The Cycle

Before the cycle was a theory it was a set of facts, and the facts came from counting. In the decades around the turn of the twentieth century a generation of patient empiricists — Wesley Mitchell foremost among them, and the National Bureau of Economic Research he helped to found — set themselves to measure what everyone could feel: that business does not proceed at an even pace but moves in long swells of prosperity and depression. They catalogued the turning points, dated the expansions and contractions, and gave the recurring shape its anatomy. An expansion, in which output, employment, and trade climb together; a peak, where the climb exhausts itself; a recession, in which activity falls — a fall the unlucky call a recession and the very unlucky a depression; a trough, the bottom; and a recovery that begins the round again. The rule of thumb names a recession after two quarters of falling output, but the fuller definition the daters use is richer: a significant decline in activity, spread across the economy, lasting more than a few months.

The most arresting of the facts is the comovement. It is not one industry that slumps while others thrive, nor one region; it is nearly all of them, together. Output and employment, investment and consumption, freight and retail and the hiring of hands — series that have no obvious reason to share a fate move up and down in rough unison, as though some unseen conductor set the tempo for the whole orchestra. Within the chorus there is structure: some series lead the turn (new orders, housing starts, the stock market), some lag it (unemployment, which keeps rising for a while after the trough), and investment swings far more violently than consumption, a clue the next plates will follow. But the dominant fact is the synchrony itself — that the parts rise and fall as one.

This synchrony is the puzzle the whole volume exists to explain. An economy is struck constantly by small, unrelated disturbances — a failed harvest, a clever invention, a firm's good or bad luck. Were the cycle merely the sum of these, the disturbances would mostly cancel in the aggregate, and the whole would glide smoothly while every part jiggled. That the parts instead move together tells us the cycle is not a sum of independent jiggles but a property of the system that binds them — of the web of spending and lending and expectation through which each actor's fortune is tied to every other's. The figure draws the basic picture: actual output as a wave around a rising trend, cresting at a peak, breaking into a recession, bottoming at a trough, recovering — the shape that recurs.

The deepest obstacle to understanding the cycle is a habit of mind so natural we scarcely notice it: the assumption that what is true of one is true of all. The economists call its violation the fallacy of composition, and the macroeconomy is its great theatre. One person can leave a crowded theatre quickly; if all rush the exit at once, none can. One firm can win customers by cutting its prices; if every firm cuts, the advantage vanishes and only the lower prices remain. One household can grow richer by saving more of its income; and here the fallacy turns from a curiosity into the engine of the slump. For a household's saving is some firm's lost sale, and a firm's lost sale is some worker's lost wage, and that worker is another household whose income has fallen. What is sound for the one can be calamitous for the all, because in the aggregate my spending is your income and your spending is mine.

This is the paradox of thrift, glimpsed by Mandeville and Hobson and made central by Keynes. Suppose every household, prudently, resolves to save more. Each cuts its spending; but that spending was income to others, so incomes fall across the economy; and as incomes fall, households find they can save no more than before — perhaps less — having succeeded only in making everyone poorer. The collective attempt to save more, in a slack economy, can leave total saving unchanged and total output diminished. Thrift, the homeliest of private virtues, becomes in the aggregate and at the wrong moment a public vice. The paradox is not a trick; it is what follows once one takes seriously that the economy is a circle, each person's outlay another's receipt, and that the circle can shrink as easily as it can grow.

What the paradox reveals is that the whole has a behavior the parts do not — that aggregate demand, the total of all spending, is a thing in its own right, and that it can be too little. The output a nation produces must be bought by someone's spending: by households consuming, firms investing, governments purchasing, foreigners importing. When that spending falters and no force gathers it up again, output falls to meet it, and the economy comes to rest below what its factories and its workers could produce. This is the systemic character of the cycle. It is not that each part fails on its own; it is that the parts, each behaving sensibly, fail to coordinate — that the system lacks an assured mechanism to lift them together. The whole is not the sum of optimizing parts; it is the parts plus the web of dependence among them, and the cycle lives in the web.

An expansion is not merely a period in which good things happen; it is a process in which good things cause further good things, and the first mechanism of that self-reinforcement is the multiplier. Richard Kahn and then Keynes saw its logic. Let some new spending enter the economy — a burst of investment, a public works program, a rise in exports. The income it creates goes to someone, who spends a fraction of it — the marginal propensity to consume — and saves the rest; that spending is fresh income to another, who again spends a fraction; and so on, in a converging series. The total rise in output exceeds the initial injection, multiplied by a factor that grows with the fraction re-spent: a dollar of new investment, if four-fifths of each round is spent, lifts output by five. Spending, in the aggregate, partly creates the income out of which it is paid.

