Why does every country have a central bank, and why 2% inflation? — History, Chapter 3

The one-way ratchet in your groceries

Follow your own prices across a decade and a pattern shows up that feels less like economics than like weather. The rent, the groceries, the coffee, the haircut, the bus fare: nearly all of it drifts upward, a little each year, and it almost never drifts back. The exceptions stand out precisely because they are exceptions. The television that cost a month’s wages now costs a weekend shift, and the software that came in a box is free, while a semester of college and a night in a hospital have raced away from everything else. Why those particular things got cheaper while others climbed is a question with its own home, in the chapter on why healthcare and schooling outrun televisions (Frameworks, on cost disease). This chapter takes the more basic half of the puzzle: why the general level of prices drifts up at all, year after year, and why that drift is not an accident or a failure but the declared aim of the institution that manages a country’s money.

The previous chapter ended on a promise: that money is mostly created by banks when they lend, and that some institution stands over the banks and steers the total. That institution is the central bank, and it steers toward a slow, deliberate rise in prices, most often 2% a year. Both halves of that arrangement, the standing behind the banks and the aiming above zero, were built out of hard experience rather than pure theory, and both are worth taking in order.

Why money needs a backstop at all

Start with why a country needs a central authority over money in the first place, because the answer explains the institution’s oldest and least controversial job. A banking system that creates spendable deposits by lending is efficient in calm and fragile in fear. The deposit in an account is a promise the bank will pay cash on demand, but the cash to honour every promise at once does not exist, because the deposits were lent out to make them (Chapter 2). In normal times almost no one asks, so the arrangement works. Let a rumour spread that a bank is unsound, though, and every depositor has reason to demand cash first, before the till runs dry. The rush that follows can empty a bank that was perfectly solvent the day before, and because banks lend to and hold deposits with one another, a run on one can become a run on all.

In a panic like that, what the system needs is someone able to supply cash without limit against the assets banks actually hold, so that a liquidity scare does not turn into a wave of failures. No ordinary bank can play the part, because the whole trouble is that no ordinary bank keeps enough cash idle. The role falls to an institution that can create the money to lend, which is to say a central bank, and this is its founding function: the lender of last resort. The rest of what central banks do, including the steering of inflation this chapter turns to, was layered on top of that backstop over time.

The layering shows in the founding dates. The two oldest central banks were not born from panics at all. Sweden’s Riksbank, the oldest, opened in 1668, assembled by parliament from the wreckage of a failed note-issuing bank whose collapse had burned the public. The Bank of England followed in 1694, chartered not to rescue anyone but to lend its founding capital to a king, William III, who needed to fund a war against France. Only later, across the financial crises of the nineteenth and twentieth centuries, did the crisis-backstop role become the thing a central bank was chiefly for. The United States assembled its central bank last among the great economies, and did so directly out of a panic.

Figure 3.1 The central bank assembled over three centuries, shown in order rather than to scale. The two oldest were founded for other reasons, a failed bank and a war, and only grew into backstops later; the doctrine and the clearest crisis-born institutions came with the financial panics of the nineteenth and twentieth centuries. The through line is that a country tends to build or empower its central bank in the wreckage of a crash. Dates: Sveriges Riksbank 1668 and Bank of England 1694 (Bank of England Museum); Bagehot’s Lombard Street, 1873; US Federal Reserve Act, 1913 (Federal Reserve history); Bank of Canada Act, 1934 (Bank of Canada). Placed in sequence; horizontal spacing is even, not proportional to time.

The American case is the cleanest illustration. In the Panic of 1907, with banks failing and no public authority able to supply cash, the financier J. P. Morgan personally organised a rescue, locking bankers in his library until they agreed to pool funds and stop the bleeding. That a private banker had to play the part a state institution plays elsewhere was taken as a national embarrassment, and it set off the process that ended in the Federal Reserve Act of 1913. Canada built its central bank the same way a generation later, chartering the Bank of Canada in 1934 in the depths of the Great Depression, when the existing banks were widely blamed for deepening the slump. The institution that now steers your prices was, in most countries, the settlement after a financial disaster.

