Chapter 12 New Keynesian Economics and the Modern Monetary-Policy Consensus

Introduction

The counter-revolution of the previous chapter ended with a victory and an open question. Monetarism, rational expectations, and the real-business-cycle program had won the methodological war: after the Lucas critique, no model of stabilization policy could be taken seriously unless its agents formed expectations rationally and its equations descended from explicit optimization. But the same program had thrown out the substance the Keynesian tradition had been built to defend. If markets clear continuously and money is neutral, recessions are efficient responses to real shocks and there is no welfare case for stabilization at all. The question this chapter answers is the one the counter-revolution left on the table: once you concede the method, what is left of the case for active monetary policy? The answer is New Keynesian economics — the program that accepted every methodological demand the New Classicals made and then beat them on their own terrain by adding one well-documented friction. That move produced the modern monetary-policy consensus: the Taylor rule, inflation targeting, central banks that act because output gaps are real welfare losses. The Great Moderation looked like its vindication; 2008 exposed what it had still left out. What follows traces the synthesis, the consensus it operationalized, the apparent vindication, the labor and econometric-methodology lineages that share the period, a working non-Western monetary framework that runs without the interest-rate instrument the consensus depends on, and the reckoning that extended the framework rather than breaking it.

12.1 The Synthesis on the Victor's Foundations

The starting point is a concession. The chapter on the counter-revolution closed with the New Classical program holding the methodological high ground: the Lucas critique had shown that the parameters of the old Keynesian models were not invariant to policy, because they bundled together expectations that would shift the moment policy did; Kydland and Prescott had shown that a model with rational agents, optimizing firms, and continuous market-clearing could generate realistic business cycles out of nothing but technology shocks, with no nominal frictions and no role for money. The real-business-cycle world was a world in which stabilization policy was not merely ineffective but unwarranted: cycles were the efficient response of optimizing agents to real disturbances, and smoothing them would destroy welfare rather than create it. To answer that program, it was not enough to point out that recessions feel like waste. The Keynesian tradition had to meet the New Classicals where they stood — accept rational expectations, accept that every behavioral relation must come from optimization — and show that the substantive conclusions could be rebuilt on those foundations anyway. New Keynesian economics is the name for the research program that did exactly this.

The move turned on a single friction. The real-business-cycle model assumed that prices are perfectly flexible — that when a firm faces a change in demand or costs, it adjusts its price instantly and at no cost, so the price level always moves to clear markets and nominal disturbances have no real effects. The New Keynesians asked what happens if that one assumption is relaxed: not abandoned wholesale, not replaced with the old ad hoc wage-and-price rigidity the Lucas critique had discredited, but derived from optimization like everything else. The friction they reached for was the cost of changing a posted price. N. Gregory Mankiw's 1985 paper "Small Menu Costs and Large Business Cycles" made the case at its sharpest. A menu cost is the small fixed cost a firm pays to change a price it has already posted — reprinting a menu, re-tagging shelves, reprogramming a system, renegotiating a contract. Individually these costs are trivial, a few dollars against the firm's revenue. Mankiw's result was that triviality at the level of the firm is consistent with large effects at the level of the economy. A firm facing a small menu cost will leave its price unchanged in response to a small shock, because the private gain from adjusting is smaller than the cost of adjusting. But when many firms leave their prices unchanged, the aggregate price level fails to adjust, a fall in nominal demand becomes a fall in real demand, and output contracts. The social loss from the recession is large; the private cost that produced it is small. The aggregation of a trivial friction generates a non-trivial business cycle.

The intellectual force of this is easy to understate. Mankiw was not asserting that prices are sticky as a brute empirical fact and asking to be believed — the New Classicals would have called that exactly the kind of unmicrofounded assumption the Lucas critique had ruled out of bounds. He was showing that price stickiness is what fully rational, optimizing firms produce when changing prices is even slightly costly. The friction is a consequence of optimization, not a violation of it. And menu costs are not a stipulation: surveys of firms' actual price-setting behavior, most influentially Alan Blinder's interview studies in the 1990s, found that firms change prices roughly once a year, cite the cost and disruption of changing prices among their reasons, and think in terms of the kind of fixed adjustment cost the menu-cost model formalizes. The single friction New Keynesian economics added was the one friction the data most clearly supported.

With that friction in place, the consequences cascade, and they run exactly opposite to the real-business-cycle conclusions. If prices do not adjust instantly, then a change in the money supply — or, more precisely, in nominal demand — is not absorbed entirely by prices. Some of it falls on quantities. Money is non-neutral: a monetary contraction lowers real output, a monetary expansion raises it, even though agents are fully rational and understand what the central bank is doing. This is the result the real-business-cycle program had been built to deny, recovered without abandoning rational expectations. And because the recession that follows a nominal contraction is the product of prices being stuck rather than of any real shock to technology or preferences, it represents output below the level the economy would produce if prices were free to adjust — a level the New Keynesians call the flexible-price or efficient level. The gap between actual output and that efficient level is the output gap, and in the New Keynesian framework it is a genuine welfare loss: resources sitting idle not because that is efficient but because a coordination problem in price-setting has prevented the economy from reaching the allocation it would otherwise reach. That last point is the hinge of the entire chapter, and the next section develops it; for now the structure is what matters. One friction, derived from optimization, flips the verdict.

