Your goal: Follow money from stone discs and wooden sticks, through the accidental rise and violent fall of the gold standard, to the pure fiat world we live in now — and the MMT argument about what a money-issuing government really can and cannot do.

Felix Martin opens his book with a test he once sprang on a friend over a drink — a successful entrepreneur in financial services. Asked what money actually is, the friend reconstructed, from scratch and without hesitation, the very theory economists have told for centuries. That is how deep the intuition runs.

Start with the intuition almost everyone shares. Direct swapping is clumsy because it needs a double coincidence of wants: to trade, you must find someone who has exactly what you want and happens to want exactly what you have, at the same moment. A baker who needs shoes has to track down a shoemaker who is also hungry for bread today. Money is supposed to be the fix — one thing everyone accepts, so you can sell bread to anyone and hold that thing until you find your shoemaker.

This barter-origin story has an old pedigree: the earliest treatment is in Aristotle's Politics, and Locke and Smith later built on it.

Its logic picks one commodity to serve as a medium of exchange — ideally one that is durable, malleable, portable, and rare, which is why metals kept being chosen.

On this commodity theory, money is fundamentally a thing, and its value flows from the material it is made of. That is why gold — durable and rare — became the archetype, and why Adam Smith himself thought gold and silver the natural monetary metals. Hold that assumption in mind; almost everything later in this course turns on whether it is right.

There is an awkward hole in the story. When anthropologists went looking for the barter economies that supposedly existed before money, they came back empty-handed.

By the 1980s the leading anthropologists of money considered the verdict in: no society has been documented running its everyday trade on barter.

Barter does happen — between strangers, or after a monetary system collapses — but as a stage before money it appears to be a myth. That is awkward for a theory whose whole premise is that money arose to cure barter.

Felix Martin's counter-claim is that money is not a commodity at all but a system of credit and clearing — a running record of who owes what to whom.

On this view the coin, note, or stone is only a token that tracks the underlying web of credit; it is not itself the money.

The credit view has its own distinguished lineage. The monetary scholar Alfred Mitchell Innes put its heart in a single line: “The eye has never seen, nor the hand touched a dollar.” A dollar is a unit of credit — a measure, like a metre or a kilogram — not a thing you could ever pick up.

This is not a small academic quarrel. If money is a thing, its quantity is limited by how much of the thing exists, and a government can "run out." If money is credit, its limits are completely different — a distinction that will decide what you make of the gold standard, and of MMT, by the end of the course.

The best argument for the credit theory comes from the strangest currency ever documented. At the beginning of the twentieth century, an American visitor found a tiny Pacific island using money that no commodity theorist could love.

On the island of Yap, in the Caroline Islands, the currency was the fei: huge carved discs of limestone, some taller than a person.

The observer was William Henry Furness, who had trained as a doctor before turning to anthropology, and who at first thought the stone discs bizarre.

What makes the fei devastating for the commodity theory is that the limestone was worthless on Yap — it was not even local. It had to be quarried on other islands and hauled back across open ocean by canoe, so a fei’s worth came from the labour and danger of getting it there, not from any use the stone had at home.

Then comes the detail that settles it. Furness was told of a fei that no living person had ever seen.

One family's fei had slipped off a canoe and lay for generations at the bottom of the sea, yet the whole island still counted it, unseen, as that family's wealth.

Think about what that means. If money were the physical object, a stone rotting on the seabed would be worth nothing. Its value survived because everyone on Yap simply agreed on who owned it. The object was gone; the money remained.

Martin's point: what mattered was never the stone but the shared record of who owned it — money as the island's collective memory of credit.

The account was read by the economist John Maynard Keynes, then at the British Treasury, who judged Yap's money system profound rather than primitive.

That verdict matters: one of the century’s greatest economists looked at stone discs and inconvertible sunken wealth and saw not a curiosity but a clear window into what money everywhere actually is.

And he turned the mirror on us. Martin quotes the famous jibe that for a century the “civilized” world treated as the height of its wealth metal “dug from deep in the ground, refined at great labor, and transported great distances to be buried again in elaborate vaults deep under the ground.” Set beside a stone at the bottom of a Yap lagoon, a pile of gold locked in a vault looks no more like money — and no less.

If money is really credit, why do we instinctively picture coins? Partly an accident of survival: metal lasts, so coins are what archaeologists dig up, while the credit records rot away. But coins have their own revealing history.

The earliest known coins were minted in Lydia, in present-day Turkey, in the early sixth century BC.

They were struck from electrum, a natural alloy of gold and silver.

A coin looks like the commodity theory made solid — value you can bite. But the coin quietly smuggles in the state. Someone has to certify the weight and stamp the metal, and whoever does that gains a lever.

Because a ruler controlled the mint, they could quietly cut a coin's precious-metal content — a trick called debasement — to stretch their own spending.

Debasement is the first hint of a theme that will dominate the whole course: the people who issue money are rarely neutral, and they can change its value for their own ends. Keep that in mind when central banks arrive.

For the purest case that money need not be valuable stuff at all, England ran its royal finances for centuries on notched sticks of wood.

The Exchequer tally was a hazelwood stick notched to record a debt owed by or to the Crown.

The stick was split lengthwise so each party held a matching half, turning the tally into a transferable IOU that could be spent on to someone else.

That last step is the crucial one. Because a tally recorded a debt the Crown would honour, and because it could be handed on, people accepted tallies in payment — a plain wooden stick circulating as money purely on the strength of the credit it represented.

Tallies were officially abolished by an Act of Parliament of 1782; the old sticks were later burned in a House of Lords stove, and that fire spread and burned down Parliament itself.

Coins made money look like a commodity; tallies show it was credit all along. With that settled, we can watch one commodity — gold — climb, almost by accident, to rule the world.

Before gold reigned alone, most countries ran on both gold and silver at once — bimetallism. The catch is obvious once you see it: you are trying to fix a permanent price ratio between two metals whose market values keep moving.

The representative statute was the French monetary law of 1803, which fixed the mint's ratio of silver to gold at 15½ to 1.

That meant the mint would turn either metal into coin at 15½-to-1. It worked smoothly only as long as the market price of the two metals stayed near that official ratio.

Suppose the market moves so that gold is worth more — 16 ounces of silver per ounce of gold instead of 15½. Now there is free money: import 15½ ounces of silver, have the mint coin it, swap it for a gold coin, export that gold, and sell it abroad for 16 ounces of silver. Repeat, and every gold coin drains out of the country.

That drain is Gresham's Law: the metal overvalued at the mint drives the undervalued metal out of circulation.

Small gaps didn't break the system, because coining carried a fee called brassage — in France about one-fifth of 1 percent — which, with shipping costs, left a band where arbitrage didn't pay.

And this was no blackboard exercise. When gold strikes in California in 1848 and Australia in 1851 produced a tenfold surge in world gold, gold cheapened, poured into France to be coined, and French silver drained away to the Far East. Bimetallism had turned an entire country into the arbitrage machine — absorbing the cheap metal and spitting out the dear one on a continental scale.

Gold did not win because someone proved it best. It won by a mistake — made by the most famous scientist alive.

Britain slid onto gold by accident in 1717, when Sir Isaac Newton, as master of the mint, set the price of gold too high against silver and drove full-bodied silver coin out of circulation.

