The most consequential intellectual revolution in twentieth-century economics — and what the discipline made of it. The chapter walks what Keynes inherited from Wicksell, Malthus, and Marshall, what the General Theory argued at its strongest form, what Hicks’s IS-LM preserved and what it lost, the neoclassical synthesis Samuelson built on top of the formalization, Schumpeter as the great contemporary alternative the synthesis could not absorb, and the three readings of what Keynes really meant that the post-1937 literature has been arguing over since.
What happens when the losing side of an old debate turns out to have been asking the right question? Chapter 3 §3.4 walked the Malthus-Ricardo glut debate of 1815–1820 as the road the classical school did not take. Malthus argued that aggregate demand could fall short of aggregate supply, that gluts were possible, and that the system did not automatically employ everything offered to it. Ricardo answered with what the discipline would later call Say’s Law: supply creates its own demand, savings are spent on investment, and a general glut is a category mistake. The profession sided with Ricardo. The argument went away. A hundred and thirty years later, in 1936, John Maynard Keynes opened the General Theory of Employment, Interest and Money by saying Malthus had been right about the wrong thing in the right way. The chapter starts here because the inheritance is not a footnote. It is the pattern the chapter is built around: a revolution assembled from materials a previous generation had set aside, and from one major source (Wicksell) the previous generation had treated as continuous with the orthodox program.
Three intellectual debts are load-bearing for what Keynes built. The first runs to Knut Wicksell. Wicksell’s Interest and Prices (1898) had argued that the economy contains two interest rates: a natural rate, the rate at which the marginal productivity of capital equals the marginal cost of waiting, and a market rate, the rate banks actually charge. When the two coincide, the price level is stable. When the market rate sits below the natural rate, lending expands, investment runs ahead of saving, and prices rise; when it sits above, the reverse. Wicksell’s framework is monetary-disequilibrium analysis, and its policy implication is that central banks should manage the gap. Keynes absorbed the framework and inverted its conclusion. The two-rate apparatus stayed; the disequilibrium it described changed. For Wicksell, the source of trouble is the gap between the natural and market rates, with prices as the symptom. For Keynes, the source of trouble is that the rate at which savings would equal investment at full employment may be lower than any rate the monetary system can deliver, with output as the variable that adjusts. Same machinery, opposite policy payload.
The second runs to Thomas Malthus. Malthus’s effective demand argument, the claim that capitalist economies can sustain insufficient aggregate demand because the propensity to consume out of profit income falls below the propensity to consume out of wage income leaving a residual the system cannot dispose of unless something replaces it, was the live concept Ricardo had buried. Keynes recovered the concept and rebuilt it. The Malthus-Ricardo replay is the pattern: not “Malthus was right and Ricardo was wrong,” but the more interesting structure where Malthus was wrong about why effective demand could be deficient (his explanation routed through landlord consumption and a quasi-physiocratic theory of luxury) and right that it could be deficient at all. Keynes kept the conclusion, threw away the route, and built a different mechanism (liquidity preference, animal spirits, the volatility of investment) that delivers the same result on grounds Malthus could not have stated. The replay is not vindication. It is the structural fact that an argument the discipline had retired carried a question the discipline still needed to answer, and that the answer would arrive by a route the original disputants had not anticipated.
The third runs to Alfred Marshall, the most complicated of the three. Marshall was Keynes’s teacher at Cambridge; Keynes inherited the Cambridge tradition of monetary analysis (the cash-balances approach to the quantity theory, the partial-equilibrium toolkit of ch. 5 §5.3, the master-pupil discipline of textbook-driven training) and built on it. Marshall’s Money Credit and Commerce (1923) had treated money as one entry in the partial-equilibrium framework, and Pigou’s extensions had given Cambridge a working monetary apparatus by the 1920s. Keynes was both heir to and rebel against this tradition. The cash-balances framing, the demand-for-money emphasis, the methodological discipline of writing for an educable profession: all of it Keynes carried forward. What he broke from was the assumption underneath the tradition: that money was a veil over a real economy that, given enough time, would clear at full employment. Marshall was incubator and target. Without Cambridge there is no Keynes. Without Keynes’s break from Cambridge, there is no General Theory.
What Keynes was reacting against was not only classical political economy. The marginalist consolidation that ch. 5 walks (Walras’s general equilibrium, Marshall’s partial-equilibrium scaffolding, Pigou’s welfare economics, the simultaneous-equation method that defined the discipline by 1930) supplied the methodological orthodoxy the General Theory set itself against. Keynes was inside that orthodoxy as a Cambridge-trained economist, and the book’s opening argument (that the “classical economists” had taken full employment as a special case and were building the science of an economy that did not exist) reads as an attack on the marginalist program as much as on Ricardo. The opponent is not a single chapter; the opponent is the consolidated apparatus the previous chapter narrated.
The chapter’s working position is stated upfront. The General Theory was a deeper intellectual break than the IS-LM apparatus that carried it into textbooks. Hicks preserved the demand-driven mechanism but lost the qualitative content that made Keynes’s framework different from a marginalist-with-rigid-wages. The neoclassical synthesis then stitched the formalization back together with classical long-run equilibrium, producing the postwar consensus and the macro curriculum it left behind. The synthesis worked, in the sense that it gave governments a toolkit, departments a curriculum, and the discipline an organizing framework. The question is what kind of working it was: faithful elaboration, productive compromise, or domestication that lost what was radical. Each of the three readings the chapter walks in §10.6 is a different answer.
Definitions used loosely above will be tightened as they appear. Effective demand means aggregate spending in the system as a whole, considered as a quantity that can fall below the level of output offered. The natural rate of interest is Wicksell’s name for the rate at which planned saving equals planned investment without monetary distortion. The Cambridge tradition names the school of monetary analysis that ran from Marshall through Pigou through Robertson, treating money demand as a portfolio decision (cash balances) rather than as a transactions residual. The relational landscape of these inheritances and what came after is anchored at the keynesian revolution era cluster of the timeline. What Keynes built from these materials was something none of his sources had imagined.
