The Level

The price level is not the price of any one thing but the price of everything at once. Since one cannot literally add the price of bread to the price of haircuts to the price of steel, the measure is built from a representative basket — a fixed assortment of the goods a household buys — whose total cost, tracked over time and set against the cost in some base year, becomes a single number: a price index. Inflation is the rate at which that index climbs; deflation, the rate at which it sinks. The index reduces a million distinct, jostling prices to one figure, the better to see a movement that no single price reveals — the slow swelling or shrinking of money prices in general, the drift of the level itself.

And here is the strange heart of the matter: the price level is the value of money inverted. To say that prices have doubled is to say that money buys half as much — that each unit of currency is worth half what it was. Money's own value is never quoted directly; one cannot ask the price of a dollar in dollars, for the answer is always one. Its worth is revealed only obliquely, in the mirror of all other prices, rising as they fall and falling as they rise. The figure draws the two as the mirror images they are: the price level ascending, the purchasing power of money descending, crossing in the middle, each the other turned upside down. This is the yardstick problem in its purest form. Money is the rod by which a society measures the value of all things — and the rod itself changes length, lengthening in deflation and shrinking in inflation, so that every measurement taken with it is suspect.

Even the measuring of the rubber rod is treacherous. The basket is held fixed while the world it prices shifts beneath it: people buy less of what grows dear and more of what grows cheap, so a fixed basket overstates the rising cost of living; goods improve, so that a phone or a medicine priced today is not the thing it was a decade ago; entirely new goods appear that no base-year basket contained. A price index must somehow reckon with all of this, and every method of reckoning is a compromise. The number is indispensable — pensions and wages and debts are tied to it, central banks steer by it — and it is, irreducibly, an approximation of a thing that does not sit still to be measured.

Suppose that tonight, while everyone sleeps, a miracle of arithmetic is worked: every banknote and every bank balance doubles, and so does every price tag, every wage, every debt, every contract. You wake to find twice the money in your pocket and twice the price on everything you might buy with it. What has changed? Nothing real. You can purchase exactly the basket you could have purchased yesterday; the baker bakes the same loaves; the worker works the same hours for the same bread. The great doubling was a pure relabeling — a change in the unit of measure, not in any quantity the unit measures, like restating every distance in inches where before it was in feet. This is the neutrality of money, and it is the foundation stone of all thinking about the price level.

From this thought experiment the classical economists — David Hume stated it with the most lasting clarity — drew a great separation, the classical dichotomy. Real variables, they held, are settled by real forces: what is produced, who is employed, the relative prices of goods, the real wage that labor commands, all flow from tastes and technology and the scarcity of resources. Nominal variables — the price level, the money value of everything — are settled by the quantity of money. The two run on separate tracks, and to understand what is really happening in an economy one must look through the monetary surface to the real machinery beneath. Money is a veil; or, to borrow the image this plate is named for, money is the mask, and the real economy is the living face behind it, whose expressions the mask merely covers without changing.

The dichotomy is true in the long run and false in the short, and its long-run truth is as profound as it is easily forgotten. Over years and decades money is genuinely neutral: nations that have endured great inflations and nations that have kept stable currencies are left, when the dust settles, with the same real economies, growing at rates set by the real forces of the previous volume and indifferent to how many zeros their banknotes carry. This is why sustained inflation cannot, in the end, purchase prosperity — why the printing press is no substitute for the slow accumulation of capital and ideas. And yet the overnight experiment cheats in exactly one respect: it assumes that every price moves at once, in the same instant, by the same proportion. In the waking world they do not. Prices and wages are set in contracts, revised on schedules, anchored by habit, and they move at different times and different speeds. That single, realistic failure of the thought experiment is the crack through which the short-run power of money pours — and from it flows the whole of the plates to come.

The classical intuition has a precise expression, the equation of exchange, set down in its modern form by Irving Fisher: the quantity of money, multiplied by its velocity — the number of times each unit is spent in a period — equals the price level multiplied by the volume of real output. Written compactly, MV = PY. The left side is total spending: all the money there is, times how often it changes hands. The right side is the money value of everything produced: the price level, times the quantity of real goods. As it stands the equation is an identity, true by the meaning of its terms, telling us nothing on its own. It becomes a theory of the price level the moment one makes an assumption about the two terms on the side of money — velocity and output.

