Mad About Models

Even experienced performers can get nervous when a leading light is observing. Imagine hooping in front of Michael Jordan, singing in front of Taylor Swift, or acting in front of Meryl Streep. Trying to chin the high bars that they have set and fearing their subsequent judgments would be very stressful.

Imagine, then, the apprehension that must have filled the halls at the Bank of England (BoE) this month. Ben Bernanke, a Nobel Laureate with a place on the Mount Rushmore of central bankers, had been engaged to opine on the efficacy of the BoE’s forecasting process. His critique was not kind, but the lessons he offered are worthy of reflection far beyond central London.

Forecasts are foundational to the conduct of monetary policy. Because changes in interest rates (or balance sheet strategy) work with a lag, projections of where growth and inflation will be in future quarters are essential.

It goes without saying that forecasting economic outcomes is hard. The underlying data used in this endeavor can suffer from imprecision; inflation is a particularly challenging concept to quantify. Relationships between variables are not completely stable, even in more settled times. Nonetheless, models using thousands of equations authored by armies of PhDs have been employed to anticipate the future.

Inflation YoY and Inflation Forecast Errors