Model

Bagehot’s rule for a panic

The doctrine that still governs a central bank in a crisis was set down by Walter Bagehot, editor of The Economist, in Lombard Street (1873), written after the London panic of 1866. His prescription has three parts. In a panic the central bank should lend freely, in whatever quantity frightened firms demand; it should lend only against good collateral, the sound assets a solvent firm can pledge, so it rescues the illiquid rather than the bankrupt; and it should lend at a high, penalty rate, so that only those who genuinely need the help take it, and they give it back once calm returns. Lend freely, against good security, at a high rate: the backstop that lets a banking system survive a scare without a government having to guess which banks deserve to live.

W. Bagehot, Lombard Street: A Description of the Money Market (1873); the doctrine is still cited by central banks describing their lender-of-last-resort function.

Why not zero, then

A central bank that can create money to rescue the banks can also create too much of it, and the same institution that stops panics is trusted to keep the quantity of money in a sensible relation to the quantity of goods. Here the modern central bank does something that surprises people when they first notice it: it does not aim to hold prices steady. It aims, quite deliberately, for prices to rise at a slow and predictable pace, and across the rich world the chosen pace has converged on about 2% a year. That convergence is why the ratchet in your groceries is so uniform. It is chosen, and it is chosen at roughly the same setting almost everywhere.

The number is precise enough to look as though it fell out of an equation. Its actual origin is more human, and worth knowing, because it separates two questions that get run together: where the specific figure came from, and why the target sits above zero at all.

Ledger

Where the 2% came from

The world’s inflation target was born in New Zealand, and not in a model. In an interview on 1 April 1988, with inflation finally falling into single digits, the finance minister Roger Douglas said on television that he would like to see it come down to “around zero to one per cent.” The remark was partly a device to stop the public expecting the anti-inflation drive to ease off, but it stuck. The Reserve Bank picked up the idea, the Reserve Bank of New Zealand Act of 1989 made price stability the bank’s legislated goal, and the first Policy Targets Agreement, in 1990, wrote down a band of 0 to 2%. The governor, Don Brash, was to reach it by the end of 1992, and got there a year early. From there the target spread: Canada in 1991, the United Kingdom soon after, and the United States Federal Reserve, last of the large central banks to name a number, in January 2012. The figure that became the world’s anchor, 2%, was neither the number Douglas named on air nor the midpoint of that first band. It settled as a convention, as one central bank after another took up a low positive target, and it spread by imitation rather than by derivation.

Reserve Bank of New Zealand histories of the Policy Targets Agreement; the Douglas anecdote follows Governor Don Brash’s own recollection, and is remembered variously as “zero to one” and “zero to two.” US Federal Reserve, Statement on Longer-Run Goals, 25 January 2012.

The lesson of that history is that the exact figure is contingent. Had the New Zealand band been written a point higher or lower, the world might have settled somewhere else. But the decision that the target should sit above zero rather than at it is not contingent, and does not depend on the number. It rests on three reasons that central banks state openly, and that together explain why aiming at no inflation is treated as more dangerous than aiming at a little.

The case for a little inflation

The first reason is the one central banks talk about most, and it concerns the room they have to fight a recession. A central bank fights a downturn by cutting the interest rate it controls, which lowers borrowing costs and coaxes spending back. But a nominal interest rate cannot be pushed far below zero, because anyone charged to hold money can simply hold cash instead, which pays zero. That floor near zero is called the zero lower bound, and it sets a limit on how much a central bank can cut before it runs out of conventional ammunition.

The height of the interest rate in normal times depends on the inflation rate the bank is targeting: aim for higher inflation and normal interest rates sit higher, which leaves more distance to cut before hitting the floor. A target of zero holds normal rates low and leaves the bank close to the floor before a recession even begins, so a serious slump can exhaust its room almost at once. A target of 2% lifts the whole ladder and hands the bank several extra points of cutting room. The inflation buffer is, in this sense, self-insurance against the next downturn.

Figure 3.2 Why a zero target leaves a central bank short of ammunition. Normal interest rates sit roughly at the targeted inflation rate plus a real return, so a higher target lifts the whole ladder and leaves more distance to cut before rates hit the floor near zero. Under a zero target the bank starts close to that floor and can exhaust its room in a single serious recession; the buffer a 2% target buys is the extra room on the right. Schematic. Illustrative levels assume a real return of about 2%; the point is the relative room to cut, not the exact figures.