Savor the irony: the man who mis-set the ratio was Isaac Newton himself, and Eichengreen dryly notes that his reputation for brilliance survived it untarnished. The father of modern physics, running the Royal Mint, nudged an empire onto gold by getting a single price slightly wrong.

Britain only formally acknowledged gold later, when silver lost legal-tender status for large payments in 1774 and for small ones in 1821.

Here is the twist that matters: Britain became the workshop and banker of the nineteenth-century world. Once the leading trading nation was on gold, its money became the one everyone else found convenient to match.

Bimetallism persisted, and then gold spread, because of network externalities: sharing your trading partners' monetary system made trade and borrowing easier, so no one wanted to switch first.

It took large shocks — industrialization and Germany's switch to gold after the Franco-Prussian War — to break the bimetallic bloc and tip the world toward gold.

In 1865 Belgium, France, Italy, and Switzerland formed the Latin Monetary Union to harmonize their silver coinage on a 0.835-fineness standard.

Within a few decades the argument was over: the industrial world had tipped decisively onto gold. So how did the gold machine actually run day to day?

First, plainly, what the thing is — because we have been circling it for two modules without a definition. A country is on the gold standard when its government fixes what its money is worth in metal — so many grains of gold to the pound — and stands ready to swap the one for the other at that price, on demand, for anyone who asks. Everything else follows from that single promise. Because every country on the standard makes the same kind of promise, their currencies are pinned to each other through the metal, which is what makes money move predictably across borders at all.

On paper such a system is beautifully automatic. The classic account dates to the philosopher David Hume in 1752, and our guide to what became of it is Barry Eichengreen, the economic historian whose Golden Fetters is the standard modern account of how the gold standard actually behaved — and of how badly it eventually failed.

The textbook model is the price-specie-flow mechanism of David Hume — specie being simply coined metal, gold and silver money: when gold flows out of a country its money supply shrinks, prices fall, its exports get cheaper, and gold flows back — a self-correcting loop.

Two steps in that sentence deserve spelling out, because Hume runs them together and everyone since has copied him. First, why does gold leave at all? Because a country that buys more from abroad than it sells has to settle the difference somehow, and in a world of coin there is nothing to settle it with except coin. The metal sails as payment. That is how the loop starts — not from nowhere, but from buying more than you sell.

Second, why should cheaper domestic prices be cheaper to a foreigner? This is where the fixed parity does its work. Both countries have pinned their money to gold, so they are pinned to each other: a shilling is worth a fixed number of francs, and nothing about a fall in British prices changes that. So when the price of English cloth falls in shillings it falls in francs too, and a French buyer sees a genuine bargain. Had the shilling instead been free to slide against the franc, it might simply have slid far enough to cancel the whole effect — a first hint of why fixing money to metal mattered so much.

And the gap opens from both ends at once, because the gold that leaves has to arrive somewhere: with less coin circulating at home prices fall in the deficit country, while with more coin circulating abroad prices rise in the surplus country.

Two things follow that are easy to skate past. Since it is the whole price level moving, wages move with it: a general fall in prices is a fall in what employers pay as much as in what shoppers pay, and that is what will make this medicine so bitter to swallow when we watch a country actually take it. And the correction works on both sides of the ledger at once — as home-made goods get cheaper, foreign ones look correspondingly dear, so the country does not merely sell more abroad, it also buys less from abroad.

Notice, too, that nothing in the argument depends on which way the imbalance runs. Take a country in surplus and turn the whole thing around: gold arrives, money expands, prices and wages climb, its goods price themselves out of foreign markets while foreign goods start to look cheap — and the surplus erodes exactly as the deficit did. Under Hume’s machine neither position is stable. Both are pulled back toward balance, which is precisely what makes it a machine rather than a story about one unlucky country.

The machine rests on a stark simplification, Eichengreen notes: in Hume’s world only gold coin circulated and the role of banks was negligible — so every imbalance is settled in metal, and metal is the only thing that moves.

It is worth pausing on the author. David Hume — the great Scottish philosopher — sketched this mechanism in 1752, and Eichengreen remarks that its most striking feature is sheer durability: nearly three centuries later it is still how the gold standard gets taught on the first day.

The lever in the whole story is the money supply: gold moving in or out changed how much money circulated, and therefore the price level.

The step to watch there is the one from money to prices, because Hume takes it for granted and so does everyone who repeats him. The thought is this: if there is less money in circulation but the same quantity of goods to buy, then each unit of money must buy more than it did — which is the same as saying prices, measured in money, have fallen. That is the engine of the whole loop. Everything downstream — cheaper exports, a corrected trade balance, gold coming home — follows from prices moving, so if that step fails nothing else in the machine turns.

If you have been paying attention through Modules 1 to 3, something here should be nagging. You were shown at length that money is not the metal — the fei at the bottom of the sea, the notched stick, the debased coin that still passed. Now gold leaves the country and the money supply is said to shrink with it. Has the course quietly changed its mind?

No — and the place where it does not is worth being exact about, because the whole module turns on it. That gold leaving squeezes the money supply is not in dispute; both theories agree on the direction. What they disagree about is how tight the link is. In Hume’s world money simply is the coin, so the money supply falls by exactly the weight of gold that sailed: an identity, automatic, with nobody deciding anything. Put a layer of bank credit on top and it stops being an identity and becomes a decision — the bank watches its reserve fall and chooses how hard to pull in its lending, which is a question of judgement, and of how much spare reserve it was sitting on.

And they sat on a great deal. The law set only a minimum reserve, and nothing stopped a central bank holding more — so when gold was presented and shipped abroad, it no longer followed that the money supply had to fall by the amount of the gold losses, as it would under a textbook gold standard. The bank could simply absorb the loss and leave the money alone.

Hold onto that, because it is the answer to the question this whole course keeps circling. If money were really the metal, its quantity would be chained to the hoard and no one could do anything about it. Because money is credit, the quantity is a thing a bank can hold steady while gold walks out of the door — and the moment that is true, someone has a choice to make. Everything from here on, the discount rate included, is about who makes that choice and how. Hume’s machine has no such person in it, which is exactly why it describes a world that never quite existed.

There is a problem, and Eichengreen presses on it. Notice what kind of gold flow Hume’s machine requires: a large, sustained, trade-driven one. Take those three words in turn. Trade-driven, because in his world metal moves for one reason only — to pay for goods bought abroad; money moving to chase a better return somewhere else is a thing his model does not contain at all, and it will turn out to matter enormously. Large, because the model routes the whole imbalance through metal: if a country runs a deficit of a million pounds, a million pounds of gold is what settles it, so the shipments should be the size of the imbalances themselves. And sustained, because the draining does not stop when it starts to bite — it must continue until prices have fallen far enough to close the gap, and prices are slow, stubborn things. If balance were really kept that way, the world should have seen enormous, constant shipments of the metal. It did not: the gold that actually crossed borders came, in Eichengreen’s assessment, to “but a fraction” of the trade deficits and surpluses it was supposed to be settling — while the flows of investment money across borders were often substantially larger than the trade balances themselves. (Economists call the running tally of what a country owes the rest of the world against what the world owes it its balance of payments; settling that tally is the job gold was supposed to be doing.)