The second mechanism is the accelerator, and it concerns investment. A firm invests to add capacity; but capacity is added only when output is growing. To keep producing at a steady rate a firm need only replace worn machines; to produce more it must build. Investment therefore depends not on the level of demand but on its rate of change — and this couples investment violently to the cycle. When output is rising, investment is high; when output merely stops rising, investment collapses, for there is nothing new to build. Paul Samuelson showed that the multiplier and the accelerator, set to interact, can generate cycles of their own accord: rising output spurs investment (accelerator), investment raises output further (multiplier), the faster growth spurs yet more investment — a self-feeding ascent, until the growth of demand slows, the accelerator throws investment into reverse, and the same logic runs downhill.

Above the mechanical there is the psychological, which is where the boom becomes most dangerous. Rising profits and prices breed confidence; confidence loosens the purse and the credit line; the spending and lending that confidence unleashes validate the optimism that prompted them, and so good times manufacture the belief in good times. The expansion thus carries, invisibly, the cause of its own reversal. Capacity is built ahead of the demand that will use it; debts are taken on against expected gains; expectations climb ahead of anything the economy can deliver. The boom is the phase in which the seeds of the bust are sown — in over-building, over-borrowing, and over-belief — though no one, in the warmth of it, can say which boom is the sound kind and which the kind that ends in ruin.

How can disturbances that are themselves irregular produce something as regular as a cycle? The answer, framed by Ragnar Frisch and anticipated by Eugen Slutsky, separates two things a careless eye runs together: the impulse and the propagation. Slutsky's demonstration of 1927 was startling and remains so. Take a series of purely random numbers — the economic equivalent of unrelated, meaningless shocks — and form running sums or averages of them. The smoothed series, built from pure randomness, displays cycles: swells and recessions of no fixed period but unmistakable shape. Order, it seemed, could be manufactured from disorder merely by the act of accumulation.

Frisch supplied the image that has organized thinking ever since. Picture a rocking horse, or a pendulum, struck at random intervals by a child wielding a club. The blows are irregular, arriving without pattern; yet the horse does not jerk arbitrarily — it rocks, smoothly, at its own natural frequency, each blow setting off a swing that decays until the next blow renews it. The shocks supply the energy; the structure of the horse supplies the rhythm. So with the economy. The impulse is the shock — where the disturbance comes from; the propagation is the machinery by which the economy's own structure receives a one-time blow and rings with it over years, converting an instantaneous disturbance into a drawn-out, oscillating, fading response. The cycle's character — its length, its shape, its persistence — is the economy's own, imposed upon whatever happens to strike it.

The propagation mechanisms are the very couplings the earlier plates described — the multiplier and accelerator, the slow adjustment of the capital stock, prices that lag, debts that take years to work off, the spread of disturbance along the chains that link supplier to buyer. Through them a single shock becomes a wave; and because the wave is damped — fading as the structure dissipates the energy of the blow — a cycling economy needs a continuing stream of fresh shocks to keep it moving, which the figure shows as a single impulse decaying toward stillness. The impulses themselves come from the three great sources the volume will take in turn: real shocks, to technology and harvests and the price of oil and the productivity of work; monetary and financial shocks, to the quantity of money, the supply of credit, the willingness to lend; and shocks of belief, the waves of confidence and fear that are, uniquely, capable of bringing about the very conditions they anticipate.

The classical tradition answered with a doctrine of striking elegance: money is a veil. Double every price in the economy and double the stock of money to match, and nothing real has changed — the same goods are made, the same hands employed, the same bargains struck, only the figures on the tags are larger. Money, on this view, is a unit of account and a lubricant of exchange, not a force on the real machinery. David Hume put it as a thought experiment: let the money in every pocket double overnight, and in time every price will double, leaving the real economy exactly where it was. The quantity theory made it formal — that the money stock times its velocity equals the price level times real output — and drew the classical dichotomy: real things (output, employment, relative prices) are set by real forces; the price level alone is set by money. Inflation is a monetary phenomenon, and money is neutral.