The friction needed a tractable formal home, and Guillermo Calvo supplied it in 1983. Modeling each firm's discrete, occasional, optimally-timed price change is analytically forbidding; Calvo's device was to assume instead that in each period a firm gets the chance to reset its price with some fixed probability, drawn at random, independent of how long it has been since the firm last reset. A fraction of firms reset their prices each period; the rest keep last period's price. This is Calvo pricing, and its appeal is entirely tractability: it captures the central fact that prices are reset infrequently and at staggered times, without forcing the modeler to track when each firm last adjusted. The random-reset assumption is not meant as a literal account of how firms decide — it is the formalization that makes the menu-cost intuition usable in a general-equilibrium model, and the technical machinery of how the Calvo parameter maps into the dynamics of inflation lives in the economics textbook's New Keynesian chapter rather than here. What Calvo pricing delivers is the central New Keynesian relationship: the New Keynesian Phillips curve, in which current inflation depends on expected future inflation and on the current output gap (or, equivalently, on the gap between firms' actual prices and the prices they would charge if they could reset freely). The crucial word is expected. Unlike the Phillips curve of the postwar synthesis — a stable, exploitable menu trading inflation against unemployment — and unlike the expectations-augmented Phillips curve of the counter-revolution, which made expectations adaptive and the long-run curve vertical, the New Keynesian Phillips curve is forward-looking. Firms setting a price they will be stuck with for a while care about where costs and demand are headed, so inflation today is driven by the entire expected future path of the output gap. The Phillips curve survives, but it is a different object: a forward-looking pricing equation, not a policy menu.

The program was a collective achievement, and its self-naming moment came in 1991, when Mankiw and David Romer edited the two-volume collection New Keynesian Economics, which gathered the menu-cost, staggered-contract, efficiency-wage, and coordination-failure literatures under a single banner and announced that a coherent research program now existed. David Romer — the author of the standard graduate text Advanced Macroeconomics and co-editor of that defining collection, and not to be confused with Paul Romer of endogenous-growth theory, who belongs to an entirely different lineage — gave the program much of its textbook articulation. Olivier Blanchard contributed foundational work on the aggregate-demand externalities and nominal rigidities that make the menu-cost result general rather than a special case, showing that one firm's failure to cut its price imposes a cost on others by depressing real demand, so the private decisions that produce price stickiness are collectively suboptimal in a way that gives stabilization policy something to correct. The common thread across all of it is the one move: take the New Classical apparatus as given, add a microfounded nominal friction, and watch the Keynesian conclusions reappear. New Keynesian economics did not refute rational expectations or microfoundations. It conceded both completely, and won anyway.

It is worth being precise about what kind of victory this was, because the framing invites a mistake. New Keynesian economics did not bring Keynes back. The synthesis it produced is microfounded where Keynes was not, forward-looking where the old models were adaptive, and grounded in optimizing agents who understand the policy regime they live under — none of which describes The General Theory or the IS-LM models that followed it. What survived was a set of conclusions, not a method: that money is non-neutral, that demand shortfalls produce real losses, that stabilization policy has a warranted role. The method that now carried those conclusions was the New Classicals' own. The drama of the synthesis is that the Keynesian substance survived by joining the program that had been built to bury it, rather than by resisting it — and that to see the move clearly you have to watch what the single friction does. The interactive below lets you do exactly that: push the share of sticky-price firms from zero to one and watch the verdict travel from the real-business-cycle world to the New Keynesian world.

12.2 Woodford and the Workhorse

A research program becomes a paradigm when someone writes the book that states it whole. For New Keynesian economics that book is Michael Woodford's Interest and Prices (2003), and its subtitle — Foundations of a Theory of Monetary Policy — announces the ambition. Woodford's achievement was to take the scattered New Keynesian results and assemble them into a single, internally consistent theory of how a central bank should set interest rates, derived end to end from the optimizing behavior of households and firms. The book's most striking feature, captured in a phrase that became its signature, is that it is a theory of monetary policy without money. There is no demand for money in the central equations, no quantity equation doing analytical work; the central bank is modeled as setting an interest rate directly, and the price level and output are determined by how that rate responds to conditions. This was not a quirk. It was a recognition that the modern central bank does in fact operate by setting a short-term interest rate rather than by controlling a monetary aggregate, and that the theory should describe what central banks actually do.

The deepest contribution of Interest and Prices was to put the case for stabilization on a welfare-theoretic footing. The output gap of the previous section is asserted there to be a welfare loss; Woodford proved it. By taking a second-order approximation to the utility of the representative household, he showed that household welfare in the New Keynesian model can be written, to that order, as a loss function — a weighted sum of the squared output gap and squared inflation. The two things a central bank should hate, deviations of output from its efficient level and deviations of inflation from target, drop out of the microfoundations as the things that actually reduce the welfare of the agents in the model. This is the formal content of the claim that output gaps are genuine welfare losses rather than efficient adjustments. In the real-business-cycle world there is no such loss function, because output is always at its efficient level and there is nothing to stabilize; the loss function exists in the New Keynesian world precisely because sticky prices drive a wedge between actual and efficient output, and the size of that wedge is what the welfare cost measures. Stabilization policy is warranted not as a matter of taste but because it raises the welfare of the agents the model is built from.