The quantity theory makes exactly that assumption. If velocity is roughly stable — fixed by the slow-moving habits of payment, the calendar of wages and bills — and if real output is determined by the real forces of the previous volume rather than by money, then the price level has nowhere to move but with the money supply. More money, the same goods, and prices must rise to absorb it; inflation becomes simply money growth running ahead of output growth. The long-run evidence, gathered across many countries and many decades, is the figure: money growth and inflation line up close to one-for-one, the points scattered along a single rising line, the high-money-growth nations carrying high inflation and the disciplined ones low. It is this regularity that Milton Friedman compressed into his famous verdict — that inflation is always and everywhere a monetary phenomenon — meaning that wherever and whenever prices rise persistently, an excess of money lies behind it.

The qualifications matter as much as the rule, for the quantity theory is a truth of the long run and the large sample, not of the month and the margin. Velocity is not in fact constant: it drifts with interest rates, with the inventions of finance, with confidence — and in a panic, as the volume on the cycle showed, the demand to hold money can leap and velocity collapse, so that even a steady money supply yields falling prices. Output, too, is not fixed in the short run, for money moves it. So over short horizons the link between money and prices is loose, noisy, and an unreliable guide to policy; it is across years, and across the vast range that runs from stable currencies to raging hyperinflations, that the quantity theory shows its iron. The equation is a frame within which to think, not a dial by which to steer from one quarter to the next.

A government that issues money can buy real things with paper that costs almost nothing to make, and the difference between what the money will purchase and what it cost to print is a revenue — seigniorage, the ancient profit of the issuer. It was once the prince's privilege, exercised by clipping and debasing the coinage, melting good coin into baser metal and spending the difference; it is now the modern state's, exercised through the central bank that creates the base money. Seigniorage is, at root, a way of commandeering resources without raising a tax or selling a bond — and so it is the resort of every government that finds it cannot tax enough, and cannot borrow enough, to pay for what it has resolved to spend. When the ordinary sources fail, a state can always, as a last device, print.

But to print money beyond the growth of output is to cause inflation, and inflation is itself a tax — perhaps the purest tax there is. It is a levy on everyone who holds money, whose real value the rising price level quietly eats away. Hold cash through a year in which prices rise a tenth, and you have paid a tenth of it in tax as surely as if an assessor had knocked at your door — except that no assessor called, no legislature voted, and the burden fell, without distinction or appeal, on whoever happened to be holding money as prices climbed. This is the inflation tax: its base is the real quantity of money the public chooses to hold, its rate is the rate of inflation, and its singular feature — the feature that makes it so tempting to the cornered state — is that it need never be legislated, debated, or even named.

There is, however, a ceiling on what this tax can raise, and the ceiling is the doorway to catastrophe. For the inflation tax erodes its own base: as inflation rises, people contrive to hold less money — they spend it faster, switch into goods, flee to a foreign currency — and the real stock of money that the tax falls upon shrinks. Beyond some rate, faster inflation yields less revenue, not more, the same Laffer-shaped peak that haunts every tax pushed too far, now traced in the demand for money. A government that has come to lean on the inflation tax discovers, as its base flees, that holding revenue steady requires inflating faster, which drives the base away faster still. Phillip Cagan, studying the great hyperinflations, mapped precisely this dynamic — the desperate chase of a fleeing base by an accelerating inflation — and in mapping it described the mechanism of the next plate, where the chase runs away entirely and money itself is destroyed.

Hyperinflation — by the usual convention, inflation above fifty percent a month — is the disease in which money destroys itself. Prices double in weeks, then in days; a wage is worth spending the very hour it is paid; banknotes are carried in baskets and wheelbarrows, printed in ever larger denominations, and at last simply refused. The historical cases are notorious: Germany in 1923, where a loaf of bread came to cost billions of marks; Hungary in 1946, the worst ever recorded, with prices doubling in a matter of hours; Zimbabwe in the first decade of this century, and a grim handful of others. The figure traces the signature they share: the price level, drawn even on a logarithmic scale — a scale on which mere exponential growth would be a straight line — still curving upward, bending toward the vertical, because the inflation is accelerating, feeding upon itself.