The General Theory of Employment, Interest and Money (1936) begins with a demolition. The book’s first task is to clear away what Keynes called the “classical” framework, by which he meant a single composite tradition running from Ricardo through Mill, Marshall, and Pigou, identified by its commitment to Say’s Law, the doctrine that aggregate supply creates its own aggregate demand. On the classical view, production generates exactly the income needed to purchase the output, savings flow into investment through the interest rate, and any departure from full employment is a friction (rigid wages, monopoly distortion, intervention) the system would clear given time. Keynes’s rejection is structural, not empirical. The book is not arguing that Say’s Law fails sometimes. It is arguing that Say’s Law is the wrong frame. Income spent on consumption purchases consumption goods; income saved does not automatically purchase investment goods, because the decision to save and the decision to invest are made by different agents on different grounds. The interest rate that classical theory assigned the job of equilibrating the two cannot do the job once the savings-investment identity is recognized as an accounting equality holding ex post rather than a behavioral mechanism guaranteeing it ex ante.
From the rejection of Say’s Law follows the chapter’s first substantive concept. Involuntary unemployment is unemployment that persists even though workers are willing to work at the prevailing wage. The classical position is that this cannot happen except as a transient friction; flexible nominal wages clear the labor market, and any persistent unemployment must be voluntary (wages above market-clearing because of union power, regulation, or worker preference) or frictional (search, mismatch, churn). Keynes’s claim is that involuntary unemployment is a normal state of capitalism, not a friction. The mechanism is aggregate: when effective demand is insufficient to absorb the output a fully employed labor force would produce, output adjusts downward and unemployment is the result. Wage flexibility does not solve the problem because cutting wages cuts the income that finances aggregate demand, and the demand-deficiency the wage cuts were supposed to address worsens. The economy can sit at an underemployment equilibrium for extended periods because the mechanism that would push it back to full employment in the classical framework (falling prices and wages restoring real money balances and reviving demand) either does not run or runs perversely.
The next move supplies the monetary mechanism that makes the demand-deficiency story coherent. Liquidity preference is Keynes’s theory of money demand. People hold money for three reasons: transactions (the cash needed to mediate ordinary spending), precaution (a buffer against unexpected expense), and speculation (a hedge against uncertainty about the future course of interest rates). The speculative demand is the load-bearing piece. Holders of wealth choose between money (which earns no interest but can be deployed instantly) and bonds (which earn interest but lose value if rates rise). At low interest rates, the loss from being wrong about future rates can exceed the interest forgone by holding cash; at very low rates the bond becomes a one-way bet, and money demand becomes effectively infinite (the liquidity trap). The interest rate, in this account, is not the price that equates saving and investment. It is the price that clears the money market, set by the interaction of money supply and the public’s liquidity preference. Saving and investment are equated by the level of income, not by the interest rate; income adjusts so that, at the prevailing interest rate, investment matches saving.
Liquidity preference connects to the volatility of investment through the most contested concept in the book. Animal spirits is Keynes’s phrase for the impulse driving investment decisions: “a spontaneous urge to action rather than inaction, and not as the outcome of a weighted average of quantitative benefits multiplied by quantitative probabilities.” The reasoning is structural. Investment commits resources today against returns delivered over years or decades. The probability distributions over those returns are not knowable; the conditions that will determine the returns (technology, competition, demand, policy, war) cannot be enumerated, let alone weighted. The relevant uncertainty is what Keynes (and independently Frank Knight in Risk Uncertainty and Profit (1921)) called fundamental uncertainty: uncertainty that cannot be reduced to calculable probability. Investors form conventional expectations, hold them with varying confidence, and revise them in jumps when confidence breaks. Animal spirits is not irrationality. It is the analytical name for what investment decisions are made of when the framework that would make them calculable is not available. Investment volatility, on this account, is not noise around a stable function. It is the structural feature of an economy in which the future cannot be pinned and where collective swings of confidence drive expansion and contraction.
The mechanism that translates demand changes into output changes is the multiplier. Richard Kahn had developed the apparatus in “The Relation of Home Investment to Unemployment” (1931): an initial increase in spending raises the income of those who receive it; some fraction of the new income is spent on consumption goods; the recipients of that consumption spending see their incomes rise; some fraction of that is spent again; and the process iterates, with each round smaller than the last. The total increase in income is a multiple of the initial spending, with the multiplier’s size determined by the marginal propensity to consume. Keynes appropriated Kahn’s analysis and made it the transmission channel for fiscal policy. Government spending, by injecting demand the system did not previously have, drives output up by a multiple of the injection. The formal derivation lives in economics ch. 8, which carries the IS-LM model and its multiplier as a working tool. The intellectual point here is not the algebra. It is that Keynes had a theory of how policy could move output, with a quantitative mechanism the classical framework did not supply.
Read “the multiplier” as a formula and it is an inert ratio; drive it and it becomes a cascade. Move the marginal propensity to consume and watch the planned-expenditure line rotate against the 45° line, the equilibrium income slide, and the round-by-round re-spending populate below. Step an autonomous injection (Kahn framed it for investment; Keynes appropriated it for fiscal policy) and the total income change is a multiple of the injection — larger the higher the propensity to consume, vanishing as it falls.
Figure 10.3 (interactive). The Keynesian cross. Equilibrium income sits where planned expenditure $E = A + \text{MPC}\cdot Y$ crosses the 45° line. An autonomous injection shifts $E$ up; the gap between the injection and the resulting income change is the multiplier, and the round-by-round panel shows where it comes from. Drag MPC; step the injection; flip the channel.
Spend a dollar into a factory town and the worker who earns it spends most of it at the next shop; that shopkeeper spends most of that; and the ripple keeps going, each round smaller than the last. Add up the whole geometric ripple and a single injection produces several dollars of income — the higher the share people re-spend (MPC), the longer the ripple runs and the larger the total. Kahn built the mechanism for investment in 1931; Keynes turned it into the case for fiscal policy.
The injection re-spends as a geometric series $\Delta Y = \Delta A\,(1 + \text{MPC} + \text{MPC}^2 + \cdots) = \dfrac{\Delta A}{1-\text{MPC}}$, so the multiplier is $\dfrac{1}{1-\text{MPC}}$.
Equilibrium income solves $Y = A + \text{MPC}\cdot Y$, i.e. $Y^\* = \dfrac{A}{1-\text{MPC}}$. Raising the autonomous component $A$ by $\Delta A$ moves $Y^\*$ by $\dfrac{\Delta A}{1-\text{MPC}}$ — the same multiple. $\Delta G$ and $\Delta I$ enter $A$ identically; the channel changes the story, not the arithmetic.
