Research Discussion Paper – RDP 2025-02 Boundedly Rational Expectations and the Optimality of Flexible Average Inflation Targeting

Expectations play a central role in the transmission of monetary policy, but how people form expectations is widely debated. We revisit optimal monetary policy design in a model with behavioural expectations that nest rational and adaptive learning beliefs as special cases, and approximates several alternative expectations theories, such as myopic or level-k expectations. Optimal policy is a weighted average inflation target plus adjustments for belief persistence and constraints faced by the central bank (information frictions and the zero lower bound). Optimal policy is well-approximated in all cases by flexible average inflation targeting, where we make explicit what average and flexible mean.