a consumer lending instrument with a payment convenience attached. Household debt in the industrial world rose accordingly, and a new industry grew up to decide who should be permitted to borrow — credit scoring, pioneered by a small California firm founded in 1956 and standardized into the general-purpose score that dominates American lending from 1989. Somewhere in a database there is a three-digit number attached to each person, calculated from their payment history, that determines the price of their mortgage and, in some jurisdictions, their access to an apartment or a job.
That is the deeper change, and it has almost nothing to do with plastic. For most of history, being trusted with credit meant being known by a lender — a banker who knew your family, a merchant who knew your business. The card industry replaced that with a statistical estimate of trustworthiness computed from records, portable, instantly queryable by strangers, and applied to hundreds of millions of people who have never met anyone at the institution deciding about them. The ledger has always been the money. What changed in the second half of the twentieth century is that the ledger became a machine, and it began keeping a record not only of what you owe but of what you are likely to do.
Chapter Twenty-Seven — The Money Nobody Governs
A dollar deposit in a bank in London is not a dollar in America. It is a claim, in dollars, on a bank outside American jurisdiction, created by that bank when it lends, subject to no American reserve requirement, insured by no American agency, and rescuable by no American authority. Trillions of these exist. They are called eurodollars, the prefix having nothing to do with the European currency of the same name, and they constitute the largest pool of money in the world that no government issues or controls.
The origin story is politically delicious. After the Second World War the Soviet Union held dollar balances, which it needed for trade, and which it preferred not to keep in New York, where they might be frozen in a crisis — a concern that hardened considerably after the United States froze Chinese assets during the Korean War. The solution was to keep the dollars in Soviet-owned banks in Europe: the Moscow Narodny Bank in London and, in Paris, the Banque Commerciale pour l’Europe du Nord, whose telex address was Eurobank. The market took its name from a communist bank’s cable address, which is the sort of thing that ought to be more widely known.
Two pieces of regulation then turned a curiosity into an ocean. American banks operated under a rule capping the interest they could pay on deposits, so when market rates rose above the cap, depositors with any sophistication moved their dollars offshore where no cap applied. And in 1957 Britain, defending a weak pound, restricted the use of sterling to finance trade between third countries — whereupon London’s merchant banks, having lost the ability to do their traditional business in their own currency, simply started doing it in dollars. The City reinvented itself as the offshore center for a currency it did not issue, and became the capital of international finance again on that basis.
The consequence is a genuine oddity in the architecture of the modern world. Banks outside the United States create dollars by lending them, in exactly the way domestic banks create domestic deposits, and no American institution is counting. Estimates of the size of the market are estimates because there is no register. It funds global trade, corporate borrowing, and the balance sheets of banks on every continent. When people speak of the dollar’s dominance, this is largely what they are describing: not the notes in circulation, but a vast offshore credit system denominated in a currency that most of its participants have no political relationship with whatsoever.
Money without a lender of last resort is fragile in a specific way, and the market demonstrated it early. In June 1974 the German authorities closed a mid-sized bank in Cologne, the Bankhaus Herstatt, at the end of the German business day. The bank had taken in deutsche marks that morning from counterparties expecting dollars in return that afternoon in New York, where the day was still going. The marks had been paid; the dollars never came. Because the world’s time zones do not overlap, foreign exchange settlement contained a gap in which one side had paid and the other had not, and a failure inside that gap propagated instantly across the international system. It is still called Herstatt risk. The response was the formation, at the Bank for International Settlements in Basel, of a committee of banking supervisors from the major economies — the first serious attempt to regulate international banking, prompted by a mid-sized German bank being shut at the wrong hour of the afternoon.
The market also became the transmission channel for the largest debt crisis of the century. When oil prices quadrupled in the 1970s, producing states deposited their revenues in the international banks, and those banks, awash in dollars and short of borrowers, lent enormous sums to governments in Latin America and elsewhere at floating rates. The loans looked safe on the reasoning that countries do not go bankrupt — a formulation attributed to the head of the largest American bank of the era, and one of the more expensive sentences ever uttered. Then Volcker raised interest rates to break American inflation, the floating rates on all those loans repriced upward, commodity prices fell, and in August 1982 Mexico informed Washington it could not pay. The lost decade that followed across Latin America was made in an offshore market by a monetary decision taken in a different country for entirely domestic reasons.
The pricing of all this ran for decades on a benchmark that turned out to be a courtesy. The London Interbank Offered Rate was not a measured market rate; it was a daily average of estimates submitted by a panel of banks about what they thought they would have to pay to borrow. Hundreds of trillions of dollars of contracts — mortgages, corporate loans, derivatives — referenced this number. In 2012 it emerged that traders at several banks had been asking their submitters to nudge it, sometimes to benefit a position, sometimes, during the crisis, to make their institution look healthier than it was. The scandal produced fines in the billions, a handful of prosecutions, and the eventual replacement of the benchmark. The interesting part is not the dishonesty; it is that the central price of the global financial system was, for forty years, a poll.
The final demonstration of who actually governs this money came in 2008. Banks in Europe and Asia held enormous dollar assets funded by short-term dollar borrowing, and when that borrowing dried up they needed dollars they could not obtain, from a central bank that was not theirs. The Federal Reserve responded by opening swap lines — lending dollars to foreign central banks against their own currencies, in sums that ran into the hundreds of billions, so that they could lend them onward to their own banks. An American institution became the lender of last resort to the world’s banking systems, without a treaty, because the alternative was watching the offshore dollar market fail and take the domestic one with it. The arrangement was repeated in 2020 within days of the pandemic shutdowns.
So the honest description of the modern dollar is not that it is issued by the United States. It is that a great deal of it is created privately, offshore, beyond anyone’s statistics, and that the United States has accepted the role of backstopping the whole structure in emergencies because it cannot afford not to. That is an extraordinary amount of unlegislated power, and it is exercised through conference calls. What it looks like when the structure fails is the subject of the next chapter, and the striking thing about that failure is how little of it involved money in any form a previous century would have recognized.
Chapter Twenty-Eight — When the Ledger Blinked
Nothing was destroyed in 2008. No factory burned down, no harvest failed, no port was blockaded, no skilled worker forgot how to do their job. On the last day before the crisis the world contained a certain quantity of buildings, machines, knowledge, and willing hands, and on the day after it contained exactly the same. What changed was a set of numbers describing who owed what to whom, and the change in those numbers put tens of millions of people out of work. If any single episode demonstrates that money is a shared record rather than a substance, it is the one where the record went wrong and the substance was untouched.
The machinery that broke had been built for a respectable purpose. A mortgage is a long, illiquid, risky asset for a bank to hold; securitization pools thousands of them and sells claims on the combined stream of payments, which lets the bank lend again and lets pension funds hold housing exposure. The pooled claims were then sliced into layers, so that the top layer would be paid first and could be rated as safe while the bottom layer absorbed the early losses. The logic depends on one assumption: that the loans in the pool fail independently, for local reasons — a divorce here, a lost job there — rather than all at once for the same reason.
That assumption held for as long as house prices rose, which they had done nationally in America for as long as anyone in the business had been working. Underwriting standards deteriorated in the way standards always deteriorate when everybody is being paid on volume: documentation was waived, teaser rates were offered on the understanding that the borrower would refinance before they reset, and originators who sold the loan onward within weeks had no reason to care whether it was ever repaid. The ratings agencies, paid by the issuers whose products they rated, applied models fitted to a period in which the thing that was about to happen had never happened.
The first crack appeared not among the borrowers but in the plumbing. Modern banks do not fund themselves chiefly with retail deposits; they borrow short-term from other institutions, often overnight, secured against collateral in what is called the repurchase market. When doubts arose about what mortgage-backed collateral was actually worth, lenders demanded more of it against the same loan, or refused it entirely. This is a bank run, conducted between institutions, at wholesale, by professionals, with no queues outside any building — and it moves at the speed of a phone call rather than the speed of a crowd.
There were queues, once. In September 2007 depositors of Northern Rock, a British mortgage lender that funded itself in the wholesale market and had just been refused, formed lines outside its branches, crashed its website, and produced the first run on a British bank since 1866. Photographs of orderly British people waiting on pavements for their savings went around the world and did more to communicate the situation than any amount of commentary about repo markets.
The following year removed the remaining doubt in stages. Bear Stearns was sold in March at a price that amounted to a supervised failure, with public money supporting the transaction. The two enormous American mortgage agencies were taken into government conservatorship in early September. And on the fifteenth of September, after a weekend of negotiations in which no buyer could be found and no public support was offered, Lehman Brothers filed for bankruptcy — the largest in American history, with over six hundred billion dollars of assets.
What happened next was not predicted by the people who allowed it. A money market fund called the Reserve Primary Fund held Lehman commercial paper, and the loss was enough to push the value of its shares below a dollar. Money market funds were understood by everyone who used them to be as safe as cash, and this one had broken the buck. Investors began pulling money out of every such fund in the country, which meant those funds stopped buying the short-term paper that ordinary corporations use to make payroll. A general industrial company with no connection to housing suddenly could not roll over its financing. The contagion had crossed from mortgages to the entire economy through a channel nobody had listed as a risk.
The rescue that followed was on a scale that broke the vocabulary. Government guarantees were extended to money market funds; the central bank became a direct buyer of commercial paper; capital was injected into banks whose executives were summoned to Washington and told they would take it whether they wanted it or not; the swap lines discussed in the last chapter poured dollars into foreign central banks; and, when interest rates reached zero and could go no lower, central banks began buying government bonds and mortgage securities in enormous quantities under the name quantitative easing. Balance sheets that had taken a century to reach one size quadrupled in a few years.
The intellectual content of all this was two centuries old. Bagehot’s rule — lend freely against good collateral — was followed this time, at a scale he could not have imagined, in markets he would not have recognized, and the world avoided a repeat of the 1930s. Output fell sharply and recovered; the banking system did not disintegrate; deposit insurance meant no ordinary saver in the major economies lost money. Measured against 1931, it was a triumph of applied history.
Measured against anything else, it left a residue that has not settled. The institutions rescued were largely the ones that had caused the damage, and the households that lost homes were not rescued at all; in America alone something like ten million families lost theirs. Almost no senior executive faced criminal consequences. The clearest lesson available to the public was that a sufficiently large institution operates under a different set of rules, and that lesson has been an ingredient in every populist movement of the following decade. There is a straight line — not the only line, but a real one — from the rescue to the political convulsions that followed it, and to the arrival of a proposal for money that no government could rescue, print, or seize, published seven weeks after Lehman fell.
The deepest lesson, though, is the one stated at the beginning. Most of the money in a modern economy is not issued by anybody. It is created when commercial banks lend, and it is destroyed when loans are repaid or written off, which means the money supply of a developed country is a byproduct of millions of private credit decisions rather than a quantity anyone sets. The Bank of England published a paper in 2014 saying so plainly, to the surprise of a good many people who had been taught otherwise. In 2008 that private money creation went into reverse simultaneously across the world, and the state discovered that it was the guarantor of a system it neither owned nor fully understood. The ledger blinked, and everything that depended on the ledger — which is everything — blinked with it.
Chapter Twenty-Nine — The Speed of Disbelief
The hundred-trillion-dollar note is a genuine artifact of the Reserve Bank of Zimbabwe, issued in January 2009, and it can be bought today as a novelty for a few actual dollars, which is more than it was ever worth in groceries. By the time it was printed the country had abandoned its own currency in practice, and within a month it would abandon it in law.
The Zimbabwean collapse is instructive because the mechanism was unusually well documented and unusually simple. Following the land seizures that began in 2000, agricultural output — the country’s export base and tax base — fell sharply. Government revenue collapsed while obligations did not, and the central bank was directed to fund not only the budget deficit but a range of activities that a central bank has no business funding: subsidized loans to farmers, foreign exchange support for favored importers, allowances for veterans. These quasi-fiscal operations were financed by creating money, which is a deficit conducted through the printing works rather than the treasury, and it produced the second-highest inflation ever recorded, with prices at the peak in November 2008 roughly doubling each day.
What happened next was not a policy. Zimbabweans simply stopped using their currency. Shops began pricing in American dollars and South African rand; wages were negotiated in dollars; the informal economy converted first and the formal economy followed. By the time the government legalized the use of foreign currencies in early 2009, the population had already made the switch, and the effect was immediate: inflation stopped, shelves refilled, and the economy grew for several years. A country cured its hyperinflation by abolishing its own money, which is the most extreme demonstration available that a currency is a service and its users are customers who can leave.
The sequence in which people leave is worth noticing, because it repeats everywhere. The unit of account goes first: sellers begin quoting prices in the foreign currency while still accepting local notes at a daily rate. Then the store of value goes, or rather it went long ago, since nobody with savings holds a depreciating currency for longer than an afternoon. The medium of exchange holds out longest, because small transactions need small denominations and dollars come in awkward sizes — Zimbabwean shops famously gave change in sweets, pens, and vouchers, an improvised small-change crisis of the kind that has attended every dollarization. Money dies from the top down.
Venezuela ran the same play with better resources and worse timing. An economy dependent on oil revenue for the overwhelming majority of its exports, running expansive social programs financed from that revenue, met a collapse in oil prices from 2014 combined with the deterioration of its own production capacity. The response included price controls, which produced the shortages that price controls always produce, and multiple official exchange rates, which produced the arbitrage that multiple exchange rates always produce, and monetary financing of a very large deficit. Inflation reached the millions of percent. The currency has been redenominated three times since 2008, striking off a total of fourteen zeros, and a person who saved a large sum in 2007 has, arithmetically, nothing. The country also launched a state cryptocurrency notionally backed by oil reserves in the ground, which attracted essentially no users, since a state that has destroyed one currency does not become credible by issuing a second one with a longer name.
Argentina deserves a place here as the specialist. It has defaulted on sovereign debt roughly nine times since independence, has run inflation above fifty percent in more decades than not, and has replaced its currency five times, lopping off thirteen zeros in the process. The most instructive episode is the convertibility regime of the 1990s, when the peso was fixed by law at one to the dollar and the central bank was required to hold dollar reserves against the currency it issued. It worked: inflation fell from thousands of percent to nearly nothing, and it held for a decade. It also removed the country’s ability to adjust when the dollar strengthened and Brazil devalued, and the adjustment had to happen through wages and employment instead. By 2001 the arrangement was breaking, deposits were fleeing, and the government imposed the corralito — the little fence — restricting withdrawals to a small weekly sum. Riots followed, the president left the roof of the palace by helicopter, the country went through five presidents in two weeks, defaulted on around a hundred billion dollars, and broke the peg. Provinces issued their own emergency scrip to pay employees, and for a period something like a dozen quasi-currencies circulated inside a single country.
Lebanon supplies the most recent and in some ways the strangest case, because the currency held its peg for twenty-two years and the banks failed anyway. The arrangement depended on a continuous inflow of deposits from the diaspora, attracted by high interest rates and lent onward to a government that spent them, in a structure whose critics used a rude word for it and whose defenders had no better one. When the inflows slowed in 2019, the banks simply stopped letting depositors withdraw their own dollars, without any legal authority to do so and without any formal declaration. Savings of a lifetime became unreachable numbers on a statement, convertible only at punitive official rates. The currency lost more than ninety-five percent of its value, and the phrase that entered common use for the trapped deposits translates roughly as banknote-money that isn’t.
Across all these cases the same structural signature appears. The deterioration takes years and is fully visible: the deficit, the arrears, the parallel exchange rate widening in the street. Officials deny it, because acknowledging it accelerates it. Then a threshold is crossed and the change is not gradual at all — shopkeepers reprice daily, then hourly; depositors who waited become depositors who queue. Nothing in the physical world moves on the day it happens. What moves is the belief that other people still believe, and that quantity has no intermediate states.
The consolation, such as it is, is that the same discontinuity works in the other direction. Confidence returns as abruptly as it left, provided the change offered is real and visible — a Rentenmark with a closed printing press, a bank holiday with an audit, a dollarization that removes the discretion entirely. What cannot be done is a gradual restoration by reassurance, because reassurance is exactly what a failing issuer always offers, and everybody knows it.
Chapter Thirty — The Phone Became the Bank
In 2007 a Kenyan mobile operator launched a service allowing customers to convert cash into electronic value at a corner kiosk, send it by text message to any other subscriber, and convert it back into cash at a kiosk near the recipient. It was called M-Pesa, from the Swahili for money. Within four years a majority of Kenyan adults were using it. Within a decade the value moving through it each year was comparable to a substantial fraction of the country’s entire economic output, and the ordinary way to pay a plumber, split a restaurant bill, buy vegetables, or pay school fees in Nairobi was to send a text.
The origin was almost accidental. A pilot funded by a British development agency had intended the platform for microfinance loan repayments. Testing revealed that borrowers were using it for something the designers had not anticipated: sending money to relatives. Kenya had a large internal migration, with workers in the cities supporting families in the countryside, and the existing options for getting money home were an expensive formal transfer, a bus driver willing to carry an envelope for a fee, or a personal journey. The pilot was redesigned around the behavior the users had invented, which is a rare and admirable thing for a project to do.
What made it possible was an absence rather than a presence. Kenya in 2007 had few bank branches outside the cities, low card penetration, and a population that had never been served by conventional finance — and consequently no incumbent industry with the political weight to strangle a competitor, and no installed base of habits to overcome. The regulator, notably, allowed the service to operate before writing rules for it, on the condition that customer funds were held in trust at commercial banks and could not be lent. Countries with mature banking systems took far longer to arrive at anything comparable, because they already had something that worked well enough for the people who mattered.
The technology was deliberately primitive: text messages on basic handsets, no internet, no smartphone, no card. The genuine infrastructure was the agent network — tens of thousands of small shopkeepers who exchanged cash for electronic value and back, and who had to keep enough of both on hand to serve their customers. Managing the liquidity of that network across a country, so that agents in remittance-receiving districts have cash and agents in sending districts have electronic float, is the hard operational problem, and it is a logistics business rather than a technology one.
The consequences ran further than convenience. Studies tracking Kenyan households over years found measurable reductions in poverty associated with access to the service, with the effects concentrated among female-headed households, and a shift of women out of subsistence agriculture into small business. The mechanisms are unglamorous: a household hit by illness or crop failure can receive help from a wider network of relatives, faster and at lower cost, which means fewer assets sold at distress prices. Informal risk-sharing has always been how people without insurance survive shocks. The phone extended its range.
It also made theft harder in a way that changes daily life. Carrying cash on a bus, keeping savings in the house, or paying a supplier in person all involve risks that fall most heavily on those with the least. A matatu driver who does not carry the day’s takings home is a driver less likely to be robbed. Salary payments moved to the system for the same reason.
The model spread across East Africa, into South Asia, and beyond, though the results have depended entirely on local structure rather than on the technology, which is available everywhere. India took a different route, building a public digital payments layer on top of a national identity system, with an interface any bank or app can connect to and no fees on ordinary transfers; the volume of transactions running through it now exceeds the card networks of most countries. China arrived at something else again, with two enormous private platforms absorbing the payment behavior of most of the population and building lending, investment, and credit scoring on top of it, until the state grew uneasy about a payments monopoly it did not run and brought them under a firmer hand.
Those three routes — the telecom operator, the public utility, the private platform — are the live options for the future of retail payment, and the differences between them are political rather than technical. Who holds the balances. Who sees the transaction record. Who may be excluded, and by whose decision. A society that pays by text message has handed a telecom company a detailed and permanent account of its economic life; one that pays through a state-run interface has handed it to the state; one that pays through a platform has handed it to a company that also knows what its users read and whom they message.