Out of these foundations comes the object every graduate student now learns as the canonical model of monetary policy: the three-equation New Keynesian model. The first equation is a dynamic IS curve, in which the current output gap depends on the expected future gap and on the real interest rate the central bank engineers — the demand side, microfounded from household consumption-smoothing. The second is the New Keynesian Phillips curve from the previous section, the forward-looking supply relation linking inflation to expected inflation and the output gap. The third is a policy rule describing how the central bank sets the nominal interest rate in response to inflation and the gap. Three equations, three endogenous variables — output gap, inflation, interest rate — and a complete account of how a monetary economy responds to shocks and to policy. The algebra of solving the system, the conditions for a determinate solution, and the optimal policy that minimizes Woodford's loss function are the economics textbook's territory; what matters for the history of thought is that the New Keynesian program had, by the early 2000s, produced a compact and complete workhorse, and that Jordi Galí's textbook Monetary Policy, Inflation, and the Business Cycle (2008) codified it into the form in which it is now taught everywhere.

The workhorse did not stay in the seminar room. Its full-scale empirical descendant is the dynamic stochastic general equilibrium model — DSGE — and the canonical specification is the one Frank Smets and Rafael Wouters estimated for the euro area and then the United States in the mid-2000s. The Smets-Wouters model takes the real-business-cycle real side as its backbone — optimizing households, capital accumulation, technology shocks — and layers the New Keynesian nominal frictions on top: sticky prices, sticky wages, and a battery of additional rigidities and shock processes tuned so the model fits the data. The result was a model rich enough to be estimated against the actual time series and used to forecast and to evaluate policy, and it spread through the world's central banks with remarkable speed. By the late 2000s the New Keynesian DSGE model in something close to the Smets-Wouters form was the standard analytical apparatus inside the Federal Reserve, the European Central Bank, the Bank of England, and most of their peers. This is the sense in which New Keynesian economics is not a fringe repair but the deployed mainstream: the model central banks reach for when they ask what a shock will do or what a rate change will accomplish is, in its bones, the three-equation model with an estimable real backbone bolted on.

Return now to the interactive of the previous section, because it is in this workhorse that the one-friction move pays off. When you pushed the share of sticky-price firms to zero, the model collapsed to the real-business-cycle backbone — money neutral, no stabilization role, the output gap a non-thing. When you pushed it back up, the same backbone acquired a forward-looking Phillips curve, a determinate role for the interest rate, and a welfare loss for the central bank to minimize. The DSGE workhorse is that toggle made operational at full scale. The reason every major central bank runs a model of this family is not loyalty to Keynes; it is that the model accepts the methodological discipline the profession agreed on after the Lucas critique — rational expectations, microfoundations, an explicit shock structure — and still delivers a central bank with something to do. The synthesis won the seminar and then won the institution.

12.3 The Taylor Rule and Inflation Targeting

A theory that says output gaps are welfare losses and that the central bank should minimize a loss function is still an abstraction until it becomes a rule a central bank can actually follow. The bridge from theory to practice is the most influential single equation in modern monetary economics, and it arrived from an unexpected direction — not from a theorist deriving an optimal policy but from an empiricist describing what the Federal Reserve was already doing. John Taylor's 1993 paper "Discretion versus Policy Rules in Practice" pointed out that the Fed's behavior over the preceding years could be summarized startlingly well by a simple reaction function: set the nominal interest rate equal to a baseline, then raise it when inflation runs above target and when output runs above potential, by fixed amounts per point of each. The Taylor rule is that reaction function. Its power was that it fit — a single, transparent formula reproduced the Fed's actual decisions — and that it gave the abstract New Keynesian prescription "respond to inflation and the output gap" a concrete, communicable, checkable form.

Inside the Taylor rule sits a condition that turns out to be the whole game. Call the amount the nominal rate rises per point of inflation the inflation-response coefficient. The Taylor principle is the requirement that this coefficient exceed one: when inflation rises by a point, the central bank must raise the nominal rate by more than a point, so that the real interest rate — the nominal rate minus expected inflation — actually goes up. If the bank raises the nominal rate by less than the rise in inflation, the real rate falls when inflation climbs, which loosens policy exactly when it should tighten, and inflation spirals. If it raises the nominal rate by more, the real rate rises, demand cools, and inflation is pulled back. The single inequality — the inflation-response coefficient greater than one — is the difference between a monetary policy that anchors inflation and one that lets it run away, and it is no exaggeration to call it the inequality at the heart of modern central banking. Everything the consensus claims about why central banks can control inflation depends on it. The interactive below lets you cross the threshold and watch the inflation path flip from anchored to explosive as the coefficient passes one.

The rule needed an institution to live in, and the institution was inflation targeting. A central bank that follows the Taylor principle has solved a technical problem; it has not yet solved the credibility problem the counter-revolution diagnosed. The earlier chapter laid out the time-inconsistency result: a central bank that retains discretion has, every period, a temptation to engineer a little surprise inflation to push output above potential, and because the public anticipates the temptation, the economy ends up with higher inflation and no extra output — an inflation bias built into discretion itself. The remedy is commitment: tie the bank to a rule, or to an institutional structure that mimics one, so the temptation is removed and expectations can anchor. Inflation targeting is that institutional structure. The central bank commits publicly to a numerical inflation target, is held accountable for meeting it, and conducts policy transparently against it. New Zealand's Reserve Bank adopted the first formal inflation target in 1989; the Bank of England followed after the UK's exit from the Exchange Rate Mechanism in 1992 and was granted operational independence to pursue it in 1997; over the following two decades the regime spread across the developed and much of the developing world, until an inflation-targeting, operationally independent central bank became the default model of how a country runs monetary policy.