In practice those large trade-driven shipments mostly failed to appear: much balance-of-payments adjustment was achieved in the absence of significant gold flows at all.

Two different things are being claimed in those last two sentences, and it repays pulling them apart. To settle an imbalance is simply to pay for it: somebody must hand over value for the goods that came in. To adjust it is to correct the underlying gap so that it stops recurring — which in Hume’s telling is the work the falling prices do. His machine fuses the two, and that is its elegance: the very gold that settles the account is the gold whose departure shrinks the money supply and moves the prices. One flow, both jobs.

Prise them apart, though, and you can have either without the other — which is exactly what happened. Settlement drifted onto paper: claims were set against each other and cancelled, in the clearing Module 1 described, so that only a small remainder ever needed paying in metal. And adjustment stopped being automatic and became something somebody did, with a deliberate hand on a lever we have not met yet. Neither job any longer required much gold to move, which is why the shipments Hume’s model predicts are missing from the record without the imbalances themselves going uncorrected.

Part of the confusion is that two different models get run together. Hume’s bare 1752 version had no banks in it at all — only gold coin. The version that later hardened into the textbook ideal quietly added them, imagining central banks dutifully reinforcing every gold flow, moving in lockstep with the metal. It is that obedient picture — not Hume’s original — that turns out to be a retrospective ideal more than a record of what bankers actually did.

That idealized picture was later christened playing by the rules of the game, even though there was never any actual rule book.

So the self-correcting machine, as Hume drew it, is not what actually kept the gold standard in balance. Something quieter did — and to see it we need two things Hume’s model simply does not contain.

The first is the real tool. Central banks did not sit back and let gold flow; they steered, using an instrument called the discount rate — subtler and more powerful than shipping metal, and almost universally misunderstood. The next module takes it apart.

The second is the institution. That tool needed a hand to wield it and a reputation to make it credible — the central bank, which over the nineteenth century grew from a private bank into the guardian of the entire system. That strange history is Module 9. And the deepest answer of all, we will find, is not mechanical but psychological: the machine ran, in the end, on belief.

Module 6 promised that seeing how the real machine worked would take two things: the tool central banks steered with, and the institution that wielded it. Here is the tool — and our guide to it is the man who described the Victorian money market better than anyone.

The raw material was the bill: a merchant’s written promise to pay a fixed sum on a fixed future date — commonly sixty or ninety days out — the everyday IOU on which English trade actually ran.

A merchant who had sold goods held a stack of these promises, each worth (say) £100 in ninety days. But he often wanted cash now. So he took the bill to a bank, which handed him cash for it — a little less than £100 — and kept the bill.

Handing over cash for a bill before it matured was called discounting it — and, crucially, it was a purchase, not a loan: in Bagehot’s words the bill “becomes our property … we keep it and lock it up until it falls due.”

So whoever discounts a bill is a buyer, not a lender: the bill becomes the buyer’s property, and its original signer — the merchant — stays the one who must pay at maturity. The buyer has simply swapped cash now for a larger sum later.

And discounting was no rarefied act: it was the daily trade of the whole money market. Country banks, London banks, and the specialist bill brokers all did it, buying and re-selling the same bills up a chain, each taking a small turn of profit.

One more thing about that chain, because Module 9 leans on it. The merchant remains the one who owes the money — but a house that re-sells a bill signs its own name on the back, an endorsement, and thereby stands behind the bill too: if the merchant fails to pay at maturity, the holder can come after the signatories instead. This is not a technicality. It is what made a stranger’s bill worth buying at all — you need not know the merchant in Bombay, only the London house whose name is on the back — and it is why every re-sale made the paper safer to hold.

An endorsement is a real obligation, not a formality: Bagehot's bank balance sheets count the bills a house has put its name to among its liabilities, alongside its deposits and its note circulation — money it may yet have to find.

So a bill carried a row of names, each of them good for the money. That is what made the market deep — and it is also, as Module 9 will show, how one failure could reach right across it.

Now the piece that trips everyone up. Whenever a bill is discounted, the “rate” and the “price” are two readings of one deal, not two separate things. Once the bill’s face value and maturity are fixed, naming the price fixes the rate, and naming the rate fixes the price. They are not the same number — the rate is a yearly percentage, the price is the cash you hand over today — but each pins the other down. That yearly percentage the buyer earns has a name of its own: his yield. One caution, because the market did not quote the yield. A discount rate is reckoned against the bill’s face value — the £100 it will pay — while the buyer’s yield is reckoned against the smaller sum he actually laid out. So the yield always runs a shade above the quoted rate. The two move together and nothing below turns on the gap, but they are not the same number.

Everything so far is just the private money market: dealers buying and re-selling bills, each deal struck at its own price. Now add the one player that changes everything. The central bank was the biggest and most dependable buyer in that market, and it did what no ordinary dealer could: it posted a single rate and stood ready to buy any eligible bill at it, in any quantity, all day long.

Eligible meaning what, exactly? The Bank would not buy just any paper. A bill qualified if it fell due soon — a matter of weeks or a few months, not years; if it had arisen from a real trade rather than being written up to raise cash; and if the names signed on it were good ones. Short, real, and well-signed. Keep those three tests in mind: they decide who the Bank can rescue, and that turns out to matter enormously.

Bagehot describes this as the biggest dealer in a market naming its terms and the rest having to trade around them. He puts it in selling language — the Bank of England “lays down the least price at which alone it will dispose of its stock” — and it is worth pausing on why that is the same act as buying bills. The Bank’s stock in trade is cash, and discounting a bill is selling cash for it. Naming the least price at which it will part with cash and naming the discount rate are one utterance, seen from the two ends.

That is the duality of the last section doing its work at scale: one posted rate is a standing price for every eligible bill in the country at once.

But posting a price is not the same as making the market obey it. Why should the rest of the market heed the Bank’s rate rather than trade wherever it likes? Because that standing offer to buy sits behind every private deal as a fallback — an outside option that clamps the price from both sides.

The seller’s side sets a floor. No one will sell a bill for less than the Bank would pay, since he can simply carry it to the Bank instead — so a buyer cannot lowball him to grab a fatter yield.

The buyer’s side sets a ceiling, by a different route. You will not overpay for any one bill, because eligible bills were abundant — every trade in the country was drawing fresh ones, so a seller asking too much was simply passed over for the next. And the market as a whole could not float its prices above the Bank’s for long, for a reason worth spelling out. A discount house lived on turnover, not on sitting on money, so it held almost no idle cash: what it did not have out in bills it lodged at the Bank, where the whole market’s spare cash accordingly ended up. (That is why Module 9 calls the Bank the bankers’ bank.) A market holding almost no cash of its own cannot buy up every bill on offer. Here is the line from Bagehot that captures the first step of it: “they seldom can get them discounted very much cheaper, for if they did everyone would leave the Bank.” The desertion is his; the rest is the arithmetic of it. A market everyone has piled into has only its own cash to buy with, and that runs out, so the surplus bills come back to the Bank — at the Bank’s rate.

That is the machine, complete: a market in short-dated promises, a single posted rate that prices every one of them, and a bank whose standing offer the rest of the market cannot escape. Stand back and notice what it hands whoever holds the lever — the power to make credit dear or cheap across a whole economy, in a morning, without asking anyone’s permission.

The next module puts that power to work on the job it was famous for: holding a currency at its promised value in gold.

