What Is Money, and Why Does It Matter?

On a small island in the Pacific, a stone that nobody had seen for generations still counted as a family’s wealth. It lay at the bottom of the sea. Everyone agrees it is worth something, and everyone knows whose it is.

This essay starts there and follows the same question from the first clay tokens to Bitcoin: what makes people trust a thing enough to call it money? By the end, you will be able to look at any form of money, a coin, a bank balance, a stablecoin or Bitcoin, and ask the two questions that explain it.

Prologue. Why can a stone nobody sees be worth something?

• Memory

On Yap, in the western Pacific, the most valuable money used to be stone. The islanders quarried great discs of limestone, called raiRaiLimestone discs quarried on Palau and carried to Yap; the largest were taller than a person., on the islands of Palau, and brought them home across open ocean. Some were taller than a person. They were far too heavy to carry to market, so they mostly stayed where they stood. When one changed hands, nobody moved it. The village simply remembered who owned it now.

One story from Yap, recorded by the anthropologist William Henry Furness in 1910, goes further. A great stone was lost overboard on the way home and sank. No one living had ever seen it. It went on working as money anyway, because everyone agreed it existed and knew whose it was.

It is a legend, and it may be the most accurate description of money there is. The stone was not valuable because of what it was made of. It was valuable because a group of people shared a memory about it, and trusted that memory enough to settle their debts with it.

Hold on to that picture. Every form of money in this essay, from barley to gold to bank balances to Bitcoin, answers the same two questions the village of Yap answered with a stone: who keeps the memory of what everyone owes, and why should the rest of us believe them?

To follow those questions, we will look at money through four lenses that come back throughout:

• Memory: who keeps the record.
• Reach: how far, and for how long, the trust can travel.
• Power: who writes the rules, and who can take the money back.
• Value: whether money is something to keep or something to spend.

How did a sack of barley become a sign?

• Memory
• Value

Long before coins, before alphabets, before anyone wrote a poem or a law, people in the Near East were counting grain.

Farming changed the arithmetic of life. A village that grows more barley than it eats has a surplus, and a surplus raises questions a hunting band never faces. Who brought what to the common store? Who is owed a share at harvest? How much is left for the lean months? A few families can keep that in their heads. A town cannot.

The answer, according to the archaeologist Denise Schmandt-Besserat, was a handful of clay. From around 8000 BC, sites across the Near East used small modeled tokens, each shape standing for a unit of a particular good. In her reading, a cone stood for a small measure of grain, a sphere for a larger one, an ovoid for a jar of oil. To record that a family had delivered three small measures of barley, you kept three cones.

Between about 3500 and 3300 BC, accountants began sealing tokens inside hollow clay balls, so nobody could add or remove one. That created a problem: to check the contents, you had to break the ball. So they pressed the tokens into the outside of the envelope before sealing it. Schmandt-Besserat’s verdict is simple: “These markings were the first signs of writing.”

Once the marks were on the outside, the tokens inside became redundant. By around 3200 BC, scribes were pressing the shapes directly into flat clay tablets: a cone left a wedge, a sphere left a circle. Around 3100 BC came the decisive step. Instead of impressing a token, a scribe traced a picture with a stylus: an ear of barley, a jar of oil. And the quantity separated from the thing counted. Thirty-three jars of oil were no longer thirty-three jar-shaped marks. They were three circles meaning tens, three wedges meaning units, and one drawn jar. In Schmandt-Besserat’s reading, the number had become abstract, free to count anything. Other specialists point out that scribes still used different counting systems for different goods, so fully abstract numbers came later.

Other signs followed the same logic. In ration lists, a human head next to a bowl meant a ration to eat. The sign for barley looked like an ear of grain.

The largest archive of these early tablets comes from Uruk, in today’s southern Iraq, a city that may have held 40,000 people. About 5,000 proto-cuneiformProto-cuneiformThe earliest stage of Mesopotamian writing, around 3350 to 3000 BC: signs pressed or drawn in clay. tablets survive from roughly 3350 to 3000 BC. As the Assyriologist Robert Englund notes, the figures usually cited are averages: 85 percent administrative and 15 percent lexical lists, the word lists scribes used to learn their craft. In the oldest layer, fewer than one percent are anything other than accounts. They record grain, beer and bread rations counted in 30-day months, livestock sorted by age, and dependent workers.

85%of Uruk’s proto-cuneiform tablets are administrative accounts, on average. In the earliest layer, fewer than 1% are anything else.

One tablet has become famous. It records about 135,000 liters of barley to be delivered over 37 months to the official in charge of a temple brewery, and it carries a sign that reads Kushim. It is sometimes called the first recorded name of a human being. Scholars are more careful: Kushim may have been a person, a job title or an institution. Either way, one of the oldest names we have sits on a receipt.

Did writing, then, come from accounting? In Mesopotamia, the evidence points that way, and even Schmandt-Besserat’s critics accept that cuneiform began with numerical signs. Her full chain, from 8000 BC tokens to writing, is more debated. Specialists have faulted her for over-interpreting the shapes, and newer studies ask whether some early “tokens” were counters at all. And writing was invented more than once. In China, the earliest texts, the oracle bones of around 1250 to 1050 BC, record divination, not grain. The Inca, meanwhile, ran an empire’s taxes and censuses on knotted cords, the khipu, without writing in our sense at all.

The careful conclusion is still striking. In Mesopotamia, one of the first places where writing appeared, it appeared to keep track of what people owed each other. Before people there wrote down their gods, their kings or their stories, they wrote down their debts.

Before coins, barter or debt?

• Memory
• Value

Most of us learned a tidy story. First people bartered: my shoes for your wheat. Barter was clumsy, because the shoemaker had to find a farmer who wanted shoes at the same moment. So people settled on one commodity everyone accepted, and money was born. Adam Smith made the story famous in 1776.

