A postdoc is a temporary research appointment taken after the PhD, formally a training position and functionally the workforce of academic science. Federal counts through the National Center for Science and Engineering Statistics put the US postdoc population in the tens of thousands — with biomedical fields dominating — and the position's defining bargain has frayed: in exchange for modest pay and years of additional training, a postdoc was long presumed to be the reliable path to a professorship. The presumption no longer matches the arithmetic. Biomedical sciences produce far more PhDs than faculty positions, meaning most postdocs will not get the job the training was supposed to lead to — a structural imbalance that surveys of doctoral career outcomes have documented for over a decade.
This is an explainer on research-career economics, not individual career advice.
What does a postdoc actually pay?
The reference point is the NIH National Research Service Award stipend scale, which many institutions adopt even for non-NIH-funded postdocs. After a landmark raise, the entry-level NRSA stipend stood at $61,008 for fiscal 2024 — a figure that followed years in which the starting stipend hovered in the low $50,000s. Context matters: postdocs typically hold doctorates plus several years of specialized training, often work 50-plus-hour weeks in expensive metropolitan research hubs, and commonly defer employer retirement contributions because universities classify many postdocs as trainees or term staff. Comparisons circulate regularly: new bachelor's graduates in industry data roles frequently out-earn postdocs with PhDs, and unionization drives have made the pay gap the central public fact of the position.
How long does the stage last, and where does it end?
Officially two to three years; in practice often longer, with many researchers accumulating multiple postdocs across labs and countries. The end matters more than the length. Long-running analyses of NSF's Survey of Doctorate Recipients show that only a minority of biomedical PhDs hold tenure-track faculty positions — estimates around one in ten circulate for US biomedicine — and faculty hires skew toward graduates of a handful of elite departments who typically completed their own postdocs at similarly elite institutions. The rest of the cohort spreads into industry research, government, data science, publishing and teaching-track appointments. None of that is failure by the numbers, but the pipeline's promises were never revised to match the distribution, which is where the grievance concentrates.
Why does the system persist?
Because every participant has a locally rational incentive. PIs need skilled, motivated labor that grant budgets can afford, and a postdoc costs less than staff scientist with the same skills. Universities get research output without permanent commitments. Early-career researchers accept the deal because the professoriate is the profession they trained for and because the alternative signals — leaving for industry — feel irreversible at 29 in a way they rarely are at 39. The result is a queue that allocates itself by patience and family resources as much as by scientific merit, a pattern documented in studies of academic career persistence across multiple countries.
Related stories: Peer Review Runs on Free Labor — and the Bill Is Enormous · One Advisor, Five Years, No Recourse: Academia's Most Concentrated Job.
Who becomes a postdoc — and what does that shape?
The position filters demographically in ways that compound. Because postdoc years are precisely the years of peak family formation, researchers without independent wealth, a partner with stable income, or family support face the steepest trade-offs, and attrition studies show women and first-generation academics leave research careers at higher rates through this window — frequently citing pay and location constraints rather than scientific interest. National Academies reports on the postdoctoral experience have made the point explicitly since 2000: a training system that selects partly on the capacity to absorb low pay reshapes who remains in academic research. That consequence is measurable in longitudinal data and is the strongest argument the reform camp has: the current economics do not merely underpay individuals, they bias the composition of the future professoriate.
What is changing?
More than in previous decades. Postdoc unions at major universities have negotiated substantial minimum salaries — several private institutions have crossed the $65,000-70,000 floor for all postdocs through collective bargaining. NIH raised the NRSA scale sharply and encourages career-stage policies that cap postdoc duration. Institutions have built permanent staff-scientist tracks that convert skilled postdocs into stable, better-paid research careers without pretending they are trainees. And PhD programs increasingly publish placement data, changing what applicants can demand to know. None of these adjustments rebuilds the pipeline's original promise, but each converts a hidden subsidy — cheap expert labor — into a priced position, the precondition for any honest decision by the people inside it.
How should the numbers be read?
Honestly, in both directions. The postdoc is a genuinely bad deal measured against its traditional promise — years of below-market pay for odds that have fallen to roughly one in ten in biomedicine. It remains a genuinely good deal measured against its alternative uses: two to four years of funded research independence, co-authorship on papers that anchor a CV, and time to test whether the academic life fits before leaving it. The facts to watch for anyone evaluating the stage are three: the current NRSA stipend level, the fraction of a program's graduates who reach faculty jobs, and whether the institution offers a real staff-scientist alternative. Where those three numbers are good, the squeeze is survivable; where they are not, the position's name — training — is doing a lot of unpaid work. Read that way, every reform on the table is really the same reform: making the deal legible enough that a 29-year-old with a doctorate can price it accurately before signing.
