Central Banks' Flawed Models Risk the Next Crisis

How Central Banks Failed to Predict the 2021–23 Inflation Surge

In early 2021, the Federal Reserve's preferred inflation measure was running below 2 percent. Within eighteen months, it had surged past 9 percent. The same trajectory played out across the eurozone, the United Kingdom, and most of the developed world. Central banks — institutions whose singular institutional purpose is price stability — did not see it coming. This was not bad luck. It was a central bank inflation forecast failure embedded in the architecture of modern monetary policymaking.

The post-mortem is now well underway. Among the most authoritative voices demanding accountability is Otmar Issing, the former Chief Economist of the European Central Bank and one of the principal architects of the ECB's monetary policy framework. His critique cuts to the bone: central banks abandoned a crucial warning signal — rapid growth in the money supply — in favor of sophisticated but ultimately deficient forecasting models, and they are not yet doing enough to correct the underlying problem.

The argument is not merely academic. The Federal Reserve's M2 money supply grew at an annualized rate of roughly 25 percent through parts of 2020 and 2021, a pace without modern precedent in peacetime. The ECB's broad money aggregates followed a similar trajectory. Central banks had this data in real time. They chose, systematically, to underweight it.

The Structural Weaknesses of Inflation Targeting

Inflation targeting became the dominant monetary policy framework of the 1990s and 2000s for defensible reasons. It anchored expectations, provided institutional accountability, and delivered two decades of relative price stability across advanced economies. The 2008 financial crisis and its prolonged aftermath only deepened central banks' commitment to the framework — in a world of chronically subdued inflation, a 2 percent target seemed like the horizon that always receded.

That experience bred a particular form of institutional complacency. The models used to generate inflation forecasts — dynamic stochastic general equilibrium frameworks and their variants — were calibrated on data from the Great Moderation era. They encoded the assumption that supply and demand shocks would be temporary, that inflation expectations would remain well-anchored, and that the transmission of fiscal and monetary stimulus into consumer prices would be slow and modest.

The BIS and IMF both documented, in their 2021 and 2022 assessments, that major central banks were systematically underestimating near-term inflation across multiple forecast rounds. This was not a one-quarter miss; it was persistent, directional error. The forecasts were not wrong by noise — they were wrong in the same direction, for the same structural reasons, quarter after quarter.

The deeper problem is that inflation-targeting frameworks optimize for communicating credibility rather than detecting turning points. When an institution's primary accountability mechanism is hitting a number, the institutional incentive is to forecast continuity rather than disruption. Central banks are not unique in this pathology — it is endemic to large bureaucracies — but the costs when monetary authorities get it wrong are borne across entire economies.

Money Supply: The Alarm Bell Central Banks Ignored

In the 1990s and early 2000s, the ECB maintained a formal "monetary pillar" alongside its economic analysis pillar. Broad money growth had an explicit role in the policy framework. It was, in Issing's framing, a cross-check — a signal from a different information set that could flag when the main models were missing something.

That pillar was progressively downgraded in the years after the financial crisis. The intellectual justification was not unreasonable on its face: the relationship between money supply growth and inflation — the velocity of money — had become unstable and difficult to model. Quantitative easing pumped reserves into the banking system without generating proportionate price increases. The case for treating money growth as a reliable leading indicator looked weak.

But the 2020–21 episode exposed the flaw in that reasoning. The fiscal-monetary combination deployed during the pandemic was categorically different from post-2008 quantitative easing. Central bank reserves were not merely sitting on commercial bank balance sheets; expanded money supply was finding its way into household accounts through direct transfers, enhanced unemployment benefits, and loan guarantee programs. The transmission mechanism had changed. The old dismissal of money supply as a policy signal no longer held.

The alarm bell was ringing. M2 growth in the United States hit levels last seen during the Second World War. Eurozone broad money aggregates were expanding at multi-decade highs. Central banks, having dismantled or sidelined the analytical apparatus that would have flagged these readings as meaningful, missed what the signal was saying. The central bank inflation forecast failure of 2021–23 was, in part, the consequence of an earlier intellectual decision: the decision to stop listening to the money supply.

Why the Same Mistakes Could Happen Again

The institutional response to the inflation episode has been, at best, incomplete. Central banks have acknowledged that their forecasts were wrong and that their initial characterization of inflation as "transitory" was mistaken. Some have commissioned internal reviews. None of the major central banks has yet announced a fundamental revision to the modeling frameworks that generated the errors.

This matters because the conditions that enable forecast failure are structural, not episodic. The same inflation-targeting frameworks remain in place. The same models — recalibrated but not reconceived — continue to drive policy decisions. The monetary pillar remains dormant at the ECB. Money supply aggregates continue to receive limited weight in formal policy deliberations at the Federal Reserve.

Meanwhile, the macro environment continues to shift in ways that models calibrated on the Great Moderation may systematically misread. Fiscal policy has become more activist in most major economies. Supply chains are being restructured along geopolitical rather than efficiency lines. Climate transition spending represents a novel demand shock whose inflationary profile is genuinely uncertain. Each of these features creates conditions where rapid monetary expansion could translate into price pressure in ways that pre-2020 models would underestimate — again.

The BIS has warned repeatedly, most explicitly in its annual reports since 2022, that central banks face a more volatile inflation environment than the prior two decades would suggest. The forecasting machinery has not been rebuilt to match that environment.

Rethinking Central Bank Frameworks Before the Next Crisis Arrives

The critique Issing advances is not a call to abandon inflation targeting or return to crude monetarism. His authority on this point derives from decades of practical framework design, not theoretical elegance. The argument is more precise: no single indicator, no single modeling approach, can bear the full weight of macroeconomic surveillance. When central banks allowed inflation-targeting frameworks to crowd out the monetary pillar, they reduced the diversity of their information set at exactly the wrong moment.

The corrective is conceptually straightforward, even if institutionally difficult. Central banks need explicit, formal mechanisms for monitoring money supply aggregates and flagging rapid deviations — not as a policy rule, but as a cross-check that forces deliberate engagement when the main models diverge from the monetary signal. The goal is not a return to Bundesbank-era monetary targeting. It is a more honest accounting of model uncertainty.

Central banks should also publish, systematically, post-mortem analyses of significant forecast errors — not to assign blame, but to discipline the institutional tendency to rationalize misses rather than learn from them. The repeated underestimation documented by the BIS and IMF through 2021 and 2022 should have generated comprehensive methodological reviews. The pace of that introspection has been insufficient.

Most critically, policymakers need to resist the temptation to treat the 2021–23 episode as a once-in-a-generation anomaly that the existing framework can absorb without structural change. It was not anomalous. It was the result of a policy environment — large-scale fiscal expansion accompanied by extraordinary monetary accommodation — that is likely to recur. The frameworks that failed to detect it once will fail again under similar conditions.

Central banks earned their institutional credibility over decades. They cannot afford to treat that credibility as inexhaustible. The models that missed the last inflation surge need to be rebuilt before the next one arrives — not after.

Source: Project Syndicate