It is hard to find in the record. In 1985 the Cambridge anthropologist Caroline Humphrey wrote: “No example of a barter economy, pure and simple, has ever been described, let alone the emergence from it of money; all available ethnography suggests that there never has been such a thing.” David Graeber built a book on the point, Debt: The First 5,000 Years (2011): credit came first, coins thousands of years later.

Mesopotamia fits that account. Value was measured in silver and in barley. In the Ur III period, around the 21st century BC, the official ratio was one shekel of silverShekelA unit of weight, about 8.3 grams of silver, used to value goods long before coins existed., about 8.3 grams, for one gur of barley, roughly 300 liters. Silver was the yardstick in which “all commodities” were valued, but a debt could be counted in silver without any silver changing hands. People owed, they were owed, and scribes kept the tally. The laws of Eshnunna and of Hammurabi, from the 19th and 18th centuries BC, even set standard interest rates: 20 percent a year on silver, 33 and a third percent on barley. Money, in other words, was first a unit of account and a record of debts, and only later a thing in your hand.

The debate is not closed, and the other side deserves a fair hearing. The economist George Selgin points out that a lack of evidence for barter economies is not evidence against Smith, and that simple credit only works between people who deal with each other again and again. That is the key, and it lets both sides be right. Inside a community, where everyone knows everyone, obligations can be remembered and settled later. Between strangers, who may never meet again, that memory is missing, and something else has to fill the gap.

Anthropologists who followed Humphrey added one more observation that will matter later in this essay: where barter dominates today, it is rarely among people who never had money. More often, it is among people who know money but cannot get it or rely on it, because the money has broken down.

Why cattle, shells and metal?

• Value
• Reach

If money is a shared memory, why did so many societies attach it to objects?

Because memory does not travel. A village can remember that you are owed a share of the harvest. A trader three valleys away cannot. An object can carry the promise with it.

Our vocabulary still remembers the first candidates. The Latin pecunia, money, comes from pecu, cattle. The English word fee is related to Old English feoh: livestock, property, money. In the Iliad, Homer prices armor in oxen: a golden set worth a hundred oxen, a bronze one worth nine. In China, cowrie shells were prized under the Shang dynasty, from around 1600 BC, and bronze imitations of them appeared later, though historians debate whether they were true currency or prestige goods. China later used bronze spades and knives as money, before the Qin standardized a round coin in 221 BC.

What makes an object good for carrying a promise? Economists list a handful of properties. We will look at them closely in Part III. For now, notice the logic. Cattle are valuable to everyone but hard to divide. Shells are portable and countable, as long as nobody can gather them by the boatload. Metal is durable, divisible and hard to fake, as long as someone vouches for its weight and purity.

That last condition gave us coins. The earliest coins we know were struck in Lydia and the neighboring Greek cities of Ionia, in today’s western Turkey, from electrumElectrumAn alloy of gold and silver found naturally in the rivers of Lydia. The royal mint adjusted the mix, adding silver., an alloy of gold and silver. Ninety-three early electrum coins were found under the Temple of Artemis at Ephesus, offered in the later seventh century BC. A stamp on a lump of metal did something new: it let an issuer, often a king, vouch for the weight in advance, so strangers did not have to weigh and test every piece. The coin was a promise with a signature on it.

A myth to set aside. You may have heard that the word salary comes from Roman soldiers being paid in salt. No ancient source says so. Pliny the Elder links the word salarium to salt, and to public office and military service, but never says anyone was paid in salt; the story of salt wages was added by dictionary makers in the 18th and 19th centuries. Roman soldiers were paid a stipendium, in coin. The etymology may be real. The payroll of salt is not.

Why do we trust a stranger in a uniform?

• Reach

For most of human history, trust had a face. You trusted your family, your friends, your neighbors, the people you had watched keep their word. Among friends, nobody needs a coin: someone pays for dinner, someone else gets the next one, and a loose tally of favors lives in everyone’s head, close enough to trust and just loose enough to drift. The anthropologist Robin Dunbar famously suggested that the human brain can maintain stable relationships with around 150 people. The exact number is contested: a 2021 reanalysis found the statistical range so wide, from a handful to several hundred, that its authors called specifying any one number “futile”. But the intuition survives the critique. Personal trust does not scale.

And yet today you hand your savings to a bank whose employees you have never met, board a plane flown by a pilot you will never see, and accept a salary paid in notes issued by an institution in another city. You trust strangers every hour of every day.

The economist Paul Seabright made this puzzle the subject of The Company of Strangers (2004). Humans evolved in small, wary groups. In roughly ten thousand years they learned to cooperate with people they would never know, and institutions like money, prices and banks let strangers treat each other almost as friends.

Money is at the center of that trick, and economics has a precise way to say why. In 1998 the economist Narayana Kocherlakota published a paper titled “Money Is Memory”. He showed that anything an economy can achieve with money, it could also achieve with a complete shared record of everyone’s past dealings. Nobody has that record. Money stands in for it. When a stranger accepts your coin, they do not need to know your history. The coin carries the proof that someone, somewhere, gave you something of value.

The sociologist Anthony Giddens described the same shift from another angle. Modern life, he argued, lifts relationships out of local settings and stretches them across time and space. It runs on what he called abstract systems: expert systems we cannot check ourselves, and symbolic tokens that circulate between people who never meet. His main example of a symbolic token is money. We renew our trust in those systems at what he called access points: the bank counter, the cockpit door, the person in the uniform. A uniform is not a guarantee of honesty. It is a sign that an institution stands behind the person wearing it, and that the institution can be held to account.

Long-distance trade shows how far that borrowed trust could reach, long before modern banks. In the first century, Pliny the Elder complained that India drained the Roman Empire of at least 50 million sesterces a year. Historians read the figure as rhetoric, not bookkeeping, but Roman coins, mostly silver with some gold, have indeed been found in hoards in southern India, many of them in Tamil Nadu, and a Tamil poem describes foreign ships that came with gold and returned with pepper. The merchants who took those coins had no reason to care which emperor was a good one. They trusted the weight of the metal, whatever face was stamped on it: many coins found there bear cuts made to test it.