The doctrine is true — in the long run. Its failure in the short run is the hinge on which modern macroeconomics turns, and the failure has a name: prices and wages are sticky. They do not all leap at once to their new level when money changes; they are bound by contracts, by the costs of changing posted prices, by the difficulty of coordinating who moves first, by a stubborn habit of reckoning in money rather than in goods. And if prices do not adjust at once, then a change in money is not immediately swallowed by a change in all prices — so it spills into real quantities. A monetary expansion, meeting sticky prices, raises real spending, real output, real employment. The central bank, by this leverage, can lift a slack economy or cool an overheated one. The figure assembles the standard apparatus: a downward-sloping aggregate-demand schedule; a short-run aggregate-supply schedule that slopes upward because, with prices sticky, firms meet higher demand by producing more; and a vertical long-run supply line at potential output, where in the fullness of time every price has adjusted.

The reconciliation is the figure's lesson. Let aggregate demand shift rightward — a monetary easing, a surge of spending. In the short run the economy slides up the sticky-price supply line: output rises above potential and prices rise too, a genuine boom in real activity. But the boom is borrowed. As wages and prices catch up to the new quantity of money, the short-run supply line gives way, output drifts back to potential, and the economy comes to rest on the vertical long-run line — at the same real output as before, but a permanently higher price level. Money bought a temporary expansion, paid for in lasting inflation, not in lasting prosperity. Knut Wicksell had supplied the dynamic logic a generation earlier: when the interest rate the market charges falls below the natural rate that would balance saving and investment, spending and prices climb in a cumulative process, and the central bank's whole art is to keep the one rate tracking the other.

Finance is not a neutral conduit through which saving flows to investment; it is an amplifier, and it amplifies in one direction at a time. Credit is pro-cyclical: it swells in booms and shrivels in busts, magnifying the swing it ought, in a tranquil world, to smooth. The reason lies in collateral and confidence. In good times asset prices rise, so the property and securities that back a loan are worth more; borrowers look creditworthy, defaults are rare, and lenders — flush, competitive, lulled — extend more on easier terms. In bad times the machinery runs backward: asset prices fall, collateral shrinks, the same borrowers look fragile, and credit is withdrawn precisely when it is most needed. Ben Bernanke and Mark Gertler gave the mechanism its modern name, the financial accelerator: because the willingness to lend depends on the borrower's net worth, and net worth rises and falls with the cycle, a modest shock to fundamentals is magnified through balance sheets into a large swing in credit and spending.

Hyman Minsky drew the deeper and more unsettling lesson: that stability is destabilizing. A long stretch of calm and prosperity does not entrench itself; it breeds the confidence that undoes it. As good times persist, the memory of disaster fades, risk comes to seem a fool's worry, and borrowers and lenders alike grow bolder. Minsky charted the drift through three postures of finance. In hedge finance, a borrower's income comfortably covers both interest and principal. In speculative finance, income covers the interest but not the principal, which must be perpetually rolled over. In Ponzi finance, income covers neither, and the borrower survives only so long as asset prices keep rising to permit refinancing. Over a long boom the financial structure migrates, unseen, from the safety of hedge toward the fragility of Ponzi — and the system grows brittle in exact proportion to how secure it feels.

Then comes the moment the fragility is called. Some event — a rise in interest rates, a prominent default, a stumble in the asset whose ascent the whole edifice required — punctures the belief that prices will keep climbing. The realization spreads that debts cannot be served, and the rush to sell and to deleverage begins: the Minsky moment, after which the financial accelerator runs in reverse at full speed, asset prices collapsing, collateral evaporating, credit freezing. The figure draws the asymmetry that is the signature of the credit cycle — leverage climbing by slow, quickening stages through the boom, then crashing far more steeply than it rose. And because financial busts impair the balance sheets of borrowers and banks alike, the recessions they bring are the deepest and most stubborn — the slow, grinding recoveries of a balance-sheet recession, in which everyone is repairing their finances at once and no one is spending.

Every boom carries the cause of its own ending, and the ending, when it comes, is swifter than the rise. The fragilities accumulate quietly through the expansion — capacity built ahead of demand, leverage drifting toward Ponzi, prices and expectations climbing past anything the economy can deliver — until they reach a limit, and some trigger reverses the belief that had sustained them. The trigger may be small and almost arbitrary: a rise in interest rates, a single conspicuous default, a piece of news that punctures the story. What matters is the asymmetry it sets off. Expansions are long and gradual; contractions are short and violent. The wave crests slowly and breaks fast — because fear travels faster than confidence, and because the mechanics of the downturn, unlike those of the boom, are self-accelerating in a way that feeds on speed itself.

