Macroeconomics expectations and labor supply in AI era

Participants are randomly assigned to receive one of several information treatments providing recent, factual macroeconomic statistics. The information describes current or recent U.S. macroeconomic conditions, specifically: (i) CPI inflation rate; (ii) hourly earnings growth; (iii) aggregate uncertainty index; (iv) control group: local air quality index (irrelevant information), designed to control for information exposure without affecting economic beliefs.

2025-12-01

2027-05-31

Macroeconomic Expectations (Expected CPI inflation over the next 12 months, Expected nominal wage (hourly earnings) inflation over the next 12 months, Unemployment rate over the next 12 months, Reservation wage for a short, one-time task (single-period horizon), Reservation wage for a recurring task lasting multiple months (multi-period horizon), Reservation wage for a single task in 6 months (delayed task, when inflation is more salient) and desired duration of working with us in months.

The project compares macroeconomic expectations and labor supply preferences (reservation wages for a single-period, multiple-period, and delayed HITs) before and after the information treatment. We elicit reservation wages by asking the lowest reward rate at which respondents would be willing to work with us for a similar 10-min task that they were currently completing. We also ask for how many months they would be interested in working with on a monthly frequency at specified pay. These two labor supply preference variables are used as the main dependent variables.

Subjective Probability of Labor Market Transitions (from employment to a) employment with the same employer, b) different employer, c) self-employment, d) unemployment, e) out of labor force; from unemployment to a) employment b) self-employment, c) unemployment, d) out of labor force). Minimum required raise relative to current pay to switch jobs (for employed). Minimum required pay raise or cut relative to the previous pay to accept a new job (for the unemployed). Responses to hypothetical scenarios: responses of average wages, unemployment, and individual labor supply to inflation. Self-reported AI use status.

The primary outcomes capture online labor supply preferences. The secondary outcomes describe offline labor supply preferences, including hypothetical scenarios of inflation changes. They enable comparison of similarities and differences between labor supply responses to inflation expectations across settings. Self-reported AI use status, along with data quality check questions, will enable analysis of how AI impacts data quality in online surveys.

This study is a randomized controlled online experiment conducted on Amazon Mechanical Turk and Prolific. Participants are randomly assigned to receive information about one macroeconomic variable (price inflation; wage growth; or economic policy uncertainty index) or irrelevant information. The experiment measures how information-induced changes in expectations causally affect reservation wages for short- and longer-horizon online labor supply decisions as well as offline labor supply preferences. Randomized information provision serves as an instrument for generating exogenous revisions of macroeconomic expectations, allowing identification of causal effects despite expectation endogeneity.

Not available

Randomization is administered within the survey after completion of the screening task using Qualtrics functionality.

Individual

No

2,800 individuals

2,800 individuals

About 700 individuals per arm