Still fit for purpose? Upgrading our economic models for an age of uncertainty

The failure was not in the assumptions — it was in the transmission

When ECB economists reran their projections using what actually happened — real commodity prices, real interest rates, real exchange rates — the models still underestimated the inflation surge. That result isolates the problem cleanly: bad luck with initial conditions does not explain the shortfall. Something inside the modelling architecture itself broke. Specifically, the linearised equations that govern how shocks propagate through the simulated economy could not accommodate the non-linear dynamics that defined the post-2022 environment. A 50% jump in energy costs does not produce twice the effect of a 25% jump; the relationship is convex, and the pass-through intensifies when it arrives on top of already elevated inflation. Linear models, by design, flatten that curve. The result: a systematic underestimation of price pressure at precisely the moment accuracy mattered most.

Why central bank models matter beyond central banks

Macroeconomic models are not academic exercises. They are the small laboratories — to borrow the ECB's own framing — in which policy committees simulate economies, stress-test assumptions, and generate the forecasts that anchor market expectations. When those laboratories misread the propagation mechanism, the error cascades: rate decisions lag reality, forward guidance diverges from outcomes, and the credibility gap between institutional projections and lived economic experience widens. The ECB's candour here is notable. Rather than attributing the failure to exogenous shocks alone, the bank is pointing to structural limitations in its own toolkit — a feedback loop between model specification and policy calibration that had gone unexamined for too long.

What comes next — and what to watch

The upgrade signals a shift toward modelling frameworks that can accommodate regime changes, non-linear pass-throughs, and the kind of persistent uncertainty generated by geopolitical fragmentation and energy transition. No specific timeline or technical architecture has been disclosed. What matters for the rest of us is the direction: a major central bank is formally conceding that the analytical infrastructure inherited from a lower-volatility era is no longer fit for purpose. Watch for whether other institutions — the Federal Reserve, the Bank of England — follow with comparable admissions. If the ECB's models were misaligned, the structural assumptions embedded in competing frameworks deserve the same scrutiny. The age of linear certainty is over. The question is how quickly the replacement toolkit arrives — and whether it arrives before the next structural break.