The roots of hyperinflation are, almost always, fiscal — it is a fiscal catastrophe wearing a monetary mask. It begins where a government's spending so outruns its power to tax or to borrow — in the wreckage after a war, a defeat, a revolution, a collapse — that it finances the gap by printing, and finds it cannot stop. The inflation tax of the previous plate is driven past its Laffer peak; the base flees; the government prints faster merely to stand still; prices accelerate. The monetary explosion is the proximate cause, the thing one sees, but the fiscal hole the printing fills is the ultimate one — which is why a true hyperinflation can never be cured by monetary sternness alone, but only by closing the budgetary chasm that compels the printing in the first place. To tell a government in this state simply to stop printing is to tell it to stop paying its army and its pensioners; the printing is a symptom of a fiscal disease.

What turns a high inflation into a hyperinflation is the collapse of the demand for money — the velocity term of the equation grown monstrous. As people come to expect prices to rise without limit, they race to convert money into anything that holds value before it melts in their hands; the racing itself drives prices up faster, which confirms the expectation and quickens the next flight, in a loop that tightens upon itself. Money sheds its functions in order: first it ceases to be a store of value, then a unit of account — merchants begin quoting prices in a foreign currency or in goods — and finally a medium of exchange, as barter creeps back and the currency is abandoned where it stands. The currency dies. Its death is ordinarily ended only by a sharp and credible fiscal reform, often a wholly new money, and — as the plates ahead will insist — by the restoration of the single thing a hyperinflation annihilates most completely, which is belief.

In 1958 the engineer-turned-economist A.W. Phillips plotted nearly a century of British data and found a relation of arresting stability: years of low unemployment were years in which wages rose quickly, and years of high unemployment were years in which wages rose slowly or fell. Drawn in the plane of inflation and unemployment, the points traced a smooth downward-sloping curve — the figure of this plate. The mechanism seemed plain enough: a tight labor market, with jobs chasing few workers, bids wages up; a slack one, with workers chasing few jobs, holds them down. Paul Samuelson and Robert Solow soon carried the relation across the Atlantic and from wages to prices, fitting it to American data, and in doing so drew the moral that would dominate a decade of policy. Here, they suggested, was a menu — a stable and exploitable trade-off, from which a society might simply choose the combination of inflation and unemployment it preferred.

If the curve were stable, the implication for policy was irresistible. Monetary and fiscal management faced a choice along it: accept a little more inflation and be repaid in a little less unemployment, or wring inflation down and pay for it in idle workers. For a decade this framing organized the central debate of macroeconomic policy — not whether to manage demand, but where on the curve to come to rest, how to weigh the public's distaste for rising prices against its dread of joblessness. The trade-off seemed to promise that resolute demand management could purchase permanently lower unemployment for a modest and permanent inflation — a humane bargain, and to the politically bold an apparently free lunch, a standing offer that only timidity would decline.

But the menu rested on an assumption that no one had thought to examine: that the trade-off would hold still while policymakers leaned upon it. The Phillips curve was an empirical regularity in want of a theory, and the theory it lacked was precisely the one that would explain why inflation and unemployment should be linked at all. That missing theory — which the next plate supplies — locates the link not in inflation as such but in the gap between actual inflation and the inflation people expect. Workers and firms, it insists, care in the end about real wages, about what money will buy; and once they come to expect inflation, they build it into the wages they demand, so the gap closes and the trade-off evaporates. The menu, in other words, was a snapshot taken while expectations sat still — and the act of ordering from it repeatedly was exactly the thing that would set expectations moving and dissolve the offer. The curve, leaned upon, moves.