The deepest claim takes the longest to state. The monetary-production economy is Keynes’s name for an economy in which money is not a neutral medium of exchange laid over a real barter economy, but enters the production decision itself. Entrepreneurs borrow money to acquire inputs and pay wages; they produce in the expectation that the money receipts from selling output will exceed the money outlays of production; and the properties of money (its liquidity, its zero or near-zero elasticity of production, its function as the unit in which contracts are written and debts must be repaid) shape what gets produced and what does not. Money is not a veil. The decision to hold money rather than to invest in productive capacity is a real economic choice with real consequences for output and employment. This is what makes the speculative demand for money matter: an increase in liquidity preference draws purchasing power out of the goods market and into the money market, depressing investment and output without any change in “real” preferences or technologies. The classical framework, by treating money as a neutral medium, could not see this mechanism because the framework had abstracted from it at the start. The growth-theory side of the framework, the question of what happens to the demand-output story when the capital stock is changing over time, was taken up by Roy Harrod (1939) and Evsey Domar (1946) in the Harrod-Domar model. The formal Harrod-Domar apparatus, including its knife-edge instability and its successor in Solow’s growth theory, lives in economics ch. 13; the General Theory framework here gets only the static demand-output cut.
The four moves — rejection of Say’s Law, involuntary unemployment as normal, liquidity preference as the monetary mechanism, the monetary-production economy as the setting — hang together as a single argument about what kind of system a capitalist economy is. The economy operates with effective demand as its forcing variable; effective demand is unstable because investment depends on conventional expectations under fundamental uncertainty; the monetary system can fail to deliver the interest rate that would equate saving and investment at full employment, because the interest rate is set by liquidity preference rather than by saving-investment equilibration; and the result is that involuntary unemployment is not a special case but a frequent feature, with the economy capable of resting at output below capacity for periods that matter politically and economically. The Depression was the empirical setting in which this framework was written; economic-history ch. 12 walks the event the framework was responding to, and the institutional collapse of the gold standard, the banking failures of 1930–1933, and the chronic mass unemployment of the 1930s gave the classical framework an empirical embarrassment the marginalist consolidation could not absorb. Keynes and the General Theory as a work sit relationally on the timeline. The modern policy debates that descend from this framework live in Walkthrough 01 (the multiplier and fiscal policy) and Walkthrough 08 (the demand-deficiency explanation of recessions); both Walkthroughs carry the contemporary version of arguments the General Theory started.
You just drove the multiplier (IX1). The modern $1T-stimulus argument is this 1936 mechanism, still contested.
Where the modern fiscal-multiplier debate started: Kahn's 1931 multiplier → Keynes's appropriation of it as the transmission channel for fiscal policy → the Hicks IS-LM tradition the Krugman-side argument still runs through. The size of the multiplier — whether a dollar of spending buys more or less than a dollar of output — is the hinge the whole stimulus debate turns on, and the apparatus this section walks (with IX1) is its origin.
The General Theory gave the first workable demand-side answer — and left a gap the later schools filled.
The Keynesian revolution supplied the first workable theory of recessions: a demand shortfall, amplified by the multiplier, drives output below capacity and holds it there. What it explained — why a fall in spending need not self-correct — and what it left open: price-stickiness is asserted, not derived, a gap the New Keynesian program (and the counter-revolution) would press on.
This is the founding move of the thread — the synthesis/Phillips operationalization is a later stop (redraft ch.9).
What Keynes broke: involuntary unemployment as a real category, not a friction — idle workers willing to work at the going wage, held idle by a shortfall the labor market cannot itself fix. That is the thread's origin node. NOTE: the synthesis/Phillips-curve operationalization of the idea routes to the postwar-synthesis chapter (redraft ch.9), not here.
Keynes's animal spirits is the origin node of the expectations thread that runs forward to behavioral economics.
Keynes's animal spirits — expectations as conventional beliefs held with varying confidence, not probability distributions — is the thread's origin node. Investment under fundamental uncertainty cannot be reduced to a weighted average of calculable returns; the convention holds until confidence breaks, and then it shifts in a jump. The thread runs from here to Kahneman-Tversky and prospect theory.
Keynes reframed unemployment as a demand-side, money-side phenomenon — the founding move this walkthrough opens on.
The General Theory introduces involuntary unemployment as a demand-side phenomenon: the shortfall is in effective demand, mediated through liquidity preference and the interest rate the money market sets — the founding move that made “unemployment is a money problem” coherent rather than a category error. Whether the right lever is the labor market or the monetary system is the question the walkthrough runs forward.
One context piece is worth pinning. The intellectual project behind the Bretton Woods institutions of 1944 (the IMF, the World Bank, the dollar-anchored fixed-exchange-rate system) was substantially Keynes’s, and Keynes himself was the British negotiating principal. Bretton Woods as institutional event is hist ch. 13’s territory. The Keynesian intellectual project that produced the framework the negotiators were operating inside is this chapter’s territory. The two are connected, and the cross-link is reciprocal. What this section has done is establish the framework. What the next section will show is what the framework looked like once Hicks formalized it, and what the formalization changed.
In 1937, one year after the General Theory, John Hicks published an article that would become more influential than the book it summarized. “Mr Keynes and the Classics: A Suggested Interpretation” appeared in Econometrica. It was twelve pages long and contained four equations, two diagrams, and a proposal: that Keynes’s framework could be written as a system of simultaneous equations, with the goods market summarized by an IS curve (the locus of income-interest-rate combinations at which investment equals saving) and the money market summarized by an LM curve (the locus at which money demand equals money supply), and the equilibrium found at their intersection. The IS-LM model, as the apparatus came to be called, became the textbook Keynes. The book it modeled was set aside, and the model became the tradition.
Hicks’s 1937 article had an ambition Keynes’s book did not. Keynes had written a verbal argument with mathematical interludes; Hicks proposed a translation into the simultaneous-equation language the marginalist consolidation had made the discipline’s working medium. The translation was generous; it preserved more of Keynes’s substance than the “classical” treatment Hicks set against it. What was preserved is worth naming. Demand-driven output determination survived the translation: in IS-LM, output is determined by the intersection of goods-market and money-market equilibrium, and demand shocks shift the equilibrium in the direction Keynes’s framework predicted. The liquidity-preference interest-rate mechanism survived: the LM curve is the liquidity-preference relation written as a function of income, and the interest rate is set by the money market rather than by saving-investment equilibration. The multiplier survived as an equilibrium concept embedded in the IS curve’s slope. By any reasonable accounting, IS-LM kept the operational core of Keynes’s policy framework. Demand management worked inside it; fiscal policy moved output; monetary policy moved interest rates. As a working policy tool, the apparatus delivered.