The clearest case of that self-acceleration is the run, which the volumes on money and on the spread have met from the inside. When belief reverses, the rush for the exits becomes the catastrophe: everyone tries at once to sell the asset, to withdraw the deposit, to call in the loan, and the simultaneity is fatal. Liquidity — the power to sell without moving the price — evaporates exactly when it is most wanted; fire sales drive prices below any sober estimate of worth; the solvent are dragged down with the insolvent because no one can tell them apart in the panic. A run is a coordination failure in its purest form, of the kind Diamond and Dybvig modeled: each depositor is rational to flee because the others are fleeing, and the flight of all destroys the institution that, had they stayed, was sound. The crisis is the financial accelerator running in reverse at maximum velocity.

Irving Fisher, who had lost a fortune in the crash of 1929, drew from it in 1933 the cruelest mechanism of all: debt-deflation. In an economy thick with debt, debts are fixed in money terms while prices are free to fall — and when prices fall, the real weight of every debt grows heavier, for each dollar owed must now be earned from cheaper goods. So the distressed sell assets to pay down what they owe; the selling drives prices lower; the lower prices raise the real burden of the remaining debt; which forces yet more selling. Fisher's terrible summary was that the more the debtors pay, the more they owe — the collective stampede to deleverage deepening the very distress it flees, a paradox of deleveraging that is kin to the paradox of thrift. By this spiral an ordinary downturn becomes a depression, as it did in the years he was writing, deflation and debt devouring each other while the real economy starved.

Here is the scandal that called modern macroeconomics into being. An economy can settle into a lasting state in which its factories run below capacity and its workers stand involuntarily idle — willing to work at the going wage, and unable to find anyone to hire them — and it can stay there, with no automatic force returning it to full employment. This was the claim with which Keynes, in The General Theory of 1936, broke a century of orthodoxy. The classical doctrine had held that markets clear: that a glut of anything, labor included, drives down its price until the glut is gone, so that unemployment must be voluntary, frictional, or brief. Keynes denied it for the economy as a whole. A slump, he argued, is demand-deficient — there is simply not enough spending to buy what full employment would produce — and the shortfall can persist, because the price system that clears a single market has no sure way to summon the demand the whole economy lacks.

The mechanism is the Keynesian cross, the multiplier seen from the side of the slump. When there is slack, output is governed by aggregate demand: firms produce what they can sell. Let planned spending fall — investment collapses, or households resolve to save more, the paradox of thrift made flesh — and output falls with it, until the lower output generates just enough saving to match the lower investment. But that resting point can lie well below full employment. The figure draws it: a forty-five-degree line along which expenditure would equal output, and a flatter aggregate-expenditure line whose slope is the propensity to consume; where they cross is the equilibrium the economy actually reaches. That crossing can sit to the left of full-employment output, and the gap between them — the recessionary gap — is idle capacity and idle people, output that could be produced and is not. Nothing in the geometry forces the gap to close on its own.

Why does the economy not cure itself? The classical answer was that wages and prices fall until employment is restored. Keynes's reply was twofold and devastating. First, wages and prices may simply not fall — they are sticky, bound by contracts and resistance and the difficulty of all moving together. Second, and more profound, even if they did fall it might not help, and could hurt: a general cut in wages cuts incomes, and incomes are demand, so the cure drains the very spending the economy needs; and falling prices raise the real burden of debt, igniting Fisher's deflationary spiral. So wage-and-price flexibility, the classical remedy, can deepen the slump rather than lift it. The economy is stuck not because any individual behaves foolishly but because they cannot coordinate a return — each firm waits to hire until demand revives, each household waits to spend until employment is secure, and the demand that would justify the hiring is the spending that waits upon the jobs. Involuntary unemployment is the human face of that failure: the output gap, measured in idle lives.

If the slump is a shortfall of demand that the private economy cannot make good, the remedy follows with a certain logic: supply the demand. This is the practical heart of the Keynesian revolution, and its boldest claim is that the state, which sees and acts upon the whole, can coordinate what the parts cannot. Fiscal policy is the first instrument. When households and firms will not spend, the government spends in their place — building, employing, transferring — or it cuts taxes to put spending power in private hands, and the multiplier makes each dollar of injection raise output by more than a dollar. In a deep slump the spending need not come at the expense of private activity, for the resources it employs were standing idle; the worker hired onto a public project was not otherwise at work. Monetary policy is the second instrument. The central bank cuts the interest rate, easing the terms of credit, coaxing forward the investment and consumption that a lower cost of borrowing makes worthwhile. Between them, fiscal and monetary stabilization are the great practical fruit of the revolution — and the reason, perhaps, that the world has not since seen a depression of the old severity.

