Researchers use what people think will happen to earnings instead of what actually did — and get different answers about job stability

A new method for measuring earnings and job changes uses survey data about what workers expect to earn rather than tracking what they actually earned. This removes a major blind spot: when people switch jobs or get laid off, it's hard to tell whether they had unstable work or just made a choice — the data alone can't separate those. Using expectations data instead sidesteps that problem and reveals that individual earnings are less volatile than older studies claimed, but that differences in ability and job-fit matter much more than economists thought.

For decades, labor economists have built models of earnings by watching people's actual paychecks go up and down. But those paychecks are scrambled together — some volatility comes from genuine instability in the job market, and some comes from people strategically leaving bad jobs for better ones. You can't untangle those stories from the numbers alone. This method uses what workers say they expect to earn (and their odds of accepting job offers) to separate genuine market turbulence from deliberate choice. That distinction matters because it changes what policymakers should worry about. If earnings bounce around because people are finding better matches, that's different from earnings bouncing around because jobs are precarious. The numbers flip the conventional wisdom: prior models overstated how chaotic individual earnings are, but understated how much variation comes from differences in who you are and how well you fit your job.