Module 7 built the lever and left it sitting inside one country: bills, dealers, a posted rate in London. The gold standard, though, was an arrangement between countries — so to connect that lever to Module 6’s gold we have to step across a border, and that needs one piece of vocabulary the course has so far taken for granted.

The exchange rate is simply the price of one country’s money in another’s — how many francs a pound fetches. Like any price it is set by supply and demand: when foreigners want pounds, they must buy them, and buying pushes the price of a pound up. Hold that: demand for a currency lifts its exchange rate. It is the hinge of everything that follows.

Now the part that sounds like a contradiction, and is the key to this whole section. The gold standard is famous for fixed exchange rates — yet we are about to watch exchange rates move. Both are true, because the pound and the franc were never pinned directly to each other. Each was pinned to gold, and so to each other only at one remove — and there is play in that joint. Here is where the play comes from. Anyone needing to turn pounds into francs has a second route besides the market: buy gold at the Bank’s fixed price, ship it to Paris, sell it at France’s fixed price. That route pays a fixed amount, nailed down at both ends by the two parities, so it becomes worth taking only when the market has turned bad enough to offer less. And it does not become worth taking the moment the market dips. If the pound sags a little against the franc, you cannot profit by turning pounds into gold, shipping the metal to Paris and selling it for francs, because freight and insurance eat the difference. Only once the sag is bigger than the cost of moving metal does that trade pay.

So the exchange rate was not rigid but free to drift inside a narrow band, whose edges — the points where shipping bullion finally became profitable — were called the gold points. Inside them, imbalances were settled in paper claims and the metal sat still.

The gold points are the walls of the cage; most of the time the system paces about inside them without touching the bars. That narrow freedom is not a flaw in the fixed rate — it is the room in which the discount rate does its work.

A promise to convert is only as good as the metal behind it, and the bank holds only a finite reserve. Pressure builds on it whenever the country pays out more abroad than it takes in, or investors move money somewhere it earns more: all of that is selling pounds, and selling pushes the exchange rate down. Most of the time the rate simply drifts a little inside the band and not an ounce stirs. But drive it all the way to the gold point and the pressure stops being a price and becomes cargo — metal itself starts leaving the vault. That is a drain; let one run far enough and the promise cannot be kept: the bank suspends convertibility, the parity is gone, and that is how countries fall off the gold standard altogether.

Defending the parity means arresting that drain before it gets there. This is where the lever comes back to Module 6’s gold: raising the discount rate arrested the drain through two channels at once — and Bagehot names both in a single breath.

The fast channel is financial, and it runs through the deposit rate. When the discount rate rises, every short-term rate is dragged up with it — including, as you just saw above, what banks pay on deposits — so money parked in London now earns more, and foreign capital moves in to collect it.

But be careful about what physically arrives. Mostly it is not gold. The incoming investor buys sterling deposits and bills — claims, not metal — and that wave of buying lifts the exchange rate. The rise is what stops the drain, and the reason repays a moment’s thought, because it runs opposite to the obvious guess.

Stand in the shoes of someone holding pounds who wants francs. He has two routes. He can sell his pounds on the market and take whatever rate it is offering — a variable payoff. Or he can turn them into gold at the Bank’s fixed parity, ship the metal to Paris and sell it at the French parity — a fixed payoff, since both ends are pinned by the two parities, less the freight. The gold route cannot pay him any better than that ceiling, however the market moves.

So the gold route only wins when the market is offering less than it does — which is to say, when the pound has sagged. A dear pound does not tempt him into gold; it does the reverse, because selling on the market has just become the better of his two options while the gold payoff sits nailed where it was. That is why lifting the exchange rate stops a drain: it pulls the pound away from the very sag that made shipping metal worth doing, the arbitrage stops paying, and the bullion stays in the vault.

And if the rate climbs far enough — past the parity by more than the freight — the trade runs the other way: now a Paris holder of francs who wants pounds finds it cheaper to ship gold to London than to buy sterling on the market, and metal comes in. That is the sense in which gold moves only at the margin: it crosses a border only when the exchange rate is driven right out to one of the gold points, and the rate’s whole job is to keep it away from them. This is why a rate rise can defend the parity while barely disturbing the vault — and it is the concrete answer to Module 6’s puzzle: adjustment achieved with no significant gold flows at all.

The slow channel is the domestic squeeze — the half the tidy model left out. Because dearer discounting means fewer bills bought and less fresh cash released, credit tightens at home; spending and prices fall; cheaper exports and dearer imports turn the trade balance in your favour. That improving balance relieves the pressure on the currency — and if it is ever strong enough to drive the exchange rate out to the gold point, bullion comes your way too. In Bagehot’s words: “a rise of the value of money in Lombard Street immediately by a banking operation brings money to Lombard Street. And there is also a slower mercantile operation … Prices fall here; in consequence imports are diminished, exports are increased.”

So one move did two jobs at once: it defended the parity by pulling capital toward you, and it tightened domestic credit — the external pull and the internal squeeze are two faces of the same turn of the screw.

And notice how unlike Hume’s picture the result is. His adjustment was the metal: large, slow, trade-driven, gold hemorrhaging out until prices had fallen far enough to stop it. The discount rate works before it ever comes to that. At the first sign of a drain the bank raises the rate; capital turns around and comes toward the currency; the exchange rate strengthens back inside the gold points, and the outflow simply stops — usually with little or no bullion having moved either way. So the rate does not replace Hume’s big outflow with an equally big inflow. It heads the outflow off, settles the imbalance in claims instead of metal, and does the rest of the adjusting at home through credit. That is how the discount rate can be both an alternative to Hume’s flows and, at the margin, a lever on gold.

One thing still dangles. Often the bank did not even have to move the rate: capital flowed in on the mere expectation that it would. And that expectation was not faith from nowhere — it rested on a known commitment. Everyone understood that defending the gold parity was the central bank’s overriding priority, so everyone assumed it would raise the rate as far and as long as that defence required. Believing that, investors moved their money in ahead of the rise rather than after it, for a plain reason: the pound was cheap now and would be dearer once the defence began, so waiting meant buying the very same pounds at a worse price — and that pre-emptive inflow steadied the parity before the bank had lifted a finger. Why a mere credible promise could do the work of the whole machine is the deepest part of the story, and it belongs to the institution that made the promise — which is where the next module ends up, though it starts somewhere that will look unrelated: with banks collapsing.

A gold currency does not manage itself, and a credit economy stacked on top of it is a serial cripple without a guardian. The nineteenth century learned this the hard way, panic by panic.

Britain was rocked by banking crises in 1825, 1837 and 1857, and the United States by panics in 1837, 1857, 1896 and the great crash of 1907. The telling difference was not how often they came but who was there to stop one: Britain had an institution that could act, and the United States, with no central bank at all, had nobody.

Why is a banking system panic-prone in the first place? Because of one structural fact, and it is worth naming plainly.

A bank takes in deposits repayable on demand but keeps only a fraction of them as cash, lending the rest out — fractional reserve banking. Its promises to pay are therefore always larger than the cash it holds to honour them.