Some of the stabilization, and not the least of it, happens with no hand on any dial. The automatic stabilizers are built into the fiscal structure: when the economy slumps, tax receipts fall of their own accord and unemployment benefits and welfare payments rise, cushioning household incomes and propping demand without any decision being taken; when the economy booms, the same machinery runs in reverse, receipts swelling and supports receding, quietly restraining the expansion. Because they require no forecast and no vote, the automatic stabilizers act at once, where discretionary policy must wait on recognition and decision; they are stabilization made structural — a thermostat rather than a hand repeatedly adjusting the flame. A great part of what keeps modern downturns shallower than their forebears is this silent, continuous leaning against the wind.

The case for active stabilization rests on a premise the next plate will contest: that the cycle is a costly failure which policy, well-aimed, can improve upon. Where the instruments still bite, the case is strong. But there is a circumstance that blunts the chief of them — the zero lower bound, the floor below which the central bank cannot easily push interest rates, reached precisely in the deepest slumps when stimulus is most wanted. There the conventional lever loses its grip, and the burden falls back upon fiscal policy and upon unconventional measures — the large-scale asset purchases of quantitative easing, the promises about future policy that go by the name of forward guidance — improvised at the frontier of what the tools can do. The slump of the 1930s, the response to the crisis of 2008, the emergency of the pandemic: each in turn tested the toolkit, extended it, and left the question of stabilization's reach an open and urgent one.

Against the confidence of the stabilizers rose a counter-revolution, and its opening argument was about time. Milton Friedman insisted that monetary policy works with long and variable lags: the effect of a change in policy arrives not at once but months or years later, and the delay is itself unpredictable. A stimulus aimed at today's slump may land in tomorrow's boom, adding fuel where none is wanted and amplifying the very cycle it meant to damp. From this Friedman drew a conclusion hostile to the activist hand: that discretion is more likely to destabilize than to stabilize, and that an economy is better served by a steady, announced rule — money growing at a constant rate, the k-percent rule — than by the well-meant interventions of those who cannot see the lag they act across. And the apparent tradeoff that had seemed to give policymakers a menu to choose from — the Phillips curve, the inverse relation between inflation and unemployment — was, Friedman and Edmund Phelps argued independently, a mirage that vanishes the moment one looks at the long run.

Their reasoning produced the expectations-augmented Phillips curve, the figure of this plate. The short-run tradeoff is real, but it exists only because inflation is unexpected. A burst of surprise inflation, by raising prices before wages catch up, briefly cheapens labor and so lowers unemployment — but only until workers and firms perceive the inflation and build it into their wage demands and price-setting. Once the inflation is expected, the spell breaks: unemployment returns to its natural rate, now accompanied by higher inflation. There is, in consequence, no permanent tradeoff to exploit; the long-run Phillips curve is vertical, anchored at the natural rate of unemployment, and a policymaker who tries to hold unemployment below it merely ratchets inflation upward with each turn. The stagflation of the 1970s — high inflation and high unemployment together, an impossibility on the naive curve — confirmed the critique in the most public way and shattered the old confidence.

The deeper blow came from the rational-expectations revolution of Robert Lucas and Thomas Sargent. If people form expectations rationally — using all the information they have, including their understanding of how policy behaves — then any systematic policy is anticipated and, being anticipated, neutralized; only surprises move real output, and surprises cannot be manufactured to order. The Lucas critique sharpened the point into a methodological scandal: the historical relationships on which policymakers rely will themselves shift when policy shifts, because expectations adjust, so a model that treats those relationships as fixed will mislead exactly when it is used to guide a change. Finn Kydland and Edward Prescott added the problem of time inconsistency: a government free to choose afresh each period will promise low inflation and then, once expectations are set, be tempted to inflate for a quick boom — and being foreseen to yield to that temptation, will be saddled with an inflation bias it never wanted, an argument for binding rules and an independent central bank over unfettered discretion. The same pair pressed the most radical claim of all in their real-business-cycle theory: that the cycle, in its essentials, is the efficient response of a well-functioning economy to real shocks — chiefly to technology — and is therefore no failure at all, and nothing a wise policy would try to smooth away.