And in every version the money itself has become entirely informational for its users. Nobody in Nairobi thinks of the balance on their phone as a claim on a trust account at a commercial bank, though that is precisely what it is. They think of it as money, they treat it as money, and they are right to, because that is what money has been in every chapter of this book: an entry in a ledger that enough people believe. The remaining question, which occupied a small group of cryptographers at the same time all this was being built, was whether the ledger needed to be kept by anybody in particular at all.
Chapter Thirty-One — Satoshi’s Puzzle
Copying a digital file is free and perfect, which is wonderful for photographs and fatal for money. If a coin is a piece of data, its owner can spend it and then spend the identical data again, and there is no way for the second recipient to know. Every previous solution to this problem worked the same way: a trusted institution kept a list of who owned what and refused to process the second transaction. The list was the money, and somebody had to hold it.
Attempts to escape that requirement have a longer history than most people assume. A cryptographer named David Chaum built a system in the 1980s using blind signatures, which allowed a bank to certify a payment without learning who had made it — electronic cash with genuine privacy, offered to banks who were not interested, and to a public that in 1995 did not yet shop online. His company went bankrupt in 1998. Around the same time others were circulating proposals for currencies whose issuance would be governed by computational work rather than by an institution, and an anti-spam scheme called hashcash had established the mechanism: require the sender to perform a costly calculation, easy to verify and expensive to produce.
On the last day of October 2008, six weeks after Lehman Brothers failed, a nine-page paper was posted to a cryptography mailing list under the name Satoshi Nakamoto. Its proposal was to maintain the list of who owns what as a public record, replicated by everyone, with new entries added in blocks by whichever participant first solves a difficult computational puzzle, and with the rule that the valid history is the one carrying the greatest accumulated computational effort. Rewriting the past would therefore require redoing all that work faster than everyone else is adding to it. Participants are paid for the effort in newly created units, which is simultaneously the security budget and the issuance mechanism — an elegant piece of design, since the thing being protected pays for its own protection.
The first block was produced on the third of January 2009, and embedded in it, in a field that could hold arbitrary text, was a line from that morning’s edition of a London newspaper about the Chancellor being on the brink of a second bailout for the banks. This was both a timestamp proving the block was not created earlier and an unmistakable statement of what the project was against.
The monetary rules were fixed at the outset and are unusually strict: a total supply capped at twenty-one million units, with the rate of new issuance halving roughly every four years until it reaches zero. This is deflationary by construction, which its designers considered a feature and which most monetary economists consider the single worst property a currency can have, since money expected to appreciate is money people hoard rather than spend. The most famous transaction in the system’s history illustrates the problem with painful clarity: in May 2010 a programmer in Florida paid ten thousand units for two delivered pizzas, an exchange that both parties considered fair and that is commemorated annually by people calculating what those pizzas would cost today.
The technical achievement is real and should not be diminished by anything that has been built on top of it. Getting a large number of mutually distrustful participants, with no shared authority, to agree on a single ordered history of events is a genuinely hard problem in computer science, and it had not been solved practically before. Whether the solution should be applied to money is a separate question, and the answer after fifteen years is complicated.
As a payment system it has been largely bypassed. Transaction throughput is low by design, confirmation takes minutes, fees fluctuate, and volatility makes pricing anything in it impractical — a currency that can move twenty percent in a week is unusable as a unit of account, which is why virtually nobody quotes prices in it, including the businesses that accept it and convert immediately to dollars. El Salvador made it legal tender in 2021, built a state wallet, and found that most citizens used the application for the sign-up bonus and then stopped; the policy was substantially unwound in 2025 as a condition of an international loan. As a store of value it has performed spectacularly for early holders and ruinously for many who arrived later, and the pattern of arrival is the familiar one from an earlier chapter: a real innovation, credit-fueled speculation, buyers who bought because the price was rising, and periodic collapses of eighty percent.
The custody problem turned out to be the recurring disaster. A holder who keeps their own keys can lose them irrecoverably — a Welsh engineer has spent years petitioning a council for permission to excavate a landfill containing a hard drive with a fortune on it — and a holder who uses an exchange has simply reintroduced a trusted third party, with none of the regulation or insurance that governs banks. The largest exchange of the early years, which at one point handled most of the world’s trading, collapsed in 2014 having lost hundreds of thousands of units to theft and mismanagement. The pattern repeated in 2022 with a firm that was, at the time, the industry’s most respectable face, and whose founder was subsequently convicted of fraud. A technology built to remove the need for trusted intermediaries has spent its entire existence being intermediated by some of the least trustworthy institutions in modern finance.
The most philosophically interesting failures, though, are the ones where the human governance showed through. In August 2010 someone exploited a bug to create billions of units out of nothing; the developers patched the software within hours and the network was rolled back, erasing the transaction. A software update in 2013 caused the chain to split, and the resolution was a conversation in a chat room among a handful of people who agreed which version everyone should abandon. In 2017 a dispute over capacity split the community permanently into rival currencies. Each episode demonstrated that the rules are not enforced by mathematics alone; they are enforced by a rough consensus among developers, miners, and exchanges about which version of the software constitutes the real one. The trust was not eliminated. It was relocated — from a central bank with a published mandate and an accountable governor to an informal coalition with neither.
That relocation is the honest summary. Every monetary arrangement in this book required somebody to be trusted about something: the assayer, the mint, the goldsmith, the clearing bank, the central bank. What the cryptographic approach offers is not the absence of trust but a different distribution of it, together with a genuine and valuable property that no state currency has — that the rules cannot be changed by a government in an emergency. Whether that is a virtue depends entirely on whether one expects the next emergency to be an unjustified confiscation or a bank run that a central bank ought to stop. The historical record contains a great many of both, and no technology has yet been proposed that distinguishes between them.
Chapter Thirty-Two — Stablecoins, Central Bank Money, and Programmable Cash
A currency too volatile to price a sandwich is not much use for buying a sandwich, and the crypto industry discovered this early. The solution was to issue tokens that promise to be worth exactly one dollar, backed by dollars held somewhere, redeemable on demand. Hundreds of billions of dollars of these now exist, they settle an enormous volume of transactions, and they are — in every economically meaningful respect — private banknotes, issued by unregulated institutions, circulating at par on the strength of a promise about reserves.
Anyone who has read the chapter on the free banking era will recognize the entire structure, including the failure modes. The value of the note depends on the quality of the assets behind it and on the issuer’s willingness to redeem promptly. The largest issuer spent years declining to publish a full audit and was fined by American authorities over misstatements about its reserves. A well-regarded competitor briefly traded below its peg in March 2023 because a portion of its cash reserves sat in a bank that had just failed — a modern note falling to a discount because of doubts about a specific counterparty, exactly as the bank note reporters used to record. The parallel is not an analogy; it is the same instrument with better fonts.
The catastrophic version arrived in May 2022. An arrangement calling itself a stablecoin maintained its peg not with reserves but with an algorithm: the token could always be exchanged for a dollar’s worth of a companion cryptocurrency, whose value derived largely from the demand for the token. Attached to it was a lending protocol paying around twenty percent, which is a yield that in any other context would be described as a warning label. When large withdrawals began, the mechanism did precisely what its critics had predicted: redeeming the token created more of the companion coin, which lowered its price, which required creating still more of it. Something like forty billion dollars of nominal value evaporated in under a week, and the resulting failures propagated through crypto lenders and funds for the rest of the year. It was John Law’s machine, rebuilt without any of the excuses available in 1720, and it collapsed for the same reason: two assets each supporting the value of the other is not a foundation, it is a mirror facing a mirror.
The response of states to all this has been to consider issuing digital money themselves. A central bank digital currency is a claim on the central bank, held directly by the public, in electronic form — which sounds like a small change and is not, because at present the only central bank money the public can hold is physical cash. Everything in a bank account is a claim on a commercial bank, which is why bank failures matter and why deposit insurance exists. Digital central bank money would give ordinary people access to the safest asset in the system, and in doing so would raise an awkward question about what commercial banks are for, since in any crisis depositors would move their money to the central bank instantly and by phone. Proposals therefore include holding limits, which is an admission that the product must be made deliberately inconvenient to avoid destroying the banking system.
Progress has been mixed and instructive. A small Caribbean state issued the first one in 2020 and uptake has been modest. Nigeria launched one in 2021, achieved adoption in the low single digits despite considerable pressure, and watched its citizens continue to prefer agent banking and, increasingly, dollar stablecoins. China has run the largest pilot by far, distributing its electronic currency through the existing payment apps and municipal giveaways, with volumes that are large in absolute terms and small relative to the private platforms it was partly intended to check. The European debate has run for years and centers on questions that are entirely political: whether transactions would be visible to authorities, whether an offline mode preserving cash-like privacy is technically achievable, and who is liable when a phone is stolen.
Underneath the institutional questions sits the property that makes digital money genuinely new, and it is not speed. It is programmability. Money that exists as an entry in a system with rules attached can carry conditions: it can expire on a date, be spendable only on certain categories of goods, be restricted to a geographic area, be released only when a delivery is confirmed, or be blocked from certain recipients. Some applications are obviously useful. Chinese municipal stimulus vouchers with expiry dates were designed to be spent rather than saved, and they were. Escrow, tax withholding, and conditional aid disbursement all become automatic.
The same property, viewed from a different angle, is the ability to make money conditional on behavior, and history offers no reassurance about how such powers get used. Company scrip redeemable only at the company store was programmable money implemented with paper, and it was a mechanism of control rather than convenience. When Canadian authorities in 2022 directed banks to freeze accounts associated with a protest, using emergency powers and without court orders, the accounts belonged to ordinary participants and small donors, and the point was made vividly on both sides: to supporters, that the state must be able to cut off funding for illegal activity, and to critics, that the ability to exclude a person from the payment system is a power that had never previously required a trial. Nothing about that episode needed new technology. It required only that money already be a set of ledger entries under institutional control, which it has been for some time.
So the coming argument is not about whether money will be digital, since it overwhelmingly already is, nor about whether it will be programmable, since the capability exists and will be used. It is about defaults and limits: what a payment system is permitted to refuse, who may issue an instruction to it, what record is kept and for how long, and whether an anonymous option survives at all. Cash is a strange, inefficient, criminally convenient technology whose one irreplaceable property is that it does not report to anyone. Nothing currently proposed replicates that, and the decision about whether anything should is being taken, in most countries, without much public conversation — which is how monetary decisions have generally been taken, and one of the reasons this book exists.
Chapter Thirty-Three — What Money Actually Is
Five thousand years of evidence permit a definition, and it is short. Money is a record of obligation that a community treats as transferable. Everything else — the metal, the paper, the shell, the notch, the database entry — is the medium in which the record happens to be kept, and the choice of medium is a matter of engineering, fashion, and politics rather than of monetary principle.
The definition has three components, and separating them explains most of the confusion in the subject. There is a unit, which is a scale for measuring obligations and which need not exist physically at all — the Mesopotamian shekel of silver, the ancient French livre in which prices were quoted for centuries although no coin of that denomination was ever struck, the modern unit of account in which a bank statement is denominated. There is a ledger, which is the record of who is owed what, kept in the memory of a village, in cuneiform, on split hazel sticks, in double-entry books, or on servers. And there is an enforcement mechanism, which determines what happens to a person who declines to honor the record — gossip, exclusion, a court, a debt-bondage contract, a repossession, a credit score.
Almost every argument about money is really an argument about one of those three, conducted as though it were about the first. Disputes over gold are disputes about who controls the unit and whether the control should be removable from human hands. Disputes over central bank digital currency are disputes about who holds the ledger and what they can see in it. Disputes over debt forgiveness, from the Mesopotamian clean slates to modern student loans, are disputes about enforcement — about whether a recorded obligation is a fact of nature or a social arrangement that can be revised when its consequences become intolerable.
A second conclusion follows from the whole survey and is worth stating flatly: trust is never eliminated, only relocated. Coinage moved trust from the individual assayer to the issuing authority. The bill of exchange moved it from the traveler to a network of correspondent bankers. The gold standard tried to move it from politicians to a metal, and succeeded only in moving it to the credibility of the promise to keep converting. Deposit insurance moved it from the individual bank to the state. Cryptographic ledgers moved it from institutions to software and to the informal coalitions that maintain the software. In every case something still has to be trusted, and the useful question is never whether a system is trustless but rather who is being trusted, how visible they are, and what happens to them if they abuse the position.
The third observation concerns failure. Monetary systems do not degrade gracefully. They hold, absorbing strain that is fully visible to anyone watching, and then they reorganize suddenly, because every participant is deciding based on what they expect other participants to do. This is the reason the histories all read the same way: years of warnings, official denials, and then a fortnight in which everything changes. It also explains why repairs work the way they do. Gradual improvement does not restore a collapsed currency, because gradual improvement is indistinguishable from the reassurances that failed. What works is a visible, costly, irreversible commitment — a printing press stopped by law, a bank holiday with an audit, an interest rate raised past twenty percent and held there while tractors circle the building. The commitment does not need to be large in metal. It needs to be expensive to fake.
A fourth point is easy to miss because it is uncomfortable. Every monetary arrangement distributes as well as measures. There is no neutral option. A currency that holds its value rewards those who hold savings and punishes those who owe; one that loses value does the reverse. A deflationary currency transfers wealth to the patient and destroys the borrower; an inflationary one transfers wealth to the borrower and quietly taxes anyone with a bank account. Who is permitted to create money determines who receives it first, and receiving new money before prices adjust is one of the oldest privileges in economic life. The technical language of monetary policy exists partly because the underlying decisions are distributional and stating them plainly would make them harder to take.
There is also a durable illusion worth naming, since every generation falls for it. Whatever money one grew up with feels natural, and everything before it looks primitive while everything proposed after looks reckless. The Roman who trusted silver found paper absurd; the Victorian who trusted gold found managed currency irresponsible; the modern citizen who trusts a central bank finds cryptographic tokens ridiculous; and the crypto enthusiast finds a currency that can be created by a committee obviously insane. All four positions are internally coherent, and all four are describing arrangements that worked for a while and then had a bad century. The sensible attitude to one’s own monetary furniture is the attitude one ought to have toward any inherited institution: it is not natural, it was built by people with interests, it has failure modes, and it will be replaced eventually by something with different ones.
So what should be said to the person who asked, at the beginning of this book, what money is, and got a shrug? Something like this. Money is society’s memory of who has done what for whom, written down in whatever medium is convenient and enforced by whatever institution is credible. It is not a substance and never was, which is why removing the gold changed nothing fundamental and why the technology can change completely without the thing itself changing at all. It works because a sufficient number of strangers behave as though it works, and that behavior is stable across decades and then, occasionally, is not. It is the most useful fiction our species maintains, and the only reason it does not feel like a fiction is that we are all inside it, using it to buy lunch.
That leaves the question of what a person is supposed to do with this information, which is a different question from what money is, and which deserves its own final word.
Conclusion
A reasonable reader, having followed several thousand years of shells, notches, debasements, panics, and cryptographic experiments, is entitled to ask what any of it is for. Not the money — the reading. Knowing that the Yapese kept a stone at the bottom of a lagoon does not lower anyone’s rent. So this last section is an attempt to say what the history is good for, which is chiefly a matter of recognizing shapes.
The first shape is the one that recurs most often and is hardest to see from inside. A monetary system generates an obligation it cannot meet, and the shortfall is closed by the least visible method available. A Roman emperor puts less silver in the coin. A revolutionary assembly prints against land it has not sold. A colonial power imports cheaper shells. A treasury issues tallies against a revenue that will not arrive. A modern government allows inflation to run slightly above target for a decade and reduces the real burden of its debt without ever holding a vote on the matter. The methods are unrecognizably different and the structure is identical: a promise larger than the resource, closed by a transfer from whoever is least able to notice or object. Learning to see the transfer is the single most practical thing this history offers, because it is happening somewhere at all times and it is always described in technical language.
The second shape is the discontinuity. Confidence in money is not a quantity that slides; it is a configuration that holds and then reorganizes. The chapters on Tabriz, on the Haarlem auction, on the bank runs of 1931, on the repo market in 2008, and on the shopkeepers of Harare are all the same event told in different centuries. Nothing physical changes on the day it happens. What changes is everyone’s estimate of what everyone else is about to do, and because each person is watching the others, the estimate flips rather than drifts. Anyone who wishes to know whether a monetary arrangement is in danger should therefore stop watching the price and start watching the behavior of the people closest to the mechanism — whether the government accepts its own money for taxes, whether shopkeepers quote prices in it, whether officials are making unprompted statements about how safe it is. The last of these is a particularly reliable indicator, and it always arrives too late to be actionable.
A third shape concerns repair. Broken money is repaired by commitment rather than by resources, and the commitment must cost something. The Rentenmark had ludicrous backing and worked because the printing visibly stopped. The Athenian restoration worked because a public tester stood in the marketplace where everyone could see him. Roosevelt reopened the banks after an audit and an explanation, not after finding more gold. Volcker fixed American inflation by accepting a recession he could have avoided. In each case the mechanism was the same: an action expensive enough that faking it would have been more painful than doing it. This is why so many attempts at monetary reassurance fail. Announcements are cheap, and everybody knows they are cheap, including the announcer.
Beyond the shapes, the history settles a few arguments that continue to be conducted as though they were open. It settles the question of whether money must be backed by a commodity: it need not, it frequently has not been, and the arrangements that failed did so because their issuers could not stop issuing, not because of what was or was not in a vault. It settles the question of whether fiat money is a modern aberration: it was invented in eleventh-century Sichuan and abandoned by the fifteenth century Ming, which is a longer track record than most alternatives can claim. It settles the question of whether a metallic standard prevents crises: it did not prevent 1873, 1893, 1907, or 1931, and in the last of these it actively transmitted the disaster from one country to the next. And it settles, in the other direction, the question of whether discretionary management is safe: Weimar, Zimbabwe, Venezuela, and half a dozen others answer clearly. Both of the available systems have killed people. Choosing between them is a choice between different failure modes, not between failure and safety, and any account that presents it otherwise is selling something.
What can be said with more confidence is that the thing which distinguishes the successes from the failures is rarely the design and almost always the institution. A currency survives where the issuer can be constrained — by law, by independence, by an external commitment, by a rule that is politically expensive to break — and it fails where the issuer is also the party facing the bill. The Bank of England worked because Parliament could refuse. The Ilkhanate’s paper failed because there was nothing between the treasury and the printing block. This is a boring conclusion. It is nevertheless the one the evidence supports, and it explains why the same technology produces triumph in one polity and catastrophe in the next.
As for what comes next, the honest position is modest. Anyone writing about the future of money in 1965 would have predicted a stronger gold link and would have missed the credit card, the eurodollar market, and the fact that the entire system would abandon metal within a decade. What can be observed is direction rather than destination. The physical instrument is disappearing from daily life in most wealthy countries, faster than anyone planned and faster than the arrangements for those who depend on it. Money is becoming programmable, which means the rules of the payment system are becoming a policy instrument rather than plumbing. The number of parties who can see a person’s complete transaction history is rising, and the number who can block a transaction is rising with it. And a large offshore system of dollar credit continues to operate outside the control of the country whose currency it is denominated in, which is the sort of arrangement that persists for decades and then does not.
None of that is a prophecy of doom, and this book has tried to avoid the genre of monetary apocalypse, which has been wrong every year since it was invented and shows no sign of correcting. The modern arrangement, for all its strangeness, has coexisted with the largest reduction in absolute poverty in human history and has survived two crises that would have destroyed the interwar system. Managed fiat currency in a country with functioning institutions has been, empirically, one of the better monetary regimes ever operated. That is a defensible statement and a fragile one, and both halves matter.
