Read the first pages of The World Today

Stop for a moment and consider what that means. A committee of a dozen or so people, none of them elected by popular vote, changed in a single week what it cost to borrow money in the largest economy in the world: for families taking out a mortgage, for companies meeting payroll, for the state itself. It passed no law. It pulled a lever: the price of money. Forty years earlier, around 1980, Paul Volcker had pulled the same lever in the opposite direction: he raised rates to 20 percent to break the inflation the United States had been dragging along since the 1970s, at the cost of a brutal recession and a debt crisis in Latin America.

That lever is called monetary policy: everything a central bank does to decide how much money costs and how much of it circulates. Few tools of a modern state move so much while being seen so little. This chapter goes over how it works in normal times and, above all, what happens in a crisis: what a central bank does when the price of money cannot fall any further, what it achieves by doing so (and what it does not), and who should have a hand on the lever.

2

Interest rates: the economy’s thermostat

Before going on, a quick review of the chain. An interest rate is the price of using someone else’s money: whoever borrows 100 and pays back 105 has paid 5 percent to have it for a year. That price reaches people through three links. At the top is the central bank, which lends money to banks and pays them for what they leave on deposit with it, and the price at which it does so is the policy rate. In the middle are the banks, which lend to families and firms at a little more than money costs them, and pay savers a little less. At the bottom are the mortgage, the car loan, a company’s credit line, or the interest on a savings account. When the central bank moves its price, everything below moves after it: not at the same pace or by the same amount, but in the same direction.

For decades a central bank had one main tool: that policy rate. It works like the thermostat in a house. When people and firms spend more than the economy is able to produce, prices start to climb fast; this is what is called an overheated economy. Then the central bank raises rates. Mortgages get more expensive, firms think twice before borrowing to invest, and saving pays better than spending. People spend less and prices calm down. When the opposite happens and the economy cools, with firms closing and people losing their jobs, the central bank lowers rates so that borrowing and spending pay again.

3

Volcker applied this logic in its most extreme form, and his case teaches something the thermostat leaves out. Raising rates makes loans more expensive, yes, but above all it sends a message: the central bank takes inflation seriously and will do whatever is needed to stop it.

Why does a message matter so much? Because inflation feeds largely on what people expect. If everyone believes prices will rise 10 percent next year, workers ask for 10 percent more pay, and firms, which set their prices with an eye on what wages and suppliers will cost them tomorrow, raise theirs just in case. In the end prices rise 10 percent because everyone expected it. If instead people believe the central bank when it says inflation is going to fall, unions moderate their demands, firms hold back, and inflation really does fall. The expectation becomes reality, one way or the other. This is called anchoring expectations. It lies behind the inflation targets that almost every central bank has announced since the 1990s, and it is perhaps the most important concept in modern monetary policy: a large part of a central bank’s job is managing what an entire economy believes is going to happen.

When rates hit zero and the liquidity trap appears

The thermostat has a problem: it cannot go below zero, or at least that is what people believed for a long time. If a bank charged its customers for holding their money, they would take it out in cash and put it in a safe, which charges nothing. So when the policy rate reaches zero, the lever has no more room to move down, just when the economy most needs a push.

4

The United States had already hit that floor in the 1930s, during the Great Depression, when the interest the government paid on its loans of a few months spent much of the decade stuck near zero. Keynes described at the time the extreme case in which pushing the interest rate lower no longer does any good. Japan was the first to hit it again after the war, with the bursting of its bubble in the early 1990s, rates at zero in 1999, deflation (prices that fall year after year instead of rising), and the lost decades.

It was Paul Krugman who, in 1998, revived that idea of Keynes’s to explain what was happening to Japan, and it has been linked to him ever since. It is called the liquidity trap, and one question explains it: what do you lose by keeping money idle? Normally, something. Whoever leaves 1,000 in a drawer gives up the interest they would earn by lending it or buying a bond with it, and that loss is what pushes money to move. With rates at zero, the loss disappears: keeping money pays the same as lending it, which is to say nothing, and it carries no risk. That is why it does not matter how much money is pumped into the system: people and firms keep it instead of spending or investing it.