What did the formalization keep? Demand-driven output and the liquidity-preference interest rate — the curves themselves. Drag the LM curve (money supply) in the normal region and the rate falls while output rises, as Keynes's framework predicts. Then flip the liquidity trap on: the LM curve goes flat at the floor, more money slides along it without moving the rate or output — monetary policy is dead — yet dragging the IS curve (fiscal/autonomous demand) still moves output. That is Keynes's verbal “only demand support works in a slump” made structural.
Figure 10.4 (interactive). IS-LM equilibrium. The IS curve (goods market) slopes down in $(Y,r)$; the LM curve (money market) slopes up, floored at $r_{\min}$ in the trap. In the trap the equilibrium sits on the flat LM segment: money-supply shifts no longer move $r$ or $Y$, while IS shifts still move $Y$. Drag each curve; flip the trap.
In a deep slump the interest rate is already on the floor, and the public will hold any extra money rather than chase a bond that can only fall — so printing more money is “pushing on a string,” it just sits idle. The one lever left is to spend directly into the goods market, which is why the IS curve still moves output when the LM curve cannot. The model shows the structural result Keynes argued in words: when monetary policy is impotent, demand support is the only thing that works.
Goods market (IS): $Y = C(Y) + I(r) + G$, giving a downward-sloping locus in $(Y,r)$. Money market (LM): $M/P = L(Y, r)$, giving an upward-sloping locus. Equilibrium is their intersection.
The trap is $L(Y,r)\to\infty$ elasticity at $r_{\min}$: money demand becomes flat, so $\partial r/\partial(M/P)=0$. A rise in $M/P$ shifts along the flat segment — $\Delta Y\approx 0$ from money — while a rightward IS shift (higher $G$) still raises $Y$. Monetary multiplier $\approx 0$; fiscal multiplier large.
The losses are where the chapter’s argument lives. Four dimensions of the General Theory’s qualitative content did not survive the translation. The table below sketches the divergences as a navigation aid; the prose walks each at depth.
| Dimension | Keynes’s General Theory (1936) | Hicks’s IS-LM (1937) |
|---|---|---|
| Expectations under uncertainty | Conventional beliefs held with varying confidence; fundamental uncertainty cannot be reduced to probability | Calculable expected returns; agents optimize over known functional forms |
| Investment behavior | Animal spirits — “a spontaneous urge to action,” not parameter-stable | I(r) — investment as stable function of the interest rate |
| The role of money | Monetary-production economy; money enters the production decision; liquidity shapes structure | Portfolio asset; LM curve is choice between bonds and money, not a production input |
| Status of unemployment | Normal state of capitalism; aggregate demand chronically insufficient | Deviation from equilibrium; long-run tendency to full employment is recovered |
The table above shows the four trades side by side; here you perform them. Each row starts on Keynes's verbal framing. Flip a row and it swaps to Hicks's formalized version — and a one-line consequence appears: what the substitution bought, and what it cost. Flip all four to read the whole compression in one motion. The point is not that Hicks was wrong; it is that each loss is a deliberate, individually legible trade, made for the tractability the next paragraph explains.
Keynes: expectations are conventional, held with varying confidence (General Theory ch.12) — fundamental uncertainty that cannot be reduced to probability.
Hicks: expectations are known probability distributions agents optimize over.
Consequence: the economy gains a determinate, steerable equilibrium — at the cost of the regime-shift dynamics Keynes built the framework to capture.
Keynes: investment springs from “a spontaneous urge to action,” not from calculable return — animal spirits, not a parameter.
Hicks: investment is I(r), a stable function of the interest rate.
Consequence: investment becomes predictable, so the model closes — and the boom-bust dynamics driven by collapsing conviction drop out.
Keynes: money enters the production decision; liquidity shapes the economy's structure — a monetary-production economy.
Hicks: money is a portfolio choice between bonds and money (the LM curve).
Consequence: money becomes a veil in the long run — the production-side role Keynes was arguing for is put back exactly where he objected.
Keynes: unemployment is a normal state — aggregate demand is chronically insufficient.
Hicks: unemployment is a deviation from equilibrium that policy can correct.
Consequence: full employment becomes the default the economy returns to — the synthesis's “Keynesian short run, neoclassical long run.”
Figure 10.1 (interactive). The same four dimensions as the table above, made operable. Each row flips from Keynes's verbal framing to Hicks's formalization, revealing the one-line consequence of the trade. Flip rows independently, or flip all four at once.
Dimension 1: fundamental uncertainty becomes calculable probability. Chapter 12 of the General Theory, “The state of long-term expectation,” is the densest passage in the book. Expectations about returns from long-lived investments are not, on Keynes’s account, probability distributions. They are conventional beliefs held collectively with varying degrees of confidence, sustained by the implicit agreement that today’s convention will likely continue into tomorrow. The convention can break, and when it does the change is not a Bayesian update on new information; it is a regime shift in which the prior framework collapses and a new one assembles. IS-LM has no place for this. The simultaneous-equation method requires expected returns to be specified as numbers (or distributions over numbers) that agents take as inputs to optimization. Hicks did not write the assumption explicitly (the article keeps expectations in the background) but the apparatus he proposed cannot run without it. Once expectations are calculable, the structural feature Keynes identified (that conventional beliefs can shift in jumps without any underlying parameter changing) drops out of the framework as a thing the framework cannot represent.
Dimension 2: animal spirits becomes a stable investment function. Inside IS-LM, investment is written as I(r), a stable function of the interest rate. Higher r reduces investment, lower r raises it; the function’s parameters are stable across the time horizon the analysis covers. This is the operational content of Keynes’s investment story translated into a form the simultaneous-equation framework can accept. What it loses is the claim Keynes was making. Animal spirits, in the General Theory, is the analytical name for the volatility of investment that comes from confidence-driven shifts in conventional expectations: the swings the framework explicitly cannot model as parameter movement, because the parameters themselves are what the swings shift. To replace this with I(r) is to translate “investment depends on confidence we cannot model” into “investment depends on the interest rate, which we can.” The translation makes the framework workable. It also neutralizes the qualitative point. An IS-LM economy with stable I(r) does not have boom-bust dynamics driven by collapsing conviction; it has policy-correctable departures from equilibrium driven by parameter shocks. This is a different economy from the one Keynes wrote about.
