Forecasting the post-COVID inflation surge in Switzerland
Summary
The surge in inflation after the COVID-19 pandemic caught forecasters and policy makers by surprise: shocks treated as transitory turned out to be persistent, producing large forecast errors. Watson (2026) asks whether the U.S. surge could instead have been recognised as persistent in real time using Unobserved-Components Stochastic Volatility (UC-SV) models. This paper replicates and extends that exercise for Switzerland, using (i) total Swiss CPI inflation and (ii) sectoral inflation for 22 expenditure groups - both extended with an explicit seasonal component since the Swiss CPI is not seasonally adjusted. Both models attribute the bulk of the 2021-2023 Swiss surge to the permanent trend rather than to transitory noise and reach that reading from data available in real time. In the sectoral model, the sector-specific trends account for most of the variance, whereas the common inflation trend plays only a minor role.
- Issue:
- 14
- Pages:
- 42
- JEL classification:
- E31, E37, C11, C32
- Keywords:
- Inflation persistence, Unobserved components, Stochastic volatility, Bayesian Gibbs sampling, Seasonal adjustment, Switzerland, COVID-19
- Year:
- 2026