Economists crack a 30-year-old math problem that lets them measure job dynamics from messy real data

A new mathematical approach solves a longstanding problem in econometrics: how to extract reliable estimates from panel data when individual differences are hard to measure. This means researchers can now build more accurate models of how employment patterns, wages, and other individual behaviors actually change over time using real survey data instead of simplified assumptions.

For decades, economists studying individual behavior over time (whether someone stays employed, changes jobs, gets a raise) have been stuck with a tradeoff: either make unrealistic assumptions to get clean answers, or admit their data won't give them reliable estimates at all. This paper removes that tradeoff by connecting an obscure math problem to a practical solution. The immediate payoff is cleaner evidence on labor market dynamics. The longer payoff is that dozens of other econometric problems built on similar logic may now have solutions too.