Medieval merchants pushed the idea further. The Italian banking houses settled trade across Europe with bills of exchangeBill of exchangeA written order to pay a sum at another place or date, which merchants could pass on to others., paper orders to pay that let a merchant in one city collect funds in another without carrying a single coin across a dangerous road. The papers of Francesco Datini, a merchant of Prato who died in 1410, include around 150,000 documents, most of them business letters. Across much of the Muslim world and South Asia, the hawala system moved value between distant cities on the word of brokers; the IMF describes it as depending “more on absolute trust between the participants than on legal documents”.

Whether the carrier is a gold coin, a bill of exchange or a broker’s word, the pattern is the same. Money lets trust leave the village.

Why can nobody make a pencil alone?

• Reach

In 1958 the writer Leonard Read published a short essay in the voice of an ordinary wooden pencil. Its boast was simple: “not a single person on the face of this earth knows how to make me.” The cedar comes from one place, the graphite from another, the rubber from a third, the brass ferrule from a fourth. Millions of people, the essay says, contribute a sliver of knowledge to the pencil, and none of them holds the whole recipe.

Adam Smith had made a similar point in 1776 with pins. In a small workshop, he wrote, pin-making was divided into about eighteen distinct operations. Ten workers produced upwards of 48,000 pins a day. Working alone, each might not have made twenty, “perhaps not one”. And, in his third chapter, Smith drew the conclusion that matters for us: the division of labor is limited by the extent of the market. Specialization only pays when you can trade with enough people.

In 2009 a design student at the Royal College of Art, Thomas Thwaites, tested the idea literally. He took apart a cheap electric toaster, bought for £3.94, and tried to rebuild it from raw materials: mining the ore, smelting the iron, making the plastic. The project took him the better part of a year and £1,187.54 of direct spending, and the result barely worked. A product that thousands of strangers make cheaply together is nearly impossible for one determined person to make at all.

Now scale that up. The machines that print the most advanced computer chips, built by the Dutch company ASML, contain about 100,000 parts provided by more than 5,000 suppliers, according to a 2021 industry report. A single chip can go through 400 to 1,400 manufacturing steps. Apple says its products rely on thousands of supplier facilities in more than 60 countries. The laptop or phone you are reading this on is the work of more strangers than you will meet in your lifetime.

How do those strangers coordinate without a central plan? In 1945 the economist Friedrich Hayek gave the classic answer. The knowledge needed to run an economy, he wrote, never exists in one place; it is scattered in “dispersed bits of incomplete and frequently contradictory knowledge” held by separate individuals. Prices, expressed in money, compress that scattered knowledge into a signal anyone can read. His own example was tin: when it gets scarce, its price rises, and people who have never heard why start to economize on it.

Money has become that signal almost entirely. SWIFT, the messaging network banks use for international payments, was founded in 1973 by 239 banks from 15 countries and now connects more than 11,000 institutions, carrying tens of millions of messages a day. In the twelve months to September 2025, 329 billion payments and cash transactions carried the Visa brand. Most money today is not metal or paper. It is information, moving between databases, telling strangers who has delivered what.

The results are visible in the long run. According to the Maddison Project, world GDP per person, adjusted for prices, rose from about $1,128 in 1820 to about $16,677 in 2022, roughly fifteen times. Exports and imports together went from about a quarter of world GDP in 1970 to more than half. Money did not cause that growth on its own. Energy, science, law and institutions all played their part, and historians still argue about their weight. But none of it could have been coordinated among billions of strangers without a way to keep score.

There is a price for this reach. Long chains are fragile. When one link breaks, as supply chains did during the pandemic, the whole system can stall, and the knowledge to rebuild locally may no longer exist. The same trust that lets strangers build a microchip together makes all of them dependent on that trust holding.

Why is money we keep no longer money we spend?

• Value

Textbooks give money three jobs. It is a medium of exchange, something people accept in payment. It is a unit of account, the common measure we use to price things. And it is a store of value, something you can keep today and use later.

These jobs can pull against each other. When people expect a money to gain value, they are tempted to keep it rather than spend it.

Thomas Gresham, an adviser to Elizabeth I, warned about one version of this in the sixteenth century, and in 1858 an economist named the rule after him: Gresham’s lawGresham’s lawWhen two kinds of money must be accepted at the same legal value, people spend the worse one and keep the better one.. When two kinds of money must be accepted at the same official value, people spend the worse one and hoard the better one. Bad money drives good money out of circulation. The law needs that condition of a fixed official value to work, but the instinct behind it is common.

The clearest modern demonstration came from a prisoner of war camp. In 1945 R. A. Radford, who had been held in prisoner of war camps in Italy and Germany during the Second World War, described their economy in the journal Economica. Cigarettes from Red Cross parcels had become money, used “as a unit of account, as a measure of value and as a store of value”. They were uniform, durable and easy to carry. And the prisoners rediscovered several classic monetary troubles. Cigarettes “could be clipped or sweated by rolling them between the fingers”, taking a little tobacco out before passing them on. Hand-rolled cigarettes drove machine-made ones out of circulation. And when large deliveries of cigarettes arrived, prices in cigarettes rose.

John Maynard Keynes described a related trap in 1936. If everyone tries to save more at the same time by spending less, when demand is already weak, incomes fall, and “the attempt necessarily defeats itself”. Money everyone keeps is money no one earns.

This tension will return when we reach Bitcoin, which many holders treat above all as something to keep. For now, notice that every monetary system is a compromise between keeping and spending, and that societies move along that line when their trust moves.

What happens when the promise breaks?

• Value
• Memory

Economists list the properties of good money in slightly different ways. The Federal Reserve Bank of St. Louis names durability, portability, divisibility, uniformity, limited supply and acceptability. History suggests reading them as promises made to everyone who uses the money, and six of those promises have broken often enough to be worth watching.

