Research

Harvard Study: GLP-1 Weight Loss Linked to More Jobs and Marriages for Women

GLP-1 Companion · 9 min read

Quick answer

A new Harvard working paper by economist Rebecca Diamond followed women using GLP-1 medications over 18 months. The headline numbers — 27 and 29 percentage points — are large, but the study is careful about what they do and do not prove.

Most conversations about GLP-1 medications focus on weight, blood sugar, and side effects. A 2026 working paper from Harvard economist Rebecca Diamond, titled "GLP-1–Induced Weight Loss and the Female Obesity Penalty," asks a different question: do these medications change the parts of life that sit downstream of weight — the jobs people get hired into and the relationships they form?

Diamond compared three groups of women with high BMIs over an 18-month follow-up: women who started a GLP-1, women who wanted one but did not start, and women who were not interested. The headline numbers are big. The study's own caveats are even bigger.

What the Study Looked At

The paper drew on a large lifestyle-survey dataset and isolated three matched groups of women with elevated BMI:

Women who started a GLP-1 medication during the study window.
Women who said they were interested in starting a GLP-1 but had not (often citing cost).
Women who were eligible but uninterested.

Diamond controlled for variables including self-rated health, race, and baseline household income. The "interested non-users" group is the most important comparison, because they share the motivation profile of GLP-1 starters — the only thing that separates them is access to the drug. Comparing those two groups isolates the effect of being on the medication, not the effect of wanting to lose weight.

Outcomes were tracked for 18 months across two domains: labor-market participation (whether women got hired, hours worked, earnings) and household formation (whether women started cohabiting, married, or separated).

The Headline Findings

Jobs

Among women who were not employed at baseline, those who started a GLP-1 were 27 percentage points more likely to have begun working within 18 months, compared with the interested-but-untreated group.

For women who were already employed, the picture was different: no significant change in hours, earnings, or job-change rates. The effect concentrated entirely at the hiring step.

Relationships

Among single women, those on a GLP-1 were 29 percentage points more likely to have started cohabitation or marriage during the follow-up. As with jobs, partnered women on a GLP-1 saw no measurable change in relationship stability.

There was one notable asymmetry: in mixed-sex couples where the man started a GLP-1, the man was more likely to leave the relationship. The paper notes this finding but does not develop a causal explanation.

The Researcher's Framing

"The markets that respond are the ones where someone forms a fresh impression of a woman's body weight." — Rebecca Diamond, Harvard

In other words: the effects show up at first impressions — interviews, dating — and not inside relationships or jobs that already exist. That pattern is what Diamond and others have called the "obesity penalty" — a measurable cost imposed by other people's reactions to body weight, separate from any health effect of the weight itself.

What the Numbers Do Not Show

The size of the employment and relationship effects is striking, but the study is honest about what is missing.

Self-rated health barely moved between groups. The women on GLP-1s did not rate themselves as substantially healthier than the comparison group during the same window.
Life-satisfaction scores edged slightly downward, not up.
Depression and loneliness scores showed no meaningful change.

That gap matters. If hiring and partnering improved because the women felt healthier, more confident, or happier, the mental-health scores would have moved with the labor-market scores. They did not. The most parsimonious explanation Diamond entertains is that the changes are driven by how other people respond to a lower body weight — not by how the women themselves feel.

Diamond is careful to add: this is not "clear causal evidence of weight discrimination." Working-paper findings need replication, alternative explanations, and peer review before they earn that language. But the pattern fits the discrimination hypothesis better than it fits a health-improvement hypothesis.

The Cost Asymmetry

A separate finding inside the paper has its own implications. Roughly 40% of GLP-1 users in the dataset paid out-of-pocket — about $300 per month. Those self-paying users sat in the highest household-income bracket of the three groups. The interested-but-untreated group — the women who wanted the medication but could not access it — had the lowest incomes.

Put bluntly: if Diamond's effects are real, then access to GLP-1 medications is also a channel for widening the gap between higher- and lower-income women. The women most likely to face the labor-market and relationship penalty Diamond describes are also the ones least likely to be able to pay $300 a month to escape it.

Whether insurance broadens coverage of anti-obesity medications — and whether public payers like Medicare and Medicaid follow — directly affects who has the option this study is describing.

What This Study Is, and Is Not

A working paper is the first draft of an idea, not the final word. The limitations matter.

Not peer-reviewed. The findings have not yet survived independent review and replication.
Observational. The comparison group is matched on observable characteristics, but unmeasured differences between women who do and do not start GLP-1s could explain some of the gap.
Self-reported outcomes. Survey data on employment and household status is reliable in aggregate but noisier at the individual level than administrative records.
Specific to women with high BMI. The size of the effect for women in the healthy BMI range, or for men, is not what this paper measures.

None of those caveats erase the result. They explain why it is a working paper, not a press release.

Why the Result Is Worth Taking Seriously Anyway

Two things make the finding harder to dismiss than typical "Ozempic changes lives" anecdotes.

First, the comparison group. "Interested non-users" — women who wanted a GLP-1 but did not get one — share the demographic and motivational profile of GLP-1 starters. A simple users-vs-population comparison would not control for the type of person who pursues this medication; this one does.

Second, the asymmetry between first-impression markets (hiring, dating) and ongoing markets (existing jobs, existing relationships). That asymmetry is what you would predict if the mechanism is appearance-based judgment. It is harder to fit to a "the women felt healthier" story.

What This Means for People on a GLP-1

For an individual on a GLP-1, the practical takeaways are narrower than the headlines suggest:

The medication is doing something measurable. The weight loss the study tracks is real and large enough to register in life outcomes.
The mental-health benefits people often expect from weight loss are not automatic. Mood, life satisfaction, and loneliness did not move in this dataset. If you are starting a GLP-1 expecting depression or anxiety to improve as the scale moves, plan for them as separate problems with separate care.
Muscle preservation matters more than ever. Larger life shifts make it tempting to chase a faster scale number. The clinical evidence is clear that the women who hold weight off after stopping are the ones who protected lean mass on the way down — adequate protein, resistance training, and a structured plan.

If you want to see whether the changes in your life track with the data you have on your own treatment — dose timeline, weight curve, side-effect clusters, protein intake — that is what a tracker is for. Nuvo logs all of those in one place so the patterns are visible to you and to your clinician.

The Bigger Picture

Diamond's paper is part of a larger shift in how researchers are looking at GLP-1 medications. The first wave of evidence focused on weight and metabolic health. The second wave is now mapping downstream effects: cardiovascular outcomes, kidney function, cognition, addiction-related behaviors, biological aging, and — with this paper — labor-market and household-formation outcomes.

These secondary effects can be larger than the medical community expected. They also raise harder questions about access, equity, and what we are actually treating when we treat obesity. None of those questions are settled by one working paper. But the questions themselves are now hard to ignore.