In calm times that is harmless, because depositors never all come at once. But it means a bank can be perfectly solvent — its loans are good — and still be destroyed by a demand for cash it cannot meet today: its loans are sound but do not fall due for months, and in a panic nobody will buy them at anything like their worth. Worse, the fear is self-fulfilling: if you think others will withdraw first, your rational move is to run too. That gap between promises payable on demand and reserves actually held is the permanent crack in a credit economy, and it is what a guardian exists to cover.

Contrast that with being handed a coin — and be careful about what the contrast is and is not. It is not that a coin is real money while a deposit is only a promise; Modules 1 and 3 spent their length dismantling exactly that idea, and a coin takes its worth from being accepted, not from its metal. The difference is narrower and sharper. Holding a coin, you hold no claim on any particular institution: there is nobody whose failure wipes you out, and no queue you might be too late to join. The coin-holder is not safe — a ruler can debase the coinage under him, as Module 3 showed — but debasement is slow, falls on everyone alike, and nobody gains by rushing to spend first. A deposit is the reverse: a claim on one bank that can fail, where being early is everything. That is what makes a run a race — and races are what a guardian exists to stop.

And a run does not stay where it starts, which is the fact the rest of this module depends on. Two things carry it outward. The first is plain exposure, and it works through the endorsements of Module 7. Bills travelled by being bought and re-sold up a chain, and every house that passed one on had signed it and so stood behind it, so a great firm’s signature ended up on paper in drawers all over the City. Let that firm fail and its guarantee is worthless: every holder relying on that name takes the loss at once — and, worse, nobody can see who is holding how much of it.

The second is inference, and it is the more dangerous. A depositor cannot look inside a bank. If a house everybody believed unshakeable turns out to be rotten, the reasonable conclusion is not “that one was unlucky” but “I have been wrong about how safe these places are” — and because withdrawal is a race, you need not even believe your own bank is unsound. It is enough to suspect that your neighbours might, and that they will move before you do. That is how one failure becomes a general panic, and why a single house going down could threaten the lot.

At which point a fair objection: why can the bank not simply sell something? Module 7 showed a market of country banks, London houses and brokers buying paper all day long, with a central bank behind it standing ready to take any eligible bill in any quantity. Why does that market not rescue the solvent bank that is merely short of cash today?

Three reasons, and they compound. Most of what a bank owns is not eligible paper at all — eligible meant short, real and well-signed, remember, and the bulk of a bank’s assets are ordinary advances to ordinary customers meeting none of those tests. What is eligible can indeed be sold; the rest cannot. Second, the buyers in that market are the very houses now scrambling for cash themselves — at the moment everybody needs to sell, nobody is left wanting to buy. And third, the good names that made paper acceptable are exactly what is now in doubt. The market of Module 7 works beautifully right up to the moment it is needed, and then it is not there.

Which is the case for a guardian in a single line. What a panic requires is a buyer whose ability to pay does not depend on the market having any cash left. Every private house must find its cash before it can spend it; there was one institution that did not, because what the market treated as cash was largely its own notes, which it could issue — a monopoly this module comes to shortly. It could create the means of payment rather than having to find it. There is only one institution in the country that can do that.

Each panic taught the same lesson: a credit economy needs an institution to hold reserves and stop the run — a role taken on by central banks, usually privately owned and only gradually public.

Pause on that phrase — “privately owned.” It is the oddest fact about the whole institution. We keep calling the central bank the guardian of the entire system, yet for most of its life it was a private, profit-making company — and no one ever quite decided it should be the guardian at all.

In return for that loan the Bank received a monopoly on issuing bank notes and the running of the government’s debt. That was the original bargain of central banking: private capital lends to the state, and the state grants privileges.

It is worth being clear what a bank note actually was, because it will matter greatly later. In this period a note is not yet ordinary money in our sense: it is a piece of paper by which the issuing bank promises to pay the bearer, on demand, a stated weight of gold. A monopoly on issuing them is therefore a licence to put your own promises into circulation as the country’s currency — and it means every note in the land is a claim on the Bank’s gold. Which is precisely why Parliament eventually put a legal limit on how many of them there could be.

That limit was the Bank Charter Act of 1844 — Peel's Act — which tied Britain's note issue to the gold actually in the vault. Note how much stricter that is than the minimum-reserve rule of Module 6: a minimum puts a floor under the reserve and leaves the bank free to hold more, while Peel's Act capped the notes themselves. Britain had bound its own hands tightest — a legal ceiling that bites hardest in the next module, where the Bank has to rescue the banking system and honour that ceiling at the same time.

From there a public role crept up on it. Because it banked the government and its notes were the most trusted, every other bank came to keep its reserve with it — until, in Bagehot’s words, on the directors of “that one Joint Stock Company” — a company owned by shareholders, exactly as an ordinary business is — hung “whether England shall be solvent or insolvent.”

And here is the rub Bagehot could not stop worrying about. This guardian of the nation’s money answered to private shareholders, who at “almost every meeting” pressed for a fatter dividend — which meant lending money out, not letting gold sit idle as a reserve for everyone else. A profit-seeking firm had been saddled with the duty of protecting the whole system, its two roles quietly at war.

Strangest of all, nobody had ordered it to take the job on: the great public duty, Bagehot noted, “was cast upon” the Bank, yet “no distinct resolution of Parliament” ever required it. Guardianship had simply accreted onto the largest private bank in the land.

Notice the thread back to Module 7, and the order of it. Dominance came first: the Bank was the largest dealer, the government’s own banker, the issuer of the most trusted notes. Because of that, the other banks lodged their reserves with it — which is what makes it the bankers’ bank of this module’s title, the bank at which the other banks themselves keep their money — and that, in turn, is what gave its posted rate the grip Module 7 described, since a market whose spare cash sits in your vault must come back to you on your terms. Public authority grew out of private dominance, not out of any design.

The cure was named by Bagehot — whom we met in Module 7 anatomizing the discount market. He saw that the financial stability of London — and, he added, of the world — rested on the Bank of England, run by a board he tartly called “quiet serious men … (who) have a good deal of leisure.”

Be precise about that third clause, because the word “penalty” is doing exact work. A penalty rate means a deliberately high one — the Bank lifts its posted rate well above the easy level of normal times rather than meeting the panic with cheap money. (Module 7 showed the Bank’s rate leads the whole market, so this is not the Bank undercutting or overcharging against some separate market rate; it is the Bank dragging the price of money up for everyone, itself included.) That is the whole anti-abuse device. If the Bank rescued at cheap rates, banks would run permanently thin on reserves and lean on the rescue as ordinary funding, since it would be their cheapest source of cash. Setting it that high means no one touches it while private money can be had at all: you go to the lender of last resort only when there is genuinely no other resort. The help is always there, and always expensive.

This is what it means to act as lender of last resort: lending freely into a panic to halt it, at the cost of the bank's own reserves.

And notice how the Bank pushed that cash out of the door — it is the machine of Module 7, thrown wide open. Ordinarily the Bank bought eligible bills at its posted rate. In a panic it went on buying, now taking in paper that no one else in the market would touch — still perfectly good paper, note, but paper nobody else at that moment had the cash or the nerve to take, which is the difference between an asset that is unsellable and an asset that is bad — at that deliberately punishing rate; alongside this it made advances — straight loans — against good securities, meaning sound assets pledged and returned, which is lending in the strict sense. Both counters were open, and the Bank’s own account of how it broke the 1825 panic — “by every possible means and in modes we had never adopted before” — describes using them together rather than choosing between them. So “lender of last resort” names only half of what the Bank did: it was the buyer of last resort just as much — the same window of Module 7, opened to all comers precisely when every other buyer in the market had vanished.