• Scarcity: nobody can cheaply make more of it.
• Legitimacy: the group recognizes it as money.
• Portability: it can reach you.
• Verifiability: you can tell it from a fake.
• Uniformity: my unit is worth exactly the same as yours.
• Stability: it will still hold its value tomorrow.

Money holds as long as those promises hold. History is a catalog of what happens when one of them breaks.

Scarcity, broken by metal tools. Wampum were shell beads made by Native peoples of the American Northeast, rich in diplomatic and ceremonial meaning. Dutch traders brought them into the New England trade in 1627, and in 1637 Massachusetts made them legal tenderLegal tenderMoney that a creditor is required by law to accept in payment of a debt.. Part of their value came from how hard they were to make. Metal drills and awls brought by Europeans made them far easier to produce, and colonists produced them in quantity. Quality fell, the official rates were eventually cut, and in 1661 Massachusetts ended their legal tender status, citing “much inconvenience”.

Scarcity, broken by ships. Cowrie shells from the Maldives had long circulated as money across large parts of West Africa. European traders shipped them in bulk, importing what the historian Jan Hogendorn describes as “tens of thousands of tons” over the era of the Atlantic slave trade, and imports rose and fell with that trade. In the nineteenth century, German and French firms flooded the coast with a cheaper cowrie species from Zanzibar. “A great inflation ensued,” Hogendorn writes, and colonial authorities banned the imports around 1900. It is sometimes said that colonizers bought a continent with shells. The real mechanism is subtler and darker: when the thing a society uses as money can be produced far away, cheaply, by outsiders, the money stops being a shared agreement and becomes a lever.

Legitimacy, and the stones nobody authorized. Back on Yap, an Irish-American captain named David O’Keefe arrived in the nineteenth century. Rai had been quarried by hand with shell and stone tools and carried home by canoe, which is why they were rare. O’Keefe used a steamship and iron tools to help islanders quarry on Palau, and brought back stones in quantity. His stones were bigger and more finely finished. According to the Bank of Canada Museum, they were worth less than many smaller, cruder traditional ones, because they “weren’t authorized by any chief, were far less labour-intensive to make and gained no history”. Scarcity, it turns out, is not enough. A money also needs the group’s recognition.

Portability, broken by weight. In the seventeenth century Sweden, rich in copper, issued money as copper plates. The largest, a 10-daler plate from 1644, weighed about 19.7 kilograms. Large payments traveled by cart. It is probably no coincidence that Stockholm’s first bank issued Europe’s first real banknotes in 1661. When money cannot travel, people invent a promise that can.

Verifiability, broken by scissors. Many silver coins circulating in seventeenth-century England were old hand-struck pieces with irregular edges, easy to clip: shave a little silver from the rim, spend the coin at face value, keep the shavings. By 1695 nearly half of the silver was missing from coins in circulation. The Great Recoinage of 1696 withdrew them, and trade nearly froze while new coins were minted. Isaac Newton was appointed Warden of the Royal Mint that year, later became its Master, and pursued counterfeiters to the gallows.

Uniformity, broken by too many issuers. In the United States between 1837 and 1863, more than a thousand state banks issued their own notes; the Office of the Comptroller of the Currency counts more than 7,000 different ones. A dollar from a distant or shaky bank traded at a discount, and merchants needed printed guides to know what each note was worth. The deep dive below tells that story.

Stability, broken completely. Hyperinflations usually start when a government prints money to pay its bills. Then people stop believing the money will hold its value and spend it as fast as possible, which pushes prices up faster and confirms their fear. At its peak in October 1923, prices in Germany rose 29,500 percent in a month and doubled every 3.7 days; the crisis ended with a new currency, the Rentenmark, worth one trillion old marks. In Zimbabwe, in November 2008, prices doubled every 24.7 hours. The record belongs to Hungary in July 1946, when prices doubled every 15 hours.

Remember the barter that textbooks place at the beginning of money? It often reappears at the end. When the paper money of China’s Yuan dynasty collapsed after printing ran out of control, paper money was scrapped when the dynasty fell in 1368 and, in the words of the economic historians Hanhui Guan, Nuno Palma and Meng Wu, “barter was observed in all prefectures and counties”. Barter is less where money begins than where money goes when trust dies.

Near or far, now or later: how do we judge a money?

• Reach
• Value

A good money is accepted far from where it was made, and long after it was made. Having both is rare.

Rome managed it for a while. Its coins were valued from Spain to southern India, even where they were treated as plain metal. But extending a currency across an empire takes roads, garrisons and armies, and armies are expensive. The more a state spends to make its money accepted far away, the more strain it puts on the finances that make that money trusted over time. That trade-off between reach in space and reach in time is not a law of nature, but it is a pattern worth keeping in mind as we turn to empires and their currencies.

How do we know whether people still trust a money? They rarely say so. They show it. There are at least five places to look.

What people save in. From 1997 until the peg broke down in 2019, the Lebanese pound was peggedPegA fixed exchange rate that a central bank promises to defend. at 1,507.5 to the dollar. By February 2023 it had lost more than 98 percent of its pre-crisis value, according to the World Bank, and the economy running on dollar cash had reached 45.7 percent of GDP in 2022. In Argentina, the national statistics agency counted about $230 billion in foreign cash and deposits held by residents at the end of 2020: in accounts abroad, in safe deposit boxes, under mattresses.

The gap between the official price and the street price. In July 2022, the gap between Argentina’s official exchange rate and its “blue” street rate reached about 150 percent. A dollar on the street cost about two and a half times its official price.

The price of borrowing. In February 2012, Greece’s ten-year government bonds yielded 29.24 percent on average. Germany borrowed for ten years at 1.85 percent that month. The difference largely reflected the market’s doubt that Greece would keep its promises in full.

The speed of the exit. On March 9, 2023, depositors pulled $42 billion out of Silicon Valley Bank in a single day, and the Federal Reserve’s own review found management expected more than $100 billion to follow on March 10. Together, that was about 85 percent of its deposits. In 2008, Wachovia took eight days to lose $10 billion. Trust used to leave through the front door. Now it leaves through a phone.

