Graduate career progression illustrated as a journey through professional development, skills and opportunities.

Graduate Career Progression: Beyond the First Job

Share

Graduate career progression is how a graduate's work changes over years, not months: whether employment becomes stable, whether the job comes to match the qualification, how earnings grow, whether further study pays off, and whether the graduate judges the career a success. The first job is a weak guide to all of these. National data from Australia show full-time employment and salaries rising sharply in the three years after graduation, and research shows that early disadvantages can fade, persist or reappear. Comparing graduates' salaries without adjusting for prices, place, occupation, field and selection tells you little about university quality.

Why the first destination is only a starting point

Most graduate outcome reporting measures a single moment: a survey a few months after completion, or a destination snapshot about a year later. These measures help to spot transition problems, but a graduate on a short contract in the right field and one in a job that does not need a degree may both appear as "employed".

Progression asks a different question: what happens next? Five components are worth separating, because they move at different speeds and do not always move together:

  • Employment stability: moving from short, casual or involuntary part-time work into secure employment, and staying there.
  • Employment quality: whether the job uses the graduate's skills and education, and the conditions of the work itself.
  • Earnings growth: how pay develops over time, not only its level at one point.
  • Further education: postgraduate study, which delays earnings in the short run and can change the career path later.
  • Subjective career success: how graduates themselves judge their careers, which research treats as distinct from objective markers such as pay and position.

In a review in the Journal of Management, Spurk, Hirschi and Dries (2019) report that research has tended to explain objective career success with factors such as human capital and performance, and subjective career success with factors such as stable personal traits, and they recommend that research better account for the distinction. A salary figure and a career satisfaction figure answer different questions; neither substitutes for the other.

How first-destination employment rates are defined, and why they are hard to compare, is covered in the companion article Graduate Employment Rates: Why the Numbers Need Context.

What follow-up data show about the first three to ten years

Australia: the same graduates, surveyed twice

Australia's Graduate Outcomes Survey - Longitudinal (GOS-L), part of the government-funded Quality Indicators for Learning and Teaching program, re-surveys graduates about three years after they answered the short-term Graduate Outcomes Survey at four to six months. The 2026 national report (updated October 2026) covers graduates who completed their courses between March 2022 and February 2023. It received 43,665 completed surveys from 125 institutions, a response rate of 53.6 per cent of the in-scope sample.

For domestic undergraduates, the report shows full-time employment rising from 80.2 per cent in the short term (2023) to 91.8 per cent in the medium term (2026), and the median full-time salary rising from $71,400 to $93,500, an increase of 31.0 per cent. Of the 14.4 per cent of undergraduates who were working part time and wanted more hours in the short term, 73.5 per cent of those available for work were employed full time or satisfied with their hours three years later. Around 28 per cent of graduates across all study levels felt that their job did not fully use their skills and education three years after completion, but the most common reason given by undergraduates was that they were satisfied with their current job (26.8 per cent), which the report says does not necessarily indicate an unsuccessful outcome.

Three cautions apply. The report notes that the labor force questions changed in 2025, so medium-term full-time employment rates from 2025 onward are not directly comparable with earlier years. It also shows that international undergraduates working in Australia had a medium-term median salary $8,500 below that of domestic undergraduates. And, in our reading, the medium-term sample consists of graduates who answered both surveys, so it may differ from all graduates.

England: tax records rather than surveys

The Department for Education's Longitudinal Education Outcomes (LEO) statistics link education records to tax and benefit records, so they do not depend on survey response. The 2023-24 release (published 25 June 2026, updated 14 September 2026) reports that five years after graduation, first-degree graduates who graduated in 2017-18 had median earnings of £33,300, and 88.3 per cent were in sustained employment, further study or both.

The release is explicit about its limits. Its outcomes are "raw" and do not control for differences in student characteristics. Annualized earnings do not scale up pay for people who worked part-time hours, earnings are reported only for graduates in sustained employment who are not also studying, and London had the lowest share of graduates in sustained employment or further study but the highest median earnings.

An earlier, now superseded LEO release for tax year 2018-19 illustrates a subtle reading problem. It reported median earnings of £20,800, £24,500, £27,400 and £31,100 at one, three, five and ten years after graduation. These four figures look like a career path, but they come from four different graduating cohorts observed in the same tax year. They describe a cross-section of cohorts, not the progression of any one group of graduates, and differences between them mix career growth with differences between cohorts.

What research shows about how careers diverge after the first job

Entry conditions can leave long shadows

Von Wachter (2020, Journal of Economic Perspectives) summarizes research on young workers who enter the labor market during recessions. According to the abstract, the literature finds large initial effects on earnings, labor supply and wages that tend to fade after ten to fifteen years, accompanied by changes in occupation, job mobility and employer characteristics. It also reports evidence of later effects on health and mortality in middle age, with a reopening of earnings gaps. Two cohorts from the same program can therefore have very different early outcomes for reasons unrelated to the program.