Read the rule again and the real trick appears: the promise, if believed, mostly prevents the panics it insures against. And the mechanism is the one this module already gave you, run backwards. A run is a race — you withdraw because others will get there first. A guarantee that cash will be available to whoever turns up, however late, destroys the advantage of being early; and once being early buys you nothing, there is no reason to queue. The demand never forms, so the reserve is never tested, so the promise costs nothing to have made. That is why belief alone does the work — and it is exactly the material that quietly ran the gold standard.

And here, at last, is the answer to the puzzle Module 6 left open. Adjustment happened with little gold moving not only because the discount rate pulled capital in, but because markets believed a country would defend its parity — so stabilizing capital flowed in on its own, pre-emptively — before a drain had ever forced the rate up at all. (Recall from Module 8 what would otherwise set the rate moving: selling pressure driving the exchange rate down toward the gold point, at which stage metal really would start to leave — so the bank raises the rate to pull the currency back up before it gets there.) The gold standard ran on the same credibility as the lender of last resort: belief doing the work that reserves and gold would otherwise have to.

So the guardian is in place and the panics have an answer. Credit money is elastic — it expands when banks lend and contracts when they stop, unlike a fixed hoard of metal — and that is what makes it both powerful and panic-prone, and what makes an institution willing to lend into a panic indispensable rather than merely useful.

But notice what the answer costs, because the next module is about nothing else. Module 8 said that a rate rise defends the parity by squeezing credit — dearer money means fewer bills brought in and less cash released. Bagehot has just told the Bank that when the banks are failing it must release cash without limit. The two agree about the price of money and disagree flatly about the quantity, and in a crisis both instructions arrive at the same institution at the same moment.

Tighten to save the currency, loosen to save the banks. That is the contradiction the last module closed on, and this one is about nothing else. Module 8’s rule and Module 9’s rule are each sound on their own: raise the rate when gold is draining, lend freely when the banks are collapsing. They arrive at one institution, and in a crisis they arrive together.

It is worth being exact about why those two pull apart, because the link is not obvious. When the Bank lends and discounts freely it puts more of its own notes and deposits into the world — and every one of those is a claim convertible into gold at the fixed parity. The gold in the vault has not grown; only the claims on it have. So the cover behind each claim thins, and anyone who suspects the Bank cannot honour them all has reason to present notes for gold first — draining the very reserve. Two further channels push the same way, and both are reasons the rescue must not be cheap. Rescue money lent at easy rates drags domestic rates down, so capital goes hunting a better return abroad — selling the currency, pushing the exchange rate toward the gold point, and only there turning into metal that actually sails. And the extra money raises domestic prices, worsening the trade balance, which is Hume’s outflow all over again. And on top of all this sat the legal ceiling Module 9 flagged: Peel’s Act tied the note issue to the gold in the vault by statute, so the Bank could not simply print its way through a rescue even if it judged that wise. The one instrument the crisis called for was the one Parliament had capped. Rescuing the banks, in short, manufactures claims on a fixed hoard — and does it under a legal ceiling.

Although — and this matters for everything that follows — the ceiling turned out to be liftable, and everyone knew it.

In the last stage of every panic since the Act was passed — 1847, 1857 and 1866 — Peel's Act was suspended, and Bagehot reports that no such occasion had ever arisen in which it was not suspended: the world, he writes, confidently expects and relies that in all similar cases it will be suspended again.

So the statutory cap was not quite the wall it looks like. It bound the Bank in the early stage of a panic and was lifted in the late one — which means the real constraint on a rescue was never the letter of the Act but the gold behind it, and the willingness of a government to say the word. Keep that in mind: it is a promise resting on expectation, exactly like everything else holding this system up.

Those two names are worth fixing here, because the rest of this module and all of the next turn on the difference. An external drain is metal actually leaving the country — Module 8’s case, where the exchange rate has been pushed to a gold point and bullion sails. An internal drain is domestic: a run, in which the public wants cash in hand, and where the metal need never move at all. The same reserve is at stake in both, but only one of them ships it abroad.

That has a consequence worth facing, because it cuts against what Module 9 said the penalty clause was for. If a bank in a panic will pay nearly any price, a punishing rate deters it not at all — so the penalty cannot be doing its anti-abuse work in the middle of the crisis. It does that work beforehand. A bank deciding, in the calm of an ordinary year, how thin to run its reserves is very much weighing prices, and knowing the rescue will be expensive is what stops it treating the Bank as its cheapest funding. The penalty disciplines the quiet years and protects the gold in the loud ones; it never deters the panicking.

There is still something that prescription does not explain, and it should. If lending freely manufactures fresh claims on a fixed pile of gold, how can doing it at a high rate leave the pile any safer? The answer is that the two halves are aimed at different people. The high rate speaks to foreigners, and to anyone at all deciding where in the world to keep their money: it answers the external drain by making London the place to be. The free lending speaks to a domestic bank queueing at the discount window, and what that bank wants is notes to hold, not bullion to ship abroad. Money created to quiet a panic at home largely stays at home, in tills and under mattresses.

Which needs defending, because two paragraphs ago this module said the opposite about the same money: that a holder who doubts the Bank has reason to present notes for gold. Both happen, and which one dominates is the whole question. The answer is that they are different fears. Presenting notes for gold is what you do if you doubt the Bank; hoarding notes is what you do if you doubt your own bank and merely want cash in hand. A domestic panic is overwhelmingly the second — in 1866 the queues formed at ordinary banks, not at the Bank of England’s bullion counter — and notes will do perfectly well for that. The Bank only faces the first if the rescue is large enough or clumsy enough to put its own soundness in question, which is precisely the line Bagehot’s rule is drawn to stay inside. And note what follows for the price channel above: money that sits in a till is not money bidding prices up. In a panic that channel is slow and weak, exactly as Module 6 warned when it said frightened people hold on to their money and it changes hands more slowly. It is a danger for the months after, not the week of the crisis.

So the two remedies do not cancel each other out. But notice that this is a matter of degree rather than a guarantee, and the module should not pretend otherwise. The rule works while the sums stay within what the reserve can carry and while the promise is believed. Let the panic run big enough, or the reserve be thin enough, or the belief fail, and the arithmetic stops working: the new claims outrun the metal and the bank faces exactly the choice it was trying to avoid. Bagehot describes a way through the bind in favourable conditions; he does not prove the bind can always be escaped. We are going to watch it fail — though not yet, and not in the next module, which is about the four decades in which it did not.

Which leaves an obvious question, and the next module is the answer to it. If the escape only works in favourable conditions, and the contradiction never went away, how did the system carry it from the 1870s to 1914 without once being forced to choose?

Two things kept the choice from ever having to be made, and only one of them was monetary. Partly it was the credibility of Module 8: while markets believed the parity would be defended, the conflict rarely came to a head, because stabilizing capital arrived before the bank had to choose. But there was a second reason, and it was political rather than monetary.

Which should prompt an objection, if you have been reading closely. Module 9 spent a whole section insisting the Bank was a private company answerable to its shareholders, on which no resolution of Parliament ever imposed a duty. If that is so, why should it matter in the slightest who was allowed to vote?

