This paper studies the transition to high inflation during the COVID-19 pandemic, using a behavioral version of the New Keynesian model, which replaces the conventional assumption of rational expectations with subjective and heterogeneous expectations. Different shares of agents in the economy form expectations based on alternative views regarding future economic variables: (1) a share of agents forecasts that inflation and output will rapidly revert to steady state; (2) another share forms forecasts based on a model resembling the MSV solution under rational expectations; (3) a third share of agents uses an under-specified model that captures trend-following, adaptive, or extrapolative behavior. Agents learn over time the parameters of their perceived model and they can also shift across different views based on past forecasting performance. The macroeconomic model is estimated using Bayesian methods to fit realized macroeconomic variables and data on expectations from surveys. The results document an additional channel that operates through switches in agents’ perceptions and amplifies the impact of the original inflationary shocks. In response to rising inflation after COVID, agents begin shifting from the mean reversion model to the trend-following specification (with a belief about perceived inflation persistence that is simultaneously revised upward). Consequently, the impact of inflationary shocks is magnified and the effects of monetary policy attenuated.

How do geopolitical risk shocks impact monetary policy? Based on a panel of 18 economies, we develop and estimate an augmented panel Taylor rule via constant and time-varying local projection regression models. First, the panel evidence suggests that the interest rate decreases in the short run and increases in the medium run in the event of a geopolitical risk shock. Second, the results are confirmed in the time-varying model, where the policy reaction is accommodating in the short run (1 to 2 months) to limit the negative effects on consumer sentiment. In the medium term (12 to 15 months), the central bank is more committed to combating inflation pressures.

We propose a novel approach to classifying inflation-targeting (IT) economies using fractionally integrated processes. Motivated by the rising prevalence and diversity of IT, we leverage variation in the persistence of inflation rates to identify four de facto strategies, or “shades” of IT. Moving from negative orders of fractional integration, indicating anti-persistent behaviour, to more persistent long-memory processes, often associated with less credible policy frameworks, we classify countries into average, strict, flexible, and uncommitted IT. This framework sheds light on differences between declarative and actual strategies across 36 advanced and emerging economies. Most countries fall into the flexible IT category, though extreme cases, including uncommitted IT, occur quite frequently. Furthermore, we link our classification to institutional features of national frameworks using ordinal probit models. The results suggest differences across categories are related to variations in the maturity and stability of IT frameworks, with weaker connections to central bank independence and transparency.