Some early advantages shrink

Deming and Noray (2020, The Quarterly Journal of Economics) measure changes in the skill content of occupations between 2007 and 2019 using online job postings. According to the abstract, the earnings premium for college graduates in technology-intensive majors such as computer science, engineering and business declines rapidly, and these graduates move out of faster-changing occupations as they gain experience. In the authors' model, faster skill obsolescence lowers the return to experience and flattens earnings growth in fast-changing careers. A field with high starting pay is therefore not necessarily the field with the highest pay ten years later.

Poor matches can persist

Clark, Joubert and Maurel (2017, IZA Journal of Labor Economics) follow the US NLSY79 cohort and define overeducation as having more schooling than is typical for workers in the same occupation. Among workers with at least some college, overeducation fell from 62.3 to 50.4 per cent over the first 12 years of their careers, and about 66 per cent of overeducated workers remained overeducated after one year. Their model suggests overeducation was very persistent for about 30 per cent of the sample, while the rest tended to leave it quickly. In their preferred specification, past overeducation is associated with a wage penalty of 2.6 to 4.2 per cent that persists over four years. The levels reflect an occupation-based definition and a cohort that entered work decades ago; the transferable point is that early mismatch is a risk factor for later mismatch.

Gaps can open after the first job

Bertrand, Goldin and Katz (2010, American Economic Journal: Applied Economics) study MBA graduates of a top US business school. According to the abstract, men and women had nearly identical earnings at the start of their careers, but the male earnings advantage reached almost 60 log points a decade after the MBA, associated with differences in pre-MBA training, career interruptions and weekly hours, which were largely linked to motherhood. In a broader review, Cortés and Pan (2023, Journal of Economic Literature) report that in the United States more than two-thirds of the overall gender earnings gap can be accounted for by the different effects of children on women and men. A first-job measure would miss these divergences entirely.

Further study has very different payoffs

Altonji and Zhong (2021, Journal of Labor Economics) estimate returns to graduate degrees with US National Science Foundation survey data (1993 to 2015), comparing earnings before and after the graduate degree within combinations of undergraduate and graduate field. They estimate a return of 0.110 log points for an MBA, far below the simple regression estimate of 0.282, and essentially zero returns for graduate degrees in the arts and the humanities, while law and medicine show large returns. Their more tentative specifications also suggest that graduate premiums grow with years since the degree. The authors conclude that simple regression-based estimates of returns to graduate fields are often highly misleading, a warning that also applies to institutional comparisons of graduates in further study and graduates in work.

Selection explains much of the raw gap between institutions

Dale and Krueger (2014, Journal of Human Resources) link College and Beyond survey data on students who entered college in 1976 and 1989 to Social Security earnings records, estimating the effect of attending a highly selective college. According to the abstract, standard models show sizeable earnings effects of college characteristics, but in models that adjust for selection these effects mostly shrink to near zero, although they remain large for some subgroups, such as Black and Hispanic students. Higher graduate earnings at one institution than another can therefore reflect who enrolled rather than what the institution added.

Why salary comparisons mislead

Earnings are the most quoted progression measure and the easiest to misread. The OECD Job Quality Framework treats job quality as multi-dimensional, defined by earnings quality, labor market security and the quality of the working environment. Pay is one of these three dimensions, not a summary of all of them. A graduate in a well-paid but insecure or high-strain job is not necessarily better off than one in a stable, moderately paid professional role.

Even as a measure of pay, raw salary figures need adjustment before any comparison. The World Bank's International Comparison Program exists to produce purchasing power parities (PPPs) and comparable price level indexes, which are used to convert measures such as GDP per capita into a common currency; its latest completed cycle covers reference year 2021. Graduate salaries compared across countries need the same kind of price adjustment. Price differences also exist within countries, as the London pattern in the LEO statistics shows.

Before comparing graduate salaries: an adjustment checklist (editorial framework proposed by this article; not the UNIRANKS methodology)
FactorWhy it distorts a raw comparisonWhat to ask for
Price levelsThe same salary buys different amounts in different countries and regionsPPP-adjusted figures across countries; regional context within countries
Geography of workGraduates move to high-wage, high-cost labor marketsWhere graduates work, not only where they studied
Field and occupation mixAn institution with more medicine or engineering students will show higher pay regardless of qualityComparisons within field, or adjusted for field mix
Hours and contract typeAnnualized earnings may include part-time work without adjustmentFull-time-only figures, or hours-adjusted estimates
Prior attainment and backgroundSelective institutions enroll students who would earn more anywaySelection-adjusted estimates, or at least a statement that none were made
Cohort and entry conditionsGraduating into a recession can depress earnings for a decadeCohort year and labor market context for each figure
Further studyGraduates in postgraduate study have low current earnings but may earn more laterHow students in further study are treated, and at what time point
Gender and caregivingGaps often open years after graduationOutcomes by sex at several time points
Coverage and responseSurveys and record linkages miss graduates abroad, self-employed or non-respondingResponse rates and coverage limits for each figure

None of these adjustments turns earnings into a direct measure of educational quality; they only make the comparison less wrong. A university whose graduates go into lower-paid public service, teaching or creative work is not, for that reason alone, providing a worse education.