One rhythm in the record deserves a mention on the way out, because it is currently repeating. Private money keeps coming back. Whenever the official arrangement is inconvenient, expensive, or slow, somebody invents an instrument outside it — the merchant deposit receipts of Chengdu, the bills of exchange of Florence, the notes of ten thousand American banks, the offshore dollar, the token pegged to a dollar and issued by a company nobody has heard of. Every time, the innovation happens because it solves a real problem the incumbent system was ignoring. And every time, after a period of enthusiasm and a memorable failure or two, the state absorbs it: by taxing it out of existence, by chartering it, by regulating its reserves, or simply by issuing a competing version with official backing. There is no reason to expect the current round to end differently, and the interesting question is only which of the new instruments will be domesticated and which will be prosecuted.
It is worth admitting how imprecisely the subject can even be measured. Central banks publish several different totals for the quantity of money in an economy, depending on whether one counts notes, current accounts, savings that can be moved with a few days’ notice, or short-term instruments that behave like cash until the afternoon they do not. The boundaries between the categories are conventions, and they have been redrawn repeatedly, sometimes because a new product genuinely blurred them. A money market fund share was cash to everyone who held it until one of them broke the buck. This is not a scandal, but it is a caution: the object being managed is defined by behavior rather than by law, and behavior is precisely what changes during a crisis.
And it is worth noticing that nobody living through any of the transitions described in this book knew they were living through one. The Athenians restoring their silver were not aware of founding anything; the clerks of the Exchequer thought they were keeping accounts, not inventing tradable public debt; the Bank of England’s subscribers were funding a war, not building the fiscal machine that would win a century of them. Monetary history is made by people solving next month’s problem, and the structures they leave behind outlast the problems by centuries. Whatever the present arrangement is turning into, the people building it are almost certainly describing it as a payments upgrade.
There is a final thing worth saying, and it belongs to the reader rather than to the history. The invisibility of money is not an accident and it is not entirely benign. A great many arrangements persist because most people never look at them: the fee embedded in the price of everything, charged to the cash customer and rebated to the frequent flyer; the exchange rate spread on a remittance sent home by someone who cannot afford it; the deposit that earns nothing while the institution holding it earns plenty; the slow, unlegislated transfer that inflation performs on a savings account. None of this is hidden. It is merely uninspected, in the way that plumbing is uninspected, and the people who benefit from the arrangement are perfectly content for it to stay that way.
So the small practical suggestion at the end of a long book is simply to look. Not with alarm, and not with the conviction that the whole edifice is a swindle, because it is not — it is an extraordinary cooperative achievement, and the alternative to it is barter, which everyone who has tried has hated. Look in the way one looks at any inherited institution: with curiosity about how it came to be shaped this way, awareness that it was shaped by people with interests, and a readiness to notice when it starts behaving strangely. The Yapese knew exactly where their money was and precisely how it worked, including the part at the bottom of the lagoon. Most modern people, holding an instrument of vastly greater sophistication, could not say who creates it or what stands behind it.
It is a rectangle of printed cotton, or a number on a screen, and it is worth what it is worth because a very large number of strangers have decided to behave as though it is. That decision has held, in various forms, for five thousand years, which is a considerable run for anything imaginary. It will keep holding for as long as the people who manage it can be constrained and the people who use it keep paying attention. Both halves of that sentence are the responsibility of somebody, and the somebody, in a society that gets a vote, is not entirely elsewhere.
Case Studies
Case Study One — The Bank of Amsterdam and the Secret That Lasted a Century
Background. By the early seventeenth century Amsterdam was the commercial capital of Europe, and its merchants were drowning in coin. Some eight hundred varieties of foreign silver and gold circulated in the Dutch Republic, of wildly varying weight and fineness, and the money changers who assessed them had an obvious incentive to shade their assessments. Worse, the good coins were constantly being culled from circulation and melted or exported, leaving the worn and clipped behind. A merchant settling a large bill of exchange could not know with confidence what he had been paid.
The challenge. The problem was not scarcity of money but the absence of a reliable unit. Contracts were denominated in guilders, but a guilder in practice meant whatever coins the payer chose to hand over, and disputes over quality were endemic. For a city whose entire business was the settlement of international trade, this was a structural handicap. The municipal authorities also faced a related nuisance: the cashiers who held merchants’ coin were unregulated, occasionally absconded, and were suspected of lending out deposits they had promised to keep.
The solution. In 1609 the city established the Wisselbank, the Bank of Amsterdam, as a municipal institution. Merchants deposited coin and bullion; the bank assessed the metal content and credited an account in bank money, a unit defined by weight of silver rather than by any circulating coin. Bills of exchange above a modest threshold were required by law to be settled through accounts at the bank, which meant every significant merchant in the city needed one. Transfers between accounts were made by written order, free of charge, with no coin moving at all. Crucially, the bank was initially forbidden to lend, holding full reserves against its deposits.
The effects were immediate and larger than anyone had planned. Because bank money was of certified quality, it traded at a premium over circulating coin — the agio — which typically ran a few percent and became the reference price quoted across Europe. Amsterdam had created a currency that existed only as ledger entries, was not legal tender, could not be withdrawn as such, and was nevertheless the most trusted money on the continent. Settlement of international trade was concentrated in one set of books in one building, and the cost of moving value between the major commercial houses of Europe fell to the price of a clerk’s ink.
Complications. The prohibition on lending did not survive contact with opportunity. From the middle of the century the bank began advancing funds quietly — first to the city, then to the Dutch East India Company, which was both the largest commercial enterprise in the world and a persistent borrower. These loans were not disclosed. On the books shown to the public, the bank still held metal against every guilder of deposits; in reality an increasing fraction of its assets were claims on a trading company operating on the other side of the world. This is the earliest well-documented case of a bank running a fractional reserve while advertising a full one, and it worked for roughly a hundred and fifty years.
The bank also invented, in 1683, an instrument of real ingenuity: receipts. A depositor of coin received both a credit in bank money and a transferable receipt entitling the holder to withdraw the metal within six months on payment of a small fee. The receipt could be traded separately, and its price moved with the market’s appetite for metal. This split the deposit into a monetary claim and an option on the underlying, which is a piece of financial engineering that would not look out of place in a modern derivatives desk, arrived at by a municipal bank in the seventeenth century to manage its bullion stock.
The results. The end came when the concealed lending grew too large to hide. The Fourth Anglo-Dutch War, beginning in 1780, went badly; the East India Company’s finances deteriorated toward its eventual collapse; and the bank’s advances to the city rose. In 1790 the bank announced that it would pay out only a portion of certain balances, and the agio, which had been positive for nearly two centuries, went negative — bank money now traded at a discount to coin, which is the market announcing that it had worked out what was in the vault. The city took over the bank’s obligations, and the institution was wound up in 1819 after two centuries of operation.
Relevance today. Three lessons transfer directly. First, the most valuable thing a monetary institution can offer is not credit but certainty about quality — the agio was the market’s price for the removal of doubt, and it was substantial. Second, an institution that is prohibited from lending will eventually lend anyway unless somebody is checking, and the checking must be independent of the institution’s owner, which in this case was the city that was also its largest borrower. Third, and most modern of all: the Bank of Amsterdam ran for a century and a half on an undisclosed gap between its published and actual balance sheet, and nothing happened until a war made the gap too wide. Opacity is not the same as stability; it merely postpones the date on which the discovery is made, and concentrates the consequences into a shorter period. Every contemporary argument about bank disclosure, stablecoin attestations, and central bank transparency is an argument the Amsterdam city council had already lost.
A detail worth adding is how the bank handled the one thing it could not certify: foreign exchange. Because bank money was defined by weight of fine silver rather than by any coin, the exchange rate between it and every circulating currency in Europe could be calculated precisely rather than negotiated. Amsterdam consequently became the place where those rates were quoted, and the quotations were published in printed lists that traveled with the post. A merchant in Livorno or Danzig could look up what his coin was worth in Amsterdam bank money and therefore what it was worth against any other currency on the list. The bank had accidentally created the world’s first reliable exchange rate reference, and the informational value of that was arguably greater than the settlement service itself. Financial centers acquire their position less by holding assets than by being the place where prices are discovered, which is why they are so difficult to displace once established, and why London retained its role for a century after Britain ceased to be the largest economy.
Case Study Two — Copper Plates, Palmstruch, and Europe’s First Banknote
Background. Sweden in the seventeenth century was a great military power with an unusual monetary handicap: it had copper rather than silver. The mine at Falun was the largest copper producer in Europe, and the Crown, seeking to support the price of its principal export, put the metal into circulation as money. Since copper is worth a fraction of silver by weight, a coin of meaningful value had to be enormous. The Swedish plate money of the period was exactly that: rectangular sheets of copper, stamped at the corners and center, with the largest denominations weighing close to twenty kilograms.
The challenge. A currency that requires a wheelbarrow for a substantial payment is not a currency; it is a logistical problem with a face value. Merchants transporting money hired carts. Payments were made by weight and required scales, cranes, and strong assistants. Storage was a difficulty, theft was inconvenient rather than easy, and any transaction across a distance involved shipping a significant tonnage of metal along roads that were, for much of the year, impassable. The value of the plates also fluctuated with the copper market, so a merchant’s cash holdings changed in worth with the fortunes of an export commodity.
The solution. A merchant of Latvian birth named Johan Palmstruch obtained a royal charter in 1657 to establish a bank in Stockholm, on terms that gave the Crown half the profits. The bank accepted deposits of plate money and, from 1661, began issuing credit notes — kreditivsedlar — printed on paper, in round denominations, payable to the bearer, requiring no named depositor and carrying no interest. These were the first true banknotes issued in Europe: not receipts for a specific deposit of a specific person, but standardized bearer instruments intended to circulate.
The reception was enthusiastic to a degree that ought to have alarmed somebody. A piece of paper that could be carried in a pocket and exchanged for goods was, to anyone accustomed to twenty-kilogram plates, an almost miraculous improvement. The notes were accepted readily, circulated at par, and were in some respects preferred to the metal they represented. Sweden had discovered the convenience of paper money by having the least convenient metallic money in Europe, which is a reliable pattern — innovation arrives where the incumbent system is most painful.
The failure. Palmstruch’s bank issued more notes than it held plate money to redeem. The motive was ordinary: notes could be lent, lending was profitable, and demand for redemption was low so long as confidence held. It also had a specific trigger. The copper price fell, plate money was recoined at a lower weight, and holders of notes calculated that they would prefer metal after all. Redemption demands rose, the bank could not meet them, and by 1667 it had suspended payment. Palmstruch was tried, convicted of mismanaging the bank, and sentenced to death — a sentence commuted to imprisonment, from which he was released shortly before his death in 1671.
The results. What is remarkable is what Sweden did next. Rather than abandoning the concept, the Riksdag took the bank over in 1668 and reconstituted it under parliamentary rather than royal control, precisely so that no king could raid it. That institution, the Riksens Ständers Bank, is the Sveriges Riksbank, and it is the oldest central bank still operating. Note issue was suspended for decades after the failure, and when it resumed it was under public supervision. A private experiment failed, the failure was investigated, the lesson was that note issue required accountability rather than prohibition, and the institution built on that conclusion has now operated for over three and a half centuries.
Relevance today. The arc is the standard one for financial innovation and is worth having in mind whenever a new payment instrument appears. A genuine inconvenience creates an opening; a private issuer solves it and is rewarded; the issuer over-issues, because the constraint is voluntary and the incentive is not; the failure destroys the issuer but not the idea; and the state institutionalizes the innovation with the constraint made mandatory. Sweden has since traveled the whole distance, incidentally: the country that produced Europe’s first banknote is now one of the closest to eliminating cash entirely, and its central bank has spent years studying whether a digital successor is needed for exactly the reason Palmstruch’s notes were needed — because the existing instrument has become inconvenient for ordinary use.
Sweden supplies one further detail that ought to be better known, because it inverts the usual assumption about who resists monetary innovation. The strongest opposition to paper notes came not from suspicious peasants but from the mining interests and the Crown officials tied to copper, who understood perfectly well that paper reduced the demand for the metal that underwrote both the export trade and their own revenues. Monetary conservatism is very often a commercial position wearing a philosophical coat. The same alignment can be traced in nineteenth-century silver politics, in twentieth-century arguments over the gold standard, and in the present decade, where the loudest institutional objections to new payment systems tend to originate with the incumbents whose fee income they threaten. Following the money in an argument about money is unglamorous and almost always productive.
Case Study Three — The Panic of 1907 and the Private Central Bank of One Man
Background. The United States in 1907 had no central bank, having abolished two of them, and a banking system fragmented into thousands of institutions. Alongside the national and state banks operated trust companies — institutions that took deposits and made loans much as banks did, while operating under lighter reserve requirements and outside the clearinghouse arrangements that gave banks mutual support in a crisis. Trust companies had grown rapidly and held a substantial share of New York deposits.
The challenge. In October a pair of speculators attempted to corner the stock of a copper company, using borrowed money channeled through banks and trust companies they were associated with. The corner failed, the share price collapsed, and depositors began to wonder which institutions had lent to the scheme. Runs began on the associated banks, were contained, and then jumped to the Knickerbocker Trust Company, the third largest in the city, whose president had been connected with the speculators. Knickerbocker paid out eight million dollars in three hours on the twenty-second of October and closed its doors. The panic then spread to the Trust Company of America and beyond, and because the trust companies were outside the clearinghouse, there was no mechanism to support them.
The solution. With no public institution capable of acting, the response was organized by J. Pierpont Morgan, then seventy years old, from the library of his house on Madison Avenue. Morgan assembled teams of accountants to work through the books of the threatened institutions overnight and to sort them into those that were merely illiquid and those that were insolvent — Bagehot’s distinction, applied under extreme time pressure by young men with ledgers. He then summoned the presidents of the major banks and trust companies and required them to contribute to a pool to support the sound institutions. On one celebrated occasion in early November he locked the trust company presidents in the library until they agreed, keeping the door key in his pocket.
The interventions extended well beyond the banks. The New York Stock Exchange faced closure when brokers could not obtain call money; Morgan raised twenty-five million dollars from the banks in minutes and had it lent on the floor at a fixed rate. The City of New York was unable to meet its payroll and was rescued with a bond purchase. The Treasury Secretary deposited federal funds with New York banks. And in a decision still argued about, Morgan arranged for United States Steel, which he had assembled, to acquire a large coal and iron company whose shares were being used as collateral by a failing brokerage — obtaining an assurance from President Roosevelt, a committed trust-buster, that no antitrust objection would be raised.
The results. The panic was contained within about three weeks. The cost was a sharp recession, a fall in industrial production of over ten percent, and several thousand bank and business failures nationally. The deeper consequence was political. The spectacle of a private banker performing the functions of a central bank — deciding which institutions would live, allocating emergency credit, and negotiating with the President — was intolerable to a great many Americans on both the left and the right of the argument. A congressional commission was established, spent years studying European central banking, and its work led through a series of political compromises to the Federal Reserve Act of 1913. The regional structure of the Federal Reserve, with twelve district banks rather than a single institution in New York, was a direct concession to the fear of concentrated financial power that the events of 1907 had inflamed.
Relevance today. The episode is the clearest illustration available of why a lender of last resort exists and what happens without one. Everything the Federal Reserve was later designed to do was done in 1907 by a private individual with no mandate, no accountability, and an obvious personal interest in the outcome — and he did it competently, which is precisely what made it intolerable. The modern versions of the same problem are visible whenever a crisis reaches institutions outside the regulated perimeter: money market funds in 2008, or the private credit and stablecoin sectors that have grown rapidly under lighter supervision since. The trust companies of 1907 were not shadow banks in the technical sense, but they were doing bank-like things without bank-like protections, and the crisis found them first. It generally does.
One last observation about 1907 concerns the mechanism the banks used among themselves, which is largely forgotten. Clearinghouse associations in American cities issued loan certificates during panics — instruments that member banks could use to settle balances with each other in place of cash, backed by collateral pooled at the clearinghouse and guaranteed jointly by the membership. In effect the banking system issued its own emergency currency, and in 1907 some of it leaked into general circulation as ordinary businesses accepted it for payroll. It was technically illegal, everyone knew it, and no prosecutions followed, because the alternative was worse. Private emergency money appears reliably whenever official money becomes unobtainable, from clearinghouse certificates to the depression scrip issued by hundreds of American municipalities in 1933, and its appearance is one of the more reliable signals that an official system has stopped functioning.
Case Study Four — Cigarettes, and the Economy That Assembled Itself
Background. In 1945 a young British economist named Richard Radford, recently released from a German prisoner-of-war camp, published a short article describing the economic organization of the camps he had lived in. It remains one of the most cited papers in the discipline, for the excellent reason that it documents a monetary system forming from nothing among people with no capacity to produce anything, which is as close to a laboratory as the subject ever gets.
The challenge. Prisoners received roughly equal rations from the camp authorities and roughly equal parcels from the Red Cross, containing tinned food, chocolate, sugar, jam, and cigarettes. The distribution was equal; the preferences were not. Some men did not smoke, some disliked tinned fish, Sikh prisoners would not eat beef, and everyone had different intensities of craving. There were gains from trade available immediately, and no currency in which to make them.
The solution. Trade began as barter and became monetary within weeks, without discussion or design. Cigarettes emerged as the medium of exchange because they had the right properties, which the camp discovered empirically: they were durable, portable, homogeneous enough to be counted rather than assessed, divisible into small units, and in constant demand from smokers, which gave them a floor of intrinsic value. Prices came to be quoted in cigarettes throughout the camp. A notice board appeared where offers were posted. A shop was eventually organized where goods could be left for sale at stated prices, with the seller paid in cigarettes on completion.
What followed is the part that startles readers who assume monetary phenomena require institutions. The camp developed a price system that converged across compounds; arbitrageurs walked between huts profiting from price differences, and the differences narrowed as they did so. Credit appeared, with goods sold for payment after the next parcel delivery. A chaplain was recorded circulating with a tin of cheese and five cigarettes, trading his way around the camp and returning with a substantially larger stock, having provided a genuine service by moving goods to those who valued them most.
It also developed everything else. There was inflation, when a large parcel delivery flooded the camp with cigarettes and prices rose across the board. There was deflation in the weeks before deliveries, when cigarettes were smoked and the money supply literally went up in smoke, causing prices to fall and trade to become sluggish. There was debasement: cigarettes were opened, some tobacco was removed, and they were rolled again, so that hand-rolled cigarettes drove machine-made ones out of circulation and into hoards — the good money vanishing exactly as it had in Athens. There was a currency crisis when the German authorities confiscated tinned goods, and a period when a paper currency was attempted, based on a food account, which failed when confidence in the underlying stock wavered.
The results. Radford’s central observation was that the economy was not planned by anyone, was not the product of any cultural tradition, and arose among men who mostly had no interest in economics. It appeared because a group of people with differing preferences and no coercive authority will trade, and trade will select a commodity to serve as the unit. He also noted, with some feeling, the moral reaction it provoked: prisoners who profited from arbitrage were resented, price rises were blamed on the traders rather than on the parcel schedule, and there were periodic demands for fixed prices, which produced shortages and black markets on the smallest scale imaginable.
Relevance today. Every self-contained community with restricted access to official currency reproduces this experiment. Cigarettes served the same role in occupied Germany after the war, in the Soviet camps, and in prisons everywhere until smoking bans arrived. The paper is also a useful corrective in both directions: it is cited by those who wish to argue that money emerges spontaneously from exchange, which it does under these very specific conditions — strangers, no shared future, no enforcement, goods arriving from outside — and it is quietly ignored by the same people when it demonstrates that such an economy immediately generates inflation, hoarding, resentment of intermediaries, and demands for price control. The monetary problems of a great power and of a hut containing forty men with tinned jam are, in structure, the same problems.