$42 billionleft Silicon Valley Bank in a single day, on March 9, 2023.

What central banks buy. The institutions that issue money are hedging against the money of others. According to the World Gold Council, central banks bought more than 1,000 tonnes of gold a year in 2022, 2023 and 2024, more than double their average from 2010 to 2021.

There is also a reverse test. After the Gulf War, the Kurdish north of Iraq kept using the old “Swiss” dinar, printed before the war, while Saddam Hussein’s government printed new notes at will. No government stood behind the old notes, and nobody could print more of them. Mervyn King, then Governor of the Bank of England, described in 2004 how they rose to be worth about 300 of Saddam’s dinars, and how their value tracked expectations about the regime’s fall. A currency with no one left to print it can be trusted more than one whose issuer nobody believes.

“The essence of good money has always been trust in the stability of its value,” the Bank for International Settlements wrote in 2018. Every figure in this chapter is that sentence, measured.

Was gold the real money?

• Power

On December 19, 1912, the most powerful banker in America sat before a subcommittee of the US House of Representatives investigating the “money trust”. Its counsel, Samuel Untermyer, asked J. P. Morgan whether the basis of banking was credit. Morgan answered: “Not always. That is an evidence of banking, but it is not the money itself. Money is gold, and nothing else.”

That line is often quoted in a sharper form, “Gold is money, everything else is credit”, which does not appear in the transcript. What the transcript does contain, a few pages later, is more interesting. Asked whether commercial credit was based primarily on money or property, Morgan said: “No, sir; the first thing is character.” Before money or property? “Before money or anything else. Money can not buy it.”

On the same day, the same man said that money is gold, and that credit rests on trust in a person. Between those two answers lies the whole history of the gold standard.

If trust in money depends on whether the issuer keeps its promises, one solution is to tie the issuer’s hands. Britain drifted into doing so. In 1717 Isaac Newton, still at the Mint, set the value of the gold guinea at 21 shillings, a rate that pushed silver out of circulation. Gold became Britain’s legal standard after the Napoleonic wars. From the 1870s Germany, France, the United States and others joined, and until 1914 much of the world’s money was defined as a weight of gold. Over the period from 1880 to 1914, average US inflation was close to zero.

The discipline had a cost. The supply of money depended on the supply of gold, and the supply of gold depended on mines. A country could not create money to fight a crisis. The First World War ended the system. In 1925 Winston Churchill, as Chancellor of the Exchequer, returned Britain to gold at the prewar parity, over the objections of John Maynard Keynes, who argued the pound was now overvalued by about 10 percent. Britain was forced off gold on September 21, 1931. Later research by Barry Eichengreen and Jeffrey Sachs, and by Ben Bernanke, found that countries that abandoned gold earlier in the Great Depression recovered sooner.

The sharpest lesson came from the United States. On April 5, 1933, Executive Order 6102 required Americans to deliver most of their gold coins, bullion and gold certificates to the banking system by May 1, at the official price of $20.67 an ounce. Refusing could mean a fine of up to $10,000, ten years in prison, or both. In January 1934 the Gold Reserve Act let the president reset the price, and the next day he set it at $35. The dollar was anchored to gold, and the government that held the anchor moved it.

In July 1944, delegates from 44 nations met at Bretton Woods, New Hampshire, and rebuilt the world’s money around the dollar: other currencies were pegged to the dollar, and the dollar was convertible into gold at $35 an ounce for foreign governments. In 1960 the economist Robert Triffin pointed out the flaw. The world could only get the dollars it needed if the United States ran deficits, and those deficits would eventually undermine confidence that the dollars could really be exchanged for gold. On August 15, 1971, President Nixon closed the gold windowGold windowThe US commitment to exchange dollars held by foreign governments for gold at $35 an ounce.. By 1973 the major currencies were floating.

Should the world go back? In a 2012 survey of leading academic economists by the University of Chicago’s IGM Forum, none agreed that a gold standard would improve price stability and employment for the average American. Gold’s defenders answer that the point was never average outcomes but discipline over governments. Both sides can learn from the record. An anchor only works if the people holding it cannot cut the rope. For most of the gold standard’s life, they could.

Who creates money today, and for whom?

• Power
• Memory

Ask most people where money comes from and they say the government prints it. Most of the money you use was never printed.

In 2014 the Bank of England published a paper titled “Money creation in the modern economy”. Its central sentence is plain: “Whenever a bank makes a loan, it simultaneously creates a matching deposit in the borrower’s bank account, thereby creating new money.” Banks do not wait for savers to deposit money and then lend it out. Lending creates the deposit. When the loan is repaid, that money disappears.

The same paper says the “money multiplier” found in many textbooks, in which banks can only lend a fixed multiple of their reserves, “is not an accurate description of how money is created in reality”. What limits lending is profitability, regulation, how much households and firms want to borrow, and the interest rate set by the central bank. In the United States, reserve requirements have been zero since March 26, 2020.

90%or more of the euro area’s broad money is not cash but deposits and similar balances, created mostly when banks lend or buy assets (July 2026).

The scale is striking. The Bank of England estimated in 2014 that bank deposits made up 97 percent of broad money in the UK. In the United States, in July 2026, physical currency was about $2.4 trillion out of $23.2 trillion of M2M2A broad measure of money: cash, checking and savings deposits, small time deposits, and retail money market funds., the Federal Reserve’s broad measure of money: about 10 percent. Most of the rest is bank deposits, private promises recorded in bank ledgers.

Central banks create a different kind of money: cash, and the reserves commercial banks hold with them. They steer everything else by setting the interest rate on those reserves. In a crisis they can also buy assets on a massive scale, which is what quantitative easingQuantitative easingLarge-scale purchases of bonds by a central bank, paid for with reserves it creates. means. The Federal Reserve’s balance sheet grew from about $4.16 trillion in February 2020 to about $8.97 trillion in April 2022.

