Inflation expectations: the impact of priming in survey questions (WP/26/11)

Over the past decade, there have been increased efforts by central banks to communicate with the wider public and greater use of microdata to understand the way in which inflation expectations are formed and improve communication with the public. This has renewed interest in survey design. In this paper, we evaluate the impact of priming in the Bureau for Economic Research’s (BER’s) inflation expectations survey, where the survey respondents are given historical inflation numbers before they are asked to provide their inflation expectations for a series of horizons. We explore the extent to which this priming might influence the survey responses and possibly make the reported expectations more backward-looking (informed by historical inflation numbers given to the respondents) than those of the public in reality. Analysis of 2000–2024 BER survey data indicates that priming exerts a modest but significant influence on the expectations of firms and trade unions, while financial analysts appear to use the information in a more active and selective manner. A complementary 2024 survey experiment on business decision-makers finds an overall treatment effect close to zero, though larger firms, and those moderately sensitive to interest rates, report slightly higher inflation expectations when primed with a previous year’s inflation.