Measuring GLP-1 users against their own earlier selves erases most of the employment and marriage gap

If the adults filling GLP-1 prescriptions are also the ones with steadier jobs and better insurance, a comparison of users against non-users measures who gets prescribed. Kaestner and Schiman instead tracked change inside each person over the panel ^[4]. That holds fixed everything about a patient that does not vary over time, though it does not control for whatever else changed when the prescription started.

A null result is also a statement about precision. The University of Chicago summary of the paper, dated September 17, 2026, does not include sample sizes or confidence intervals, only the direction of the results ^[17]. Changes large enough to detect in this sample did not appear ^[2]. Within-person estimation costs precision too, since it is identified off the patients who actually start treatment during the years they are observed ^[4].

The measures are coarse against the question people ask about these drugs. Depression symptoms, psychological distress, self-rated health, employment and marital status ^[8] are slow-moving states, and the authors looked after roughly a year of use and again over two years ^[7]. Kaestner made that point about his own outcome list. "We measured outcomes like whether someone became employed or unemployed, whether they got married or divorced, and standard indicators of mental health," he said. "Those are important measures, but they don't necessarily capture changes in self-esteem, relationship quality, or other day-to-day experiences." ^[10]

Insurers and government programs are considering whether to broaden access, and the medical benefits are already well documented while it is less clear whether the drugs produce social or economic gains that would add to the case for wider coverage ^[12]. "We know these drugs improve health outcomes, and there's been a great deal of research documenting those benefits," said Kaestner, a research professor. "What hadn't really been examined was whether those improvements extended into other parts of people's lives." ^[11] Nothing in the paper disturbs the clinical results; for adults with diabetes, Kaestner said the clearest measurable benefits remain glycemic control, weight loss and better physical health ^[13].

Kaestner also emphasized that the study does not prove the drugs have no effects outside physical health, and that some changes may be subtler than the study's measures can capture ^[9]. The sample is adults with diabetes, drawn from the Medical Expenditure Panel Survey across the 12 calendar years from 2012 through 2023 ^[3][16]. People now taking these drugs for weight management are a different group, and the authors expect more chances to test for broader social effects as use grows beyond diabetes ^[14]. "This is still a very new area of research," he said. ^[15]