Ask Me Differently, I Expect Differently: Survey Framing and Inflation Expectations

This paper studies how the wording and format of survey questions shape measured inflation expectations during a period of exceptionally high inflation. Using a randomised experiment, embedded in the 2022 Italian Survey of Household Income and Wealth (SHIW), we compare open-ended questions with guided-choice options using an interval regression framework that jointly models point and interval-valued responses without arbitrary harmonisation. We find that question format substantially alters the level but not the dispersion of reported inflation expectations: open-ended responses yield markedly higher average expectations. These results remain robust after controlling for a comprehensive set of socio-demographic and psychographic characteristics. Our results show that the choice of survey instrument is not methodologically neutral: it can lead to systematic over or underestimate of expectations, with direct implications for monetary policy communication and households' economic decisions.