Governments, meanwhile, finance themselves by issuing debt. On September 10, 2026, total US public debt stood at just over $40 trillion. When a government spends what it borrowed, the money lands in bank deposits. When a central bank buys government bonds, it pays with reserves it creates. That is why central bank independence matters: a government that could order its central bank to buy all its debt could create money at will. In 1951 the US Treasury and the Federal Reserve formally agreed to “minimize monetization of the public debt”. Some economists, under the banner of Modern Monetary Theory, argue that a country borrowing in its own currency faces a limit set by inflation rather than solvency; when the IGM Forum asked leading economists in 2019 whether such countries need not worry about deficits because they can always create money, none agreed, and the theory’s supporters say that question simplified their position.

The whole structure rests on a public guarantee. Since 2008 the FDIC has insured US deposits up to $250,000 per depositor, per bank, for each type of account ownership, a limit made permanent in 2010. The balance in your account is a promise from your bank, created when someone borrowed, and made credible by the state.

That design has a consequence. Whoever creates new money decides who receives it first, and the first recipients spend it before prices adjust, an effect economists associate with the eighteenth-century banker Richard Cantillon. When banks create most of the money, banks decide, loan by loan, who gets it first.

So what happens when the bank decides to lend to itself?

Private money creation is not new: the American free banking era, met in Chapter 7, already showed what happens when each bank issues its own dollars.

The deeper risk is structural. When the bank that creates money belongs to a group, it can create money for that group. In 2003 Rafael La Porta, Florencio Lopez-de-Silanes and Guillermo Zamarripa found that in Mexico in the mid-1990s, loans to parties related to banks’ own owners made up 20 percent of commercial lending, and defaulted far more often. The authors did not soften it: “related lending is a manifestation of looting”. The rules that cap such lending exist because the temptation is built into the machine: the institution that can create money is the same one that decides who receives it.

Why does the whole world need dollars?

• Power
• Reach

If a bank’s money is a promise from the bank, a country’s currency is a promise from the country. And one country’s promise is held far beyond its borders.

In the first quarter of 2026, the US dollar made up 57 percent of the world’s allocated foreign exchange reserves, according to the IMF, against about 20 percent for the euro and 2 percent for China’s renminbi. In foreign exchange markets, where $9.6 trillion changed hands every day in April 2025, the dollar was on one side of 89 percent of all trades, according to the Bank for International Settlements. It carried about half of international payments by value on SWIFT in June 2026. More than $1 trillion of dollar banknotes, around half of all those outstanding, were held outside the United States in 2025.

In the 1960s the French finance minister Valéry Giscard d’Estaing called this America’s “exorbitant privilege”. A country whose currency everyone needs can borrow in it cheaply, and others must earn it before they can use it. After Nixon ended gold convertibility in 1971, his Treasury Secretary John Connally reportedly told European counterparts that the dollar was “our currency, but your problem”.

The privilege is also power. In early 2022, after Russia’s invasion of Ukraine, Western governments immobilized roughly $300 billion of the Russian central bank’s reserves and cut major Russian banks off SWIFT. In December 2025 the European Union stopped having to renew the freeze every six months, keeping about €210 billion of those assets blocked until Russia ends its war and compensates Ukraine. For many central banks, a reserve currency stopped looking only like a safe asset and started looking like something that could be switched off. Central banks bought more than 1,000 tonnes of gold a year for three years running from 2022, more than double the average of the previous decade.

Is the dollar’s dominance ending? The evidence is mixed. China has built its own cross-border payment system, and the BRICS countries speak often of reducing their dependence on the dollar. But by the Atlantic Council’s reading, their public ambitions have been significantly scaled back since 2024, and the dollar’s share of currency trading has even edged up. The euro, launched as book money in 1999 and as notes and coins in 2002, remains a strong second; Bulgaria became its 21st member on January 1, 2026. It came close to breaking in the 2010 to 2012 crisis, until the ECB’s president Mario Draghi said in July 2012 that the bank was “ready to do whatever it takes to preserve the euro. And believe me, it will be enough.”

A currency’s reach, then, is not only a matter of economics. It depends on the depth of a country’s markets, the rule of law, and the willingness of others to trust that the issuer will not turn its money against them.

Why are you paid in money?

• Power
• Value

Most people meet the monetary system first through a paycheck. It feels like the most natural thing in the world. It is not very old.

In ancient Mesopotamia, dependent workers were paid in rations, and the tablets of Uruk count barley, beer and bread by the month. In Egypt, the craftsmen of Deir el-Medina, who built the royal tombs in the Valley of the Kings, were paid in grain. Sometime in the 1150s BC, when their rations arrived late, they downed tools and staged protests, in what is often called the first recorded strike; a papyrus preserves their complaint that they were hungry. A wage, in its oldest form, was a share of the stores, and a strike was a complaint about the ledger.

Being paid in money, rather than in goods, had to be fought for. During the Industrial Revolution, some British employers paid workers in tokens redeemable only at the company’s own shop, often at inflated prices, a practice known as the truck system. The Truck Act of 1831 required wages to be paid “in the current Coin of the Realm”, and versions of that rule stood for more than 150 years. In the United States, coal towns in Appalachia long paid miners in company scrip. A wage in real money is freedom of a specific kind: the freedom to spend your pay with anyone, not only with the person who owes it to you.

Today, according to International Labour Organization estimates, about 54 percent of the world’s workers earn a wage or salary, up from about 48 percent in 1991. Nearly half the world still works for itself, on farms, in family businesses, in informal trades. In the United States, most private employers pay every two weeks or every week.

Why must that pay be in the state’s money in particular? One answer comes from taxes. A school of economists known as chartalists, going back to Georg Friedrich Knapp in 1905, argue that a currency is demanded because the state accepts it in payment, above all for taxes. History offers a harsh illustration. In parts of colonial Africa, authorities imposed “hut taxes” payable in colonial currency, which pushed people into wage labor to earn it; in Sierra Leone, a hut tax provoked an uprising in 1898. Mainstream economists debate how far the chartalist view explains modern money, but the historical link between taxes, currency and wages is well documented.