There is a class of economic outcomes that do not merely respond to belief but are constituted by it — outcomes that occur because they are expected, and that would not occur were they not. The run is the sharpest instance: a bank is sound until enough depositors expect it to fail, and their expectation, by sending them to the door, makes it fail. The cycle has the same character on a larger canvas. If firms and households come to expect a recession, they cut their investment, their hiring, their spending — and the cutting is the recession, arriving on schedule to confirm the forecast that summoned it. Such expectations are self-fulfilling; the belief furnishes its own evidence. And to the degree the cycle runs on them, it is not a response to fundamentals at all but a creature of what the participants happen, together, to believe — a fact that unsettles every theory built on shocks to the real machinery.

The formal home of this idea is the theory of multiple equilibria. When actions are strategic complements — when each person's best move rises with what the others are expected to do, so that one invests if others invest and hires if others hire — the economy can possess more than one self-consistent resting point. There is a good equilibrium, in which all are optimistic and active and output is high, and a bad equilibrium, in which all are cautious and idle and output is low, and each is fully self-sustaining; the choice between them turns on nothing in the fundamentals but on belief alone. The figure draws the mechanism: a best-response curve that bends into an S and so crosses the line of self-consistent belief three times — a low stable equilibrium, a high stable equilibrium, and between them an unstable tipping point. Extrinsic signals correlated with nothing real — what the theorists call sunspots — can coordinate beliefs and tip the economy from one to the other. A recession, on this reading, can be a coordination upon the bad equilibrium, and curable by anything that shifts the shared belief, which is much of what stabilization, and the central banker's careful cultivation of confidence, is really trying to do.

Beneath the formalism lies the oldest name for the thing — animal spirits, Keynes's phrase for the spontaneous urge to action rather than inaction on which investment depends. Most of the consequences of an investment unfold over a future that cannot be known, not merely risked; no calculation can reach them, and so the decision to build, to commit, to wager on tomorrow rests in the end on confidence, on mood, on the conventions by which people pretend to a knowledge they do not have. When the spirits are high, investment flows and the high spirits are vindicated; when they fail, investment dries and the gloom confirms itself. George Akerlof and Robert Shiller revived the term for a whole psychology of the macroeconomy — confidence and its swings, the sense of fairness, the stories people tell themselves, the recurrent bouts of exuberance — and argued that no account of the cycle that omits them can be complete. The cycle, in this light, is partly the swinging of a collective mood, amplified by the very complementarity that makes optimism and pessimism each able to confirm itself. It is the thread that ties this volume to The Bias, where the herding and the exuberance are anatomized; to The Norm, where convention is the anchor that belief clings to; and to The Game, where coordination is the prize and the trap.

Eleven plates have circled a single object, and it is time to say plainly what the circling has found. The cycle is the recurring fluctuation of the whole — and it is not a flaw bolted onto an otherwise smooth machine, to be filed away once the right repair is made, but a property of the machine's very design. It is what a system of separate decisions, coordinated only by prices, credit, and expectation, does when it grows large and the future is unknown: it does not glide, it surges and stalls. The wave and its phases; the fallacy of composition that turns private prudence into public danger; the boom that funds its own ascent; the shock the structure rings with; the neutrality of money that holds in the long run and fails in the short; the credit that swells and vanishes; the turn and the run and the deflationary spiral; the demand-deficient slump in which willing hands stand idle; the cure the state may attempt and the limits that hedge it; the belief that makes itself true — each was a facet of one emergent thing.

Where, in the end, does the cycle come from? From three sources, and the honest answer is that it comes from all of them at once. There are real shocks — to technology, to harvests, to the price of oil, to the productivity of labor — propagated by the slow adjustment of capital. There are monetary and financial shocks — to the quantity of money, sticky against the prices that should absorb it; to the supply of credit, swelling in confidence and collapsing in fear; to the leverage that builds toward a Minsky moment and unwinds in a crash. And there are shocks of belief — the self-fulfilling expectation, the coordination upon a bad equilibrium, the swing of animal spirits. No single engine drives the cycle; it is overdetermined, the work of all three, and most dangerous where they reinforce one another — where a real shock frightens belief, and frightened belief freezes credit, and frozen credit deepens the real shock.

