Drawbacks of household panel data on inflation perceptions and expectations : non-representativeness and selectivity

By comparing the Bundesbank's BOP-HH and the ECB's CES-ECB household panel surveys focusing on inflation perceptions and expectations with omnibus surveys conducted in Germany in December 2021 and 2022, this paper examines potential sample selection bias stemming from (i) individuals' willingness to participate in a household panel, and (ii) the 'learning effect' associated with repeated survey exposure. Our findings show that focusing on newcomers to the panel does not resolve these issues, as they also exhibit selection bias and are unrepresentative of the broader population. We construct new population weights for this subgroup and identify substantial differences in opinions between inexperienced and experienced respondents during the high-inflation period of 2022. Additionally, we observe notable variation in means, medians, and variances across the three surveys for both years and inflation measures.