Lu­i­gi Bu­te­ra and co-aut­hors pu­blish in Quar­ter­ly Jour­nal of Eco­no­mi­cs

Luigi Butera’s paper titled “Beliefs About the Economy are Excessively Sensitive to Household-Level Shocks: Evidence from Linked Survey and Administrative Data” has just been accepted for publication at the Quarterly Journal of Economics. The work is coauthored with Chen Lian, Matteo Saccarola, and Dmitry Taubinsky. 

This paper combines Danish administrative register data with a long-running consumer survey that elicits, among other things, people’s forecasts about future inflation and their recollection (“backcasts”) of past inflation. The main finding is stark: inflation beliefs move far too much with household-level shocks that are mostly unrelated to actual inflation. When households experience recent income declines, or expect their own finances to worsen, they report significantly higher inflation expectations. But these same household-level income changes have essentially no comparable relationship with realized inflation. In short, people’s views about inflation appear to react too strongly to their own household circumstances, not just to the macroeconomy. 

The paper then studies why. A key result is that the same household-level shocks predict not only inflation forecasts, but also people’s recollections of past inflation. In fact, backcasts are even more sensitive than forecasts. That is hard to square with standard Bayesian models, but it fits a mechanism based on selective memory: negative household events make negative price experiences more salient, and those recalled experiences then feed into inflation expectations. 

The Emergency Room-visit results sharpen this interpretation. The authors exploit the fact that the survey is a repeated cross-section collected every month. They show that, controlling for a household’s overall propensity to use the ER, respondents interviewed in a month when someone in their household has an ER visit, report higher inflation backcasts and forecasts. The effect is larger for backcasts than for forecasts, consistent with the idea that emotionally negative events first distort recall, and then shape beliefs about the economy.

More information and access to the paper can be found here.