Because a private share register never amounted to independence from the state: on the Continent the government sat inside the privately-owned central bank.

None of which is to say the pressure got through. It mostly did not — and that is exactly the fact that needs explaining, because it is the one that will stop being true.

Eichengreen's verdict is that central banks were well shielded from political pressure but the insulation was never complete — and, decisively, that their capacity to go on defending convertibility rested on there being limits to the pressure that could be brought to bear on them. The shielding was not a fixed property of the institution. It was a political condition, and political conditions change.

And there is a second layer to it. Whether a country was on gold at all, and at what parity, was never the Bank’s decision — that was a matter of statute and government policy. The Bank’s discretion ran to how the parity was defended, not whether. So the question a widening electorate would eventually force was never “should the Bank move its rate?” but the far larger one: should we be tied to this thing at all, at this price, at this cost in jobs? That is a question governments answer, and lose office over.

Which also answers how a British electorate ever reached a private company’s rate decisions: it did not, and did not need to. No voter could instruct the Bank to lower its posted discount rate — Bank rate, as the City called it — and no government tried. What a government could do was end the obligation the rate was being raised to honour — take the country off gold — and that removes the reason for the squeeze at a stroke. The pressure never had to travel to Threadneedle Street, where the Bank stood, at all. It went to Westminster, and Westminster held the switch.

Defending the parity meant raising the rate and squeezing credit until spending fell — and less spending means jobs lost. So it meant unemployment for somebody, and before 1914 the people who paid that price were largely unable to make their objections count: the right to vote was still limited in most countries to men of property, and labour parties representing working men were only in their formative years.

That is the quiet foundation the whole edifice rested on. A government could impose a deflationary defence of the currency — one that works by forcing prices and wages down — because the workers it hurt could not vote it out. Remember this when the system breaks in the modules ahead: what changes between 1890 and the crisis of 1931 — a date that will mean something in Part III — is not the economics of the bind — that stays exactly the same — but the politics around it. Widen the franchise, organise labour, and defending the parity stops being something a government can simply choose to do.

The guardians even guarded each other: when Barings Brothers faced bankruptcy in 1890 over reckless South American loans, the Bank of England organized the rescue — and had itself at times been backstopped by the Banque de France.

Stop on that word guarantee, because it is doing something odd. Why should a mere promise stop a run, when the houses making it were fractionally reserved and just as exposed as anybody? For the reason Module 9 gave: the promise kills the race. No one needs to reach the front of a queue for money that will be there whenever they arrive, so the queue never forms and the cash is never called for.

Which is precisely why banks with thin reserves could afford to make the promise: a guarantee that is believed is one that never has to be honoured. It is Bagehot’s trick again, worked by a syndicate rather than a central bank, and it rests on the same fragile thing — that the names signing it are good enough to be taken seriously. Which is also why it had to be the Governor who organised it, and why the borrowed gold mattered beyond its amount: both were there to make the promise believable. Had the City doubted the guarantors, the promise would have failed and pulled them down with it.

That borrowed gold — the Banque de France lending bullion and the Russian State Bank pledging more, so London could save Barings — was the international solidarity the gold standard leaned on: central banks propping one another up to defend the shared system.

Solidarity is a comfortable word, and it hides two questions worth asking. If gold leaves Paris for London, has the drain not simply moved to France? And why on earth would the French agree to it?

The first answers itself once you see what the gold was for. It was lent, not given — though that is not what settles it, since a loan leaves Paris just as short while it is outstanding. What settles it is that the gold was never meant to be spent. Its job was to sit in the Bank’s reserve and be counted — the bullion did cross the Channel, but it crossed once, to be looked at rather than paid out, and went home again afterwards. That was enough to make the position look strong enough that nobody troubled to test it. What threatened London in 1890 was an internal drain — a domestic run on the banks — and that kind is not a fixed quantity of metal that has to land on somebody. It is a behaviour, and a reserve that convinces people stops the behaviour before any metal moves at all — the internal case, in Module 10’s terms, not the external one. Halt the run in London and the gold need not stir from either vault.

And the arrangement was mutual, with a history behind it: the Bank of England had borrowed gold from the Banque de France once before, in 1839, and returned the favour in 1847, while the Swedish Riksbank had borrowed from the Danish National Bank in 1882.

Behind the favour-trading lay plain self-interest. By Module 9’s logic of contagion, a London collapse would never have stayed in London. French houses held London paper, and by the endorsement rule of Module 7 that paper was only as good as the London names on it. Beyond that, the bills that financed trade between two countries that were neither of them Britain were commonly drawn on London anyway, which is the practical content of Module 5’s remark that Britain had become the world’s banker: not merely that it lent widely, but that other people’s trade was financed on paper signed in London. So a seizure in the London bill market was a seizure in everybody’s credit. Lending metal to steady the centre was simply cheaper than watching the centre fail — and cheap in the same way the guarantee was cheap, because if it works the gold comes home.

Notice what Britain did not do in 1890, though. It never raised Bank rate, and it never lent freely at a penalty: the rescue money came from a syndicate of private banks, and the gold from Paris and St Petersburg. Bagehot’s rule is the escape available in principle; what actually carried Barings through was a whip-round in the City and two foreign central banks willing to lend metal. The bind was real, and in 1890 it was escapable by other means — the sums were small enough, the City cohesive enough, and Britain’s credit good enough that neighbours would lend. Four decades later, in a global depression and with the franchise widened, none of those things held, and the same contradiction tore the system apart.

By 1914 the gold standard felt eternal — money was gold, exchange rates were fixed, and finance flowed freely across borders. Then, in a single summer, it stopped.

When war broke out in 1914, the combatant nations suspended the gold standard and turned to the printing press to pay for the fighting.

This was the credit theory’s revenge in miniature: the instant survival demanded it, governments simply detached money from gold and created it. Four years of industrial war left mountains of debt, inflation, and a world that could not go back to how money had worked before — though it would spend the 1920s trying.

The story of the interwar money wars turns on four central bankers: Montagu Norman of the Bank of England, Benjamin Strong of the New York Fed, Hjalmar Schacht of the Reichsbank, and Émile Moreau of the Banque de France.

They were an unlikely quartet to hold the world in their hands. A contemporary called the Bank of England “the citadel of citadels” and Montagu Norman “the man who governed the citadel — redoubtable”; Norman was so secretive he sometimes travelled under a false name. Benjamin Strong of New York, ablest of the four, would be dead by 1928. Their choices — above all the choice to force money back onto gold — would help turn a hard decade into the Great Depression.

After the war, going back onto gold felt like going back to normal — to stability, respectability, the world before the catastrophe. Britain led the way.

In 1925, as Chancellor of the Exchequer, Winston Churchill returned Britain to gold at the prewar parity of $4.86 to the pound.

The trouble is that Britain’s prices and costs had risen with wartime inflation, so the old parity now made the pound too expensive. British exports were priced out of world markets. The only way to defend the rate was to force domestic prices and wages back down — which meant slump and unemployment. John Maynard Keynes attacked the decision at the time.

Defending the gold parity required harsh deflationary policies — squeezing prices, wages, and jobs — to protect the exchange rate.

