Did 2008 break macroeconomics?

By the mid-2000s, macroeconomics had a workhorse, and it was the same workhorse almost everywhere. The European Central Bank ran it; the Federal Reserve ran a cousin of it; graduate programs taught it as the thing you had to know. It was a dynamic stochastic general equilibrium model — DSGE — and it was the product of two decades of convergence that had genuinely settled arguments people once thought unsettleable.

The backbone came from real business cycle theory: a representative household optimizing consumption and labor across time, an economy buffeted by technology shocks, everything derived from explicit preferences and a production function rather than the ad-hoc behavioral equations the older Keynesian models had used. Onto that backbone the New Keynesians grafted frictions — sticky prices via Calvo’s staggered-contract device, monopolistic competition so firms had prices to set — which is what gives monetary policy short-run traction. A Taylor rule closed the system: the central bank moves the interest rate in response to inflation and the output gap. The whole thing was estimated against the macro time series with Bayesian methods. The canonical specification was Smets and Wouters, and it was not a toy. It forecast as well as anything the profession had built, and it forecast better than the structural Keynesian models it replaced.

The confidence had an empirical anchor. From the mid-1980s the volatility of output and inflation in the United States had fallen sharply and stayed low — Stock and Watson documented it, and in a 2004 speech Bernanke gave it a name, the Great Moderation, and read it as a deep structural stabilization rather than luck. The economic-history spine of that era — globalization, financialization, the long calm — is History Ch.18 (Globalization and the Great Moderation). If you had been watching the data for twenty years, the policy framework looked like it worked.

The post-2008 macro toolkit looks like the pre-2008 toolkit with parts added. The additions were not random repairs; they cohere into a research program with a clear organizing idea — put back into the workhorse the things the crisis proved it could not do without. Four of those additions are load-bearing today.

The first and largest is financial frictions. Bernanke, Gertler, and Gilchrist had a financial-accelerator model back in 1999, but before the crisis it lived on the periphery; the workhorse ran without it. After 2008 the financial block moved to the center. Christiano, Motto, and Rostagno built entrepreneurial net worth and a credit channel into an estimated DSGE; Gertler and Karadi built bank leverage and a tool for analyzing unconventional monetary policy. The policy menu now ships with these blocks as standard rather than as optional extras. The second addition is heterogeneity. Kaplan, Moll, and Violante’s HANK models showed that the distribution of wealth — specifically the share of hand-to-mouth households — changes what monetary and fiscal policy actually do. The representative agent had been doing load-bearing work it should never have been trusted with, and HANK gave the distributional structure a home inside the model rather than as an afterthought outside it. The third is occasionally-binding constraints: once the interest rate hits zero, the standard linearized solution methods break, and the post-crisis literature built machinery (Guerrieri and Iacoviello’s OccBin and its descendants) that lets the model handle the zero-rate regime without leaving the framework. The fourth is narrative and microdata-driven evidence — Romer-and-Romer-style identification of policy shocks, Mian and Sufi on household balance sheets — which began disciplining the structural estimates with evidence the time series alone could not provide.

The policy chronology these patches were designed against — the QE rounds, the fiscal stimulus, the emergency lending facilities, Basel III and stress testing — is the spine of History Ch.19 (The 2008 crisis and after). The models were not built in a vacuum; they were built to make sense of decisions central banks had already been forced to make.

The intellectual home of this post-2008 retrofit — how the New Keynesian consensus absorbed financial frictions and confronted what 2008 exposed — is History of Economic Thought Ch.12 §12.7 (The reckoning: what 2008 exposed and what survived).

The post-2008 patches fixed real things. They also left a different set of gaps open — and these are not the pre-2008 gaps shrunk down, they are structurally new problems. Three of them are doing the most work.

