The number is real, but it is not current
To measure what graduates earn, someone has to observe graduates earning. The federal government does this by matching college enrollment records against tax records, years after those students leave. A ten-year earnings figure therefore describes people who enrolled about a decade ago and were measured a few years after that.
This is not a flaw in the data. It is arithmetic: you cannot know what someone earns ten years after they enrol until ten years have passed. Every college ranking that quotes earnings is quoting a figure with the same lag, whether or not it says so.
What the lag means in practice
A decade-old earnings figure is measured in the dollars of its own year. It is not adjusted for inflation on the page, because converting it would make it look like a current salary when it is not. Read it as a relative signal — how one college compares with another — rather than as a prediction of your own salary.
- Industries change. A programme that led to well-paid work in the year measured may look different now.
- The national wage level has moved. A figure that looked strong then may be ordinary now, and vice versa.
- The cohort is not you. It includes everyone who enrolled, including people who left without a credential.
The comparison that is still fair
Comparisons between colleges are still meaningful, because both sides carry the same lag. If one college’s graduates out-earned another’s in the measured cohort, that is a real difference that existed — and it is more informative than a current-year figure would be, because a current figure would capture the first year or two after graduation rather than a decade of career.
What you should not do is compare a college’s earnings figure against a salary you were quoted today. Those are different units. Compare it against other colleges, and against the benchmarks on the same page.
Coverage is not universal
Not every college has an earnings figure. Institutions with too few students in the measured cohort are suppressed to protect privacy, and some programme types fall outside the tax matching entirely. Roughly four in five institutions report a ten-year earnings figure; the rest show a blank, which means unknown — never zero.