Whatever the impulse, the shape is the economy's own. A disturbance enters, and the couplings that bind the parts convert it into a wave: the multiplier and the accelerator, the prices that lag, the balance sheets that amplify, the expectations that run ahead. This is the lasting wisdom of the framework of impulse and propagation — that the cause of a cycle and its character are different things, the one supplied by the shock and the other by the structure, so that the rhythm an economy displays is a signature of its own, imposed upon whatever happens to strike it. To understand a cycle is to understand both: where the blow came from, and why the bell rings as it does.

Over all of this runs the contest the volume has refused to settle, because the discipline has not settled it. On one side, the conviction descended from Keynes: the cycle is a failure, demand governs output when there is slack, the slump can persist, and the state should act. On the other, the conviction descended from the classics and revived by Lucas and the real-business-cycle theorists: the cycle is a response, expectations neutralize systematic policy, markets clear, and intervention is as apt to harm as to help. Between them the New Keynesian synthesis has built a working peace rather than a verdict — conceding the classical insistence on forward-looking expectations and the Keynesian insistence on sticky prices, and concluding that money and demand are non-neutral in the short run and neutral in the long, that stabilization is possible but bounded, and that the deepest question — whether a recession is a sickness or a healing — remains, beneath the technical apparatus, open.

It would be a failure of nerve, though, to end in mere irresolution, for the study of the cycle has achieved a great deal that is not in doubt. It learned to measure the thing — to build the national accounts that let a society see its own output whole, and to date the turnings that had once been only felt. It produced the framework of impulse and propagation that organizes all the rest. It came to understand the short-run power of money, and with it the central bank's capacity to lean against the wind. And above all it forged the tools of stabilization — fiscal and monetary, discretionary and automatic — that have, for the better part of a century, kept the world from any repetition of the Depression's depth. The slumps still come; but they have not, since, been suffered to become catastrophes of that order, and the difference is counted in lives not ruined. The science is real, and it has saved much.

And yet the humility the volume has practiced at every edge is not false modesty but the truth of the matter. Macroeconomics cannot foretell the turning points — the very crashes it best explains in hindsight it almost never calls in advance. The great divide is unresolved at the root. Every serious crisis surprises it, the largest of the recent age least foreseen of all. Its central magnitudes — potential output, the natural rate, the multiplier — are not measured but estimated, with errors wide enough to swallow the decisions they inform. And its models share an unnerving habit: those that perform handsomely in tranquil times tend to break precisely in the storms for which one most wants them. After a century of effort the cycle remains, in part, a mystery — understood in its mechanisms, unmastered in its arrival.

This volume opens The Aggregates, and with them the atlas's turn from the part to the whole — from the price of a thing to the level of all prices, from the output of a firm to the output of a nation, from the single market to the macroeconomy that contains it. The cycle is the short-run fluctuation; its companions in this group complete the picture of the whole. Its nearest neighbor is The Growth, the long-run trend around which the cycle merely oscillates — and the larger prize by far, since what growth does to human welfare across the generations dwarfs anything the cycle does within them. There is The Level, the price level and its inflation, the nominal companion to this volume's real concerns and the other axis of the Phillips curve met here. There is The Money, the institution that is at once the engine of the monetary cycle and the object of the run. And there are the threads backward — to The Rate, where Wicksell's natural rate first appeared; to The Spread and The Bias, where the liquidity that evaporates in a panic and the exuberance that inflates a bubble are studied at close range — and forward to The Share, where the question of how the cycle's burdens divide, falling hardest on those least able to bear them, awaits its proper treatment.

What, then, does it cost a free economy to be a cyclical one? The cycle is the price the market pays for its own freedom and interdependence — for coordinating the plans of millions through prices and credit and expectation rather than by command. The same decentralization that makes the market a marvel of allocation, achieving without a planner what no planner could achieve, is what leaves it prone to the failures of the whole: to the recession that is a coordination failure made flesh, the moment when each actor, drawing back in individual prudence, confirms by the drawing-back of all the very fear that prompted it. Whether such failures can be cured without summoning worse, whether the cycle in the end is malady or medicine, this volume has weighed and declined to decide — not from timidity but because the contest is genuine and the evidence has not closed it. The aggregate, like every price the atlas has examined before it, turns out to be a verdict on a hidden contest: here the contest between the parts and the whole, between the prudence of each and the prosperity of all, between the freedom to choose and the fluctuations that freedom brings. The argument remains open.