In 1968, independently and almost at once, Milton Friedman and Edmund Phelps saw what the menu had concealed. Workers and firms do not care about nominal wages and prices for their own sake; they care about real wages, about what a wage will actually buy. The Phillips trade-off works, when it works, only because an unexpected burst of inflation makes real wages fall — nominal wages, fixed in advance, lag behind the rising prices — so that firms, finding labor suddenly cheap, take on more of it. But this turns entirely on the inflation being a surprise. Let the inflation persist, and workers learn to expect it; they demand nominal raises that build the expected inflation in; real wages stop falling; and employment drifts back to its natural rate — the level fixed by the real frictions of the labor market, by the churn of jobs found and lost, wholly independent of how fast prices happen to be rising.

The consequence is the figure. There is not one Phillips curve but a whole family of them — a separate short-run curve for each rate of expected inflation, each sloping downward, each sitting higher than the last as expected inflation rises. A society can ride down any one of these short-run curves, buying lower unemployment, only until expectations catch up to the new inflation and the curve shifts up beneath its feet, depositing unemployment back at the natural rate but now at a higher inflation than before. The long-run Phillips curve — the locus of all the points where actual inflation has come to equal expected inflation — is therefore vertical, planted at the natural rate of unemployment. There is no permanent trade-off to be had. The attempt to hold unemployment below the natural rate purchases not a steady higher inflation but an ever-accelerating one, as policy forever chases an expectation that forever catches up; this is the accelerationist hypothesis, and it is the menu's epitaph.

The prediction was vindicated almost as soon as it was made. The nineteen-seventies brought stagflation — high inflation and high unemployment together — a combination the stable menu had declared all but impossible, and which the expectations-augmented curve explained to the letter: inflation expectations had risen, shifting the whole short-run curve upward, so that even slack labor markets now came paired with rapid inflation. Robert Lucas then pressed the logic to its sharpest edge with the hypothesis of rational expectations: if people form their expectations using all the information available — including their understanding of how the central bank behaves — then only genuinely unanticipated money can move anything real, for anticipated inflation is built in at once and alters nothing. The exploitable trade-off shrinks to the merely surprising, and any systematic attempt to use it must fail, because a systematic policy is, by its nature, foreseen. The Lucas critique of the previous plate here arrives in full force: the menu was never a structure to be exploited, only a regularity waiting to be dissolved by the act of exploitation.

The expectations story cuts in both directions. If raising inflation above what people expect can buy a temporary boom, then lowering it below what people expect must exact a temporary slump. To bring inflation down, a central bank must run money tight enough that inflation falls short of expectations — which means, by the same mechanism that drove the natural-rate argument, that real wages rise, firms hire less, and unemployment climbs above its natural rate until expectations adjust downward to the new, lower reality. Disinflation, in the world the previous plates built, is paid for in foregone output and idle workers. The recession is not an accident of the cure; it is the cure's price, the mirror image of the fleeting boom that the original inflation once bought.

The size of that price is measured by the sacrifice ratio: the cumulative percentage of a year's output that must be given up to bring inflation down by a single percentage point. The estimates vary across time and country, but the ratio is typically several — disinflation is expensive, and the deeper and faster the squeeze, the steeper the recession it requires. The episode that fixed this in the modern mind was the disinflation of the early nineteen-eighties in the United States, under the Federal Reserve then chaired by Paul Volcker, which broke a decade of entrenched inflation but at the cost of the sharpest recession since the nineteen-thirties. It was a vivid, painful demonstration that the expectations-augmented Phillips curve is no mere blackboard exercise — that wringing deeply rooted inflation out of an economy is among the most costly undertakings a central bank can attempt.

But the sacrifice ratio is not fixed, and here the argument turns toward the institutions of the plates ahead. The ratio depends on expectations, and expectations depend on credibility. If a central bank announces a disinflation and is believed, people lower their inflation expectations at once; the short-run curve shifts down without the economy having to pass through a recession to drag it there; and the disinflation is nearly painless. If the bank is not believed — if people suspect it will lose its nerve and relent — they keep demanding the high nominal raises that entrenched inflation requires, inflation stays stubbornly up, and the bank must impose the full weight of recession simply to prove that it means what it says. The cost of disinflation is thus, in large part, the cost of establishing credibility, and Thomas Sargent's study of the great interwar stabilizations found exactly this pattern: where a stabilization was credible — backed by visible, durable reform of the fiscal and monetary regime together — even enormous inflations were halted swiftly and at modest cost, while where the reform was doubted, the price ran high. Credibility is the scarce resource of monetary policy, and the plates that remain are about how an institution comes to possess it.