Radford added one observation that is rarely quoted and deserves to be. The camp economy’s activity depended heavily on morale and on the expectation of the future. During periods of good news about the war, trade was brisk and credit was extended freely; during bad periods, and particularly in the final chaotic months when the camps were being moved, the market thinned out, credit vanished, and men reverted to holding goods rather than cigarettes. Nothing about the supply of cigarettes had changed. What changed was the willingness to hold a claim rather than a thing, which is precisely the variable that determines the velocity of money in any economy and which no policy instrument reaches directly. A hut full of prisoners demonstrated, on a scale of a few hundred men, that confidence is not a lubricant applied to an economy from outside but a component of the machinery itself.
Case Study Five — Company Scrip and the Store That Owned You
Background. The coal towns of West Virginia, Kentucky, and Tennessee in the first decades of the twentieth century were frequently owned outright by the mining company: the houses, the school, the church, the doctor, the road, and the store. Miners were recruited from Appalachian farms and from Europe, arrived with nothing, and lived where they were housed. The company controlled every point of contact between the worker and the wider economy, and it also controlled the money.
The challenge, from the company’s perspective, was cash flow and labor retention. Coal was sold on terms; wages were due sooner. A workforce with cash could leave. And a company store facing competition from an independent merchant in the next valley made less money. From the miner’s perspective the challenge was different: work was irregular, the pay period was long, and a family needed food before payday.
The solution adopted across the industry was scrip — tokens or paper issued by the company, redeemable at the company store, advanced against wages not yet earned. The tokens were struck in denominations from one cent upward, in brass, zinc, or fiber, stamped with the company name and often with the word that made the arrangement legally defensible: nontransferable. A miner could draw scrip between paydays, and it functioned as an advance on earnings. At the store it was worth its face value. Anywhere else it was worth whatever a discounter would give, which was typically sixty to eighty cents on the dollar.
This was, in the vocabulary of the present day, programmable money. The instrument could be spent only in an approved location, on approved goods, by an approved person. Its issuance was tied to the recipient’s employment status. It could not be saved usefully, because it lost value the moment it left the company’s ecosystem, and it could not fund a departure, because it was worthless in the next county. Every property that is currently discussed as a novel capability of digital currency was implemented in Appalachia with stamped brass.
The results. Prices at company stores ran materially above those at independent merchants — studies of the period found differentials commonly in the range of five to twenty percent, and higher in isolated camps. Combined with rent deductions, doctor’s fees, and charges for tools and blasting powder, the effect was that a substantial number of miners received little or no cash on payday, and some received a statement showing they owed the company money. The system was not universally exploitative; some companies ran fair stores and used scrip mainly as a genuine convenience, and miners in areas with road access and competing merchants suffered less. But the design of the instrument determined how much abuse was possible, and where the geography permitted, it was extensive.
Resistance took decades and came through law rather than markets. Several states passed acts requiring wages to be paid in lawful money on demand, and companies litigated them, sometimes successfully, on the ground that scrip was an advance rather than a wage. The decisive changes came in the 1930s: federal labor legislation, the growth of the mine workers’ union, and eventually the Fair Labor Standards Act, which required minimum wages to be paid in cash or negotiable instrument. Roads and cars did the rest, since a company store cannot overcharge a customer who can drive somewhere else. Scrip survived in isolated places into the 1950s and is now a collectors’ market.
Relevance today. The historical episode is the reference case for the risks of restricted-purpose money, and it is worth keeping alongside the entirely legitimate uses of the same technique. Food assistance programs in many countries issue benefits that can be spent only on approved items, for defensible reasons, and produce recognizable side effects: recipients trade the restricted value at a discount for cash they need for rent, exactly as miners discounted their scrip. Employer-issued wage advance apps, which lend against earned but unpaid wages at fees that annualize alarmingly, have reproduced the timing problem scrip was invented to solve, and have attracted regulatory attention for the same reasons. The general principle from Appalachia is durable: the more conditions a payment instrument carries, the more the issuer of those conditions controls the recipient, and the more carefully somebody outside the arrangement needs to be watching.
A structurally identical arrangement existed far from the coalfields and is worth naming, because it shows the pattern is about power rather than industry. Agricultural labor in the American South and in parts of Latin America operated on advances against the coming crop, settled at a store owned by the landowner or merchant, with the accounts kept by the party who benefited from them and the worker frequently unable to check the arithmetic. In the Andes and the Amazon the equivalent system trapped rubber tappers in obligations that could not be worked off. In each case the instrument was credit rather than a token, and the effect was the same: a worker whose purchasing power exists only inside a system controlled by their employer is not, in any practical sense, being paid a wage. When considering any modern proposal that ties money to conditions, the useful question is not whether the conditions are reasonable but who may change them, and whether the person subject to them can go elsewhere.
Case Study Six — Six Months Without Banks: Ireland, 1970
Background. Irish banking in 1970 was concentrated in a handful of associated banks, heavily staffed, and dependent on the physical movement of paper. Industrial relations in the sector had been poor for years, with a strike in 1966 that closed the banks for twelve weeks. In the spring of 1970, following the collapse of a pay negotiation in a period of high inflation, bank officials went on strike again. The banks closed at the end of April. They did not reopen until the seventeenth of November, six and a half months later.
The challenge. The Republic of Ireland was a developed economy with a population of around three million, and it had just lost its payment system. Deposits were frozen, checks could not be cleared, and no new notes were being issued into circulation. The government had prepared to some extent: cash had been distributed in advance, and a small number of non-associated banks and the post office savings system remained open, but these accounted for a minor share of the system. Roughly eighty-five percent of the country’s deposits were inaccessible. Conventional expectation was that economic activity would contract severely.
The solution was improvised entirely by the public and it centered on the pub. Irish retail banking of the period was relationship-based to an unusual degree, but so was Irish retail commerce, and there were something like twelve thousand licensed premises serving a population of three million — one for every couple of hundred people. Publicans and shopkeepers knew their customers, knew their employers, knew their families, and knew who was good for a debt. They began accepting checks, which could not be cleared, and passing them on in payment to their own suppliers, who did the same. The checks became a circulating currency, backed by nothing but the judgment of the person accepting them about the person writing them.
The mechanics were remarkable. An undated, uncleared check written by a farmer to a publican in April might be endorsed onward to a wholesaler in June and to a distributor in August, accumulating signatures, and would finally be presented at a bank in November. Businesses kept informal ledgers of who had passed them what. Wages were paid partly in cash and partly in employer checks that employees cashed with local traders. Some firms issued their own paper. The economist Antoin Murphy, studying the episode afterward, estimated that around five billion pounds of undrawn checks were circulating by the end — a substantial multiple of the money supply that had disappeared.
The results. Economic activity did not collapse. Trade continued, output grew over the year, and the anticipated disaster did not arrive. When the banks reopened, the accumulated checks were cleared over a period of weeks, and the losses from bad paper were far smaller than anyone had expected — the estimates that survive suggest defaults amounted to a small fraction of the total, comparable to normal bad debt levels. The publicans, it turned out, had been better credit assessors of their own communities than a formal credit department would have been, because they possessed information no credit department could obtain.
There were costs, and they should not be minimized. Large transactions, international trade, and any dealing between strangers became difficult; businesses without local reputation suffered; hire purchase and new lending largely stopped, so investment was postponed; and the burden of running an informal clearing system fell on small retailers who had not volunteered for it. The system worked for six months and was plainly deteriorating toward the end, as the volume of outstanding paper grew and the quality of the credit assessment fell.
Relevance today. The episode is cited in two opposite directions and supports both. It demonstrates that money is fundamentally credit and that a community with sufficient mutual information can create a functioning payment system without any institution at all — which is a genuine finding and a striking one. It also demonstrates the strict limits of that arrangement: it worked because Ireland in 1970 was a small, dense, socially interconnected society in which the person accepting the check usually knew the person writing it. The identical experiment in a large, anonymous, urbanized economy would have failed within a fortnight. This is the same boundary that has run through every chapter of this book: personal trust scales to a village, and every monetary institution ever built is a device for extending it beyond one.
The Irish case has a coda that historians of the episode enjoy. The banks reopened in November and the country returned to normal within weeks, and the striking bank officials received a settlement broadly in line with what they had demanded. The clearing of the accumulated paper took until the following spring in some institutions. And when a further, shorter dispute closed the banks again in 1976, the same informal system reassembled itself immediately, with the participants now experienced at it. Institutions can be replaced faster than most people expect when a population has done it once before, which is a comfort in a crisis and a warning to any institution that assumes its customers have no alternative.
Case Study Seven — The Swiss Currency That Only Exists Between Businesses
Background. Switzerland in 1934 was in the depths of the Depression, and small businesses were failing for a reason that had nothing to do with their products: their customers had no money, their banks had stopped lending, and the chain of payments between them had seized. Two businessmen, Werner Zimmermann and Paul Enz, had been reading the work of a heterodox economist who argued that the fundamental problem was money being hoarded rather than circulating, and they decided to try something.
The challenge. A group of firms in a slump face a coordination trap. Each would happily buy from the others if it had sales, and each would happily sell if the others were buying. What is missing is not willingness or capacity but a medium in which the mutual exchange can be settled, since the national currency has been withdrawn into savings and bank credit is unavailable precisely when it is most needed. Any small business owner in a recession recognizes the situation: the order book is empty, the workshop is idle, and the neighboring firm is in the same position for the same reason.
Get Peter Kriger’s stories in your inbox
Join Medium for free to get updates from this writer.
The solution. Sixteen participants founded a cooperative called the Wirtschaftsring — economic circle — abbreviated to WIR, which conveniently also means “we” in German. Members opened accounts denominated in a unit nominally equal to the Swiss franc but not convertible into it. A member selling goods to another member was credited; the buyer was debited. Members could be granted credit lines, secured against property, allowing them to run negative balances and spend before they had earned. No interest was paid on positive balances — a deliberate design feature, borrowed from the theory that had inspired the founders, intended to discourage hoarding the unit rather than spending it.
The essential rule is that WIR can be spent only within the network and cannot be converted into francs. This sounds like a crippling restriction and is the source of the system’s durability. Because the unit is useless outside the circle, a member who receives it must spend it inside, which generates a sale for another member, who must do the same. The currency circulates through the network rather than leaking out of it. Most transactions are conducted partly in WIR and partly in francs, since suppliers outside the network must be paid in national money, and the mix is negotiated deal by deal.
The results. The cooperative acquired a banking license, became the WIR Bank, and has now operated continuously for over ninety years. At its broadest it counted around sixty thousand participating businesses, a substantial share of Swiss small and medium enterprises, with annual turnover in the unit running to the equivalent of one to two billion francs. It survived a legal challenge, an attempt by the founders to steer it toward broader monetary reform, and the eventual abandonment of the anti-hoarding features that had motivated its creation.
The most interesting finding about it emerged from academic work in the 2000s. Analysis of decades of data found that WIR turnover moves counter to the Swiss business cycle: it expands during recessions and contracts during booms. The explanation is straightforward. When bank credit is plentiful and orders are strong, a business has little reason to accept a restricted currency at the cost of negotiating its use. When credit tightens and the order book empties, a sale in WIR is better than no sale, and members reactivate their networks. The system functions as an automatic stabilizer, of modest size, operating without any policy decision by anyone.
It has real limits, and the honest account includes them. The unit trades at an informal discount when members try to unload it, since not all goods are available inside the network and some members accept it reluctantly. Participation requires effort: finding a supplier who takes WIR is more work than paying francs. The system has never grown beyond the small-business sector, and large firms have no use for it. And its counter-cyclical effect, while real, is small relative to the Swiss economy.
Relevance today. WIR is the most successful complementary currency ever built, and the reasons for its success are instructive for the many that have failed. It is business-to-business rather than consumer-facing, which means participants are motivated by revenue rather than by ideology and will use it whenever the arithmetic favors it. It is backed by real credit assessment and secured lending rather than by enthusiasm. It solves an identifiable problem — idle capacity and an empty order book — rather than attempting to reform the monetary system. And its non-convertibility, which appears to be a weakness, is exactly what keeps the value circulating inside the community it was designed to serve. Every subsequent local currency scheme that has tried to be convertible, consumer-facing, and morally motivated has struggled, and the contrast is not a coincidence.
Case Study Eight — Brazil Cures Hyperinflation With an Imaginary Currency
Background. Brazil in 1993 had inflation of roughly two thousand five hundred percent a year, and had been living with high inflation for two decades. The country had responded with a succession of stabilization plans — five of them in eight years, each involving a new currency, a price freeze, and in one case the confiscation of bank deposits. Each had failed within months. The population had learned to disbelieve any government announcement about money, which is a rational response to the evidence and also the principal obstacle to any solution.
The challenge. Brazilian inflation was sustained by indexation. Wages, rents, contracts, and prices were tied to past inflation and adjusted automatically at regular intervals, which meant last month’s inflation mechanically produced this month’s. The system had been built to make high inflation survivable, and it had succeeded so well that the economy could function at rates that would have destroyed other countries — supermarkets repriced daily, workers spent their salaries the day they were paid, and banks made enormous profits on the float. But indexation also meant that stopping inflation by conventional means required breaking every contract in the country simultaneously, and previous attempts to do this by decree had produced shortages, litigation, and collapse.
The solution, designed by a team of economists under Finance Minister Fernando Henrique Cardoso, was one of the most elegant pieces of monetary engineering in the historical record, and it worked by separating the functions of money that this book has kept insisting are separable. In March 1994 the government introduced a new unit called the Unidade Real de Valor — a unit of account only. No notes were printed and no coins were struck. Its value was set daily by the government, roughly in line with the dollar, and it was announced each morning in the newspapers and on television.
Businesses were required, and workers encouraged, to convert prices, wages, and contracts into this unit. A shirt that cost a fluctuating number of cruzeiros reais now cost a stable number of URVs; a salary was defined in URVs; a rent was quoted in URVs. Payment still happened in the old currency, at the day’s conversion rate, which continued to inflate. So for four months the country had a stable measuring stick and an inflating means of payment, and the entire population became accustomed to thinking in the stable unit — which is to say the indexation system was gradually re-anchored to something that did not move, without any price ever being frozen and without any contract being broken.
On the first of July 1994 the government simply issued notes denominated in the unit everyone was already using, renamed the real, at one to one. The old currency was withdrawn at the prevailing rate. There was no price freeze, no deposit confiscation, no shortage. The new money arrived as the physical form of a unit in which the country had been transacting for months, and prices in it were already stable because they had been quoted in it for months.
The results. Monthly inflation fell from around fifty percent in June to single digits within weeks, and annual inflation fell from four figures to roughly twenty percent by 1995 and to single digits thereafter. Real incomes for the poorest households rose sharply and immediately, because the poor had no access to the indexed financial instruments that had protected the middle class and had therefore been paying the full inflation tax. Cardoso was elected president on the strength of it. The plan was supported by fiscal tightening and by a period of high interest rates and an overvalued exchange rate, which produced its own difficulties and a currency crisis in 1999 — but the inflation did not return.
Relevance today. The Brazilian design is the clearest demonstration available that the unit of account is the thing that has to be stabilized, and that it can be stabilized separately from the medium of exchange. It also worked because it was voluntary and gradual in its first phase, which meant nobody had to be coerced and nobody could be blamed for a broken contract, and abrupt only at the moment when the new arrangement had already become the norm. Compare this with the coercive stabilizations described earlier in this book — price ceilings, forced tender laws, capital penalties for refusal — which addressed the symptom while leaving the expectation untouched. The Brazilian planners understood that they were not managing a quantity of money but a set of beliefs about what a price means, and that beliefs can be migrated if people are given somewhere to migrate to.
Case Study Nine — Thailand, 1997: Borrowing in a Currency You Cannot Print
Background. Thailand in the mid-1990s was the model developing economy: growth averaging close to nine percent for a decade, rapid industrialization, foreign investment arriving in volume, and a currency pegged to a basket dominated by the American dollar. The peg had held for years and was regarded domestically as a permanent feature of the landscape, which is a description that should always attract attention.
The challenge lay in a mismatch that was visible in the data and ignored in practice. Thai banks and finance companies borrowed abroad in dollars, at interest rates far lower than domestic ones, and lent domestically in baht — much of it into property development in Bangkok, where the supply of new office towers had raced ahead of any plausible demand. The government had helpfully created an offshore banking facility in 1993 that made foreign borrowing easier still. So long as the exchange rate held, the arrangement was extremely profitable: borrow cheaply in dollars, lend dearly in baht, keep the spread. If the exchange rate moved, every borrower’s debt would grow in local terms while their income did not.
By 1996 the warning signs were unambiguous. The current account deficit reached about eight percent of national output. Exports stalled, partly because the dollar was strengthening and the baht was tied to it, making Thai goods expensive against Chinese and Japanese competitors. Property vacancy rates climbed. In early 1997 the country’s largest finance company effectively failed, and the authorities arranged support rather than closure, which told the market that the problems were real and that the response would be denial.
The solution attempted. The Bank of Thailand defended the peg, and defended it with unusual determination and unusual opacity. Rather than simply selling dollars from its reserves, which would have been visible in the published numbers, it entered forward contracts committing to deliver dollars at a future date. The published reserve figures therefore remained reassuring while the actual committed position deteriorated. Something in the order of twenty-three billion dollars of the country’s reserves were committed in this way, a fact not known to the market or to most of the government. Speculators, who could see the trade deficit and the property market, attacked the currency repeatedly through the spring.
On the second of July 1997 the central bank abandoned the peg and allowed the baht to float. It fell immediately, and by the end of the year had lost roughly half its value against the dollar.
The results were brutal and rapid. Every Thai company holding dollar debt found that its obligation had doubled in local terms overnight, while its revenue had not. Finance companies collapsed by the dozen; fifty-six were closed permanently. Output contracted by over ten percent in 1998. Unemployment tripled. An assistance package of seventeen billion dollars arrived from the International Monetary Fund with conditions — fiscal tightening and high interest rates — that most economists now regard as having deepened the contraction, since the crisis was one of private balance sheets rather than public profligacy, and squeezing demand did nothing to repair a currency mismatch. The contagion spread within weeks to Indonesia, Malaysia, the Philippines, and then to South Korea, in what became the Asian financial crisis, and on to Russia and Brazil the following year.
Relevance today. Two lessons have proved durable. The first is that borrowing in a currency you cannot issue converts an exchange rate movement into a solvency event, and no amount of profitability in normal times compensates for it. This is now called original sin in the development literature, and the response of the affected countries was to accumulate enormous foreign exchange reserves as self-insurance, which they did through the following two decades on a scale that reshaped global capital flows. The second is that a fixed exchange rate is a promise, and a promise defended by hidden commitments is a promise that will be broken more violently for having been concealed. The forward book was the modern equivalent of a bank publishing full reserves while lending them out, which is a story this collection has already told about Amsterdam. The technology changes; the temptation to defend confidence by managing the disclosure rather than the underlying position does not.
Case Study Ten — The Parking Garage Stock Exchange
Background. Kuwait in the early 1980s was extremely rich and financially constrained. Oil revenues had produced enormous private wealth, but the official stock exchange listed only a small number of Kuwaiti companies and was tightly regulated after an earlier speculative episode. Kuwaitis could not easily invest abroad, and the supply of approved domestic assets was far smaller than the demand for them. Money with nowhere to go is a recognized hazard.
The challenge. An unofficial market therefore assembled itself in an air-conditioned multi-story car park in Kuwait City called the Souk Al-Manakh, named after the site where camel traders had once gathered. It traded shares of companies incorporated in other Gulf states — Bahrain, the United Arab Emirates — many of which had minimal operations and existed largely as vehicles for share issuance. It had no regulator, no clearing system, no disclosure requirements, and no restrictions on who could trade.