There is a quieter assumption underneath the entire consensus, and naming it is where the chapter's evaluative discipline begins. The Taylor rule pins down the price level only on the understanding that monetary policy is in charge of it — that the central bank sets the price-level path and the fiscal authority adjusts its surpluses to keep the public debt on a sustainable course whatever that path implies. This is the active-monetary, passive-fiscal regime, and it is the regime under which the Taylor principle is the stability condition and the whole consensus story is correct. There is another regime. If the fiscal authority sets its surpluses independently and refuses to adjust them to the central bank's price-level path, then the price level is determined not by the interest-rate rule but by the requirement that the real value of government debt equal the present value of those surpluses — the fiscal theory of the price level. In that regime the Taylor-principle logic does not govern, and a determinate, well-behaved inflation path is no longer something the central bank can deliver on its own. The chapter names this boundary and stops there. Which regime actually obtains, and whether the post-2020 inflation was a monetary or a fiscal phenomenon, is a live and unsettled dispute; the regime toggle in the interactive shows the boundary firing without litigating it, and the argument itself is the territory of the monetary-or-fiscal walkthrough and the next chapter. What matters here is that the modern consensus is regime-conditional rather than universal — true under an assumption about the monetary-fiscal division of labor that usually holds and is rarely stated.

Step back and the chapter's first house call comes into view. The great macroeconomic argument of the postwar decades — Keynesians against monetarists, fiscal fine-tuning against the money-supply rule, the unstable Phillips curve against the natural rate — is often presented, in public debate and in textbooks aimed at it, as an unresolved clash of worldviews that politics is still fighting out. It is not. It was substantively resolved, and the resolution is the New Keynesian consensus. The synthesis took the monetarist insight that money matters and that inflation is, in the long run, a monetary phenomenon a central bank is responsible for, and it took the Keynesian insight that demand shortfalls produce real output losses a central bank can offset, and it housed both inside a single microfounded model in which the central bank stabilizes inflation around a target and output around potential by moving a real interest rate. The monetarists won the point that monetary policy, not fiscal fine-tuning, is the primary stabilization instrument and that credibility about inflation is decisive. The Keynesians won the point that there is something real to stabilize. The Taylor rule embodies both at once. The split that public discourse still treats as open was closed inside the profession by the late 1990s; the public debate simply did not catch up. To say so is not to flatten the disagreements that remain — about the size of fiscal multipliers at the zero lower bound, about which regime governs the price level, about how much of the cycle is demand and how much is supply — but those are arguments inside the synthesis, not the old war between schools. The schools merged. The merger is what every modern central bank runs.

12.4 The Great Moderation as Apparent Vindication

From the mid-1980s through 2007, the United States and most other industrialized economies experienced something that, viewed from the turbulence of the 1970s, looked almost miraculous: the volatility of output growth and inflation fell sharply and stayed low. Recessions became shorter and milder, expansions longer, inflation lower and steadier. The economist James Stock and Mark Watson documented the break and the name stuck — the Great Moderation. For the New Keynesian consensus the timing was irresistible. The moderation began just as central banks were learning to follow something like the Taylor principle and were moving toward explicit inflation targeting; it looked exactly like what the theory predicted good policy would produce. A profession that had just rebuilt the case for stabilization on rigorous foundations now appeared to be watching that case vindicated in the data.

The chapter does not let the synthesis take that victory lap, because the evidence does not clearly support it. The trouble is that more than one story fits the volatility decline, and they are hard to tell apart. This is the good-policy-versus-good-luck debate. The good-policy story is the flattering one: the moderation happened because central banks got better — adopted the Taylor principle, anchored expectations, stopped accommodating inflation — so the reduced volatility is the consensus working. The good-luck story is deflationary: the moderation happened because the shocks hitting the economy got smaller and less frequent over the period — fewer oil-price spikes, fewer supply disruptions, a quieter external environment — so the reduced volatility is a feature of the disturbances, not of the policy response to them. Stock and Watson's own decomposition in 2002 attributed a substantial part of the decline to smaller shocks rather than better policy. Ben Bernanke, in a 2004 speech given while he was a Fed governor, surveyed the competing explanations — better policy, structural change such as improved inventory management, and good luck — and was careful not to claim the credit entirely for monetary policy, even as he leaned toward the policy story. The honest reading is that the moderation is consistent with the consensus but does not prove it: a world in which good policy mattered and a world in which the shocks were simply calmer produce a similar-looking decline in volatility, and the data of the period cannot decisively separate them.

The confident peak of the era was reached, fittingly, by the man who had started the counter-revolution. In his 2003 presidential address to the American Economic Association, Robert Lucas declared that macroeconomics had succeeded in its central task: the problem of depression prevention, he said, had been solved for all practical purposes, and had in fact been solved for several decades. It was a reasonable thing to say in 2003. The data looked like vindication, the theory was elegant and complete, and the institutions built on it seemed to be delivering. The statement is worth quoting not to mock it — hindsight makes prophets of us all — but because it marks precisely the high-water mark of the consensus's self-confidence, and because what came five years later makes it the most evaluable sentence in modern macroeconomics. The Great Moderation looked like proof; Lucas read it as proof; and the reading was about to meet the largest financial crisis since the 1930s. The next sections walk the apparatus-origin lineages that share this period, and then the reckoning that tested the confidence Lucas voiced.

12.5 Apparatus-Origin: Labor and the Credibility Revolution

The same decades that produced the monetary-policy consensus produced two other developments that belong to the history of modern economics and that this chapter is the natural home for: a finding that overturned the textbook account of the minimum wage, and a transformation in what the discipline counts as evidence. Both are lineages whose modern formal machinery lives in the economics textbook's labor and econometrics chapters; what belongs here is where they came from and what they settled. They share a parentage — the same empirical turn produced both — and they belong side by side.