Begin with the hardest case for the prosecution: a perfectly anticipated, evenly shared inflation, in which every price and wage and contract rises together at a rate known to all in advance. Even this is not without cost. There are shoe-leather costs — the bother of economizing on cash whose value is steadily melting, the extra trips to the bank, the effort of holding wealth in things rather than in money. There are menu costs — the real resources consumed in constantly re-pricing, reprinting catalogs, re-tagging shelves, resetting the machines. There is the inflation tax on whatever cash people must still hold to transact. And there is the corrosive interaction with a tax code written in nominal terms: inflation drags incomes into higher brackets that were never meant for them, taxes purely paper capital gains as though they were real, and distorts saving and investment in ways no legislature ever chose. These costs are present even when the inflation surprises no one at all.

The graver costs come from inflation that is not foreseen. Debts are written in nominal terms, and so unexpected inflation transfers wealth wholesale from creditors to debtors — the lender is repaid in money worth less than the money he parted with — while an unexpected disinflation runs the transfer cruelly in reverse. Across a whole economy, surprise inflation is a vast and arbitrary redistribution: capricious, unlegislated, rewarding the indebted and punishing the thrifty, silently eroding the long-dated contracts — the pension, the bond, the lifetime of savings — on which provision for the future is built. It is the arbitrariness, far more than the average level, that does the deepest damage here. A known and steady inflation can be planned around, written into contracts, neutralized; a variable and surprising one cannot, and it poisons the very act of making promises across time.

Beneath all of these lies the subtlest cost, and perhaps the worst. Prices are signals — this is the lesson the whole atlas has pressed — and a rising relative price says, with great economy, this thing is scarce: make more of it, use less of it, look for substitutes. But inflation makes all prices rise at once, and in the general ascent it becomes genuinely hard to tell a relative-price change, which carries real information, from a merely monetary one, which carries none. Is this good dearer because it has grown scarce, or only because everything is dearer? The signal is buried in noise; the price system — the great decentralized machine by which a market economy computes what to make and where to send it — works less well; and resources are misallocated because the prices meant to guide them have been blurred. High and variable inflation degrades the very mechanism through which a market coordinates itself, which is why even a moderate inflation, if it is erratic, can exact a cost out of all proportion to its average rate, and why the central banks of the next plate labor so hard to keep inflation not merely low but steady.

Falling prices sound at first like nothing but a blessing: your money buys more, the cost of living drops, the household budget stretches further. And a mild deflation that springs from rising productivity — goods becoming genuinely cheaper to make — can indeed be benign, a gentle sharing-out of plenty. But a deflation that springs from collapsing demand — too little money chasing the goods, the exact mirror of inflation — is among the most dreaded conditions in all of macroeconomics, for reasons the naive view misses entirely. The danger lies not in the cheapness of things but in what a generally falling price level does to debt and to the central bank.

The first danger is the one Irving Fisher named, surveying the wreckage of the nineteen-thirties: debt-deflation. Debts are fixed in nominal terms, and when the price level falls, the real burden of every debt rises — the farmer who borrowed against dollar wheat must repay in fifty-cent wheat, owing twice the real amount he meant to. Debtors, crushed by a weight that grows heavier as their incomes and asset prices fall, cut their spending to the bone and sell whatever they can to service what they owe; the selling drives prices down further; the further fall raises the real burden of debt again; and the economy corkscrews downward in a vicious circle, in which the very struggle to pay down debt makes the debt, measured in real things, larger still. The harder the debtors labor to escape, the deeper the whole economy sinks.