Dimension 3: the monetary-production economy becomes money-as-portfolio-asset. The LM curve treats money as one of two assets in a portfolio choice, the other being bonds. Households and firms allocate wealth between the two on the basis of the interest rate the bonds pay, with money demand falling as r rises. This captures the speculative motive in Keynes’s liquidity-preference framework, and the LM relation does the work the speculative-demand argument did in the General Theory. What it discards is the deeper structural claim. In Keynes’s account, money is not just one asset among others; it is the asset in which production is financed, contracts are written, and debts must be repaid. The properties of money (its liquidity, the zero or near-zero elasticity of its supply, its monopoly position as the medium of contract) shape the production decision before the portfolio decision arises. Inside IS-LM, money has none of this. It is a holding, exchanged against bonds at a market-clearing rate, and the goods-market story (the IS curve) runs without reference to money’s production-side role. The veil that the classical framework had drawn over money has been lifted just enough to put money on the asset side of the household balance sheet, and put back exactly where Keynes was arguing against on the production side.
Dimension 4: involuntary unemployment becomes deviation from equilibrium. The most consequential transformation. Inside the General Theory, involuntary unemployment is what the system normally exhibits because aggregate demand is chronically below the level full employment would require. Inside IS-LM, the system has an equilibrium — the IS-LM intersection — and unemployment is a state from which the equilibrium can be reached by appropriate policy or, given enough time, by the economy’s own adjustment mechanisms. The framework has implicit room for full employment as the long-run state, and the textbook synthesis §10.4 will walk made this room explicit. Keynes’s qualitative claim, that the economy can rest indefinitely at output below capacity, that there is no automatic mechanism returning it to full employment, that involuntary unemployment is a normal feature of capitalist labor markets, is replaced by the proposition that involuntary unemployment is a deviation policy can correct. The political and intellectual stakes change with the substitution. A world in which unemployment is normal demands a different framework of policy and politics from a world in which unemployment is a correctable deviation. IS-LM delivers the second world; the General Theory was about the first.
Why did Hicks make these choices? Not because he had misread Keynes, and not because he was a bad-faith translator. The answer is methodological. The simultaneous-equation framework Hicks was working inside, the one the marginalist consolidation of ch. 5 had built and that defined the discipline’s working medium by 1937, demands tractability conditions Keynes’s verbal framework did not provide. To write a system of equations and find an intersection, the analyst needs known functional forms (so the equations can be specified), stable parameters (so the system has a determinate solution), and equilibrium conditions (so the intersection means something). Fundamental uncertainty resists known functional forms; animal spirits resists stable parameters; the monetary-production economy resists the closed-system specification a tractable model requires; the claim that involuntary unemployment is normal resists the equilibrium framing that gives the model a determinate answer. The four losses are not coincidental. They are what the verbal-to-formal translation costs when the verbal framework refuses the conditions formalization requires. Hicks’s choices were the price of formalization, not error. The chapter’s argument is not that Hicks should have known better; it is that the choices were real choices, that the price was paid, and that the discipline did not always remember what it had bought and what it had sold to make the purchase.
Hicks himself came to recognize this. Forty-three years after “Mr Keynes and the Classics,” in “IS-LM: an explanation” (1980), Hicks wrote a retrospective essay arguing that his own framework had been taken too literally by the profession. The General Theory, he argued, was about an economy of uncertain expectations and conventions held with limited confidence; IS-LM was a useful pedagogical device for organizing some of Keynes’s comparative-static results, but it was not a complete representation of what Keynes was about, and the discipline had been treating it as if it were. The essay is brief and the verdict is mild. Hicks did not retract the model; he relocated it. IS-LM was, in his late phrasing, a tool for short-period analysis under stable expectations — useful where those conditions held, misleading where they did not, and usable only with awareness that the framework had a domain. The arc — Hicks made a tractability trade-off in 1937, the profession took the result literally, and Hicks himself in 1980 recognized that the literalism had gone too far — is not a Post-Keynesian polemic. It is the inventor’s own correction, and it gives the “something was lost” reading support that does not depend on heterodox loyalties.
One contemporary debate is worth marking. The Hayek-Keynes argument of the 1930s, conducted between LSE and Cambridge over the proper diagnosis of the 1929 crash and the proper response, was the Austrian side of a contestation Keynes won institutionally without refuting on the terms Hayek set. Chapter 6 walks the Austrian framework Hayek brought to the debate; the relevant point here is that Keynes’s project was contested even before Hicks’s formalization domesticated it, and that the contestation came from a tradition that would not appear inside IS-LM at all because Austrian capital theory cannot be written as a simultaneous-equation system. The Austrian contestation and the Hicks formalization are different kinds of pressure on the same project. The first attacks the project on different grounds; the second translates the project into a form that loses what made the project distinctive. Both are arenas in which the General Theory’s argument was contested, and the chapter’s position is that neither was decisive, but that the second produced the working version that the postwar discipline inherited.
You just read the contestation Keynes won institutionally. This stop is the Keynesian framework at full strength.
The Treatise→General Theory arc, the Cambridge Circus that hammered the argument into shape, and Hicks's IS-LM as the channel that carried it into the postwar mainstream — the Keynesian side of the duel at full strength. By 1931 the short rate was near zero and firms would not borrow; on Keynes's diagnosis only demand support could work, and the Austrian liquidationist alternative (clear the malinvestment, let the slump run) would deepen the very shortfall that caused it.
The formal IS-LM model, with its equilibrium conditions, its multiplier formula, its policy analysis, and its open-economy extensions, lives in economics ch. 8. This chapter is asking what the formalization preserved and what it lost. Reader and writer should not confuse the two questions: economics ch. 8 carries IS-LM as a working tool; this chapter carries the intellectual transformation. Hicks and the General Theory sit relationally on the timeline. The next section walks what happened when Samuelson took the formalization and built the postwar consensus on top of it.
By the 1960s, the argument was over, or seemed to be. The neoclassical synthesis, the consensus framework Paul Samuelson articulated and the discipline taught for two decades, had reconciled the IS-LM apparatus with the marginalist long-run equilibrium the consolidation of ch. 8 had built. The reconciliation took the form of a two-run partition. In the short run, prices are sticky, output can sit below potential, and demand management works: fiscal and monetary policy can move the economy toward full employment along the IS-LM apparatus §10.3 walked. In the long run, prices adjust, markets clear, the labor market returns to full employment, and the economy reaches the marginalist long-run equilibrium where factor markets distribute output by marginal product. The phrase that captured the framework was “Keynesian short run, neoclassical long run,” and the framework was called the neoclassical synthesis: a descriptor of a consensus position rather than a proper-noun school name.