Start with the minimum wage, because it is the cleaner story and it carries the chapter's second house call. The textbook prediction is unambiguous: in a competitive labor market, a minimum wage set above the market-clearing wage reduces employment, because firms facing a higher wage hire fewer workers along a downward-sloping labor demand curve. For most of the twentieth century this was treated as settled — one of the things economics knew. Then in 1994 David Card and Alan Krueger published a study that did not behave. New Jersey had raised its minimum wage; neighboring Pennsylvania had not. Card and Krueger surveyed fast-food restaurants on both sides of the border before and after the increase, treating the state line as a natural experiment — a setting in which something close to a controlled comparison arises from circumstance rather than design, with Pennsylvania serving as the control group for New Jersey's treatment. The competitive model predicted that employment in New Jersey should fall relative to Pennsylvania. It did not. Card and Krueger found no evidence of a relative employment decline, and if anything a small relative increase. The result struck directly at one of the discipline's confident predictions, and it provoked a fight.

The fight was led, ably and persistently, by David Neumark and William Wascher, who challenged the finding on its merits. Using payroll records rather than telephone surveys, they reported employment declines where Card and Krueger had found none, and argued that measurement choices drove the contrast. This counter-position has to be taken at full strength, because the chapter's call rests on having taken it: the disagreement was real, the data were genuinely contested, and for a time it was reasonable to think the natural-experiment result might not survive. What settled it was the accumulation of evidence over the following two decades. A large body of subsequent work — more border comparisons, more states, better data, meta-analyses pooling dozens of studies — found that moderate minimum-wage increases, of the kind actually legislated, produce employment effects clustered near zero, too small and too noisy to recover the clean disemployment the competitive model predicts. The Card-Krueger finding largely survived replication. Neumark and Wascher's strand persists and the debate is not unanimous, but the center of gravity moved decisively: the simple competitive prediction does not hold cleanly at the moderate hike levels that policy actually involves.

A finding without a mechanism is fragile, and the mechanism arrived to make the Card-Krueger result coherent rather than anomalous. It is monopsony. Alan Manning's Monopsony in Motion (2003) is the synthesis. In a competitive labor market, a firm faces a flat labor supply curve — it can hire all the workers it wants at the going wage and loses its entire workforce if it offers a penny less. Real labor markets do not look like that. Search frictions, commuting costs, imperfect information, and the sheer difficulty of switching jobs give employers some wage-setting power: a firm that cuts its wage loses some but not all of its workers, and a firm that raises its wage gains some but does not face an infinite queue. The labor supply curve to the individual firm slopes up. Once it does, the competitive prediction breaks: a firm with wage-setting power hires below the efficient level, and a moderate minimum wage that forces it to pay more can, over a range, raise employment toward the competitive level rather than reducing it. Monopsony does not say minimum wages always raise employment; it says the textbook's flat prediction is the special case, and that moderate hikes in markets with employer wage-setting power need not cost jobs. The Card-Krueger non-effect, baffling under perfect competition, is exactly what monopsony predicts. The explanation gained credibility precisely because it rationalized a result the competitive model could not. The modern extensions of the labor apparatus — David Autor's work on labor-market polarization, the Acemoglu-Restrepo task framework that models how automation reallocates work between labor and capital — carry the field forward, and their formal treatment is the territory of the economics book's labor chapters when those ship.

The minimum-wage story is one instance of a larger change, and the larger change is the chapter's second house call. The natural-experiment method Card and Krueger used was not a one-off; it was the leading edge of a transformation in how empirical economics is done, a transformation usually called the credibility revolution. The phrase, from Joshua Angrist and Jörn-Steffen Pischke, names a shift in what counts as a convincing empirical claim. The older tradition estimated large structural models and relied on assumptions about functional form and exclusion restrictions to identify causal effects from observational data; the credibility revolution insisted instead on research designs in which the source of identifying variation is transparent and defensible — natural experiments, instrumental variables with a credible story for why the instrument is as good as random, regression discontinuities, and finally field experiments. Angrist and Imbens built the formal apparatus for interpreting what instrumental-variables and natural-experiment estimates actually recover, the local average treatment effect — the causal effect for the subpopulation whose behavior the natural experiment actually moves, rather than an economy-wide average. The framework made precise what a natural experiment can and cannot tell you, and it is the methodological backbone of the entire approach. The credibility revolution is not merely better technique; it is a change in the discipline's evidentiary standard — a shift in what economists will accept as having shown that one thing causes another.

The method reached its most ambitious form in development economics, in the randomized-controlled-trial program of Abhijit Banerjee, Esther Duflo, and their collaborators, who took the logic of the controlled experiment into the field: randomly assign a program — a deworming campaign, a microcredit offer, a teacher incentive — across villages or schools, and read the causal effect off the comparison between treatment and control. The randomized field experiment is the credibility revolution's purest instrument, because randomization manufactures the as-good-as-random variation that natural experiments have to find lying around. The profession ratified the turn with its highest honors: the 2019 Nobel went to Banerjee, Duflo, and Michael Kremer for the experimental approach to alleviating global poverty, and the 2021 Nobel went to Card, Angrist, and Imbens — Card for the empirical labor-economics work the minimum-wage study exemplified, Angrist and Imbens for the methodological framework that made natural-experiment evidence interpretable. The two prizes, two years apart, mark the credibility revolution as the reigning evidence standard of contemporary economics. This methodological turn has a predecessor: the time-series-econometrics revolution of the 1970s and 1980s — Granger causality, Engle's volatility modeling, Sims's vector autoregressions, Heckman's correction for selection — which is the previous chapter's territory, the rung the credibility revolution builds on and reacts against. The credibility revolution's substrate is also where several thread-walkthroughs on labor and econometric methodology pick up the lineage. The formal econometric machinery — the local-average-treatment-effect estimator, the assumptions a research design must satisfy — lives in the economics book's econometrics-foundations chapter, with a dedicated time-series chapter to come.