Finally, a salary is a promise measured over time, and inflation can quietly break it. In the United States, average hourly earnings adjusted for inflation fell about 4 percent between January 2021 and June 2022. By August 2026 they were still slightly below their January 2021 level. Paychecks grew, but prices grew faster.

Can we trust code?

• Memory
• Power
• Value

On October 31, 2008, six weeks after Lehman Brothers collapsed, someone using the name Satoshi Nakamoto posted a paper titled “Bitcoin: A Peer-to-Peer Electronic Cash System”. Its abstract named the problem it meant to solve: digital payments without “a trusted third party”. The first block of the Bitcoin blockchain, mined on January 3, 2009, carries a newspaper headline: “The Times 03/Jan/2009 Chancellor on brink of second bailout for banks.”

Bitcoin had precursors: David Chaum’s digital cash in the 1980s, Adam Back’s Hashcash in 1997, Wei Dai’s b-money in 1998, Hal Finney’s reusable proofs of work in 2004. What it added was a way for strangers to agree on a single ledger without anyone in charge of it.

Read against the history in this essay, Bitcoin is an attempt to rebuild each promise of money in software, and each lens applies.

• Memory. The ledger is public, copied across thousands of computers, and verifiable by anyone who runs the software: the village of Yap, at the scale of the internet.
• Power. No one decides how many coins exist. Issuance halvesHalvingEvery 210,000 blocks, about four years, the number of new bitcoin per block is cut in half. roughly every four years, most recently in April 2024, toward a cap of 21 million. By September 13, 2026, about 95.6 percent of all bitcoin that will ever exist had been mined.
• Value. Scarcity is enforced by code rather than by a mine, a king or a central bank, and issuing new coins once again costs real energy, as mining gold did.

95.6%of all the bitcoin that will ever exist had been mined by September 13, 2026.

The early days had their own legend. On May 22, 2010, a programmer named Laszlo Hanyecz paid 10,000 bitcoin for two pizzas. At September 2026 prices, that was roughly $770 million.

The limits are just as real, and they fall on the tension from Chapter 6. As a store of value, Bitcoin has attracted serious money: US spot Bitcoin ETFs were approved in January 2024, and in March 2025 the United States created a Strategic Bitcoin Reserve from forfeited coins that “shall not be sold”. As a medium of exchange, the evidence is thin. When El Salvador made Bitcoin legal tender in 2021, a study published by the National Bureau of Economic Research found that only 4.9 percent of firms’ sales were paid in it; in 2025, under an IMF program, accepting it became voluntary. As a unit of account, its price is volatile: it reached a record of about $126,000 in October 2025, fell to about half of that in the summer of 2026, and traded near $78,000 in September. People hoard what they expect to rise and spend what they expect to fall, which makes a rising asset a poor everyday currency.

Two more tensions deserve mention. Trust has a way of returning through the side door: many holders now own Bitcoin through funds, exchanges and companies that keep the keys for them. And the network runs on electricity, about 0.5 percent of the world’s electricity according to a 2025 Cambridge report, with just over half coming from renewable and nuclear sources among the miners it surveyed. That energy question, and what it means for countries that turn surplus power into a globally transferable asset, deserves an article of its own.

Critics at the European Central Bank have argued that Bitcoin has failed as a currency. Supporters answer that it was always meant first as hard money, and that seventeen years of operation are proof of something no central bank can offer: rules nobody can change alone. Both can be true. Bitcoin has shown that a money can exist without a trusted issuer. It has not yet shown that such a money can be what people use every day.

Private money or state money: who will keep the digital ledger?

• Power
• Value

While Bitcoin tests money without an issuer, two other forms of digital money test the opposite: money with a very clear one.

Stablecoins are tokens that promise to be redeemable one for one in a national currency, usually the dollar, backed by reserves held by a private issuer. At the end of June 2026, about $73 billion of USDC and around $185 billion of Tether’s USDT were in circulation. That is private money, and central bankers have noticed the echo. The Bank for International Settlements wrote in 2025 that stablecoins are “tagged with the name of the issuer, much like private banknotes circulating in the 19th century Free Banking era”. The United States answered with the GENIUS Act, signed on July 18, 2025 and due to take effect by January 2027 at the latest: issuers will have to hold at least one dollar of safe, liquid reserves for every dollar issued, publish monthly reports, and will not be allowed to lend the reserves out or pay interest to holders. Read against Chapter 10, those rules are the free banking era and the related-lending scandals, turned into law.

Central bank digital currencies would put state money directly into digital wallets. According to the Atlantic Council’s tracker, 146 countries and currency unions were exploring one as of May 2026, but only three, the Bahamas, Jamaica and Nigeria, had fully launched. China’s e-CNY is the largest pilot. The European Parliament voted in July 2026 to open negotiations on a digital euro. The United States went the other way: an executive order in January 2025 barred federal agencies from establishing one. The debate turns on the oldest question in this essay. A ledger kept by the state can be safe and universal. It can also be watched, and in principle programmed to decide what money may be spent on.

In the interest of transparency: Spliz, which publishes this blog, lets groups settle shared expenses in USDC, a stablecoin. Weigh this chapter with that in mind.

What if the group kept its own ledger?

• Memory
• Reach

Look back across this history and a pattern appears. Money has been organized around two scales. At one end, the individual who holds a coin, a note or an account. At the other, the institution that keeps the ledger: the temple, the king, the Exchequer, the central bank, the commercial bank, now the protocol.

Between the two sits the scale at which a great deal of life is actually lived and paid for: the household, the roommates, the friends on a trip, the colleagues organizing a send-off, the family helping one of its own through a hard year. Money has rarely been designed for them. They improvise.