5

You might ask why, if central bank money is free, banks do not borrow all of it and lend it out. In fact, they do borrow it: between late 2011 and early 2012, the European Central Bank lent banks close to a trillion euros for three years at a minimal interest rate. But a bank does not lend because it has money to spare; it lends because it finds someone who wants to borrow and will pay it back. In the middle of the crisis, families and firms did not want to take on debt, and banks did not trust those who did. A large part of that trillion ended up in government bonds and in deposits at the central bank itself, not in loans to businesses. The money reached the banks and stayed there. It is like pushing on a string. However hard you push, the string does not move forward.

In 1998 it looked like one country’s oddity. Ten years later it was the problem of half the world. After the 2008 crisis, the United States, the eurozone, and the United Kingdom cut their rates to zero or close to it, the economy still would not start, and all of them had to invent something new.

Quantitative easing, or printing money without printing it

That something new was called quantitative easing, or QE. The name is deliberately dull; put plainly (creating money to buy debt), it sounds far more alarming.

6

The mechanism works like this. The central bank buys bonds, mostly government bonds, from the banks and other investors that hold them, and pays for them with money it creates on the spot. It prints no notes: it simply records the amount in the account the selling bank holds at the central bank. It is the same entry a commercial bank uses to create money when it makes a loan, crediting the customer’s account without taking the money from anyone; the difference is that here the one typing in the figures is the central bank.

And what is the point of buying bonds? First, what a bond is: a loan in the form of a piece of paper. Whoever buys it lends money to the state, or to a company, and in return receives a fixed amount every year. That paper can also be resold to someone else, and its price rises or falls depending on how many people want it.

Picture an apartment that rents for 50 a month. If you buy it for 1,000, each month you get back 5 percent of what you put in. If the same apartment, with the same rent, costs you 1,250, those 50 are now only 4 percent. The rent has not changed; what has changed is what you had to pay to collect it. A bond works exactly the same way: the more you pay for it, the less it yields. That is the hinge of the whole mechanism.

7

Now the last step. The policy rate is the price of the shortest loans there are, overnight or one-week loans between the central bank and the banks. A ten-year loan to the government or a fixed-rate mortgage depends mostly on something else: what bonds of those maturities yield. When the state or a company wants to borrow long term, it issues new bonds, and it has to offer roughly what the ones already in circulation yield; otherwise nobody would buy its own. If the bonds on the market yield 5 percent, the new ones will have to pay 5. If they yield 4, 4 is enough. So by buying bonds in huge quantities, the central bank pushes up their price, lowers their yield, and makes long-term borrowing cheaper, even when the policy rate is already at zero. It is a way of continuing to lower the price of money when the usual lever has nothing left to give.

The Federal Reserve launched its first QE program in November 2008. Over six years it bought more than $3 trillion in Treasury bonds and in securities built from bundles of mortgages, and everything it held, mostly bonds, grew from less than $1 trillion to more than $4 trillion. The European Central Bank took longer to act: Mario Draghi launched its purchase program in 2015, three years after his famous promise to do “whatever it takes” to save the euro.

With the pandemic, everything sped up. On top of the Federal Reserve’s purchases, the European Central Bank launched an emergency program of 750 billion euros in March 2020. In three months the Federal Reserve bought almost as much as in the six years that followed 2008.

8

The effects of QE are still debated. Its supporters argue that it prevented a financial collapse in 2008 and a depression in 2020: without that injection, credit would have frozen, solvent firms would have gone bankrupt for lack of financing, and unemployment would have been much worse. Its critics point out that it pushed up the prices of shares, bonds, and housing, benefiting mostly those who already owned them, and that it helped feed the inflationary wave of 2021 and 2022. The experience of these years leaves at least one pattern: QE worked better at putting out a financial fire than at reviving a stalled economy. It was effective in 2008 and in 2020, when banks and markets were about to break and someone was needed to buy. It was much less so in the years in between, when the problem was no longer panic in the markets but that families and firms did not want to spend.

Negative rates and forward guidance

When even QE was not enough, some central banks crossed the zero line. A negative rate means that banks have to pay to keep their money on deposit at the central bank: they are charged for leaving money idle, to push them to lend it. It applies to banks and almost never to customers, because a bank cannot keep billions in cash, while a family can take its savings out of the account. The European Central Bank adopted negative rates in 2014 and the Bank of Japan in 2016. The first to try them had been Sweden’s central bank, the Riksbank, in 2009, and Denmark and Switzerland followed the same path.