The two-run partition did real intellectual work. Keynes’s 1936 framework had argued that involuntary unemployment was a normal state and that the economy did not automatically return to full employment; the marginalist long-run apparatus had argued that markets clear, including the labor market, given price flexibility. Samuelson’s synthesis kept both, separated by a time horizon. The short run was where Keynes was right; the long run was where the marginalists were right; policy operated in the short run because the short run was where most economic life happened. The partition gave the discipline a way to teach IS-LM and the long-run apparatus together without requiring students to choose between them. It also gave policymakers a framework that licensed activist demand management without abandoning the marginalist account of how a properly functioning economy distributes its output. The synthesis was, in this sense, a productive compromise: one that allowed Keynesianism to enter graduate macro and become the discipline’s working framework rather than a heterodox tradition fighting for survival outside the orthodoxy.
The synthesis’s most visible policy output was the Phillips curve. A. W. Phillips’s 1958 paper, “The Relation Between Unemployment and the Rate of Change of Money Wage Rates in the United Kingdom 1861–1957,” was an empirical study. Phillips fitted a downward-sloping curve through nearly a century of British data, finding a stable inverse relation between unemployment and the rate of money-wage growth (and, by extension, price inflation). The relation was empirical, not theoretical; Phillips offered no behavioral derivation, only the fitted curve. The 1960 reinterpretation was the consequential move. Paul Samuelson and Robert Solow, in “Analytical Aspects of Anti-Inflation Policy” (1960), took the Phillips curve and read it as a policy menu: the government could choose its preferred combination of inflation and unemployment, accepting higher inflation in exchange for lower unemployment, or accepting higher unemployment in exchange for lower inflation. The curve became a policy instrument. A government willing to tolerate a permanently higher inflation rate could buy a permanently lower unemployment rate; a government willing to tolerate higher unemployment could lower inflation. The figure below sketches the policy-menu reading of the curve.
The Samuelson-Solow reinterpretation was the synthesis’s characteristic move: take an empirical regularity, treat it as a stable structural relation, and convert it into a policy instrument. The 1960s ran on this conversion. Governments managed demand, calibrated against the Phillips trade-off; central banks targeted unemployment-inflation combinations; the political economy of macroeconomic policy was the politics of where on the menu to sit. Two features of the move are worth pinning. First, the conversion treated the relation as structural in a way Phillips himself had not. Phillips’s 1958 paper had reported a fitted curve through historical data without claiming the relation was invariant under different policy regimes; Samuelson and Solow read it as if the relation would persist when used as a policy instrument. Second, the conversion left the relation’s mechanism unspecified. Why should higher inflation buy lower unemployment? The synthesis offered no derivation; the curve was the relation, and the relation was the menu. Both features are what the natural-rate and rational-expectations counter-arguments of the 1960s and 1970s would press against. The Phillips formal apparatus, including the short-run / long-run distinction and the expectations-augmentation that ch. 10 will walk, lives in economics ch. 9. What this chapter cares about is the intellectual move: the synthesis turned an empirical observation into a policy menu, and that conversion is what the counter-revolution will attack.
On the growth side, the synthesis had its own attempt. Roy Harrod (1939) and Evsey Domar (1946) had built a Keynesian growth model on the demand-driven framework: an economy grows at the warranted rate set by the saving rate divided by the capital-output ratio, and any deviation from the warranted rate generates explosive instability rather than self-correction. The model has knife-edge instability: if the actual growth rate exceeds the warranted rate the economy expands without bound, and if it falls short, the economy contracts without bound. The empirical implication, that capitalist growth is precariously balanced, fit the Depression-era diagnosis Keynes had supplied, but the prediction did not match the postwar record of more-or-less stable growth across the industrialized world. Solow’s 1956 growth model displaced Harrod-Domar by replacing the fixed capital-output ratio with a substitution-friendly production function, restoring stable convergence to a long-run growth path. The Solow apparatus and its successors live in economics ch. 13; the relevant point here is that the synthesis’s growth-side ambition was modest because Solow’s framework, marginalist in its assumptions, won the curriculum without much resistance.
The synthesis did not just win the argument. It built the curriculum. Samuelson’s textbook, Economics (first edition 1948, in print through the 1980s in successive editions), was the institutional vehicle. Generations of students learned IS-LM, the Phillips curve, the multiplier, the two-run partition, and the long-run marginalist apparatus from the same textbook. The book was not the only macroeconomics textbook of the period, but it was the central one, and its organization of the material defined what counted as introductory macroeconomics for the postwar discipline. To know macro by the standards of the 1960s was to know the synthesis as Samuelson presented it. The institutional vehicle mattered as much as the intellectual one: the synthesis gave departments a curriculum, students a framework that could be examined and tested, and the discipline a unified language for talking about macroeconomic policy. Samuelson sits relationally on the timeline.
The empirical setting in which the synthesis worked was the postwar golden age (1945–1973): nearly three decades of high growth, low unemployment, and modest inflation across the industrialized world, supported by the Bretton Woods exchange-rate system, expanding welfare-state institutions, and the productivity gains of postwar reconstruction. The synthesis fit the period and the period fit the synthesis. Demand management appeared to work; the Phillips trade-off appeared to be exploitable; the long-run marginalist equilibrium appeared to be where the economy returned when shocks subsided. This empirical vindication is what made the synthesis the orthodoxy and the orthodoxy plausible. It is also what made the synthesis vulnerable to a regime change that came in the 1970s, when stagflation broke the Phillips trade-off and the macroeconomic framework that had ridden the golden age lost its empirical anchor. Chapter 10 takes up the counter-revolution (Friedman’s natural-rate argument, the Lucas critique, the rational-expectations assault) that attacked the synthesis on grounds the Phillips-curve-as-policy-menu had given the attackers. One major alternative, present alongside the synthesis throughout, never fit the toolkit. The next section takes that alternative up.
Keynes and Schumpeter were born in the same year (1883), published their major works within a decade of each other, and diagnosed the same economic crisis in irreconcilable ways. Joseph Schumpeter spent the 1930s and 1940s writing inside the same institutional landscape Keynes occupied (both were among the world’s most prominent economists, both held chairs at the discipline’s leading institutions, both were widely read by policymakers), and Schumpeter’s response to the Depression and the postwar reconstruction took a form Keynes’s framework could not absorb. The chapter has spent five sections on the framework that won the curriculum. This section walks the framework that lost.