12.6 Modern Islamic Finance as an Alternative Monetary Framework

Every framework this chapter has walked so far takes the interest rate as the central instrument: the Taylor rule sets it, the credit channel transmits it, the whole consensus is organized around a central bank moving a rate up and down. It is worth asking what a monetary and financial system looks like when the interest rate is removed — not as a thought experiment but as a working institutional reality serving a large share of the world's population. Modern Islamic finance is that reality. It is the most substantial non-Western framework this book engages at length, and it earns the space because it is a live, scaled, contemporary alternative built around a single binding constraint that forecloses the instrument the rest of the chapter depends on. The constraint is riba: the prohibition, grounded in Islamic law, of interest. The Qur'anic injunction (2:275) against riba has been read narrowly by some jurists as a ban on exploitative consumption-usury and broadly by others as a ban on any fixed, predetermined return on lent money — any time-rent on money as such. The broad reading is the one modern Islamic finance operationalizes, and it is the broad reading that makes the framework interesting here: if money may not earn a fixed return simply for being lent across time, then the interest-rate instrument the Taylor rule moves is, within this system, unavailable. The question is what replaces it.

The answer is a set of contracts that route the return to capital through participation in real activity rather than through time-rent on money. Classical jurisprudence had already worked them out, and modern Islamic finance revived them as the building blocks of a banking system. Murabaha is a cost-plus deferred sale: the bank buys the asset the customer wants and resells it at a disclosed mark-up payable over time — financing structured as a trade rather than a loan. Ijara is a lease: the bank owns the asset and transfers its use for a rental, the building block of much of the system's longer-term financing. Musharaka is a partnership in which two parties contribute capital and share profit and loss by an agreed ratio — contractual equity. Mudaraba is a silent partnership: one party supplies capital, the other supplies labor or expertise, profits are shared by ratio, and losses fall on the capital provider unless the manager was negligent — the structure behind Islamic investment deposits and a close analogue of venture funding. The common thread is risk-sharing or asset-backing: the financier's return is tied to the performance of a real transaction or a real asset, not promised as a fixed sum regardless of outcome. Every function conventional finance performs through interest-bearing debt has a permissible analogue that performs it through participation instead.

The capital-market instrument that carries the most analytical weight is the sukuk, often translated as "Islamic bond," though the translation obscures what is distinctive. A conventional bond is a debt claim: a promise to repay principal with interest. A sukuk is a claim on a real asset: it represents proportional ownership in an underlying asset or pool of assets, and its return is generated by that asset — the rentals from a leased property, the profits of an enterprise — rather than by a fixed coupon on a loan. The default behavior differs accordingly: a sukuk holder's recourse runs to the asset, which is what makes the instrument asset-backed in a way a debt security is not. The market's standard inflection point is the Malaysian sovereign sukuk of 2002, the first major sovereign issue priced and traded against conventional benchmarks; later sovereign issuers included the United Kingdom, Hong Kong, Luxembourg, and South Africa, none of them Muslim-majority, issuing sukuk to tap the pool of capital the framework had created. Alongside these instruments sit further constraints that shape the system: the prohibition of gharar, excessive uncertainty in contracts, which forecloses the kind of opaque, multiply-layered derivative — the collateralized-debt-obligation family — that sat at the center of the 2008 crisis; and zakat, the obligatory wealth transfer of roughly 2.5 percent, which builds redistribution into the system as a structural feature rather than bolting it on as external welfare.

As ideas, the modern framework descends from a twentieth-century revival that framed an Islamic economy as a third way between capitalism and socialism. Abul A'la Maududi argued the third-way framing in popular terms; Sayyid Qutb's Social Justice in Islam (1949) cast redistributive justice as structural rather than charitable; and Muhammad Baqir al-Sadr's Iqtisaduna (1961) is the most analytically rigorous of them, drawing a careful distinction between Islamic economic doctrine — the normative system of property and contract rules — and economic science, the positive analysis of how an economy so structured would behave. As an institutional reality the framework is recent and large. The Mit Ghamr savings bank in Egypt (1963), founded by Ahmad El Najjar, was the proof of concept; the decisive year was 1975, with the twin founding of the Dubai Islamic Bank, the first sustained commercial Islamic bank, and the Islamic Development Bank in Jeddah, an OIC-capitalized multilateral. Malaysia built the reference regulatory model — a dual banking system with statutory sharia advisory boards. By the end of 2023 global Islamic-finance assets stood at roughly 4.9 trillion dollars, with Islamic banking assets around 2.5 trillion and sukuk outstanding around 850 billion, according to the industry's own stability and development reports. This is not a curiosity at the margin of the world economy; it is a multi-trillion-dollar sector built on the proposition that the interest-rate instrument is optional.

