The Great Moderation (1990-2007): Why Inflation Disappeared

From the early 1990s to 2007, US output volatility fell by approximately 50%, inflation volatility by 70%, and the longest economic expansion in postwar history (120 months) suggested that the business cycle had been tamed.

Two decades during which inflation appeared to disappear as a serious policy concern, replaced by debates about asset bubbles and global imbalances. The episode was real but contingent: a configuration of absent shocks, disinflationary globalisation, and credibility-anchored policy that 2022 dissolved within eighteen months.

The Great Moderation is one of the most studied periods in modern macroeconomic history. Its analytical importance is dual: it represents the empirical apogee of the credibility-anchored monetary policy framework that emerged from the 1970s; and it generated the dangerous belief that inflation had become a solved problem. Reading the period correctly requires distinguishing structural improvements from contingent conditions — and recognising which dissolved when conditions changed.

The volatility decline: dating and measurement

Stock and Watson’s 2002 study “Has the Business Cycle Changed and Why?” formally dated the volatility break to approximately 1984 in US data, using structural break tests on quarterly GDP and inflation series. The reduction was substantial across measures: standard deviation of US real GDP growth fell from 2.7% (1960-1983) to 1.4% (1984-2007); standard deviation of US CPI inflation fell from 3.6% to 1.1% over the same partition. The pattern repeated across G7 economies with varying timing: the UK break around 1992 (post-ERM exit), the Eurozone aggregate around 1995, Japan exhibiting both lower volatility and lower mean inflation (the latter approaching deflation).

The economic experience of the period reflected this statistical signature. The US recorded a 120-month expansion from March 1991 to March 2001 — the longest on record — followed by a brief 2001 recession (8 months) and another expansion lasting 73 months until December 2007. Recessions, when they occurred, were short and shallow: the 1990-91 episode lasted 8 months with peak unemployment of 7.8%; the 2001 episode lasted 8 months with peak unemployment of 6.3%. The variability of household income, business investment, and asset prices all declined in lockstep.

Three competing explanations

The first explanation — “good policy” — emphasised the operational maturity of inflation-targeting central banks. Bernanke’s February 2004 speech “The Great Moderation” formalised this view: the Federal Reserve and other major central banks had learned from the 1970s, adopted credibility-anchored frameworks, and were therefore producing more stable inflation and, through the inflation-output volatility trade-off, more stable output. The empirical support included the synchronicity between independent central bank reforms (1989-1998) and the volatility break in each country’s data. The credibility infrastructure that the Volcker disinflation had built two decades earlier was the operational precondition.

The second explanation — “good luck” — emphasised the absence of major adverse shocks comparable to 1973-74 or 1979-80. Stock-Watson’s calculations, replicated by Ahmed-Levin-Wilson (2004) and others, found that the magnitude of identified shocks (oil prices, productivity, financial conditions) was indeed smaller in 1984-2007 than in earlier periods. The interpretation: the Great Moderation reflected a benign environment more than improved policy, with the implication that the next major shock would test the framework severely.

The third explanation — “good practice” — pointed to structural changes including just-in-time inventory management (reducing the inventory cycle’s amplification), financial innovation (smoothing consumption), the growing service-sector share (services being less cyclical than goods), and globalisation (with imports buffering domestic shocks). Each of these mattered marginally, none dominated. The relationship between deglobalisation and the structural inflation regime documented during the post-2020 period provides retrospective evidence that the globalisation component was significant.

The disinflationary globalisation channel

One of the most important contingent conditions was globalisation, particularly China’s integration into world trade. After joining the WTO in December 2001, China’s manufacturing exports tripled within five years, with global trade-to-GDP rising from 39% in 2000 to 53% by 2007. The mechanism operated through multiple channels: direct downward pressure on tradable goods prices (US import prices fell 0.4% annually 2000-2007 versus a 2.0% rise in services prices), wage moderation in tradable sectors of advanced economies, and a positive supply shock to global manufacturing capacity that pushed against any inflation tendency.