Schumpeter’s diagnosis of capitalism rests on one analytical move and a structural claim that follows from it. The move is to treat the economy as a non-equilibrium system whose central agent is not the optimizing consumer or the price-taking firm but the entrepreneur. The entrepreneur, in Schumpeter’s technical sense, is not a small-business owner or a manager. The entrepreneur is the agent who introduces new combinations (new products, new production methods, new markets, new sources of supply, new organizational forms) into a system that would otherwise circulate steadily through known patterns. Innovation, on this account, is exogenous to the equilibrium framework because innovation is what disturbs the equilibrium; an economy in equilibrium has no entrepreneurs in the technical sense, only managers running existing combinations. The entrepreneur creates a temporary monopoly position by being the first to commercialize a new combination, earns a rent on that position until imitation erodes it, and clears the rent with a new combination or accepts demotion to manager.
From the entrepreneur follows the structural claim. Creative destruction, Schumpeter’s term, introduced in Capitalism Socialism and Democracy (1942), is the process by which new combinations destroy incumbent firms and industries, driving long-run growth through turbulence rather than equilibrium adjustment. Capitalism, in this account, is not a smoothly clearing market system periodically disturbed by external shocks. It is a system whose internal logic produces disruption, and the disruption is the engine of the growth the system delivers. The contrast with Keynes’s framework is direct. Keynes diagnosed the Depression as demand deficiency, treatable by demand management. Schumpeter diagnosed extended depressions as the upper-rim turbulence of long innovation cycles, in which old combinations had exhausted their productive yield and new combinations had not yet matured into industries large enough to absorb the displaced labor and capital. The problem, on this view, was not insufficient demand but insufficient creative destruction at scale; the policy response, if any, was to clear the path for the new combinations rather than to prop up the old. The two diagnoses share the empirical event and disagree on what the event is an instance of.
Why did the postwar synthesis ignore Schumpeter? The question matters because creative destruction is not a fringe insight, and the entrepreneur as agent of structural change has not gone away as an analytical concept. The answer is methodological, not intellectual. Creative destruction cannot be formalized in simultaneous-equation equilibrium models, the framework on which the synthesis ran. The reason is structural. A simultaneous-equation system requires a closed list of variables, a closed list of agents, a closed list of technologies, and equilibrium conditions specified across all of them. Innovation, in Schumpeter’s sense, is the introduction of variables, agents, and technologies that were not in the system the previous period. There is no closed list to write down; there is no equilibrium for the new combinations to converge to until the new combinations exist; and what is being analyzed is the production of structure, which the framework was built to take as given. To write Schumpeter inside the framework requires either reducing creative destruction to parameter changes (which evacuates what makes the concept distinctive) or abandoning the framework. The synthesis chose neither. It set Schumpeter aside.
The set-aside has a sociological face. At the December 1936 American Economic Association meeting in Chicago, the General Theory — published earlier that year — was the dominant subject of conversation. Schumpeter, then at Harvard and one of the discipline’s most senior figures, gave a paper that was cool to Keynes’s book. He acknowledged its importance but objected to its method: the static framework, the aggregation, the absence of any account of how the productive structure changes over time. The reception was the inverse. Keynes’s book was treated as the most important publication in macroeconomics in a generation; Schumpeter’s reservations were treated as the polite dissent of a senior figure who had missed what was new. Within five years, Hicks’s formalization had given the profession a workable apparatus, the political momentum of New Deal demand management was building, and Schumpeter was being read as a historian of economic thought (which he was; History of Economic Analysis was published posthumously in 1954) rather than as a working macroeconomist with an alternative diagnosis. The choice was not just intellectual. It was sociological: Keynesian demand management gave governments a toolkit, with measurable instruments (fiscal balance, interest rate, money supply) and measurable targets (output, employment). Creative destruction gave them nothing actionable in the same idiom. The profession chose the framework that fit the institutional demand for a usable macro toolkit, and the choice held for two decades.
The set-aside was not permanent. Schumpeter’s framework was rediscovered in three waves, each from outside the postwar synthesis’s curriculum core. The first wave came in the 1980s and 1990s with the endogenous-growth literature: Paul Romer’s 1986 and 1990 papers, Philippe Aghion and Peter Howitt’s 1992 model of growth through creative destruction, and the broader research program on innovation as the engine of long-run productivity. These models embed Schumpeterian dynamics inside a tractable simultaneous-equation framework by parameterizing the innovation process (treating R&D as an investment with a known production function and innovation arrival as a Poisson process), trading qualitative content for analytical workability in much the way Hicks had done with Keynes. The second wave came from management theory: Clayton Christensen’s 1997 work on disruptive innovation translated creative destruction into a vocabulary practitioners could deploy in firm strategy, and the resulting literature on disruption has become the lingua franca of business-school discussions of technological change. The third wave came from innovation-policy scholarship in the 1990s and 2000s, where industrial-policy questions about technology, competition, and structural change drew on Schumpeter’s framework directly. The modern formal Schumpeterian growth literature lives in economics ch. 13; the chapter here treats the original framework rather than its formalized descendants.
What did the discipline lose by setting Schumpeter aside in the 1940s? The chapter takes the position that the loss was real and that the marginalization was methodological rather than substantive. Creative destruction names a structural feature of capitalist economies (that growth comes through the destruction of incumbents by innovators, that the process is internally generated rather than externally driven, that the economy’s productive structure changes over time as a feature of the system rather than as an exogenous shock) that the synthesis’s framework could not reach because the framework had abstracted from the production of structure at the start. The 1980s rediscovery does not fully repair the loss; the formal endogenous-growth models capture some of Schumpeter’s dynamics by parameter substitution, but the qualitative point about the entrepreneur as non-equilibrium agent does not survive the parameterization any better than animal spirits survived I(r). What survives best, ironically, is in management theory and innovation policy, where the formal apparatus is loose enough that Schumpeter’s qualitative content can be carried directly. The discipline that ran the synthesis kept the formal toolkit and lost the substantive insight; the disciplines that did not run the synthesis kept the substantive insight and never built the formal toolkit. Schumpeter sits relationally on the timeline. Schumpeter as a parallel anti-equilibrium tradition is the subject of chapter 7, and the cross-cutting thread through the schools is carried by the book's lineage-thread walkthroughs.