The second danger, drawn in the figure, is that deflation disarms the central bank precisely when it is most needed. Ordinarily the bank fights a slump by cutting the interest rate; but the nominal rate cannot fall much below zero, for no one will lend at a loss while they have the option of simply holding cash. When deflation sets in, the real interest rate — the nominal rate minus inflation — becomes the nominal rate plus the rate of deflation, and it can be stuck painfully high at the very moment the economy is begging for it to be low or negative. The bank, having cut its rate to zero, can cut no further; the real rate that recovery requires lies below zero and out of its reach; monetary policy is jammed against its floor and loses its grip. This is the liquidity trap of the volume on the cycle, seen now from the side of prices as a deflationary trap. Japan's long stagnation after 1990, and the rich world's brush with deflation after the crisis of 2008, made this once-academic fear vividly, expensively real — and they explain why central banks deliberately aim at a low positive inflation rather than at price stability proper: to keep a buffer of distance between themselves and the zero bound, and the trap that waits just beyond it.

The price level, left entirely to itself, is indeterminate — nothing within a pure market of relative prices pins down the absolute level of money prices, the number of dollars a thing costs. What pins it down is the monetary authority: the central bank, which controls the supply of base money and, through it, anchors both the price level and the expectations that ride upon it. This is a power of an awesome kind — to move, by adjusting a single interest rate, a number that touches every price, every wage, every debt and contract in the economy at one stroke — and the central question of monetary institutions is how so great a power ought to be exercised, and how it ought to be constrained.

Here lies a deep trap, identified by Finn Kydland and Edward Prescott: the problem of time-inconsistency. Imagine a central bank that holds discretion and announces it will keep inflation low. Suppose the public believes it and sets wages accordingly. The bank now faces a temptation, for by the Phillips logic a small surprise inflation would, at this moment, buy a welcome burst of extra employment. But the public, being no fool, anticipates that the bank will face exactly this temptation; it therefore does not believe the low-inflation promise in the first place; it builds higher inflation into its expectations; and the bank, to avoid the recession that would follow from disappointing those expectations, validates them by inflating. The outcome is an inflation bias: discretion delivers higher inflation than was promised, with no lasting gain in employment to show for it — everyone worse off, not through anyone's bad intent but through the bare logic of a promise that cannot bind the one who makes it. The disease is not malice; it is the absence of credible commitment.

The cure is to bind the authority — to put commitment in the place of discretion, so that the promise becomes credible precisely because the bank can no longer renege on it. The forms this binding takes are several, and the figure draws the most explicit of them. A rule may tie the policy rate mechanically to inflation and output: the Taylor rule of the figure raises the nominal interest rate more than one-for-one with inflation, so that the real rate itself rises to lean against any rise in prices — the Taylor principle, the steepness beyond forty-five degrees that distinguishes a stabilizing rule from a destabilizing one. Independence may insulate the central bank from the government's standing electoral temptation to inflate. And the modern synthesis, inflation targeting, may fix a public, numerical inflation goal for which the bank is held openly accountable, anchoring the expectations of all to a single number that everyone can see and check. The common thread through every form is the deliberate sacrifice of flexibility for credibility: an authority that gives up the power to surprise gains, in exchange, the far more valuable power to be believed — and a believed central bank can hold inflation low and steady at a small fraction of the cost borne by one forever suspected of being about to cheat.

The price level is the value of money inverted — the worth of the numéraire, the unit in which every price in this atlas is quoted. Across eleven plates that worth has shown itself governed, in the long run, by one thing, and entangled, in the short run, with another. We have seen the yardstick made of rubber and the mask over the real economy; the equation of exchange and the toll the printing press exacts; the spiral in which money dies; the menu that dissolves when ordered from twice and the expectation that dissolves it; the sacrifice that breaking inflation demands; the erosion it works even when foreseen; the fall that deflation brings; and the mooring by which a central bank holds the whole at rest. It remains to gather them.

In the long run, money is a veil. It sets the level of prices and touches nothing real; the price level moves with the quantity of money relative to output; and so sustained inflation is, in the end, a monetary phenomenon, and the printing press can buy no lasting prosperity. This is the soundest proposition in the subject and the most easily forgotten in the heat of any particular year — the truth that the classical economists saw and that the quantity theory, with all its qualifications, still carries: that over long horizons a society's real fortunes are written in bread and steel and ideas, not in the number of zeros on its banknotes.