The second reason is about wages. Workers, and the firms that employ them, resist outright cuts to pay in a way they do not resist a raise that merely lags prices. A nominal pay cut feels like a personal defeat and poisons a workplace, so employers avoid it even when a job’s real value has fallen. A little inflation offers a way out of that bind: it lets the real value of a wage drift down where a business genuinely cannot afford it, through a raise that trails inflation rather than a cut that shows up on the payslip. Economists call this the grease in the labour market, and the argument that low inflation can gum it up was made carefully by George Akerlof, William Dickens, and George Perry in 1996. At exactly zero inflation, the theory runs, the economy loses the quiet mechanism by which real wages adjust without anyone’s paycheck being cut in name.

The third reason is that the ruler is bent. The official price indexes that a central bank steers by do not measure the cost of living perfectly; they tend to overstate how fast it is really rising, because they are slow to credit better products and the way shoppers switch toward whatever got cheaper. In 1996 a United States commission led by the economist Michael Boskin put the overstatement at about 1.1 percentage points a year, within a range of 0.8 to 1.6. If the gauge runs high by something like a point, then a measured target of 2% corresponds to a true rise nearer 1%, and a measured target of zero would risk true prices actually falling. Aiming a little above zero keeps the economy clear of accidental deflation, which every central bank treats as the more dangerous ditch, for reasons that belong to the chapter on what happens when money breaks down (Volume V: Pathologies, on what happens when money dies).

Model

Three reasons to aim above zero

Room to cut. Interest rates cannot fall far below zero; a positive inflation target lifts normal rates and leaves the bank more room to cut in a recession before it hits that floor.

Grease for wages. People resist nominal pay cuts, so a little inflation lets real wages fall where they must without a number on a payslip going down.

A biased gauge. Price indexes overstate true inflation by roughly a point, so a measured target of 2% keeps actual prices safely rising rather than accidentally falling.

Room-to-cut and buffer arguments per Bernanke and Blanchard, Dell’Ariccia & Mauro, “Rethinking Macroeconomic Policy” (IMF, 2010); wage rigidity per Akerlof, Dickens & Perry, Brookings Papers (1996); measurement bias per the Boskin Commission (1996).

None of these three settles on 2% exactly. They argue for a target that is low, so the drift stays in the background of economic life, but positive, so the bank keeps its room and the economy keeps its grease. After the crisis of 2008, when several central banks did hit the zero floor and got stuck there, some economists argued the buffer had been set too small and the target should be higher still, nearer 4%, to buy more room against the next slump. That the argument runs in the direction of more inflation rather than less is itself a clue to why zero is not the resting point.

What the banks stuck at that floor did next left a record of its own. Unable to cut further, several of them bought assets on a large scale and paid with newly created reserves. The scale is legible in the Federal Reserve’s weekly statement of the factors supplying reserve funds: an average of $945,890 million in the week ending 3 September 2008 had become $8,992,579 million by the week ending 13 April 2022, roughly nine and a half times as much. For most of the years in between, the running complaint in the argument just described was that inflation was too low, not too high; the surge came later, with American consumer prices up 9.1 percent in the twelve months ending June 2022, the largest such rise since 1981. This chapter keeps the record rather than adjudicating it: creation on that scale arrived first alongside an argument about too little inflation, and only later alongside a surge.

The case for zero, or below

The opposing view deserves its strongest statement, because it was made by one of the century’s most formidable economists and it points the other way entirely. Milton Friedman argued that the optimal rate of inflation is not 2%, nor even zero, but slightly negative. His reasoning starts from the cost of holding money. Cash in your pocket earns no interest, so holding it means giving up the return you could have earned elsewhere, and that forgone interest is a private cost that rises with the interest rate. Yet creating money costs a central bank almost nothing. From society’s point of view, then, people are being made to economise on something that is nearly free to produce, and the waste disappears only when the interest rate on holding money is driven to zero. To get there, prices should gently fall, at about the real rate of interest, so that money held in a drawer earns its keep by buying a little more each year. On this logic, in his 1969 essay The Optimum Quantity of Money, the ideal monetary policy is a mild, steady deflation, and the modern 2% target is a standing error.