9

Negative rates did not cause the disaster some had feared, but they did not revive the economy in any spectacular way either, and in Japan and Europe growth stayed slow. The episode is now closed: the ECB abandoned negative rates in 2022, when inflation came back, and the Bank of Japan left them in 2024.

Another tool born in this era is forward guidance. The central bank announces publicly what it intends to do in the future. If the Federal Reserve announces that it will keep rates low for the next two years, long-term loans get cheaper at once, without it having moved anything. Someone lending for five years works out what money will cost them over those five years; if the central bank assures them that the first two will be cheap, they can lend more cheaply today. It is monetary policy made with words instead of money. And it works, as long as markets believe what the central bank says. If credibility is lost, the words stop having any effect.

The debate over central bank independence

Who should control this lever? It is one of the most delicate questions in modern political economy. In most developed countries, central banks are independent of the government. The European Central Bank, the Federal Reserve, and the Bank of England set rates without any minister being able to give them orders. The main reason is credibility.

10

The case for independence starts from a concrete problem. A government that controlled the printing press would be permanently tempted to cut rates before an election, hand voters a hot economy, and leave the inflation bill for later. History offers many examples of governments that financed their deficits by creating money until they ruined the currency. Taking monetary policy out of politicians’ hands, the argument goes, protects the value of citizens’ money and makes it possible to take unpopular decisions (like Volcker’s rate hikes) that no elected official would dare to sign.

The case against it raises the opposite objection. Monetary policy distributes gains and losses among millions of people: it decides who pays more on a mortgage and who watches their savings evaporate. It is a deeply political decision, and leaving it in the hands of unelected technicians who do not answer to voters clashes with democratic logic. With quantitative easing the objection gets stronger. The central bank does not lend to the state: in the eurozone the treaties forbid it, and in almost every rich country it is not done. But it buys the state’s bonds from the banks and funds that hold them, and if it buys enough, what the state pays to borrow depends on it: while it is buying, debt is cheap; if it stops, the interest rate shoots up. When a central bank props up bond prices this way, and stock prices along with them, it is making decisions that change people’s lives without those people being able to vote on them or reverse them.

11

Independence, in short, is a bargain: it gains credibility, because nobody can use interest rates to win an election, and it loses democratic control, because nobody can vote for whoever sets them. Almost every rich country made that bargain in the 1990s, and the argument reopens every time a central bank buys assets on a large scale, as it did that Sunday in March.

The limits of the lever

Monetary policy is powerful, but it has clear limits.

It can bring inflation down fairly effectively, but reviving a stalled economy is much harder. It is the string again: you can pull it, but you cannot push it. Raising rates almost always works, because if borrowing is expensive, people and firms end up spending less, whether they want to or not. Lowering them does not always work, because cheap borrowing does not force anyone to borrow. The new tools (quantitative easing, negative rates, forward guidance) help when rates are already at zero, but they do not fully replace the usual lever.

12

Its effects are also uneven, and exactly how is disputed. Every move in rates creates winners and losers between those who owe and those who save. QE added another channel, and it works as a chain. When the central bank buys bonds, their price rises, and whoever already held them is richer without having done anything. Whoever sells them receives money in exchange and has to put it somewhere; since bonds already yield very little, it goes into shares or property, and that demand pushes their prices up too. It is the apartment example in reverse: if what a bond yields falls, an apartment that brings in 50 a month is worth more than before, because the alternative pays less. And cheap mortgages let buyers offer more for the same home.

The result is that new money reaches the financial markets first, and whoever owns shares, bonds, or a home sees their wealth grow. Whoever owns none of that, whoever rents or keeps their savings in an account, barely notices, and when they do notice it is because the home they wanted to buy has become more expensive. In housing, cheap money came on top of a shortage of supply. But the same cheap money sustains employment, which is the main source of income for people who own no shares. If the first effect weighs more, monetary policy widens inequality; if the second does, it narrows it. And the studies do not agree on which of the two wins.

13