This is the heart of Eichengreen’s argument, and the central debate of the course. When the Depression hit, the gold standard became a trap: to stay on gold, countries had to keep tightening into a collapsing economy, exactly the wrong medicine.

Eichengreen argues the gold standard turned into golden fetters: the commitment to gold transmitted deflation worldwide and made recovery impossible until countries broke free of it.

If Module 10 showed central banks holding the system together through solidarity, this module shows the mirror image: how a single failure could tear it apart.

In May 1931, the biggest bank in Austria, the Creditanstalt, collapsed, taking some $200 million of depositors' funds with it.

The detail that made it unthinkable: the Creditanstalt was no fly-by-night house but the largest bank in Austria, with some $250 million in assets, and it was owned by the Rothschilds — the most storied name in European finance. If a Rothschild bank could fall, no one was safe, and depositors across central Europe bolted for the exits.

A failure in a small country should have stayed small. Instead it detonated a chain reaction. Panic spread from Austria to Germany, whose banks buckled next, and then to the pound itself. The same tight financial linkages that made the gold standard work now carried the crisis from one country to the next.

The contagion forced country after country off gold — and in 1931 even Britain, the historic heart of the system, abandoned the gold standard.

This is the flip side of everything in Modules 6 to 9. A system held together by credibility and cooperation is only as strong as the weakest link’s solvency — and when trust broke, the golden fetters snapped one leg at a time. The interwar attempt to rebuild gold was over. The question became: what replaces it?

After a second world war, the victors were determined not to repeat the interwar chaos of rigid gold and beggar-thy-neighbour devaluations. In 1944 they met in a New Hampshire hotel to design money on purpose this time.

At the 1944 Bretton Woods Conference, John Maynard Keynes and the American Harry Dexter White designed a new international monetary system.

The dollar was fixed to gold at 35 dollars an ounce, and other currencies were pegged to the dollar — a currency that, in practice, meant gold.

So gold stayed at the centre, but only the United States promised to convert dollars into it. Everyone else held dollars as reserves and kept their currency pegged to the dollar. To police the system and lend to countries in trouble, the conference created the International Monetary Fund.

Unlike the rigid classical system, Bretton Woods allowed capital controls and occasional adjustments, and it governed the West from 1945 until 1971.

It was gold, redesigned: fixed rates for stability, but with escape valves the 1920s had lacked. For a quarter-century it worked. Then its one hard promise — dollars into gold — came due.

Bretton Woods carried a flaw it could not fix, spotted early by the economist Robert Triffin.

The Triffin dilemma: to supply the world with the dollars it needed for reserves and trade, the United States had to run deficits — but the more dollars piled up abroad, the less believable its promise to convert them all into gold.

By the late 1960s foreign dollar holdings dwarfed American gold. It was a slow-motion bank run on Fort Knox: everyone could see the promise could not be kept if they all asked at once, so it was rational to ask first.

Over a weekend in August 1971, the Nixon administration closed the gold window, suspending the dollar's convertibility into gold.

That was the end. With the dollar no longer convertible to gold, the last thread tying money to metal was cut. For the first time in history the entire world ran on pure fiat money — money backed by nothing but the state that issues it and the credit of those who accept it. Money had, at last, learned to float. Which raises the question the rest of the course answers: what actually is this floating money, and what can a government that issues it really do?

Step back over the whole journey. Yap’s stones and England’s tally sticks already showed that money is really a record of credit. Coins made it look like a commodity, but debasement gave the game away. So why did money become so much more volatile and political under modern capitalism than in the ancient world?

The deep answer is the one Yap first revealed: money is fundamentally credit, a record of who owes what — and credit can expand and contract in a way a fixed lump of metal cannot.

Industrial capitalism ran on exactly this elasticity. Banks lend, and in lending they create new credit money; trade and investment surge on borrowed funds. That is enormously powerful — and enormously unstable, because credit can evaporate in a panic just as fast as it was created.

Elastic credit money is why capitalism needed active managers — central banks — to hold reserves, set interest rates, and act as lender of last resort, in a way ancient coinage never required.

That is the answer to your third question. Earlier money mostly sat there; capitalist money is a living system of credit that must be actively steered, and steering it is inescapably a matter of politics and power. The gold standard was one attempt to put that system on autopilot. When it failed, what remained was money openly run by the state — fiat. And a pure fiat currency forces a blunt question: if money is just the state’s credit, what really limits how much a government can spend?

Now that money is admittedly the state’s own credit, Modern Monetary Theory pushes the credit view to its conclusion. Start with its central move, made by economists like Stephanie Kelton.

MMT insists on the difference between a currency issuer and a currency user: a government that issues its own floating currency is nothing like a household, and cannot involuntarily 'run out' of that money.

A household must earn or borrow dollars before it can spend them. A currency issuer creates the dollars in the act of spending, and taxes them back afterward. So the familiar question "how will you pay for it?" is, on this view, the wrong question — the government is not revenue-constrained in the way a family is.

Kelton makes it concrete with an image from Warren Mosler: the government spends first and taxes afterward, the way a scorekeeper puts points on the board. When a team scores, the scorekeeper does not check whether the stadium has enough points in reserve — the points are created by the act of scoring, entered by keystroke. The Federal Reserve, Kelton says, is the scorekeeper for the dollar, and the scorekeeper cannot run out of points.

Kelton calls the idea that the federal government should budget like a household perhaps the most pernicious myth in economics.

This is where MMT is most often misread. It does not say spending is free.

The true constraint, MMT says, is not solvency but inflation: if spending pushes an economy past its real capacity, prices accelerate — so real resources, not money, are the limit.

And it reframes the deficit itself: because every dollar the government spends and does not tax back is a dollar left in someone else’s pocket, a government deficit is, by accounting identity, a surplus for the rest of us. Whether all this is a profound correction or a dangerous half-truth is the fight we turn to last.

MMT’s reframing is genuinely useful, but it is fiercely contested — and not only by its usual opponents. Three critiques, from three directions, are worth knowing.

The mainstream objection, made by Mankiw, accepts the government’s fiat budget constraint but argues that financing spending by creating money eventually spurs inflation — an inflation tax with real limits, so there is no free lunch.

The exchange-rate objection, pressed by Edwards from the record of Latin America, is that MMT quietly relies on the US issuing the world’s reserve currency; a small country monetizing deficits this way would see its currency depreciate and inflation surge.

Edwards backs the warning with a roll-call. Populist, money-financed spending sprees in Argentina, Bolivia, Brazil, Chile, Ecuador, Nicaragua, Peru, and Venezuela, he notes, all ended badly — runaway inflation, currency collapse, and precipitous falls in real wages. Chile under Allende is his set-piece: the escape hatch MMT points to, he argues, is a reserve-currency privilege that smaller, deficit-prone economies do not enjoy.

The critique from the left, by the post-Keynesian Palley, is that much of MMT is not new relative to established Keynesian economics, and that it understates the inflation challenge of full employment and the political-economy difficulties of using taxes to control it.

That is the whole arc. Money began as credit dressed up as a commodity, spent centuries chained to gold, broke those chains in 1971, and now floats freely as the state’s own credit — which is exactly why the argument has moved from "is there enough gold?" to "how much can the state’s credit safely do?" You are now equipped to follow that argument on its own terms.