The interesting question for this chapter is the one a mainstream evaluation actually has to answer: is Islamic finance a genuine alternative, or conventional finance in compliant packaging? The chapter takes the mainstream verdict and takes it with a split. At the retail level, the critique associated with Mahmoud El-Gamal's Islamic Finance (2006) is largely right. El-Gamal calls much of retail Islamic banking "sharia arbitrage": a murabaha home-purchase financing, once you trace its cash flows, converges on a conventional fixed-rate mortgage, with the mark-up benchmarked to a prevailing interest rate and the legal form rearranged to satisfy the letter of the prohibition while reproducing the economics of the thing prohibited. At the wholesale and capital-market level, the substantive-alternative argument associated with Umer Chapra (2000) and Feisal Khan (2013) retains real traction. There the asset-backing bites: an ijara-sukuk's return is genuinely tied to a real asset, default genuinely reaches that asset, the gharar prohibition genuinely forecloses the opaque-derivative structures that amplified the last crisis, and the profit-and-loss-sharing contracts genuinely implement a real-economy linkage at the level of the contract rather than as a relabeling. The chapter's call is that Islamic finance is more authentically alternative at the wholesale and capital-market layer than at the retail layer — a verdict that is neither the dismissal that treats the whole thing as cosmetic nor the advocacy that treats it as a finished replacement.

What the framework demonstrates, and why it belongs in a chapter on the modern monetary consensus, is a structural possibility the consensus tends to treat as unthinkable: an economy organized around risk-sharing rather than fixed nominal debt contracts. When interest is foreclosed, the return to capital must come either from profit-and-loss sharing on real enterprise or from a bounded mark-up over a real transaction; money does not earn a time-rent in itself, it earns participation in real activity. That is a different transmission architecture, and it changes what a monetary authority can and cannot do — the interest-rate channel the Taylor rule depends on is simply not there to move. The interactive below makes the contrast manipulable: toggle the interest channel off and watch the monetary impulse re-route through risk-sharing and asset-backing instead. The point is not that the interest-free system is better; it is that the interest-rate instrument is one institutional choice among possible ones, and a working alternative at scale is the cleanest way to see that the consensus rests on an architecture rather than on a law of nature.

The doctrine's classical and medieval origins — the jurisprudence of riba, the long silence and the twentieth-century revival of Islamic economic thought as a field — are the territory of this book's chapter on non-Western economic thought, to which this section is the modern continuation: the doctrine as a live contemporary monetary framework rather than as an intellectual history.

12.7 The Reckoning: What 2008 Exposed and What Survived

Five years after Lucas declared depression prevention solved, the global financial system seized. The crisis of 2008 originated where the consensus model was not looking — in the financial sector, in the collapse of mortgage-backed asset values, the failure of leveraged institutions, and the freezing of the markets through which banks fund one another. The canonical New Keynesian DSGE workhorse of 2007, the Smets-Wouters model and its relatives that central banks were running, had no financial sector worth the name. It had a representative household, optimizing firms, sticky prices, an interest rate set by a Taylor rule — and essentially no banks, no leverage, no balance sheets that could become impaired, no possibility of the kind of financial seizure that actually happened. The model was not built to simulate a crisis driven by financial fragility because financial fragility was not in it. When the crisis came, the profession's workhorse had no way to represent it; the financial frictions that turned out to be central had been left on the to-do list, peripheral to a model organized around price stickiness. This is the precise sense in which 2008 exposed the consensus: not that the New Keynesian framework was wrong about what it modeled, but that it had left out the thing that broke.

What the profession did next is the chapter's central evaluative claim, and the claim is that the framework was extended rather than broken. The pieces needed to represent financial fragility largely already existed; they had simply been peripheral, and the crisis moved them to the center. The financial accelerator of Ben Bernanke, Mark Gertler, and Simon Gilchrist, published in 1999, had already shown how the net worth and balance-sheet condition of borrowers amplifies shocks: when asset values fall, borrowers' net worth falls, the premium they pay for external finance rises, they borrow and invest less, output falls further, asset values fall again. After 2008 this mechanism was pulled from the periphery to the core of the workhorse. The models of Lawrence Christiano, Roberto Motto, and Massimo Rostagno and of Mark Gertler and Peter Karadi built banks, leverage, and an explicit financial sector into estimable DSGE models, and a New Keynesian model with a financial block became the new standard specification. A second extension came from the other direction. The representative-household assumption — one agent standing in for the whole distribution — had hidden the fact that the marginal propensity to consume out of income varies enormously across households, and that this distribution matters for how monetary and fiscal policy actually transmit. The heterogeneous-agent New Keynesian model, or HANK, of Greg Kaplan, Benjamin Moll, and Giovanni Violante (2018) replaced the representative consumer with a realistic distribution of households facing borrowing constraints, and showed that policy transmission runs substantially through channels the representative-agent model could not see. And the persistence of near-zero interest rates after the crisis forced the machinery of the zero lower bound and occasionally-binding constraints into the standard toolkit, where the determinacy results of the Taylor principle have to be rethought because the central bank can no longer cut the rate. None of these is a repudiation of the New Keynesian program. Each is the program absorbing what it had omitted, on its own foundations.