Borio and Disyatat’s BIS work documented the magnitude: the China effect alone may have reduced advanced economy inflation by 0.3-0.5 percentage points annually during the 2003-2007 period. When this disinflationary tailwind reversed after 2020 — through deglobalisation, supply chain reconfiguration, and the end of demographic-driven Chinese labour cost moderation — a structural source of low inflation disappeared. The relationship between structural and cyclical inflation during the Great Moderation thus reflected globalisation as a structural disinflationary force, with implications when that force reversed.

The asset price puzzle and the missing inflation

The Great Moderation also generated its own anomaly: aggregate inflation remained stable while asset prices — particularly housing and equities — exhibited substantial volatility, with the dot-com bubble (1995-2000) and the housing bubble (2002-2007) representing the largest financial cycles in postwar history. The disconnect produced an academic literature on whether central banks should respond to asset price movements (Cecchetti et al. 2000, Bernanke-Gertler 1999) and a Federal Reserve operational consensus that asset bubbles should be addressed only through their effects on inflation and employment forecasts.

The 2008 crisis tested this consensus and largely rejected it. Subsequent frameworks — macroprudential regulation, financial stability mandates added to central bank objectives, the Basel III capital and liquidity requirements — emerged from recognising that inflation stability under the Great Moderation had masked accumulating financial fragilities. The understanding that central banks miss inflation forecasts in part because they observe inflation rather than the structural drivers behind it traces directly to the post-2008 reassessment of the Great Moderation period. The relationship between inflation waves across decades shows the Great Moderation as the unusual quiet period between two more typical wave-prone regimes.

The 2007-2008 break and the missing reset

The Great Moderation formally ended with the 2007-2008 financial crisis. Output volatility immediately spiked: the 2008-2009 recession produced the largest US GDP contraction (-4.3% peak to trough) since the Great Depression, with peak unemployment of 10.0% in October 2009. The Eurozone debt crisis (2010-2012) produced sustained instability across periphery economies. Inflation volatility, however, remained low throughout 2008-2020, leading to a narrative that the Great Moderation framework had survived the crisis with respect to inflation if not output.

This narrative dissolved in 2021-2024. US CPI inflation rose from 1.4% in January 2021 to 9.1% in June 2022 — the largest 18-month increase since the early 1980s. Eurozone HICP rose from 0.9% to 10.6% over a comparable period. The pattern revealed that inflation stability during 2008-2020 had reflected weak aggregate demand and continued globalisation tailwinds rather than the structural taming of inflation that the Great Moderation interpretation had suggested. The methodological need to distinguish underlying inflation from energy and food shocks, formalised in the core vs headline inflation distinction, became operationally critical in this period. The understanding of why central banks target 2% inflation rather than higher levels reflects, in part, the institutional memory of how quickly the Great Moderation regime could collapse — an experience that the 2022 episode validated.

What the Great Moderation teaches

The episode confirms three lessons with substantial empirical support. First, the volatility reduction was real and partly attributable to improved monetary policy frameworks — the credibility infrastructure built after the 1970s genuinely reduced the propagation of shocks into expectations. Second, the magnitude was substantially overstated by the absence of major shocks during the period: the framework was less robust than two decades of stable data suggested. Third, the contingent components — globalisation tailwinds, demographic moderation of labour costs, weak aggregate demand post-2008 — were structurally finite, and their reversal would expose the framework’s limits.

The episode also illustrates a methodological problem in macroeconomic measurement: regime stability cannot be confirmed within the regime, only retrospectively after it ends. The 2002 papers identifying the Great Moderation, the 2004 Bernanke speech celebrating it, and the 2007 academic consensus that the business cycle had been tamed all looked premature once the post-2008 period unfolded. The contemporary long-run history of US real interest rates documents the secular decline that accompanied the Great Moderation — itself partly a consequence of demographic and globalisation forces that were not stable properties of the economic environment.

For the broader analytical framework on inflation regimes — measurement, mechanisms, history, and effects — see the complete guide to inflation.

Last updated — 7 May 2026