The profession chose Keynes. What did it choose? The next section walks three readings of the choice.
Three camps have argued about Keynes since 1937. Each is partly right, and each loses something the General Theory had. The chapter walks the three at strongest form before taking its position.
The Post-Keynesian reading is the most direct. Paul Davidson and Hyman Minsky are the central figures; Davidson’s Money and the Real World (1972) and his later work make the case in pure form, and Minsky’s Stabilizing an Unstable Economy (1986) carries it into the financial-fragility analysis the post-2008 literature would take up. The reading is that IS-LM betrayed Keynes. The General Theory was about an economy operating under fundamental uncertainty, in which expectations are conventional rather than calculable, in which money is a production input rather than a portfolio asset, and in which involuntary unemployment is a normal state rather than a deviation. The synthesis’s reconciliation with the long-run marginalist equilibrium threw all of this away by treating the IS-LM apparatus as a complete representation of the framework rather than as a comparative-static tool with a domain. Joan Robinson’s “bastard Keynesianism” charge, that the synthesis was an illegitimate offspring that kept the surname while losing the analytical content, is the canonical polemical statement of the position. Capitalism, on the Post-Keynesian reading, is inherently unstable because the financial system continually accumulates fragility through ordinary expansion (Minsky’s financial-instability hypothesis), because animal spirits drive volatile investment in ways no stable function can capture, and because the monetary-production economy generates structural features no marginalist framework can absorb. To read Keynes through IS-LM is to read him through machinery built to suppress what made his argument distinctive.
The mainstream reading is simpler and more operationally successful. Olivier Blanchard and N. Gregory Mankiw are the standard references; their textbooks and policy writings articulate the position that has been the working framework of central banks and finance ministries through the post-1990 period. The reading is that IS-LM captured the essential insight: aggregate demand determines short-run output; sticky prices and wages mean that demand shocks have real effects; fiscal and monetary policy can stabilize the economy around its long-run trend. Whether Keynes’s deeper claims about fundamental uncertainty and the monetary-production economy were faithfully formalized is, on this reading, beside the point: those claims belong to philosophy of economics rather than to economics, and what the discipline needs from Keynes is the demand-management framework that IS-LM and its successors deliver. The framework works, the policy tools have been productive, and the absence of Keynes’s philosophical residue is no loss because the residue was never going to do operational work anyway. This is not a hostile reading. It is the reading of a generation of economists who took Keynes’s framework, formalized what could be formalized, used what was useful, and treated the rest as the kind of philosophical surplus a working discipline can leave to historians of thought.
The New Keynesian reading is the most intellectually ambitious of the three. Michael Woodford’s Interest and Prices (2003) is the central reference; the broader research program runs through Mankiw’s 1980s sticky-price models, the dynamic stochastic general equilibrium framework, and the central-bank operational models of the 2000s. The reading is that the microfoundations project completed what Keynes started. Keynes had a verbal framework asserting that nominal rigidities and expectations matter; he could not formalize it because the analytical tools to write nominal rigidities and rational expectations into a tractable optimization framework did not yet exist. The New Keynesian program built those tools: sticky prices derived from optimizing firms facing menu costs, rational expectations replacing Keynes’s conventional-belief framework, dynamic optimization replacing Keynes’s comparative statics. The result is a set of conclusions that look like Keynes’s (demand shocks have real effects in the short run, monetary policy is non-neutral, fiscal policy can stabilize output) but rest on microfoundations Keynes himself could not provide. The bargain the New Keynesians offer is that Keynes’s policy conclusions survive in a framework whose foundations are stronger than Keynes’s own. Chapter 17 walks the New Keynesian program’s development as one of several modern macro frameworks.
The chapter’s position. The Post-Keynesian reading is the most faithful to the text of the General Theory; the philological case is substantial, and Hicks’s 1980 recantation gives the reading support that does not depend on heterodox loyalties. The mainstream reading is the most operationally successful; the framework has run economic policy in the post-1990 period, and the policy outcomes have been better than what came before. The New Keynesian reading is the most intellectually ambitious; the microfoundations program is a genuine analytical achievement, and the rebuilt foundations are more rigorous than Keynes’s own. The losses, however, accumulate across all three readings. The Post-Keynesian reading preserves the qualitative content (fundamental uncertainty, animal spirits, the monetary-production economy) but has not built a working policy framework that delivers measurable instruments and targets. The mainstream reading delivers the policy framework and discards the qualitative content as philosophical residue, accepting the cost as fair price. The New Keynesian reading delivers a microfoundations-based framework that recovers Keynes’s policy conclusions on rebuilt foundations, but the foundations replace fundamental uncertainty with rational expectations and replace animal spirits with stable optimization, which is to recover Keynes’s conclusions while jettisoning the framework that made the conclusions follow. The deepest contribution of the General Theory, the insistence that a monetary-production economy with fundamental uncertainty operates differently from any model that assumes calculable expectations and equilibrium convergence, is preserved by no reading at full strength, and the chapter takes that as the structural fact about what happened to Keynes after 1937.
The counter-revolution that attacked the synthesis (Friedman, Lucas, the rational-expectations assault) challenged not whether IS-LM was faithful to Keynes, but whether it was a valid model at all. That challenge is ch. 10’s subject. The modern multiplier debate inheriting Keynes’s fiscal-policy mechanism lives in Walkthrough 01; the modern recessions debate inheriting the demand-deficiency explanation lives in Walkthrough 08. Minsky, Woodford, Blanchard, Mankiw, and the post-Keynesian school sit relationally on the timeline.
Wicksell, Interest and Prices (1898); Marshall, Money Credit and Commerce (1923); Kahn, “The Relation of Home Investment to Unemployment” (1931); Knight, Risk Uncertainty and Profit (1921); Keynes, The General Theory of Employment, Interest and Money (1936); Hicks, “Mr Keynes and the Classics” (1937); Harrod (1939) and Domar (1946) on growth; Samuelson, Economics (1948 and editions through the 1980s); Phillips (1958); Samuelson and Solow, “Analytical Aspects of Anti-Inflation Policy” (1960); Schumpeter, Capitalism Socialism and Democracy (1942); Davidson (1972, 2007); Minsky, Stabilizing an Unstable Economy (1986); Hicks, “IS-LM: an explanation” (1980); Woodford, Interest and Prices (2003).