The mechanism that inflated it was a payment instrument rather than a security. Shares were bought with postdated checks: a buyer would hand over a check dated a year or more into the future, at a price substantially above the current cash price, the difference representing the premium for deferred payment. The seller could then use that check as collateral, or pass it on in payment for other shares. Since anyone with a checkbook could write one, and since nobody verified whether the account behind it held anything, the effective credit available to the market was unlimited. This was money creation by private individuals, unbacked, unregulated, and accelerating.
The scale became grotesque. By the summer of 1982 the outstanding postdated checks totaled around ninety-four billion dollars — a figure larger than the entire national output of Kuwait, in a market with roughly six thousand participants. A substantial fraction was attributable to a small number of individuals, the most celebrated being a passport office clerk who had written checks amounting to some fourteen billion dollars on a civil servant’s salary. Market capitalization on the unofficial exchange exceeded that of every stock market in the world except those of the United States and Japan, generated in a car park by a population smaller than that of a mid-sized city.
The collapse. In August 1982 the Ministry of Finance requested that outstanding checks be registered, and shortly afterward a check was presented and bounced. The reasoning that had sustained the market — that anyone’s check was good because prices would keep rising — reversed within days. Everyone attempted to present checks at once, and the aggregate obligations were revealed to exceed the aggregate assets of the country by a wide margin.
The solution attempted. The government established a clearing company to sort out the tangle, ordered all checks registered, and set up a tribunal. Untangling proved close to impossible, because the checks formed chains: A owed B, who owed C, who owed A, and unwinding required deciding which links were genuine. A compensation fund was created for small investors; larger participants were left to the courts. The dispute consumed the Kuwaiti financial and legal system for years. Most of the country’s banks were effectively insolvent, and all but one required state support; the government eventually purchased large volumes of bad debt from them. The affair was still being litigated when Iraq invaded in 1990, an event that overwhelmed everything else.
The results. Estimates suggest the state ultimately absorbed costs in the tens of billions of dollars. The unofficial market was closed permanently, and a properly regulated Kuwait Stock Exchange was established with disclosure requirements, settlement procedures, and limits on credit. The episode is now the standard regional example of what happens when speculation is financed by an instrument nobody is monitoring.
Relevance today. The single most transferable feature of the Souk Al-Manakh is that the bubble was created by a payments instrument rather than by an asset. Shares were merely what the checks were used to buy; if there had been no shares, the same credit would have inflated something else. Whenever a market allows participants to create their own settlement medium — postdated checks, margin extended by unregulated brokers, tokens minted against tokens, or the mutual credit arrangements that appeared in crypto lending before 2022 — the ceiling on prices is removed, because purchasing power is being manufactured at the point of sale. The second lesson concerns the chain structure of obligations. When failures happen in a market where everyone is simultaneously a creditor and a debtor to everyone else, the losses cannot be allocated by any simple rule, and the resolution takes years and generally ends with the state absorbing the residue. That is not a feature of Gulf finance in 1982; it is a feature of interlocking obligations, and it is the reason modern regulators are so preoccupied with central clearing.
Case Study Eleven — Cyprus, 2013: The Depositors Pay
Background. Cyprus is a small island economy whose banking sector had grown to roughly eight times its national output, built on attracting foreign deposits — substantially Russian — with high interest rates, low taxes, and few questions. The deposits were invested, among other places, in Greek government bonds, which offered attractive yields for reasons that were entirely visible.
The challenge. When Greek debt was restructured in 2012 and private holders took a large loss, the Cypriot banks absorbed a blow of several billion euros, against an economy of around eighteen billion. The two largest banks were insolvent. The government could not recapitalize them, because the sums required were comparable to the entire national output, and it could not print, because it used the euro. It applied to its European partners for assistance.
The solution attempted. What the eurozone authorities offered in March 2013 was ten billion euros, on condition that Cyprus raise around six billion itself. The initial proposal, agreed at a meeting that ran through the night, was a levy on all bank deposits in the country: roughly six and a half percent on balances below a hundred thousand euros and nine and a half percent above. This crossed a line that had been thought inviolable since the 1930s. Deposits below a hundred thousand euros were guaranteed by law throughout the European Union, and the proposal taxed them anyway, on the reasoning that a tax is not a default.
The reaction was immediate and severe. Banks were closed to prevent a run; cash machines were emptied; and the Cypriot parliament rejected the proposal outright, with not a single vote in favor. The wider damage was already done, because savers across southern Europe had been informed that an insured deposit was insured except when it was not.
The revised plan, agreed a week later, protected insured deposits and concentrated the losses on the uninsured. The second-largest bank was wound down, with its insured deposits transferred to the larger one and its uninsured deposits used to absorb losses. At the Bank of Cyprus, uninsured deposits above the hundred-thousand threshold were converted into equity at a rate that eventually took close to half of the balances. Depositors woke up holding shares in a bank they had not chosen to invest in, whose value they could not realize.
Capital controls were imposed within a currency union that had been designed on the principle that a euro in one member state is identical to a euro in another. Cash withdrawals were limited, transfers abroad required approval, and the export of banknotes was restricted at the airport. The controls remained, in progressively relaxed form, for two years. During that period a Cypriot euro and a German euro were not the same thing, because one could be moved and the other could not.
The results. Output fell by around ten percent over two years, unemployment rose above fifteen percent, and the offshore banking model that had sustained the island was substantially destroyed — which was, from the perspective of some of the parties involved, part of the point. Recovery was faster than expected, and the assistance program was exited in 2016. Depositors who had been converted into shareholders eventually recovered part of their value.
Relevance today. Cyprus established the template for what is now called bail-in, and it was written into European law shortly afterward: when a bank fails, shareholders, then bondholders, then uninsured depositors absorb losses before any public money is committed. This is defensible — the alternative is taxpayers paying for the mistakes of a bank’s creditors — and it changes the nature of a bank deposit in a way most people have not registered. Above the insured limit, a deposit is not safekeeping; it is an unsecured loan to a leveraged institution, ranking ahead of bondholders and behind almost nothing else. The episode also demonstrated that a shared currency does not guarantee fungibility: when a member state’s banks fail, capital controls fragment the union along national lines, and the market prices that fragmentation immediately. Anyone assessing the durability of a monetary union should watch for the moment when a euro, or a dollar, or a token, held in one place stops being freely exchangeable for the same unit held elsewhere.
Case Study Twelve — Sixty Euros a Day: Greece, 2015
Background. Greece had been in crisis since 2010, under successive assistance programs, with output down by a quarter and unemployment above twenty-five percent. In January 2015 a government was elected on a mandate to reject the terms of those programs. Negotiations with the European institutions ran through the spring and deteriorated through June, with each side convinced the other would concede.
The challenge. Greek depositors, watching this, did what depositors do. Deposits had been leaving the banking system for years, and the outflow accelerated as the possibility of an exit from the euro became a live topic — because a Greek exit would mean deposits converted into a new and immediately devalued currency, and a euro moved to a German or Cypriot account would not be. The Greek banks stayed liquid only through emergency liquidity assistance from the European Central Bank, extended at intervals and subject to the judgment of the same institution that was party to the negotiations.
On the twenty-sixth of June the Greek government announced a referendum on the creditors’ terms. Two days later the European Central Bank declined to increase the emergency assistance further. The banks had no more money to pay out.
The solution attempted. The government declared a bank holiday and imposed capital controls. Banks closed on the twenty-ninth of June and remained shut for three weeks. Cash withdrawals were limited to sixty euros per day per account, later averaged to a weekly figure because machines ran out of small notes. Transfers abroad were prohibited without approval; a committee vetted import payments; and Greeks abroad found their cards declined. The referendum went ahead on the fifth of July with the banks shut and produced a decisive rejection of the terms, after which the government accepted terms that were, by most assessments, harsher.
The results. The immediate economic damage was substantial: businesses could not pay foreign suppliers, imports of raw materials stalled, and firms reliant on international trade were paralyzed. The recovery that had begun in 2014 was reversed. The banks reopened in late July with withdrawal limits intact, and controls were relaxed in stages over four years, with the final restrictions lifted only in September 2019.
The unintended consequence is the interesting part. With cash rationed and transfers restricted, Greeks turned to cards and electronic payments in a country that had been one of the most cash-dependent in Europe. Card transactions multiplied several times over within a year, and the number of card terminals in Greek businesses rose dramatically. Since electronic transactions are visible to the tax authorities and cash transactions frequently were not, declared turnover in several sectors rose sharply, and value-added tax receipts improved in a way no enforcement campaign had achieved. A crisis measure imposed to stop a bank run accomplished, as a side effect, a substantial part of a tax compliance reform that governments had attempted unsuccessfully for decades.
Relevance today. The Greek episode illustrates two things worth carrying. First, in a monetary union without full banking union, a national banking system depends on a central bank that is not accountable to that nation, and the decision to extend or withhold liquidity is unavoidably political even when made on technical criteria. The negotiation was, in the end, settled by control of the banks rather than by argument.
Second, it demonstrates how quickly payment habits change under compulsion, and how durable the change proves. Greek card usage did not revert when controls were lifted. The same acceleration occurred worldwide during the pandemic, when cash handling was discouraged for hygiene reasons that turned out to be largely unnecessary, and the shift away from cash persisted afterward. Payment behavior appears to be sticky in normal times and highly plastic during disruption, which means the practical decisions about the future of cash are being taken during emergencies rather than in the deliberate policy debates that assume they have time.
Case Study Thirteen — Iceland Lets the Banks Fail
Background. Iceland has a population smaller than that of many mid-sized cities and, until 2003, had a banking system to match. Following privatization, three banks — Kaupthing, Landsbanki, and Glitnir — expanded abroad with extraordinary speed, funding themselves in international wholesale markets and then, when that funding tightened, by offering high-interest online savings accounts to depositors in Britain and the Netherlands. By 2008 the combined assets of the three exceeded ten times the national output of the country. This was not a large banking sector; it was a large international banking group with a small country attached.
The challenge. When wholesale funding froze in 2008, all three banks needed foreign currency they could not obtain, and the Icelandic central bank could not act as lender of last resort because the liabilities were in euros, pounds, and dollars, which it could not create. The choice was not between rescuing the banks and letting them fail. It was between failing to rescue them and bankrupting the state in the attempt. In early October the government took control of all three within a week.
The solution. Emergency legislation split each bank in two. A new domestic bank received the Icelandic deposits and the domestic loan book and continued operating, so that Icelanders could use their accounts on Monday morning; domestic deposits were guaranteed in full by government declaration. The old entity retained the foreign assets and — crucially — the foreign liabilities, and was placed into resolution. Deposit claims were given priority over other unsecured claims by law, retroactively. Bondholders, who had funded the expansion, were left with claims on the wreckage.
This provoked a diplomatic crisis. British and Dutch depositors in the online accounts were compensated by their own governments, which then demanded that Iceland repay the sums, amounting to several billion euros — a very large obligation relative to the size of the country. The British government invoked anti-terrorism legislation to freeze Landsbanki assets in the United Kingdom, which Icelanders regarded as an extraordinary insult and which is still remembered. Two agreements to repay were negotiated by the Icelandic government, and both were put to referendum after the president refused to sign them. Both were rejected by the electorate, the second by around sixty percent. The matter went to the EFTA Court, which ruled in 2013 that Iceland had not breached its obligations, since the deposit guarantee directive did not require a state to underwrite a systemic collapse.
Capital controls were imposed immediately and comprehensively, trapping foreign holders of króna assets inside the country and preventing the currency from collapsing entirely. It fell by roughly half against the euro regardless. Interest rates were raised sharply, then cut. The controls remained in place, in reducing form, until 2017 — nearly nine years — and were unwound partly by imposing a stability levy on the estates of the failed banks as the price of releasing their assets.
The results. The contraction was severe: output fell by around ten percent, unemployment rose from near zero to nine percent, and household debt indexed to inflation or denominated in foreign currency became crushing until the courts ruled many of the foreign-currency loans illegal. But the recovery was faster than in most of the countries that had chosen the alternative approach. The devalued currency made fishing and tourism competitive, and tourism grew from a modest sector into the largest export industry within a decade. Growth resumed in 2011.
Iceland also prosecuted. A special prosecutor’s office was established, and over twenty-five bankers, including the chief executives of the largest banks, were convicted of offenses including market manipulation and breach of fiduciary duty, receiving prison sentences. In no other country affected by the crisis did anything comparable occur.
Relevance today. Iceland is invoked constantly, and usually inaccurately, as proof that a country can simply refuse to rescue its banks. The accurate version is more interesting. Iceland did protect its domestic depositors in full and did rescue the domestic payment system; what it declined to do was assume the foreign liabilities of private institutions on behalf of taxpayers, and it was able to decline chiefly because those liabilities were so large that assuming them was impossible rather than merely unwise. The country also had its own currency, which could absorb the shock through devaluation — an option unavailable to Ireland and Greece, whose banking crises occurred within a monetary union.
The transferable principles are the separation of the payment system from the investment bank, which every subsequent resolution regime has adopted in some form; the priority of depositors over other creditors, now standard; and the observation that a banking sector far larger than the state supporting it is not a strength but an unfunded liability. Several jurisdictions today host banking or crypto-asset sectors of comparable relative scale, and the arithmetic that made rescue impossible in Reykjavik in 2008 has not changed since.
Case Study Fourteen — India Cancels Its Cash
Background. On the evening of the eighth of November 2016, in an unannounced television address, the Prime Minister of India informed the country that the five-hundred and one-thousand rupee notes would cease to be legal tender at midnight. Those two denominations accounted for approximately eighty-six percent of the value of currency in circulation, in an economy where the overwhelming majority of transactions by volume were conducted in cash. The measure had been prepared in extreme secrecy; most of the cabinet learned of it shortly before the broadcast.
The challenge as stated. The stated objectives were three: to destroy holdings of undeclared wealth, on the theory that hoards of illicit cash would be rendered worthless because their owners could not deposit them without explaining their origin; to invalidate counterfeit currency, some of it alleged to be produced abroad for the funding of terrorism; and to push the economy toward digital payments and the formal financial system. It was framed as a strike against corruption, and the framing was politically effective.
The implementation. Holders could deposit the old notes in bank accounts until the end of the year, or exchange limited amounts over the counter. Withdrawal limits were imposed. The difficulty was arithmetic: the country had to replace over eighty percent of its currency, in an economy of well over a billion people, using a printing capacity that had not been expanded in advance because expansion would have revealed the plan. Additionally, the new notes were of a different physical size, so the country’s two hundred thousand cash machines had to be recalibrated by hand, one at a time.
The result was months of shortage. Queues formed outside banks that ran out of cash by mid-morning; press accounts documented deaths among people waiting in line and among those unable to obtain cash for medical care. Daily-wage laborers, whose entire economy was cash, lost work. Farmers could not buy seed at the sowing season. Small traders reported sales collapsing. The informal economy, which employs the large majority of Indian workers, absorbed the shock directly, while the formal economy — salaried, banked, cashless — was inconvenienced rather than damaged.
The results. On the primary objective the outcome was unambiguous. The central bank reported that around ninety-nine percent of the demonetized currency by value was returned to the banking system. Whatever illicit wealth existed had either not been held in cash, or had been successfully laundered through the deposit process using intermediaries, family members, and shell arrangements. Some subsequent tax assessments followed from the deposit data, but the anticipated windfall of extinguished notes did not occur. Estimates of the effect on output vary; most credible assessments put the loss at one to two percentage points of growth in the affected quarters, concentrated in informal sectors.
On the third objective the effect was real and lasting. India had launched a unified payments interface earlier that year — a public infrastructure allowing instant transfers between any accounts, through any application, at no charge to the user. Adoption had been slow. The cash shortage supplied a reason to try it, merchants installed acceptance because customers had no alternative, and the habit persisted. Transaction volumes on the system have since grown into the billions per month and exceed the card networks of most countries combined. Whether this justified the cost is disputed; that it happened is not.
Relevance today. Demonetization is the clearest recent demonstration that the state can extinguish the value of its own currency by decree, instantly, and that the burden falls with brutal precision on the people furthest from the banking system. It also illustrates a repeated pattern in monetary reform: the stated objective failed, an unstated or secondary objective succeeded, and the episode is now remembered by its supporters for the success and by its critics for the failure. Several other countries have since undertaken high-denomination withdrawals for anti-crime reasons, generally with long notice periods and gradual phase-outs, which are the design choices India specifically avoided in order to prevent evasion — and it is worth noting that the evasion happened anyway, while the notice period would have saved the queues.
Case Study Fifteen — Nigeria Redesigns Its Money Before an Election
Background. Nigeria in late 2022 had roughly three and a quarter trillion naira in circulation, of which the central bank estimated that only a small fraction was actually held within the banking system; the rest sat outside it, in homes, in businesses, and in the informal economy that employs most Nigerians. The central bank announced in October that the two-hundred, five-hundred, and thousand-naira notes would be redesigned, and that the old versions would cease to be legal tender at the end of January 2023 — a window of about ten weeks. A presidential election was scheduled for February.
The stated objectives were to reduce the cash held outside the banking system, to combat counterfeiting and kidnapping-for-ransom payments, and to support the national digital currency, the eNaira, which had launched in 2021 and had achieved adoption in the low single digits despite considerable official encouragement. An unstated objective, widely assumed within Nigeria, was to disrupt the practice of distributing cash to voters, which is a well-documented feature of Nigerian elections and which requires large quantities of physical currency in the hands of political operatives.
The challenge. The new notes were not printed in sufficient volume before the deadline. Estimates suggest that of the roughly three trillion naira withdrawn, perhaps a third was replaced by the deadline. The result was a nationwide cash famine in an economy where cash is the dominant medium for the majority of the population. Bank branches ran dry. Cash machines dispensed nothing. Point-of-sale agents, who form the backbone of Nigerian retail finance, began charging premiums of twenty percent or more to supply cash, which is to say the currency traded at a discount against itself depending on its physical form.
The consequences on the ground were severe. Market traders could not buy stock. Transport operators, who deal in small cash fares, could not run. Bank branches were attacked in several states and some were burned. Protests turned violent in Lagos, Ibadan, Warri, and elsewhere. The electronic alternatives failed under load: the bank transfer systems that people turned to were overwhelmed, transactions failed while accounts were debited, and reversals took days.
The resolution. Several state governments sued the federal government at the Supreme Court, which in March 2023 ruled that the old notes should remain legal tender until the end of the year, an unusual judicial intervention in monetary policy. The central bank governor was subsequently suspended and prosecuted on unrelated charges. The election took place amid the shortage, with contested results.
The results. Electronic transaction volumes rose sharply and did not fully revert, so the policy did accelerate the shift away from cash, at a cost measured in closed businesses and lost income for people with no financial buffer. The eNaira gained little from the episode; Nigerians who abandoned cash mostly moved to commercial bank transfers and to the fintech applications that had grown rapidly in preceding years, not to the central bank’s own instrument. The proportion of currency held outside banks recovered within a year.
Relevance today. The Nigerian episode belongs beside the Indian one and adds a distinct lesson. Where India’s demonetization failed chiefly on its anti-corruption objective, Nigeria’s failed on logistics: the state withdrew a currency it had not manufactured the replacement for. This is a recurring and underappreciated risk in monetary policy, which is generally analyzed as though implementation were free. It also demonstrated the limits of pushing a state digital currency by making the alternative painful. Users deprived of cash will move to whatever electronic option they already trust, which in Nigeria meant commercial platforms; a central bank product that has not earned adoption on its merits does not acquire it by decree. Finally, the affair is a reminder that monetary instruments are political instruments. Whatever the true motive for the timing, a very large number of Nigerians concluded that their money had been withdrawn in order to affect an election, and that belief is itself a cost, paid in the currency this book has been discussing throughout.
