Working Paper: NBER ID: w31727
Authors: Daniel J. Benjamin; Kristen Cooper; Ori Heffetz; Miles S. Kimball
Abstract: Happiness data—survey respondents’ self-reported well-being (SWB)—have become increasingly common in economics research, with recent calls to use them in policymaking. Researchers have used SWB data in novel ways, for example to learn about welfare or preferences when choice data are unavailable or difficult to interpret. Focusing on leading examples of this pioneering research, the first part of this review uses a simple theoretical framework to reverse-engineer some of the crucial assumptions that underlie existing applications. The second part discusses evidence bearing on these assumptions and provides practical advice to the agencies and institutions that generate SWB data, the researchers who use them, and the policymakers who may use the resulting research. While we advocate creative uses of SWB data in economics, we caution that their use in policy will likely require both additional data collection and further research to better understand the data.
Keywords: self-reported well-being; happiness data; economic research; policy implications
JEL Codes: D60; D63; D9; I31
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| SWB measures (C21) | insights into welfare and preferences (D69) |
| interpretation of SWB data (I31) | underlying assumptions (D01) |
| SWB reflects flow utility (L90) | estimated costs of unemployment interpreted as one-time cost (J65) |
| SWB captures lifetime utility (D15) | costs of unemployment could be significantly higher (J65) |
| current SWB (F32) | current income (E25) |