ECB upgrades models for inflation, geopolitical shocks

Economists at the European Central Bank rely on macroeconomic models for forecasts and policy prescriptions.

However, increasing frequency of large shocks and prevalent uncertainty exposed limitations in the existing toolkit.

While some forecast errors stemmed from outdated assumptions, a significant shortfall remained even when projections were rerun with actual commodity prices and interest rates.

This indicated a deeper issue in the transmission mechanisms.

Linear models struggled to capture the non-linear pass-through of large commodity price shocks, especially when inflation was already high.

Furthermore, indirect and second-round effects were stronger than anticipated, with firms passing costs more aggressively and wages catching up unusually fast.

The models, largely estimated on low-inflation data, could no longer accurately capture the post-pandemic economy's key parameters.

Learning from past lessons, the ECB is developing and deploying more detailed models better equipped for unusual economic fluctuations.

A novel suite of models now adopts a disaggregated approach, splitting pre-tax energy prices into subcomponents like fuel, gas, and electricity.

These models incorporate techniques to handle outliers and extreme volatility, enhancing explanatory power and forecasting accuracy.

New data sources, such as supply constraint indicators and high-frequency activity indices, are also integrated to avoid blind spots and disentangle macroeconomic shocks.

Core projection models are also being refined with re-estimated parameters and state-dependent pricing, allowing for flexible, non-linear relationships between inflation and its determinants.