Case Study Sixteen — Mackerel, Ramen, and the Prison Economy
Background. For most of the twentieth century the currency of American prisons was tobacco. Cigarettes were durable, divisible into individual sticks, universally desired, and constantly consumed, which gave the money supply a natural drain that prevented accumulation from getting out of hand. Then, between the late 1990s and 2004, smoking was banned across the federal system and in most state systems. The money supply was abolished overnight, in an economy of well over a million people.
The challenge. Prisoners are not permitted to hold cash. Institutional accounts exist, but transfers between them are monitored and restricted, and the entire informal economy — laundry services, cooked food, haircuts, artwork, protection, contraband, gambling — requires a medium of exchange the administration does not control. Something had to replace the cigarette, and the replacement had to be available for purchase at the commissary, storable in a cell, and acceptable to everyone.
The solution, which emerged without coordination across facilities, was food. In the federal system the winner was the foil-packed mackerel fillet, universally shortened to the mack, which cost approximately one dollar at the commissary and was thereby a natural unit. It had an advantage no other item possessed: hardly anyone wanted to eat it. A currency that is consumed disappears from circulation, and a currency nobody consumes stays in circulation, which meant mackerel could accumulate as savings while still being spendable. Prisoners reported holding hundreds of packets, stored in lockers, and journalists visiting facilities found men who had never eaten one in years of holding them.
Regional and institutional variation appeared immediately, in the way monetary systems always vary when they arise locally. Postage stamps served as small change and had the advantage of being light and easy to conceal; books of stamps functioned as larger denominations, typically trading below face value because their utility inside was limited. In many state systems the dominant currency became instant ramen, which sociological work published in 2016 documented in detail — and which had displaced tobacco not only because of the smoking ban but because declining food quality and quantity in cost-cutting facilities had made calories genuinely scarce, giving ramen an intrinsic value that mackerel lacked.
The results. These currencies exhibit every phenomenon described in this book. There is inflation when commissary policy changes, and prisoners report price adjustments across the informal economy within days of a change in what may be purchased or in what quantity. There is a foreign exchange market, since a mack is worth a different number of stamps in different facilities and men transferred between prisons carry the knowledge and the goods. There is Gresham dynamics: when the commissary introduced a mackerel product that was more palatable, the edible variety was consumed and the inedible variety remained as money. There is credit, at rates that reflect enforcement conditions rather than any assessment of risk in the ordinary sense. And there is the regular monetary shock of an administrative decision — a facility that restricts mackerel purchases has, without intending to, conducted a contractionary monetary policy on a population of a thousand people.
Relevance today. The prison case is the strongest available answer to anyone who believes money requires a state. These are populations under maximum coercive supervision, explicitly forbidden to hold currency, monitored continuously, and they generate a monetary system within months of losing the previous one, complete with denominations, exchange rates, savings behavior, and credit. It is also a reminder that the choice of monetary commodity is not arbitrary and follows recognizable rules: the winner is available at a fixed official price, uniform, storable, portable, hard to counterfeit, and — the criterion most often overlooked — not too useful, because a money that people want to consume keeps vanishing from the money supply. Gold satisfied that last criterion for three thousand years. So does a fish nobody wants to eat.
Case Study Seventeen — The Largest Currencies Nobody Regulates
Background. Airline loyalty programs began in 1981 as a marketing device: fly with us, accumulate points, get a free trip. Four decades later the accumulated unredeemed balances across the industry number in the trillions of points, several programs are worth more than the airlines that own them, and a substantial share of points are earned by people who have not been on an aircraft. The programs are, functionally, private currencies with tens of millions of holders and no supervision whatsoever.
The transformation happened through banking. Airlines discovered that they could sell points in bulk to credit card issuers, who distribute them to cardholders as a reward funded out of the interchange fee described earlier in this book. The airline receives cash today for a promise to provide a seat at some point in the future, at a cost it controls. This is money creation in its purest form: the issuer prints the unit at negligible marginal cost, sells it for real money, and carries the obligation as a liability valued at its own estimate of what redemption will cost.
The scale became publicly visible in 2020, when airlines grounded by the pandemic needed emergency financing and pledged their loyalty programs as collateral. Analysts valuing those programs for the debt offerings concluded that the programs of the largest American carriers were worth more than the entire market capitalization of the airlines themselves — in some cases twice as much. The airline, in the strict financial sense, was a low-margin appendage attached to a highly profitable currency-issuing business, and the aircraft existed principally to give the currency something to be redeemed for.
The monetary dynamics are entirely familiar. Programs devalue their currency by raising the number of points required for a given seat, which is inflation imposed unilaterally by the issuer with no notice and no compensation. Devaluations have been frequent and substantial; a balance accumulated over a decade can lose a large share of its purchasing power in a single announcement. Points typically expire after a period of inactivity, which is a demurrage charge of the kind attempted at Wörgl, imposed here not to stimulate circulation but to extinguish liabilities. The industry term for points that are never redeemed is breakage, and it is booked as profit.
There is even a case of successful monetary resistance. When the operator of a large Canadian loyalty program announced in 2011 that points would begin expiring at the end of 2016, holders mounted a public campaign as the deadline approached, media coverage intensified, a class action was filed, and the company reversed the policy days before the expiry date. Several million people successfully lobbied a private currency issuer to cancel a scheduled destruction of their savings, which is more direct monetary democracy than most citizens exercise over their national currency.
Relevance today. Loyalty points are worth attention for two reasons beyond their intrinsic strangeness. First, they demonstrate that the public will accept, save in, and make plans around a currency issued by a corporation with no reserve requirement, no redemption guarantee, no regulator, and an explicit contractual right to change the terms at any time — which suggests that the willingness to trust an issuer is considerably more elastic than monetary theory usually assumes. Second, they are the closest existing model for what large-scale corporate money might look like if a major platform ever issued a general-purpose unit. The features would be the same: issuance at will, unilateral devaluation, expiry, exclusion of individual holders at the issuer’s discretion, and no appeal. Every regulatory question raised by proposals for corporate digital currencies has been running, unexamined, in the loyalty industry for forty years, at a scale of trillions of units, and the answers available from that experience are not reassuring.
Case Study Eighteen — The Economy That Runs Inside a Video Game
Background. Massively multiplayer online games create economies by accident. Players acquire goods, need things other players have, and trade; a currency emerges from whatever the game issues; and within a few years the developer discovers it is operating a central bank without having applied for the job. The most instructive case is a science-fiction game launched in 2003 in which all players inhabit a single shared universe rather than separate copies, which means there is one economy rather than thousands, and it is large enough to study.
The challenge. The developer faced problems that would be recognizable to any monetary authority. Currency entered the system through rewards paid by the game itself — the equivalent of money printing — and left it through fees and destroyed assets. If issuance exceeded destruction, prices rose; if destruction exceeded issuance, the economy stagnated. Real-world traders began selling in-game currency for actual money on grey markets, which imported an exchange rate the developer did not control and created incentives for industrialized farming operations. Fraud flourished: the game permitted deception as a design principle, so scams, market manipulation, and outright theft were legal within its rules and devastating to their victims.
The solution. In 2007 the company hired a professional economist as its first in-house economic adviser, and subsequently published quarterly economic reports on the game, containing money supply figures, price indices, trade balances between regions, and analyses of production. It introduced a sanctioned instrument allowing players to buy a token with real money and sell it in-game for currency, which brought the grey market inside the system and gave the developer a lever over the exchange rate. It ran deliberate interventions: adjusting the rate at which currency entered the world, adding or removing sinks, and occasionally intervening in specific markets when the concentration of a resource threatened the wider game.
The results include some of the most spectacular financial events ever recorded outside real markets. A player-run bank collapsed in 2009 when an administrator withdrew a substantial sum of deposits, converted it to real money on the grey market, and used it to pay for a house deposit and medical bills; the bank failed, depositors lost everything, and the perpetrator was banned but committed no crime under any national law. An infiltration operation conducted over ten months by a group of players ended with the assassination of a corporate leader and the theft of assets worth, at prevailing grey-market rates, tens of thousands of dollars. A single battle in 2014, precipitated by an accounting error when somebody forgot to pay a rent bill, destroyed equipment valued at roughly three hundred thousand dollars.
More significant than the drama is what the environment permits: controlled experiment. Economists have used game data to study the effects of monetary expansion, the formation of trade routes, the emergence of financial intermediaries and insurance, and the behavior of markets when contracts are unenforceable. The developer can change one variable and observe the outcome across a population of hundreds of thousands, which no central bank can do. Similar work has been conducted on other games, including the well-known incident in 2005 when a programming error caused a plague to escape a confined area of a fantasy game and spread through populated cities, producing behavior — flight, quarantine evasion, deliberate infection, altruistic assistance — that epidemiologists later studied as a model of human response.
Relevance today. Three points transfer. First, monetary phenomena are not a product of statehood or of scarcity in the physical sense; they arise wherever there are participants with differing endowments and a transferable unit, including inside entertainment products designed by people with no interest in economics. Second, the developer’s trajectory — from ignoring the economy, to being surprised by it, to publishing statistics and hiring an economist, to actively managing issuance — is a compressed replay of the history in this book, and it took about a decade rather than five thousand years. Third, these environments are where a generation has learned its monetary intuitions. A person who has held a virtual currency, watched it inflate, traded it for real money, and lost a balance to a scam has an experiential understanding of several concepts in this book that most adults acquire only in a crisis, and it is not accidental that the enthusiasm for cryptographic currencies drew heavily on the same population.
Case Study Nineteen — The Bristol Pound and the Limits of Local Money
Background. Bristol is an English city of around half a million people with a strong tradition of independent retail and an active community sector. In September 2012 a group of local organizers, working with a credit union, launched a local currency intended to keep spending within the city: the Bristol Pound. It was not the first such scheme in Britain, but it was by a wide margin the most ambitious, and it received support that no predecessor had achieved.
The challenge. The argument for local currency rests on the observation that money spent at a chain retailer largely leaves the area immediately, whereas money spent at an independent business circulates locally for several rounds before departing — wages to local staff, purchases from local suppliers, rent to local landlords. The multiplier is real and has been measured. The difficulty is that no individual shopper has any incentive to act on it, since the benefit is diffuse and the cost of inconvenience is personal. A local currency attempts to solve this by making the local option a distinct instrument that cannot leak away.
The solution. Bristol Pounds were issued at parity with sterling, backed one-for-one by deposits held at a credit union, and available as paper notes with elaborate locally-designed artwork or as electronic balances usable by text message and online transfer — a technically impressive arrangement for 2012, predating most mobile payment applications in Britain. Around eight hundred businesses accepted them. Critically, the city council agreed to accept payment of business rates in the currency and to pay a portion of staff salaries in it on request; the mayor elected that year took his entire salary in Bristol Pounds and said so frequently, which supplied publicity no marketing budget could have bought.
The design included the essential asymmetry: individuals could convert sterling into Bristol Pounds freely, but converting back incurred a fee, and businesses were encouraged to spend rather than redeem. This is the same insight that keeps the Swiss business circle functioning — making the exit costly keeps the value circulating — applied to a consumer-facing scheme.
The results. At its height something over a million pounds of value circulated, several thousand individuals held accounts, and the scheme became internationally known, studied by academics and imitated in other cities. It also never reached the scale required to be self-sustaining. The organization depended on grants and on a small paid staff, transaction volumes were insufficient to cover costs from fees, and the paper notes — which were the emotional heart of the project and appeared in every photograph of it — were expensive to print, distribute, and secure while accounting for a shrinking share of activity. The scheme wound down its currency in 2020 and was succeeded by an organization focused on payment infrastructure and community finance rather than on issuing a separate unit.
The reasons for the ceiling are worth stating precisely, because they apply to nearly every consumer-facing local currency ever attempted. Participation demanded effort from users and offered them no financial advantage, only a moral one, and moral advantage sustains a committed minority rather than a market. Businesses accepted the currency but often struggled to spend it, since their largest costs — wholesale stock, rent to a national landlord, tax — could not be paid in it, so the currency piled up at the end of a short chain and was redeemed rather than circulated. And the whole scheme was denominated at parity with sterling, which meant it offered nothing sterling did not, except restriction.
Relevance today. The contrast with the Swiss case examined earlier is the instructive part, and it is a contrast of design rather than of enthusiasm. A business-to-business network where members can meet a meaningful share of their input costs inside the circle produces genuine circulation; a consumer scheme where the chain terminates at the first shopkeeper produces redemption. A currency with a commercial rationale survives recessions; a currency with an ethical rationale depends on a grant cycle. None of this diminishes what Bristol accomplished — it demonstrated that a small organization could build a working payment system with local governmental cooperation, which is a considerable achievement — but the lesson for anyone contemplating something similar is that the question to answer first is not who will accept the currency. It is who will spend it onward, and to whom.
Case Study Twenty — Replacing the Money of Three Hundred Million People
Background. On the first of January 2002, twelve countries with a combined population of over three hundred million simultaneously replaced their national currencies with a new one. The exchange rates had been fixed irrevocably three years earlier, and the euro had existed since 1999 as an accounting currency for electronic transactions, bond issues, and bank balances. What happened in 2002 was the physical part: the notes and coins.
The challenge was logistical on a scale never previously attempted in peacetime. Approximately fifteen billion banknotes and fifty-two billion coins had to be manufactured, distributed, and put into circulation, while roughly nine billion national banknotes and one hundred and seven billion national coins were withdrawn and destroyed. The coins alone weighed around two hundred and fifty thousand tonnes. Every cash machine, vending machine, parking meter, ticket machine, shop till, and bank branch in twelve countries had to be converted, and the majority of them within a matter of days.
The solution. Production began three years in advance across multiple printing works and mints. Distribution to banks began in September 2001 under the term frontloading, with retailers receiving starter kits of coins in December so that they could give change from the first morning. Security for the transport operations was extensive, and the volume of cash moving on European roads in late 2001 was the largest in history. A dual circulation period was arranged during which both old and new money were legal tender, with a planned duration of up to two months; in the event, most countries found that the changeover happened far faster than expected and shortened the period, with several concluding by the end of February.
The banknote design deserves a mention as a piece of political engineering. Any real bridge, building, or monument would have belonged to one member state, so the designs depict architectural styles by period — classical, romanesque, gothic, renaissance, baroque, iron and glass, modern — rendered as structures that do not exist anywhere. Nobody could object, because nobody was excluded. The postscript is a small delight: a Dutch designer in the town of Spijkenisse persuaded his municipality to build the bridges, and seven of the imaginary structures from the reverse of the notes were constructed in reinforced concrete across a canal in the years after 2011. Fictional money produced fictional bridges which then became real bridges, which is as neat a summary of this book’s argument as anyone could ask for.
The results. The operation succeeded technically to a degree that surprised even its planners. Within a fortnight the large majority of cash transactions in the euro area were being conducted in the new currency, disruption was limited, and the anticipated chaos did not appear.
The perceptual result was entirely different. Measured inflation attributable to the changeover was small — official statistics across the area put the effect in the range of a few tenths of a percentage point. The perception among the public was of substantial price rises, so widespread that Germans coined a word for the new currency that punned on the word for expensive. Both things were true. The categories where prices did rise were the frequently purchased ones — coffee, haircuts, restaurant meals, cinema tickets — where sellers rounded to convenient new figures, while the categories that dominate the price index, such as rent and durable goods, did not move. People experience inflation through the items they buy weekly, not through a weighted basket, and the gap between measured and perceived inflation opened in 2002 and in some countries never fully closed.
There was also a loss that economists did not model. A currency carries memory: a person knows what a loaf should cost, what a fair wage is, what a house was worth when they bought it, and those anchors are calibrated in the unit they grew up with. Converting them requires arithmetic that people performed for years afterward, and older citizens in several countries were still translating prices into lira, marks, or francs a decade later. Replacing a currency deletes a population’s accumulated price intuition, and rebuilding it takes a generation.
Relevance today. The euro changeover is the reference case for any large-scale monetary transition, including the one that is now under discussion in most developed countries as physical cash declines and digital alternatives are designed. It shows that the logistics, however daunting, are solvable with sufficient lead time and coordination. It shows that public perception of the transition follows a completely different logic from the measurements, and that a technically flawless operation can nonetheless leave a durable sense of grievance which political movements will later use. And it shows that the intangible costs — the loss of price memory, the erosion of intuition, the sense that something familiar was taken away without a vote — are real, are felt by the people least equipped to absorb them, and do not appear in any of the impact assessments.
Glossary
Agent network — The shopkeepers and small businesses who exchange cash for electronic value in mobile money systems.
Agio — The premium at which one form of money trades over another. In seventeenth-century Amsterdam, the amount by which certified bank money exceeded ordinary circulating coin.
Amole — Bars of salt cut to a standard size and used as money in Ethiopia and the Horn of Africa into the twentieth century.
Andurarum — The Akkadian term for a royal proclamation cancelling debts and freeing those held in debt bondage.
Annona — The Roman system of supplying grain and other goods, used increasingly to pay soldiers and officials in kind when the coinage lost credibility.
Antoninianus — A Roman coin introduced in 215 CE, valued at two denarii but containing only about one and a half denarii of silver; the first clearly documented overvalued coin.
Arbitrage — Profiting from a price difference for the same thing in two places, which tends to eliminate the difference.
Assay — A test of the purity of precious metal.
Assignat — Paper money issued during the French Revolution, originally a claim on confiscated church lands, which lost nearly all its value by 1796.
Automatic stabilizer — Any mechanism that expands during a downturn and contracts during a boom without a policy decision being taken.
Bail-in — Resolving a failing bank by imposing losses on its shareholders, bondholders, and uninsured depositors rather than on taxpayers.
Bailout — Rescuing a failing institution with public money, usually to prevent its failure from spreading to others.
Balance of payments — The record of a country’s transactions with the rest of the world.
Bancor — The international unit of account proposed by John Maynard Keynes at Bretton Woods in 1944 and rejected in favor of an arrangement centered on the dollar.
Bank money — A balance held on the books of a bank rather than as physical currency; the form in which most money exists today.
Bank note reporter — A nineteenth-century American periodical listing the current discount on notes issued by thousands of different banks.
Bank run — A rush by depositors to withdraw their money, triggered by the fear that others will withdraw first.
Barter — Direct exchange of goods for goods without money. Common between strangers and after a currency collapse, rarely the everyday basis of any known society.
Basel Committee — The body of banking supervisors formed in 1974 to set international standards after the Herstatt failure.
Bearer instrument — A document whose value belongs to whoever physically holds it, with no name attached.
Bevel-rim bowl — A crude, standardized clay vessel mass-produced in ancient Mesopotamia, probably used to issue fixed rations.
Bill of exchange — A written order instructing a party in another city to pay a stated sum at a stated date; the backbone of international trade for six centuries.
Bimetallism — A monetary system using both gold and silver as standard money at an official ratio between them.
Bitcoin — A digital currency introduced in 2009 that maintains a shared public record of ownership without any central issuer, secured by competitive computation.
Blockchain — A record kept as a chain of batched entries, each cryptographically linked to the previous one, making later alteration detectable.
Bond — A tradable certificate of debt, promising stated payments to whoever holds it.
Breakage — Loyalty points, gift card balances, or similar obligations that are never redeemed and are therefore recorded as profit by the issuer.
Bretton Woods system — The postwar international arrangement, agreed in 1944, under which the dollar was fixed to gold and other currencies were fixed to the dollar.