The argument is clean, and central banks have not so much refuted it as declined to run the risk it carries. Friedman’s optimum assumes wages and prices that adjust smoothly and a central bank that never needs emergency room to cut. The three reasons above are, in effect, the reply that the real economy does not behave that way: wages stick, gauges mislead, and downturns arrive without warning and demand more cutting room than a near-zero anchor allows. Running the economy at the edge of deflation to capture a modest efficiency gain, the consensus holds, leaves it exposed on the side where the damage is largest. The disagreement is genuine, and it is a disagreement about which risks are worth running, not about whether the ratchet in prices is chosen. On both accounts, it is.

The 2% inflation target is just an arbitrary number.

Oversimplified Moderate confidence

Half of the claim is fair. The specific figure of 2% is historically contingent: it descends from a 0 to 2% band written in New Zealand in 1990, itself prompted by a minister’s television remark that aimed lower still, and it spread by imitation rather than by derivation. No equation singles out 2 over, say, 1.5 or 3. But the claim overreaches when it treats the whole target as arbitrary, because the decision that the target should be low and positive rather than zero rests on stated, substantive reasons: the room a buffer buys against the zero lower bound, the grease a little inflation gives to sticky wages, and the upward bias in the price indexes that makes a measured zero risk true deflation. The level is a convention; the direction is a considered choice, and calling the number arbitrary discards the second point along with the first.

Sources

Reserve Bank of New Zealand, histories of the Policy Targets Agreement — the 0–2% origin and the Douglas anecdote; supports the “contingent level” half.
G. Akerlof, W. Dickens & G. Perry, “The Macroeconomics of Low Inflation,” Brookings Papers on Economic Activity (1996) — downward nominal wage rigidity, the grease argument for a positive target.
Advisory Commission to Study the Consumer Price Index (Boskin Commission), Toward a More Accurate Measure of the Cost of Living (1996) — CPI overstates inflation by about 1.1 points a year, the case against a zero target.
O. Blanchard, G. Dell’Ariccia & P. Mauro, “Rethinking Macroeconomic Policy” (IMF, 2010) — the zero-lower-bound buffer argument, and the post-crisis case for a higher target still.
Federal Reserve, Factors Affecting Reserve Balances (H.4.1), releases of 4 September 2008 and 14 April 2022 — total factors supplying reserve funds, weekly averages of $945,890 million and $8,992,579 million; U.S. Bureau of Labor Statistics, CPI-U up 9.1 percent over the 12 months ending June 2022 — the post-crisis record kept above.
M. Friedman, The Optimum Quantity of Money (1969) — the strongest opposing reading, named per the observer rule: optimal policy is mild deflation, which would make any positive target an error.
Confidence is moderate under the rubric: the pip scores the weaker of evidence directness and construct match, and construct match binds here. The evidence for both halves is direct and well documented, but the claim mixes two things, an arbitrary level and a reasoned direction, and the verdict rejects only the first while conceding the second. A reader who means “why 2 and not 1.5” is largely right; one who means “why not zero” is not.

Where the argument goes next

The ratchet in your groceries, then, is neither an accident nor a sign that anyone is failing at their job. It is the visible trace of two deliberate arrangements: a central bank built to stand behind money in a panic, and a target that keeps prices drifting slowly up so the bank holds room to act and the labour market keeps its give. What the drift feels like from inside a household, and why the cheap televisions did nothing to console anyone about rent and tuition, is the cost-disease question next door in Frameworks (on why some prices race ahead while others fall). What happens when the discipline breaks and a state pays its bills by printing without limit is the collapse case in Volume V, on what happens when money dies; how the same machinery of credit and rates drives the recurring boom and bust is Volume V’s chapter on why boom and bust keep happening. This chapter has stayed with the normal case, where the institution works as designed.

There is one more institution in this volume built on the same public credit, and it frightens people more than inflation does. If a government can stand behind its banks and steer its money, it can also borrow in that money, on a scale no household could, and the debt it runs up is the thing politicians warn your grandchildren will have to repay. Who that debt is actually owed to, and why a government’s borrowing is so unlike a family’s, is the next chapter’s question.