Inflation Expectations, Learning and Supermarket Prices: Evidence from Field Experiments

We use a series of survey experiments to examine how consumers incorporate new information about the past to form their expectations about future inflation. Our novel experimental framework allows us to disentangle learning from different sources of information, and to separate genuine from spurious learning. To assess the role of rational inattention, the experiments were conducted with 10,000 subjects in both a low-inflation context (the United States) and a high-inflation context (Argentina). To assess the role of personal consumer experience, we assembled a unique source of data consisting of information on products purchased by the subjects immediately before being surveyed, and the historical prices for those same products. We find that individuals are highly influenced by both inflation statistics and supermarket prices when forming their inflation expectations. Our results indicate that learning rates are substantially higher in the low-inflation setting, where the stakes from ignoring inflation are lower, which is consistent with rational inattention. We also find that individuals are more influenced by information that is less costly to understand (supermarket prices) relative to information that is more costly to understand (inflation statistics). The evidence also suggests that individuals form inflation expectations using their memories of the price changes of the supermarket products they purchase, even though those memories are nearly orthogonal to the actual price changes these products experienced. This is bound to induce large biases in expectations. Finally, our results highlight the importance of accounting for spurious learning in survey experiments: half of the treatment effects can be attributed to spurious rather than genuine learning in our study.