Bulla — A hollow clay ball used in the ancient Near East to seal counting tokens inside a tamper-proof container; a probable ancestor of writing.
Capital controls — Legal restrictions on moving money into or out of a country.
Cash — Any form of payment that settles a transaction immediately and leaves no further obligation.
Cementation — The ancient refining process that separated gold from silver in electrum, used at Sardis under Croesus.
Central bank — The institution responsible for issuing a national currency, managing its value, and supporting the banking system in a crisis.
Central bank digital currency — Electronic money issued directly by a central bank and held by the public, as opposed to a claim on a commercial bank.
Chao — The paper currency of the Mongol Yuan dynasty, not convertible into metal, used across the empire in the thirteenth century.
Check — A written instruction to a bank to pay a stated sum from an account. The word descends from the Arabic sakk.
Clearing — The process of matching and offsetting mutual obligations so that only the net difference has to be settled.
Clearinghouse certificate — An emergency instrument issued by an association of banks so that members could settle with each other during a panic.
Clipping — Shaving small amounts of metal from the edges of coins, leaving them underweight while still passing at face value.
Coinage — Metal formed into standardized pieces and stamped by an authority that guarantees their weight and purity.
Collateral — An asset pledged to a lender, which may be seized if the loan is not repaid.
Commenda — A medieval Italian partnership in which one party supplied capital and the other conducted the trade, sharing profits by agreement.
Commodity money — Money made of something with value in its own right, such as grain, salt, cattle, or silver.
Company scrip — Tokens or paper issued by an employer, redeemable at the employer’s store and generally worth less anywhere else.
Contagion — The spread of financial distress from one institution or country to others.
Continental currency — Paper money issued by the American Continental Congress during the Revolution, which lost nearly all of its value.
Convertibility — A promise by an issuer to exchange its money for something else, usually a fixed weight of metal, on demand.
Corralito — The Argentine restriction of 2001 limiting bank withdrawals to a small weekly amount. The word means little fence.
Counterfeiting — Producing imitation money and passing it as genuine.
Cowrie — A small, glossy sea shell used as money across South Asia and West Africa for centuries; the closest thing to a global currency before coinage spread.
Credit — An obligation to pay in the future, especially one that can be transferred to another party.
Credit score — A number calculated from a person’s borrowing record and used by lenders to decide whether and at what price to lend.
Cross of gold — The image used by William Jennings Bryan in 1896 to attack the gold standard as a burden imposed on ordinary debtors.
Currency — The official money of a country in its circulating form, whether physical or electronic.
Debasement — Reducing the precious metal content of a coin while keeping its face value unchanged.
Debt deflation — The process by which falling prices increase the real burden of existing debts, causing defaults that push prices down further.
Debt jubilee — A general cancellation of debts declared by an authority, practiced repeatedly in ancient Mesopotamia.
Default — Failure to meet an obligation to pay.
Deflation — A sustained fall in the general level of prices, which increases the real value of money and of existing debts.
Demonetization — Officially withdrawing the legal tender status of a currency or of particular denominations.
Demurrage — A charge for holding money, designed to discourage hoarding and encourage spending.
Denarius — The standard Roman silver coin, introduced around 211 BCE and progressively debased over the following centuries.
Deposit insurance — A government guarantee that bank deposits up to a stated limit will be repaid even if the bank fails.
Devaluation — An official reduction in the value of a currency against other currencies or against a fixed standard.
Discount — The amount by which an instrument trades below its face value, reflecting doubt, distance, or the time until payment.
Dollarization — The abandonment of a national currency by its own population in favor of a foreign one, usually the dollar.
Double coincidence of wants — The requirement, under barter, that each party happens to want exactly what the other has to offer.
Double-entry bookkeeping — A method in which every transaction is recorded twice, as a debit and a credit, so that errors reveal themselves.
Drachma — An ancient Greek silver coin whose name derives from a word meaning a handful, recalling the iron spits used as money before coinage.
Dry exchange — A disguised loan structured as two currency exchanges that cancel each other out, used to charge interest where interest was forbidden.
Dual circulation — A transition period during which an old and a new currency are both legal tender.
Electrum — A natural alloy of gold and silver, of variable composition, used for the earliest coins in Lydia.
Eurodollar — A dollar deposit held at a bank outside the United States, created outside American jurisdiction and regulation.
Exchange rate — The price of one currency in terms of another.
Exorbitant privilege — The advantage enjoyed by a country whose currency is used internationally, allowing it to settle debts in money it issues.
Fiat money — Money with no commodity backing, whose value rests on confidence in the issuer and on its acceptance in payment of taxes.
Fiduciary — Acting on behalf of another and obliged to put their interests first.
Financial Revolution — The cluster of English innovations after 1688 including the funded national debt, the Bank of England, and a market in transferable securities.
Fineness — The proportion of precious metal in a coin or bar.
Fractional reserve banking — The practice of holding only a portion of deposits in reserve and lending the rest.
Free banking — A system in which any institution meeting general legal conditions may issue notes, without a specific charter for each.
Frontloading — Distributing new currency to banks and retailers before its official introduction, as was done before the euro changeover.
Funded debt — Government borrowing with no repayment date, serviced by interest payments from designated revenues and traded between holders.
Giro — A system of transferring value directly between accounts rather than by transferring cash or a written instrument.
Gold pool — The arrangement in the 1960s by which eight central banks sold gold jointly to hold its market price at the official level.
Gold standard — A system in which a currency is defined as a fixed weight of gold and is convertible into it.
Gresham’s law — The observation that when two moneys circulate at the same official value, people spend the inferior one and hoard the better one.
Hawala — A remittance system in which agents in different places pay out to each other’s customers on trust, settling their mutual balances later.
Herstatt risk — The danger that one side of a currency trade pays while the other fails before paying, because the two settlements occur in different time zones.
Huizi — A paper currency of the Southern Song dynasty, over-issued during wartime and severely depreciated by the thirteenth century.
Hyperinflation — Extremely rapid inflation, conventionally defined as prices rising more than fifty percent in a month.
Indexation — Automatically adjusting wages, rents, or contracts in line with past inflation, which tends to make inflation self-sustaining.
Inflation — A sustained rise in the general level of prices, reducing the purchasing power of money.
Inflation targeting — A policy framework in which a central bank publicly commits to keeping inflation near a stated figure.
Insolvency — The condition of owing more than one owns, as distinct from merely lacking cash on hand.
Interchange fee — The charge paid by a merchant’s bank to the card issuer on each card transaction, ultimately embedded in retail prices.
Interest — A payment for the use of money over time, expressed as a proportion of the amount borrowed.
Jiaozi — Printed deposit receipts issued by merchants in eleventh-century Sichuan, the first paper money.
Karshapana — An early Indian silver coin marked with multiple punches, often applied by successive authorities.
Koku — A Japanese measure of rice, notionally a year’s supply for one adult, used to express wealth and pay stipends.
Kula ring — A ceremonial exchange of shell valuables circulating between Trobriand islands, sustaining relationships within which practical trade took place.
Legal tender — Money that a creditor is legally obliged to accept in settlement of a debt.
Legal tender laws — Rules specifying which money must be accepted in settlement of debts.
Lender of last resort — A central bank acting to supply liquidity to solvent institutions during a panic, so that a shortage of cash does not become a wave of failures.
Liquidity — The ease with which an asset can be turned into money without a loss in value.
Mackerel — A foil-packed fish widely used as currency in American prisons after tobacco was banned, valued largely because few prisoners wanted to eat it.
Mandat territorial — The short-lived French currency issued in 1796 to replace the assignat, which lost its value within months.
Maximum — The revolutionary French law imposing legal ceilings on prices, which produced shortages and black markets.
Medium of exchange — The role of money as the thing handed over in a transaction.
Mita — The forced labor draft that supplied workers to the silver mines at Potosí, adapted from an earlier Inca obligation.
Monetary base — Currency in circulation plus the reserves that commercial banks hold at the central bank.
Money market fund — A pooled investment holding short-term instruments, widely treated as equivalent to cash until one failed to maintain its value in 2008.
Money supply — The total quantity of money in an economy, measured by several different definitions depending on what is counted.
Moral hazard — The tendency to take greater risks when protected from the consequences.
Mudaraba — An Islamic finance partnership in which one party provides capital and the other provides labor, sharing profits and allocating losses to the capital.
Multilateral netting — Offsetting many mutual obligations against one another so that only small net balances need to be settled.
National debt — The total amount owed by a government to its creditors.
Notgeld — Emergency money issued by German towns, firms, and institutions during periods of currency shortage.
Numismatics — The study and collection of coins and other money.
Offshore — Located outside the jurisdiction whose currency or customers are involved.
Open market operation — A central bank’s purchase or sale of securities to adjust the quantity of money in the system.
Original sin — The situation of a country that can borrow internationally only in foreign currency, leaving it exposed to exchange rate movements.
Panic — A sudden and general loss of confidence producing simultaneous attempts to withdraw money or sell assets.
Paper money — Currency printed on paper, whether convertible into metal or not.
Parity — An official one-to-one relationship between two units of value.
Peg — A commitment to hold a currency at a fixed rate against another currency or a commodity.
Piece of eight — The Spanish eight-real silver coin, minted from American silver and used worldwide from the sixteenth century onward.
Plate money — Large rectangular sheets of copper used as currency in seventeenth-century Sweden, the largest weighing around twenty kilograms.
Price revolution — The sustained rise in European prices during the sixteenth century, associated with the inflow of American silver.
Programmable money — Electronic money whose use can be restricted by rules attached to it, such as expiry dates or approved categories of spending.
Proof of work — A method of securing a shared record by requiring participants to perform costly computation in order to add entries.
Purchasing power — What a unit of money will actually buy.
Quantitative easing — Large-scale purchases of bonds by a central bank, used to support the economy when interest rates cannot be lowered further.
Quantity theory of money — The proposition that, other things equal, more money in circulation produces higher prices.
Rai — The large carved limestone disks used as high-value money on the island of Yap, whose ownership changed without the stones being moved.
Real Plan — The Brazilian stabilization of 1994, which ended hyperinflation by introducing a stable unit of account before introducing a new currency.
Recoinage — Calling in and reissuing a currency, usually to restore weight and fineness after clipping and wear.
Redenomination — Reissuing a currency at a new scale, typically by removing zeros, without changing its real value.
Rentenmark — The German currency introduced in November 1923 which ended the hyperinflation, backed nominally by land and effectively by a strict limit on issue.
Repurchase market — The market in which institutions borrow short term by selling securities and agreeing to buy them back, used heavily to fund modern banks.
Reserve currency — A currency that governments and institutions hold in quantity for international settlement and as a store of value.
Reserve requirement — A rule obliging banks to hold a stated proportion of deposits rather than lending them out.
Reserves — Assets held by a bank or central bank against its obligations, or a country’s holdings of foreign currency and gold.
Resolution — The orderly winding down of a failed financial institution, allocating losses according to a legal ranking.
Riba — The increase on a loan prohibited in Islamic law, generally understood as interest.
Sakk — A written payment order used in the medieval Islamic world; the origin of the word check.
Scrip — Any substitute currency issued by an employer, municipality, or private body in place of official money.
Securitization — Pooling many loans and selling claims on the combined stream of payments to investors.
Seigniorage — The profit from issuing money, being the difference between its face value and the cost of producing it.
Settlement — The final transfer that completes a payment and discharges the obligation.
Shadow banking — Institutions performing bank-like functions outside the regulated banking perimeter.
Shekel — A Mesopotamian unit of weight in silver, used to price and record obligations long before coins existed.
Silent trade — Exchange between parties who do not meet, each leaving goods and withdrawing until the other responds.
Single Whip — The sixteenth-century Ming tax reform that consolidated many obligations into a single payment in silver.
Solidus — The Roman gold coin introduced by Constantine, which held its standard for roughly seven hundred years.
Sovereign default — A government’s failure to pay its debts.
Specie — Money in the form of coin, as distinct from paper claims.
Stablecoin — A digital token designed to hold a fixed value against a national currency, typically by holding reserves against it.
Stock — Originally the longer half of a split tally stick held by a creditor; the source of the modern word for a financial claim.
Store of value — The role of money as a way of carrying purchasing power into the future.
Suftaja — A medieval Islamic instrument obtained from a banker in one city and payable by an associate in another.
Suspension of convertibility — An official decision to stop exchanging currency for the metal it nominally represents.
Swap line — An arrangement under which one central bank lends its currency to another, used to supply dollars to foreign banking systems in a crisis.
Tally stick — A notched wooden stick, split lengthwise between the parties, used to record payments to the English Exchequer for seven centuries.
Tetradrachm — The Athenian four-drachma silver coin bearing an owl, minted to an unchanged standard for roughly two centuries.
Token — A physical or digital object standing for a claim, whose value comes from what it represents rather than what it is.
Touchstone — A dark stone against which precious metal is rubbed, allowing its purity to be judged by comparing streaks.
Trust company — An American institution that took deposits and made loans under lighter rules than banks, and which failed first in the panic of 1907.
Unit of account — The role of money as the scale in which prices, debts, and wealth are measured.
Usury — Lending at interest, particularly at rates considered excessive; prohibited outright under medieval canon law and classical Islamic law.
Velocity — How rapidly money changes hands; it rises when people are eager to spend and collapses when they prefer to hold.
Wampum — Beads drilled from shells, used as records and ceremonial valuables by northeastern American peoples and adopted as currency by colonists.
Wildcat bank — A bank that satisfied the letter of a free banking law while placing its redemption office somewhere too remote to be reached.
Wind trade — The Dutch term for trading contracts on tulip bulbs still in the ground, since nothing physical changed hands.
WIR — The Swiss cooperative currency founded in 1934, usable only between member businesses and not convertible into francs.
Zero lower bound — The point at which interest rates cannot usefully be reduced further, requiring other tools to support the economy.
Timeline
c. 8000 BCE — Small clay tokens in standardized shapes are used across the Near East to count goods, the earliest known accounting devices.
c. 3400 BCE — Tokens begin to be sealed inside hollow clay balls with their shapes impressed on the outside, producing a tamper-evident record.
c. 3200 BCE — Cuneiform writing appears in southern Mesopotamia, overwhelmingly for accounting; the shekel of silver serves as a unit of account.
c. 2400 BCE — The earliest recorded debt cancellation, by Enmetena of Lagash, frees those held for unpaid obligations.
c. 1900 BCE — Assyrian merchants at Kanesh in Anatolia run long-distance trade on written partnership contracts and credit.
c. 1750 BCE — The Code of Hammurabi caps interest rates, fixes wages and prices, and voids debts after crop failure.
c. 1200 BCE — Cowrie shells circulate as money in China; the shell sign enters the writing system as an element meaning wealth.
c. 700 BCE — Iron spits circulate as money in Greece, leaving their name in the later silver drachma.
c. 620 BCE — Lydia strikes the first coins, of natural electrum, stamped with a royal mark that guarantees weight.
c. 550 BCE — Croesus refines electrum into pure gold and silver at Sardis, issuing the first bimetallic coinage.
c. 550 BCE — Punch-marked silver coins appear in northern India; cast bronze spade and knife money circulates in China.
483 BCE — A silver strike at Laurion funds two hundred Athenian warships at the urging of Themistocles.
c. 450 BCE — The Athenian owl tetradrachm becomes the reference currency of the eastern Mediterranean.
407 BCE — Wartime Athens issues gold and plated bronze emergency coinage; Aristophanes describes good money disappearing from circulation.
375 BCE — Athens appoints a public coin tester in the marketplace and restores its silver standard.
c. 350 BCE — Aristotle argues that money is barren and cannot properly breed, shaping two millennia of argument about interest.
211 BCE — Rome introduces the denarius during the war against Hannibal.
64 CE — Nero reduces the silver content of the denarius, beginning two centuries of debasement.
215 CE — Caracalla issues the antoninianus, valued at two denarii and containing about one and a half.
271 CE — Mint workers revolt in Rome against Aurelian’s coinage reform; the rising is suppressed by the army.
301 CE — Diocletian’s Edict on Maximum Prices sets legal ceilings for over a thousand goods, with death for violation. It fails.
c. 309 CE — Constantine introduces the gold solidus, which will hold its standard for roughly seven hundred years.
c. 650 — Islamic prohibitions on riba shape a financial culture built on partnership rather than interest.
c. 900 — Abbasid Baghdad supports professional bankers, deposit accounts, and written payment orders called sakk.
c. 1000 — Merchants in Chengdu issue printed deposit receipts, jiaozi, against heavy iron coin.
1024 — The Song government takes over paper money issue, establishing a state bureau with reserves and fixed terms.
1099 — The Knights Templar are founded; their network will later hold deposits and issue travel credit.
c. 1150 — Bills of exchange come into regular use among Italian merchants; the Champagne fairs clear obligations by netting.
c. 1160 — The English Exchequer standardizes tally sticks as its record of receipts.
1260 — Kublai Khan issues the chao, a paper currency not convertible into metal, across the Mongol empire.
1294 — The Ilkhanate imports Chinese paper money to Tabriz. Markets shut within days and the experiment is abandoned in about two months.
c. 1298 — Marco Polo describes money made from the bark of the mulberry, and European readers refuse to believe it.
1307 — Philip IV of France arrests the Templars, his principal creditors, and dissolves the Order.
1345 — The Bardi and Peruzzi banks fail after Edward III of England defaults.
1375 — The Ming issue the Great Ming Precious Note, refuse it for most taxes, and watch it collapse.
c. 1397 — The Medici Bank is founded on a structure of legally separate partnerships.
1454 — Portuguese ships begin carrying cowries into West Africa, joining an Indian Ocean shell trade to an Atlantic one.
1494 — Luca Pacioli publishes the first printed description of double-entry bookkeeping.
1545 — Silver is found at Cerro Rico above Potosí.
1568 — Jean Bodin attributes European price rises to the inflow of American bullion, stating the quantity theory.
1571 — The founding of Manila opens the galleon route carrying American silver directly to China.
1581 — The Ming Single Whip reform consolidates taxes into a single payment in silver.
1609 — The Bank of Amsterdam is founded, creating a certified unit of account that trades at a premium over coin.
1637 — The Dutch tulip contract market collapses in February; the damage is far smaller than legend records.
1656 — Palmstruch founds Stockholms Banco; in 1661 it issues Europe’s first true banknotes.
1668 — Sweden reconstitutes the failed bank under parliamentary control; it becomes the world’s oldest surviving central bank.
1672 — The Stop of the Exchequer suspends payment on English government tallies, ruining several goldsmith bankers.
1694 — The Bank of England is founded; its loan to the Crown is never to be repaid, creating the funded national debt.
1696 — The Great Recoinage proceeds on Locke’s principle rather than Lowndes’s, producing severe deflation. Newton joins the Mint.
1717 — Newton sets a gold-silver ratio that puts Britain on gold by accident.
1716 — John Law founds his bank in Paris; the Mississippi Company scheme follows.
1720 — The Mississippi and South Sea bubbles collapse; the word millionaire enters use in Paris.
1775 — The Continental Congress begins issuing paper dollars; the British counterfeit them from a ship off New York.
1789 — France issues assignats against confiscated church lands.
1793 — The Law of the Maximum imposes price ceilings; refusal of assignats becomes a criminal offense.
1796 — The assignat plates are broken publicly in Paris; the currency is worth well under one percent of face value.
1797 — The Bank Restriction suspends gold payments in Britain; it lasts twenty-four years. Gillray coins the Old Lady of Threadneedle Street.
1803 — Napoleon defines the franc as a fixed weight of silver, a definition that holds for over a century.
1821 — Britain formally adopts the gold standard.
1826 — English tally sticks are abolished after seven centuries of use.