So the chapter's position is that the consensus was extended, not overturned — but the qualification matters as much as the claim, and it is what keeps the verdict from being a complacent one. The post-2008 framework is not the pre-2008 framework with the gaps filled and the work finished. It is a framework with substantively different gaps. The pre-2008 toolkit's blind spot was the financial sector; the profession largely fixed that. But the post-2008 toolkit has its own open problems that the financial-friction models do not resolve: how much of the cycle is demand and how much is a genuine shift in the economy's productive capacity, whether the natural rate of interest has fallen permanently in a way that keeps economies stuck near the zero lower bound, how the distribution of income and wealth feeds back into aggregate dynamics, and whether the whole DSGE enterprise — even enriched — captures the genuinely non-linear, expectations-driven, occasionally-panicked behavior of a financial system in stress. Extended-not-broken is a claim about continuity of foundations, not a claim that the questions are settled. The framework that came out of 2008 can represent the crisis that the 2007 framework could not. It has a fresh set of things it cannot yet do.

That open frontier is where this chapter hands off. The mainstream extensions walked here — the financial accelerator absorbed into the workhorse, HANK, the zero-lower-bound machinery — are the discipline repairing its consensus from inside. Running alongside them is a different response: a set of challengers who read 2008 not as a gap to be patched but as evidence that the mainstream framework is the wrong frame. Modern Monetary Theory's claim that the monetary-fiscal division of labor the consensus assumes is itself the error; the revival of Hyman Minsky's argument that financial fragility is endogenous to tranquil expansion rather than an external shock; Thomas Piketty's relocation of distribution to the center of the analysis; the complexity-economics rejection of equilibrium itself; the renewed secular-stagnation debate about whether the advanced economies have entered a regime of chronically deficient demand. These are the heterodox challengers and the open questions, and they are the next chapter's subject — the discipline fragmenting, the frontier contested, the question of whether the synthesis holds. This chapter ends where the mainstream's confident self-repair ends and the argument about whether that repair is enough begins.

One lineage closes here rather than opening forward. The trade thread that ran from David Hume's price-specie-flow mechanism through the new trade theory has its modern terminal node in the gravity model: the empirical regularity that bilateral trade between two economies scales with the product of their sizes and falls with the distance between them, given rigorous theoretical foundations by James Anderson and, in the Ricardian framework, by Jonathan Eaton and Samuel Kortum. The gravity model is the workhorse of modern empirical trade economics, and its formal development is the economics book's trade-theory chapter; it is named here as the terminal node of a lineage this book has tracked, the point where the trade thread reaches its modern apparatus and is handed to the formal treatment.

The forward connections close the chapter. The next chapter, on the new synthesis and modern pluralism, picks up the open frontier this section names — the heterodox challengers, the secular-stagnation debate, the question of whether the discipline still has a center. The previous chapter, on the counter-revolution, is where this one began: it ended at "won the method, lost the substance," and this chapter has been the substance restored on the victor's foundations. The depth of the broken-versus-extended argument — the case that 2008 was a genuine paradigm failure rather than a gap to be patched — is the territory of the walkthrough on whether 2008 broke macroeconomics, to which this chapter contributes the mainstream's extended-not-broken side. Where the New Keynesian cluster and its post-2008 neighbors sit relationally is on the intellectual-history timeline.

Sources

Mankiw, "Small Menu Costs and Large Business Cycles" (Quarterly Journal of Economics, 1985). Calvo, "Staggered Prices in a Utility-Maximizing Framework" (Journal of Monetary Economics, 1983). Mankiw & D. Romer, eds., New Keynesian Economics, 2 vols. (1991). Blanchard & Kiyotaki, "Monopolistic Competition and the Effects of Aggregate Demand" (American Economic Review, 1987). Blinder et al., Asking About Prices (1998). Woodford, Interest and Prices: Foundations of a Theory of Monetary Policy (2003). Galí, Monetary Policy, Inflation, and the Business Cycle (2008). Smets & Wouters, "Shocks and Frictions in US Business Cycles" (AER, 2007). Taylor, "Discretion versus Policy Rules in Practice" (1993). Stock & Watson, "Has the Business Cycle Changed and Why?" (NBER Macroeconomics Annual, 2002). Bernanke, "The Great Moderation" (speech, 2004). Lucas, "Macroeconomic Priorities" (AEA presidential address, AER, 2003). Card & Krueger, "Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania" (AER, 1994). Neumark & Wascher, Minimum Wages (2008). Manning, Monopsony in Motion (2003). Autor, "Why Are There Still So Many Jobs?" (Journal of Economic Perspectives, 2015). Acemoglu & Restrepo, "The Race between Man and Machine" (AER, 2018). Angrist & Imbens, "Identification and Estimation of Local Average Treatment Effects" (Econometrica, 1994). Angrist & Pischke, Mostly Harmless Econometrics (2009). Banerjee & Duflo, Poor Economics (2011). Bernanke, Gertler & Gilchrist, "The Financial Accelerator in a Quantitative Business Cycle Framework" (1999). Gertler & Karadi, "A Model of Unconventional Monetary Policy" (JME, 2011). Christiano, Motto & Rostagno, "Risk Shocks" (AER, 2014). Kaplan, Moll & Violante, "Monetary Policy According to HANK" (AER, 2018). Anderson, "A Theoretical Foundation for the Gravity Equation" (AER, 1979); Eaton & Kortum, "Technology, Geography, and Trade" (Econometrica, 2002). On Islamic finance: Maududi, The Economic System of Islam; Qutb, Social Justice in Islam (1949); al-Sadr, Iqtisaduna (1961); Chapra, The Future of Economics: An Islamic Perspective (2000); El-Gamal, Islamic Finance: Law, Economics, and Practice (2006); Khan, Islamic Banking in Pakistan (2013); IFSB, Islamic Financial Services Industry Stability Report (2024); ICD-LSEG, Islamic Finance Development Report (2024).