They have been improvising for a very long time. Rotating savings and credit associationsROSCAMembers pay a fixed sum into a pot at each meeting; each member receives the whole pot once, in turn., in which members pay a fixed amount into a common pot and each takes the whole pot in turn, appear across the world under many names: tontines in West Africa, susu in Ghana, tandas in Mexico, hui in China, stokvels in South Africa, ekub in Ethiopia, chits in India. The World Bank described them in its 1989 World Development Report. Most run without a bank or a state, on the same foundation as the stones of Yap: a group that remembers.

The sociologist Viviana Zelizer showed in The Social Meaning of Money (1994) that people treat money socially, even though one dollar is legally interchangeable with any other. We “earmark” it by relationship and purpose: a birthday check is not meant for groceries, the vacation envelope is not the rent. A dollar is never just a dollar. We attach people and rules to it, then keep those rules in our heads because most accounts have nowhere to put them. Joint accounts and online collection pots help, but a joint account usually lets each owner withdraw everything, and a collection pot usually ends up in one person’s account.

One more fact belongs in any honest version of this argument. In the United States, before the Equal Credit Opportunity Act of October 1974, a married woman often could not get credit in her own name. For many people, the individual account was a hard-won freedom. Whatever comes next must not mean going back to money controlled by the head of a household. A group should never erase the individuals in it.

So what would money designed for a group need to be? The history in this essay suggests three answers.

• Social. A group can hold money together around a shared purpose, and every member sees the same record. That is one of the oldest forms of money we know, from the village of Yap to the tanda.
• Sovereign. The people who contribute keep control together, and no single keeper can move the money alone. Chapter 10 is the reason: wherever the keeper of the money can also lend it to itself, some keepers eventually do.
• Programmable. The group chooses rules the money follows, visible before anyone puts money in. Chapter 9 is the reason: even money tied to gold was only as reliable as rules that whoever held the vault could rewrite.

Picture twelve colleagues putting $20 each into a pot for a send-off, $240 for a gift and a round of drinks, where everyone sees what came in and what was spent. Or five friends putting $300 each into a trip pot, $1,500 in all, spent under limits they agreed on, with what is left returning to each of them at the end. None of this removes the need for trust between people. Friends still fall out, and groups can still choose bad rules. What the right tools can do is make the promise visible, and make sure the money keeps the promise the group actually made.

Epilogue. Who keeps the memory now?

David O’Keefe’s stones were bigger, smoother and easier to get than the traditional ones. They were worth less, because no chief had authorized them and they had no history.

The stone at the bottom of the sea was worth something, because a village remembered it together.

That is what money has always been: a promise that a group agrees to remember. Over five thousand years, humans have handed that memory to a series of keepers, the scribe, the king, the vault, the bank, the state, the code, and learned, often painfully, what each one protects and what each one costs.

The question has never gone away, and it is worth asking every time a new form of money appears: who keeps the memory, and why should we believe them?

Sources

• Denise Schmandt-Besserat, the evolution from tokens to writing.
• Robert Englund (CDLI), proto-cuneiform accounts at Uruk.
• The Schøyen Collection, the Kushim barley tablet.
• Eric Cripps, Cuneiform Digital Library Journal (2017), silver and barley values in Ur III.
• George Selgin, Cato Institute, the case for the barter story.
• John Kroll, Sardis Expedition, the first Lydian coins.
• Kevin Butcher and Matthew Ponting, the silver content of Roman coins.
• John Munro, University of Toronto, England’s Great Debasement.
• The Conversation, the salt salary myth.
• Narayana Kocherlakota, Federal Reserve Bank of Minneapolis, “Money Is Memory”.
• Friedrich Hayek, “The Use of Knowledge in Society”.
• Leonard Read, FEE, “I, Pencil”.
• BCG and Semiconductor Industry Association, the chip supply chain.
• Maddison Project Database 2023, long-run GDP per person.
• R. A. Radford, Economica (1945), the economy of a POW camp.
• Jan Hogendorn, cowrie shells and the Atlantic trade.
• Bank of Canada Museum, the rai stones of Yap.
• Federal Reserve Bank of New York, the Great Recoinage of 1696.
• Gary Gorton, discounts on US banknotes before 1863.
• Federal Reserve Bank of Minneapolis, the Suffolk banking system.
• Steve Hanke and Alex Kwok, Cato Journal, hyperinflation records.
• Hanhui Guan, Nuno Palma and Meng Wu, Economic History Review (2024), the rise and fall of paper money in Yuan China.
• Bank for International Settlements (2018), cryptocurrencies and trust in money.
• Federal Reserve Board, review of Silicon Valley Bank.
• Money Trust Investigation, Part 15 (1912), J. P. Morgan’s testimony.
• Bank of England (1969), a short history of the banknote.
• Executive Order 6102 (1933), the order to hand in gold.
• Federal Reserve History, the end of gold convertibility.
• Bank of England (2014), “Money creation in the modern economy”.
• Deutsche Bundesbank (2017), the role of banks in money creation.
• Banque de France (2024), how money is created.
• La Porta, Lopez-de-Silanes and Zamarripa, “Related Lending”.
• Bank of Russia (2017), captive banks.
• IMF Working Paper 12/202, the Chicago Plan revisited.
• Swissvotes, the 2018 sovereign money vote.
• IMF COFER, currency shares of global reserves.
• Bank for International Settlements (2022), dollar debt in currency swaps.
• US Bureau of Labor Statistics, pay frequency in the United States.
• Satoshi Nakamoto, the Bitcoin white paper.
• Alvarez, Argente and Van Patten, NBER, Bitcoin use in El Salvador.
• Eric Budish, Quarterly Journal of Economics (2025), the economic limits of proof of work.
• Ladislav Kristoufek, Energy Economics (2020), how bitcoin mining costs follow its price.
• Bank for International Settlements (2025), stablecoins and the monetary system.
• Congressional Research Service, the GENIUS Act.
• Viviana Zelizer, The Social Meaning of Money.