Start with the unconventional-policy regime. The workhorse can now sit the economy at the zero lower bound through occasionally-binding-constraint machinery. But the tools that actually operated during the zero-rate years — large-scale asset purchases, forward guidance, yield-curve control, the alphabet soup of emergency liquidity facilities — are reasoned about episode by episode rather than derived from first principles. There is no canonical DSGE treatment of quantitative easing that the field agrees on; the empirical literature on whether and how much QE works is large, sophisticated, and unresolved. Fifteen years in, the discipline still cannot tell you, from inside a single agreed model, what a given unconventional move will do.

Then there is secular stagnation. In 2014 Summers revived Hansen’s 1939 hypothesis: maybe the neutral real interest rate is now structurally low — demographic drag, capital-light technology, a global savings glut — so the economy keeps wanting to sit at the zero bound. The alternative reading is that the long stretch of low rates was a transient overhang from the crisis, not a new normal. The discipline has not settled which. Then the 2021–2023 inflation surge and the sharp tightening that followed complicated both stories at once, and the neutral rate’s post-pandemic level is now itself contested. The third gap is that macro-finance integration is one-sided. The financial-frictions absorption is concentrated at the back end of the cycle — fire-sale dynamics, balance-sheet constraints under stress. The front end — how stable conditions generate the leverage build that eventually destabilizes, the endogenous-instability story associated with Minsky and developed in the BIS financial-cycle work of Borio and colleagues — is not in the workhorse. The discipline has absorbed the response to financial stress. It has not absorbed the generation of it.

The live test of all this is the record itself: the 2008–2015 zero-rate years, then the 2021–2023 inflation surge and tightening cycle, all charted in History Ch.19 (The 2008 crisis and after). The post-2008 reckoning as a structural shift in what the discipline takes seriously — the macro-finance push, the secular-stagnation revival, the fragmentation-or-convergence question — is the subject of History of Economic Thought Ch.17 §17.5 (The expanding frontier).

While the structural-modeling response was building financial-friction patches, a second response was building too, on a different axis. The credibility revolution put identification — the question of what actually lets you read a causal effect off the data — at the center of empirical work. It started in micro: Card and Krueger’s natural experiment, Angrist and Krueger’s draft-lottery instruments, a general insistence that a credible research design beats a clever model fit. After 2008 it moved into macro with force. Romer and Romer identified monetary-policy shocks from the narrative record of FOMC decisions; Ramey used the timing of defense-spending news to identify fiscal multipliers; Mian and Sufi traced the Great Recession’s consumption collapse through household balance sheets in regional microdata; Nakamura and Steinsson made cross-region variation a workhorse for identifying monetary effects. What changed for macro is concrete: a paper now has to defend its identification strategy out loud, structural estimates are judged against quasi-experimental ones, and “we calibrated the model to match the time series” stopped being sufficient on its own.

This turn is not a historical episode the way the policy-chronology threads are, so it has no economic-history chapter of its own. Its place in the lineage is as one of the named structural shifts of the modern-pluralism era — the empirical, microdata-driven temper that now runs alongside the structural-modeling tradition — charted in History of Economic Thought Ch.17 §17.4 (Institutions and inequality — the empirical turn).

Here is the toolkit as it actually stands. Micro-founded DSGE with financial-friction blocks is standard in policy specifications; HANK sits in the workhorse periphery and is migrating toward its core; occasionally-binding constraints handle the zero-rate regime inside the model; narrative-identified evidence complements structural estimation; the credibility-revolution methods that conquered micro are now established in macro; the policy menu includes quantitative easing, forward guidance, lender-of-last-resort facilities, and macroprudential regulation. And the journal record is unambiguous on one point: most top-five macro papers still use some variant of micro-founded DSGE plus Bayesian estimation plus empirical-microdata identification. The structural backbone survived. The empirical overlay accelerated. The policy applications extended. That description is not in dispute across the two frames. What the frames dispute is what it means — and to see why a single description supports two readings, you have to look at the strongest version of the “nothing was broken” case. The full lineage of this consensus — from the New Keynesian synthesis through to the post-2008 reckoning — lives in History of Economic Thought Ch.12 §12.7 and the modern-pluralism map in Ch.17 §17.6 (Convergence or fragmentation?).