In a guest lecture at MIT Sloan, Czech National Bank Governor Aleš Michl reviewed the CNB's 2022 to 2025 disinflation strategy and argued that the Czech experience supports a hawkish higher for longer monetary stance after years of overly loose policy. He said inflation fell from 17.5% when he took office in July 2022, peaked at 18% two months later, reached the 2% target in February 2024 and then stayed close to target, averaging 2.4% in 2024 and 2.5% in 2025. The lecture also marked the release of a new CNB working paper using quantile regression forests with dynamic weights to forecast inflation and quantify the risk of renewed price acceleration. Michl said the CNB kept its key rate at 7% until it was confident inflation was returning to target, used communication around a strong koruna and positive real rates, monitored demand through the high frequency Rushin index, and repeatedly warned against a wage price spiral while imposing wage restraint and reducing headcount and senior management roles at the central bank in 2023. He also referred to the CNB's 2025 external review of its analytical and modelling framework, which concluded that the Bank should rely on a range of models rather than a single DSGE framework. According to the lecture, the new AI based approach outperformed standard mean and median specifications and traditional linear models on Czech data, while identifying inflation risks from food, fuel and foreign prices as well as domestic pressures such as activity, producer prices and expectations.
Czech National Bank 2026-04-30
Czech National Bank Governor argues for higher for longer rates and releases AI based inflation forecasting paper
Czech National Bank Governor Aleš Michl used a guest lecture at MIT Sloan to defend the Bank’s 2022–2025 disinflation strategy and hawkish higher-for-longer stance, noting inflation fell from 18% in 2022 to around the 2% target in 2024–2025. He highlighted keeping the key rate at 7% until disinflation was secured, strong koruna communication, wage restraint at the central bank, and use of high-frequency indicators. The lecture also introduced a new CNB working paper on an AI-based inflation forecasting model and cited a 2025 external review urging reliance on a broader suite of models beyond a single DSGE framework.