1834 — Burning obsolete tally sticks sets fire to and destroys the Palace of Westminster.
1837 — Michigan’s free banking law produces a crop of wildcat banks; note reporters and counterfeit detectors flourish.
1844 — The Suffolk Bank system holds New England notes near par by relentless redemption.
1863 — The National Banking Acts create uniform federal notes; a tax in 1865 ends state bank note issue.
1871 — Germany converts to gold with the French indemnity, pushing the world onto the gold standard.
1873 — The United States drops silver from its coinage in an act later denounced as the Crime of 1873.
1873 — Walter Bagehot publishes Lombard Street, setting out the rule for lending in a crisis.
1896 — William Jennings Bryan delivers the Cross of Gold speech and loses the election.
1907 — The Panic of 1907 is halted by J. P. Morgan operating from his library; clearinghouse certificates circulate as emergency money.
1913 — The Federal Reserve Act creates an American central bank with a deliberately decentralized structure.
1914 — Convertibility is suspended across the belligerent powers within weeks of the outbreak of war.
1923 — German hyperinflation peaks in the autumn; the Rentenmark ends it in November.
1925 — Britain returns to gold at the pre-war parity; Keynes publishes The Economic Consequences of Mr. Churchill.
1926 — The General Strike follows wage cuts in the coal industry.
1929 — The Wall Street crash begins in October.
1931 — The Creditanstalt fails in May; Britain leaves gold in September; the crisis becomes global.
1932 — A local stamped currency at Wörgl revives the town’s economy and is suppressed by the Austrian central bank.
1933 — Roosevelt closes the banks, explains banking on the radio, and reopens them; deposit insurance follows. Gold is called in.
1934 — The Swiss WIR cooperative is founded; American gold is revalued from twenty dollars sixty-seven cents to thirty-five dollars.
1944 — The Bretton Woods conference fixes the dollar to gold and other currencies to the dollar.
1945 — Radford publishes his account of the prisoner-of-war camp economy.
1946 — Hungary records the highest inflation ever measured, with prices doubling roughly every fifteen hours.
1950 — Diners Club issues the first charge card.
1957 — British restrictions on sterling financing push London banks into dollars, founding the eurodollar market.
1958 — Bank of America mails sixty thousand unsolicited credit cards to Fresno.
1959 — Robert Triffin tells Congress that the Bretton Woods system contains a contradiction it cannot survive.
1968 — The London Gold Pool collapses after a run on the official price.
1970 — Irish banks close for six and a half months; the country transacts on uncleared checks.
1971 — Nixon suspends the convertibility of the dollar into gold on the fifteenth of August.
1973 — The major currencies begin to float against one another.
1974 — Bankhaus Herstatt fails mid-afternoon, exposing settlement risk and prompting the Basel Committee.
1979 — Volcker begins the campaign that will drive American interest rates above twenty percent.
1981 — American Airlines launches the first modern frequent flyer program.
1982 — Mexico announces it cannot pay, opening the Latin American debt crisis; Kuwait’s Souk Al-Manakh collapses.
1983 — David Chaum publishes the cryptographic basis for anonymous electronic cash.
1990 — New Zealand adopts the first formal inflation target.
1991 — The Somali central bank ceases to exist; the shilling circulates anyway.
1994 — Brazil’s Real Plan ends hyperinflation using a stable unit of account introduced before the new currency.
1997 — Thailand abandons the baht peg on the second of July, beginning the Asian financial crisis.
1999 — The euro is introduced as an accounting currency.
2002 — Euro notes and coins replace twelve national currencies on the first of January.
2007 — Northern Rock depositors queue in the first British bank run since 1866; M-Pesa launches in Kenya.
2008 — Lehman Brothers fails in September; a money market fund breaks the buck; central banks open dollar swap lines. The Bitcoin paper appears on the thirty-first of October.
2009 — The first Bitcoin block is mined on the third of January. Zimbabwe abandons its currency after issuing a hundred-trillion-dollar note.
2013 — Cyprus imposes losses on uninsured depositors and capital controls within the euro area.
2014 — The Bank of England publishes a plain account of how commercial bank lending creates money.
2015 — Greek banks close for three weeks under a sixty-euro daily withdrawal limit.
2016 — India withdraws eighty-six percent of its currency by value overnight; its unified payments interface begins its rise.
2020 — Grounded airlines pledge their loyalty programs as collateral, revealing them to be worth more than the airlines.
2021 — El Salvador makes Bitcoin legal tender; Nigeria launches the eNaira.
2022 — An algorithmic stablecoin collapses in May, destroying roughly forty billion dollars of nominal value in under a week.
2023 — Nigeria’s currency redesign produces a nationwide cash shortage; a stablecoin briefly loses its peg after a bank failure.
2025 — El Salvador scales back its Bitcoin legal tender policy as a condition of international lending.
Literature
Ahamed, Liaquat. Lords of Finance: The Bankers Who Broke the World. New York: Penguin Press, 2009.
Armstrong, Wallace Edwin. Rossel Island: An Ethnological Study. Cambridge: Cambridge University Press, 1928.
Ashtor, Eliyahu. A Social and Economic History of the Near East in the Middle Ages. Berkeley: University of California Press, 1976.
Bagehot, Walter. Lombard Street: A Description of the Money Market. London: Henry S. King, 1873.
Bank for International Settlements. “Central Bank Digital Currencies: Foundational Principles and Core Features.” Basel, 2020.
Bank of England. “Money Creation in the Modern Economy.” Quarterly Bulletin, Q1 2014.
Bernanke, Ben S. Essays on the Great Depression. Princeton: Princeton University Press, 2000.
Bernanke, Ben S. The Courage to Act: A Memoir of a Crisis and Its Aftermath. New York: W. W. Norton, 2015.
Bernstein, Peter L. The Power of Gold: The History of an Obsession. New York: John Wiley and Sons, 2000.
Blustein, Paul. The Chastening: Inside the Crisis That Rocked the Global Financial System. New York: PublicAffairs, 2001.
Bodin, Jean. Response to the Paradoxes of Malestroit. 1568.
Bordo, Michael D., and Hugh Rockoff. “The Gold Standard as a Good Housekeeping Seal of Approval.” Journal of Economic History 56, no. 2 (1996).
Boyes, Roger. Meltdown Iceland. London: Bloomsbury, 2009.
Braudel, Fernand. Civilization and Capitalism, 15th–18th Century. 3 vols. Translated by Siân Reynolds. London: Collins, 1981–1984.
Bruner, Robert F., and Sean D. Carr. The Panic of 1907: Lessons Learned from the Market’s Perfect Storm. Hoboken: John Wiley and Sons, 2007.
Bryan, William Jennings. Speech to the Democratic National Convention, Chicago, 9 July 1896.
Castronova, Edward. Synthetic Worlds: The Business and Culture of Online Games. Chicago: University of Chicago Press, 2005.
Chancellor, Edward. Devil Take the Hindmost: A History of Financial Speculation. New York: Farrar, Straus and Giroux, 1999.
Chaum, David. “Blind Signatures for Untraceable Payments.” In Advances in Cryptology: Proceedings of Crypto 82. New York: Plenum Press, 1983.
Clapham, John. The Bank of England: A History. 2 vols. Cambridge: Cambridge University Press, 1944.
Coggan, Philip. Paper Promises: Money, Debt and the New World Order. London: Allen Lane, 2011.
Committee on Payments and Market Infrastructures. “Digital Currencies.” Basel: Bank for International Settlements, 2015.
Davies, Glyn. A History of Money: From Ancient Times to the Present Day. Cardiff: University of Wales Press, 1994.
de Roover, Raymond. The Rise and Decline of the Medici Bank, 1397–1494. Cambridge, MA: Harvard University Press, 1963.
Desan, Christine. Making Money: Coin, Currency, and the Coming of Capitalism. Oxford: Oxford University Press, 2014.
Dickson, P. G. M. The Financial Revolution in England: A Study in the Development of Public Credit, 1688–1756. London: Macmillan, 1967.
Eichengreen, Barry. Exorbitant Privilege: The Rise and Fall of the Dollar. Oxford: Oxford University Press, 2011.
Eichengreen, Barry. Globalizing Capital: A History of the International Monetary System. Princeton: Princeton University Press, 1996.
Eichengreen, Barry. Golden Fetters: The Gold Standard and the Great Depression, 1919–1939. Oxford: Oxford University Press, 1992.
Einzig, Paul. Primitive Money in Its Ethnological, Historical and Economic Aspects. London: Eyre and Spottiswoode, 1949.
European Central Bank. “The Euro Cash Changeover: Effects on Consumer Prices.” Monthly Bulletin, July 2002.
Feldman, Gerald D. The Great Disorder: Politics, Economics, and Society in the German Inflation, 1914–1924. Oxford: Oxford University Press, 1993.
Ferguson, Niall. The Ascent of Money: A Financial History of the World. London: Allen Lane, 2008.
Fergusson, Adam. When Money Dies: The Nightmare of the Weimar Hyper-Inflation. London: William Kimber, 1975.
Fisher, Irving. “The Debt-Deflation Theory of Great Depressions.” Econometrica 1, no. 4 (1933).
Flynn, Dennis O., and Arturo Giráldez. “Born with a Silver Spoon: The Origin of World Trade in 1571.” Journal of World History 6, no. 2 (1995).
Friedman, Milton, and Anna Jacobson Schwartz. A Monetary History of the United States, 1867–1960. Princeton: Princeton University Press, 1963.
Friedman, Milton. Money Mischief: Episodes in Monetary History. New York: Harcourt Brace Jovanovich, 1992.
Furness, William Henry. The Island of Stone Money: Uap of the Carolines. Philadelphia: J. B. Lippincott, 1910.
Galbraith, John Kenneth. Money: Whence It Came, Where It Went. Boston: Houghton Mifflin, 1975.
Galbraith, John Kenneth. The Great Crash, 1929. Boston: Houghton Mifflin, 1955.
Garber, Peter M. Famous First Bubbles: The Fundamentals of Early Manias. Cambridge, MA: MIT Press, 2000.
Gibson-Light, Michael. “Ramen Politics: Informal Money and Logics of Resistance in the Contemporary American Prison.” Qualitative Sociology 41 (2018).
Goetzmann, William N. Money Changes Everything: How Finance Made Civilization Possible. Princeton: Princeton University Press, 2016.
Goldgar, Anne. Tulipmania: Money, Honor, and Knowledge in the Dutch Golden Age. Chicago: University of Chicago Press, 2007.
Goldthwaite, Richard A. The Economy of Renaissance Florence. Baltimore: Johns Hopkins University Press, 2009.
Gorton, Gary B. Slapped by the Invisible Hand: The Panic of 2007. Oxford: Oxford University Press, 2010.
Graeber, David. Debt: The First 5,000 Years. Brooklyn: Melville House, 2011.
Hanke, Steve H., and Nicholas Krus. “World Hyperinflations.” In Routledge Handbook of Major Events in Economic History. London: Routledge, 2013.
Harl, Kenneth W. Coinage in the Roman Economy, 300 B.C. to A.D. 700. Baltimore: Johns Hopkins University Press, 1996.
Hogendorn, Jan, and Marion Johnson. The Shell Money of the Slave Trade. Cambridge: Cambridge University Press, 1986.
Hudson, Michael. …and forgive them their debts: Lending, Foreclosure and Redemption from Bronze Age Finance to the Jubilee Year. Dresden: ISLET, 2018.
Humphrey, Caroline. “Barter and Economic Disintegration.” Man 20, no. 1 (1985).
Ingham, Geoffrey. The Nature of Money. Cambridge: Polity Press, 2004.
Innes, A. Mitchell. “What Is Money?” Banking Law Journal 30 (1913).
International Monetary Fund. “The IMF and the Crises in Korea, Indonesia and Thailand.” Independent Evaluation Office Report, 2003.
Jack, William, and Tavneet Suri. “Risk Sharing and Transactions Costs: Evidence from Kenya’s Mobile Money Revolution.” American Economic Review 104, no. 1 (2014).
Jenkinson, Hilary. “Exchequer Tallies.” Archaeologia 62 (1911).
Keynes, John Maynard. A Tract on Monetary Reform. London: Macmillan, 1923.
Keynes, John Maynard. A Treatise on Money. 2 vols. London: Macmillan, 1930.
Keynes, John Maynard. The Economic Consequences of Mr. Churchill. London: Hogarth Press, 1925.
Kindleberger, Charles P. A Financial History of Western Europe. London: George Allen and Unwin, 1984.
Kindleberger, Charles P., and Robert Z. Aliber. Manias, Panics, and Crashes: A History of Financial Crises. 7th ed. London: Palgrave Macmillan, 2015.
Knapp, Georg Friedrich. The State Theory of Money. Translated by H. M. Lucas and J. Bonar. London: Macmillan, 1924. Originally published 1905.
Kroll, John H. “The Monetary Use of Weighed Bullion in Archaic Greece.” In The Monetary Systems of the Greeks and Romans. Oxford: Oxford University Press, 2008.
Kuroda, Akinobu. A Global History of Money. London: Routledge, 2020.
Lanchester, John. How to Speak Money. London: Faber and Faber, 2014.
Lehdonvirta, Vili, and Edward Castronova. Virtual Economies: Design and Analysis. Cambridge, MA: MIT Press, 2014.
Locke, John. Further Considerations Concerning Raising the Value of Money. London, 1695.
Lowndes, William. A Report Containing an Essay for the Amendment of the Silver Coins. London, 1695.
Mackay, Charles. Memoirs of Extraordinary Popular Delusions and the Madness of Crowds. London: Richard Bentley, 1841.
Malinowski, Bronislaw. Argonauts of the Western Pacific. London: Routledge and Kegan Paul, 1922.
Martin, Felix. Money: The Unauthorised Biography. London: The Bodley Head, 2013.
Mauss, Marcel. The Gift: Forms and Functions of Exchange in Archaic Societies. Translated by Ian Cunnison. London: Cohen and West, 1954. Originally published 1925.
Mehrling, Perry. The New Lombard Street: How the Fed Became the Dealer of Last Resort. Princeton: Princeton University Press, 2011.
Menger, Carl. “On the Origin of Money.” Economic Journal 2, no. 6 (1892).
Mihm, Stephen. A Nation of Counterfeiters: Capitalists, Con Men, and the Making of the United States. Cambridge, MA: Harvard University Press, 2007.
Murphy, Antoin E. “Money in an Economy without Banks: The Case of Ireland.” The Manchester School 46, no. 1 (1978).
Murphy, Antoin E. John Law: Economic Theorist and Policy-Maker. Oxford: Clarendon Press, 1997.
Nakamoto, Satoshi. “Bitcoin: A Peer-to-Peer Electronic Cash System.” 2008.
Narayanan, Arvind, et al. Bitcoin and Cryptocurrency Technologies. Princeton: Princeton University Press, 2016.
Neal, Larry. The Rise of Financial Capitalism: International Capital Markets in the Age of Reason. Cambridge: Cambridge University Press, 1990.
Pacioli, Luca. Summa de arithmetica, geometria, proportioni et proportionalita. Venice, 1494.
Peng Xinwei. A Monetary History of China. Translated by Edward H. Kaplan. Bellingham: Western Washington University, 1994.
Polanyi, Karl. The Great Transformation: The Political and Economic Origins of Our Time. New York: Farrar and Rinehart, 1944.
Polo, Marco. The Travels. Translated by Ronald Latham. London: Penguin, 1958.
Popper, Nathaniel. Digital Gold: Bitcoin and the Inside Story of the Misfits and Millionaires Trying to Reinvent Money. New York: Harper, 2015.
Quiggin, A. Hingston. A Survey of Primitive Money: The Beginnings of Currency. London: Methuen, 1949.
Quinn, Stephen, and William Roberds. “The Bank of Amsterdam and the Leap to Central Bank Money.” American Economic Review 97, no. 2 (2007).
Radford, R. A. “The Economic Organisation of a P.O.W. Camp.” Economica 12, no. 48 (1945).
Reinhart, Carmen M., and Kenneth S. Rogoff. This Time Is Different: Eight Centuries of Financial Folly. Princeton: Princeton University Press, 2009.
Reserve Bank of India. Annual Report 2017–18, chapter on the return of demonetized banknotes.
Rolnick, Arthur J., and Warren E. Weber. “New Evidence on the Free Banking Era.” American Economic Review 73, no. 5 (1983).
Ryan-Collins, Josh, et al. Where Does Money Come From? A Guide to the UK Monetary and Banking System. London: New Economics Foundation, 2011.
Sargent, Thomas J. “The Ends of Four Big Inflations.” In Inflation: Causes and Effects. Chicago: University of Chicago Press, 1982.
Sargent, Thomas J., and François R. Velde. The Big Problem of Small Change. Princeton: Princeton University Press, 2002.
Schaps, David M. The Invention of Coinage and the Monetization of Ancient Greece. Ann Arbor: University of Michigan Press, 2004.
Schnabel, Isabel, and Hyun Song Shin. “Liquidity and Contagion: The Crisis of 1763.” Journal of the European Economic Association 2, no. 6 (2004).
Simmel, Georg. The Philosophy of Money. Translated by Tom Bottomore and David Frisby. London: Routledge, 1978. Originally published 1900.
Sorkin, Andrew Ross. Too Big to Fail. New York: Viking, 2009.
Sprague, O. M. W. History of Crises under the National Banking System. Washington: Government Printing Office, 1910.
Steil, Benn. The Battle of Bretton Woods: John Maynard Keynes, Harry Dexter White, and the Making of a New World Order. Princeton: Princeton University Press, 2013.
Stein, Stanley J., and Barbara H. Stein. Silver, Trade, and War: Spain and America in the Making of Early Modern Europe. Baltimore: Johns Hopkins University Press, 2000.
Stodder, James. “Complementary Credit Networks and Macroeconomic Stability: Switzerland’s Wirtschaftsring.” Journal of Economic Behavior and Organization 72, no. 1 (2009).
Suri, Tavneet, and William Jack. “The Long-Run Poverty and Gender Impacts of Mobile Money.” Science 354, no. 6317 (2016).
Taylor, Frederick. The Downfall of Money: Germany’s Hyperinflation and the Destruction of the Middle Class. London: Bloomsbury, 2013.
Tooze, Adam. Crashed: How a Decade of Financial Crises Changed the World. London: Allen Lane, 2018.
Triffin, Robert. Gold and the Dollar Crisis: The Future of Convertibility. New Haven: Yale University Press, 1960.
Udovitch, Abraham L. Partnership and Profit in Medieval Islam. Princeton: Princeton University Press, 1970.
Van De Mieroop, Marc. “The Invention of Interest: Sumerian Loans.” In The Origins of Value. Oxford: Oxford University Press, 2005.
Velde, François R. “Government Equity and Money: John Law’s System in 1720 France.” Federal Reserve Bank of Chicago Working Paper, 2003.
von Glahn, Richard. Fountain of Fortune: Money and Monetary Policy in China, 1000–1700. Berkeley: University of California Press, 1996.
von Reden, Sitta. Money in Classical Antiquity. Cambridge: Cambridge University Press, 2010.
Weatherford, Jack. The History of Money. New York: Crown, 1997.
Wennerlind, Carl. Casualties of Credit: The English Financial Revolution, 1620–1720. Cambridge, MA: Harvard University Press, 2011.
White, Lawrence H. Free Banking in Britain: Theory, Experience, and Debate, 1800–1845. Cambridge: Cambridge University Press, 1984.
Wray, L. Randall. Modern Money Theory: A Primer on Macroeconomics for Sovereign Monetary Systems. 2nd ed. London: Palgrave Macmillan, 2015.