In the short run, money is a lever. Because prices are sticky, a change in money spills into output and employment; the price level is bound to the cycle; and the painful trades open — the Phillips menu that beckons and betrays, the disinflation paid for in slumps. And inflation, even when perfectly anticipated, is not free, while unanticipated it redistributes by caprice and, worst of all, buries the signal of true scarcity in the noise of a general ascent. Deflation is graver still, raising the real weight of every debt and disarming the central bank at the zero bound. The management of the level is therefore never the bloodless adjustment of a single number; it is inseparable from the management of output, employment, and the burden of debt.

Everything turns on expectations, and expectations turn on credibility. The trade-off between inflation and unemployment works only by surprise; the surprise cannot be repeated; the cost of disinflation is, in the main, the cost of being believed; and the inflation bias of a discretionary authority is, at root, a failure of credible promise. The scarce resource in the governance of the price level is not knowledge of what to do — that much is, by now, fairly well understood — but belief: the power of an institution to make a promise that the markets, and the wage-setters, and the lenders, will actually trust. This is the hinge on which the whole subject swings, and the reason the modern theory of monetary policy is, in the end, a theory of credibility.

And so the price level, indeterminate in itself, is anchored by an institution, and that institution's awesome power — to move the value of every price at once — must be bound if it is to be trusted: by rules, by independence, by a public and accountable target. The deliberate sacrifice of flexibility for credibility is the central design of every modern monetary regime. Here is the answer to the question with which this volume opened, the question of who should hold the power to change the value of every price at once: not a discretion free to surprise, however well-intentioned, but an institution bound, accountable, and — by virtue of being bound — believed.

What is known here is a genuine triumph, and worth naming as such. The long-run neutrality of money and the quantity theory; the expectations-augmented Phillips curve and the natural rate; the time-inconsistency problem and the case for credible rules; the inflation-targeting regimes that, built upon all of these, tamed the great inflations of the late twentieth century — together they form one of the more complete and successful bodies of knowledge in all of macroeconomics. What is not known is named with equal candor. The short-run dynamics remain contested — how sticky prices truly are, how rational expectations truly are, how flat the curve has truly become. The natural rate is unobservable and drifts, a target no one can see. The right response to the zero bound, and to the asset-price and financial risks that consumer-price targeting may miss, is unsettled. And the recent decades — a puzzling flatness of inflation, then its sudden and unforecast return — have humbled the confident on every side.

The Level is the third of The Aggregates, and the nominal companion to The Growth's real trend: the two of them, the nominal and the real, are the great trends of the macroeconomy, with The Cycle the wobble around them and The Share the question of how their fruits divide. Its bridges run throughout the atlas. To The Money, the institution whose quantity governs the level and whose very nature this volume has presupposed. Back to The Cycle, whose slumps the price level rides, and whose liquidity trap is this volume's deflationary trap seen from the other side. To The Growth, the real trend that the nominal merely shadows. To The Share, for inflation is among the great arbitrary redistributors of wealth. To The Tax, for the inflation tax is the levy that need never be voted. And to The Rate and The Spread, where the interest rate the central bank sets, and the credit through which it works, are studied in their own right.

What, then, is money worth? Exactly what the quantity of it, set against the goods it can buy, makes it worth — and that worth is anchored, in the last analysis, not by any real thing but by the credibility of the institution that issues it and the expectations of those who hold it. Money is a shared fiction that works only so long as it is believed; a measure of all value that has, itself, no value of its own; a numéraire whose own price can be asked only by inverting all the others. The price level is the value of that fiction, and its management is the management of belief. The aggregate, like every quantity this atlas has weighed, is a verdict on a hidden contest: between the real and the nominal, between money as a veil and money as a force, between the discretion that tempts and the rule that binds, between the worth a society assigns its measure of worth and that measure's own refusal to hold still. The argument remains open.