Book Description

Inflation expectations in household surveys tend to be vastly heterogeneous. The literature has been unable to distinguish empirically between alternative explanations, such as the existence of rational inattention (according to which individuals will not continuously gather costly information) and the use of information from personal shopping experiences (which can be diverse and inaccurate). To better understand the importance of these mechanisms, we use evidence from field experiments with nearly 10,000 subjects conducted in both a low-inflation country (the United States) and a high-inflation country (Argentina). We introduce a novel experimental design which, when combined with unique data sources, allows us to quantify how much weight individuals assign to each source of information about inflation. Our novel experimental framework addresses one of the most important concerns with survey experiments by separating genuine from spurious learning. We find that individuals are highly influenced by information on both inflation statistics and price changes of specific products. The results are consistent with rational inattention, since there is greater learning in a low-inflation setting where the stakes are lower (the United States), and also from information that is less costly to understand (individual supermarket prices). To further assess the importance of personal experiences, we conducted field experiments which combined data from actual products purchased by the subjects with their historical prices. We find that individuals form inflation expectations using their own memories about the price changes of supermarket products they buy, but those memories are nearly orthogonal to the actual price changes.