A progression evidence framework for universities

The table below is an editorial framework proposed by this article. It is not the UNIRANKS methodology and does not describe how UNIRANKS evaluates institutions. It matches each time horizon to what it can and cannot reasonably show.

What each time horizon can and cannot show about graduate careers (editorial framework)
Time horizonTypical evidenceWhat it can showWhat it cannot show
Up to about 15 monthsDestination surveys, such as the Graduate Outcomes survey at 15 months used in the Office for Students progression measure in EnglandTransition problems, early unemployment, early mismatch, entry into further studyStability, earnings growth, long-run match quality
About 3 yearsFollow-up surveys of the same graduates, such as the GOS-LMovement out of involuntary part-time work, early salary growth, persistence of perceived overqualificationEffects of further study, later divergence by sex or field
About 5 yearsAdministrative earnings records, such as LEOSustained employment and earnings without survey nonresponseHours, occupation and job quality (not in tax data); causes of differences
10 years and beyondAdministrative records, cohort studies, research datasetsLong-run earnings, recovery from poor entry conditions, payoff to further study, gender divergenceThe current curriculum's effect, since the graduates studied a decade or more earlier
Any horizonGraduate career satisfaction and self-assessed skill useSubjective career success, which research treats as distinct from payObjective outcomes; responses depend on expectations and response rates

Our interpretation

Three conclusions follow. First, early outcomes are a noisy predictor of later ones: the Australian follow-up shows large gains within three years, and the research shows early advantages and disadvantages that fade, persist or reverse. Reporting only the first destination rewards programs whose graduates start quickly, which is not the same as programs whose graduates build durable careers.

Second, the longer the horizon, the harder attribution becomes. A ten-year outcome reflects the economy the graduate entered, occupation, family decisions, further study and a program as it was taught a decade ago. Long-run data are best used to understand patterns and risks, such as persistent mismatch or widening gaps by sex, rather than to credit or blame a current program.

Third, progression evidence is most informative when it is broken down. Aggregate medians hide the subgroups where careers stall: international graduates, graduates who start in poorly matched jobs, and graduates whose earnings diverge after career interruptions. A university that reports these breakdowns with their limitations gives more useful information than one that reports a single high salary.

Practical steps

For university administrators and institutional researchers

  1. Report at least two time points. Pair any first-destination figure with a later one for the same cohort where national data allow it, and label which cohort each figure describes.
  2. Do not present cross-sections as trajectories. When figures for one, three, five and ten years come from different cohorts, say so.
  3. Separate the five components. Report stability, skill use, earnings, further study and graduate satisfaction as distinct measures rather than a single "success" rate.
  4. State what earnings figures exclude. Give hours treatment, coverage, response rates and whether any adjustment for field mix, prices or prior attainment was made.
  5. Break down by subgroup. Show outcomes by sex, domestic or international status and field, with suppression rules for small groups.
  6. Track early mismatch. Use graduates' reports of skill use at the first follow-up as an early-warning signal and check whether it persists at the next.

For researchers and evaluators

  1. Ask whether an earnings comparison is raw or selection-adjusted, and treat raw differences between institutions as descriptive only.
  2. Prefer outcome measures within field and cohort, and record entry-year labor market conditions alongside them.

For students comparing programs

Ask for outcomes at more than one point in time, for the field you intend to study, with the response rate and year. Ask how many graduates went into further study and how they are counted. The UNIRANKS career guidance section offers broader orientation material.

Limitations

This article draws on eight peer-reviewed sources. Two were read in full (Clark, Joubert and Maurel; Altonji and Zhong) and six from their abstracts only (Spurk, Hirschi and Dries; von Wachter; Deming and Noray; Bertrand, Goldin and Katz; Cortés and Pan; Dale and Krueger), so claims from those six are limited to what their abstracts state. Most of the research uses US data, several studies concern cohorts that entered work decades ago, and some focus on specific groups such as MBA graduates or selective colleges, so their magnitudes should not be generalized. The official statistics come from two countries and use different methods (a two-wave survey and tax records). Salary figures are in nominal local currency and are not comparable between the two countries. Several relevant studies were excluded because their publisher pages could not be opened without a bot check. The two frameworks in this article are editorial and have not been validated as measurement instruments.

Related reading on UNIRANKS

For definitions of employability and an evidence-quality approach to graduate outcomes, see How Should Universities Measure Graduate Employability? More research on graduate outcomes, internships, employer feedback and career support is collected in the Employability & Career Success research hub.

Editorial disclosure

UNIRANKS publishes university rankings, and its methodology includes an Employability & Career Success dimension. This article reviews independent research on graduate career progression after the first job. It does not describe how UNIRANKS scores institutions.

Research References

  1. Altonji, J. G., & Zhong, L. (2021). The Labor Market Returns to Advanced Degrees. Journal of Labor Economics, 39(2), 303-360. https://doi.org/10.1086/710959 (verified 22 November 2026)
  2. Bertrand, M., Goldin, C., & Katz, L. F. (2010). Dynamics of the Gender Gap for Young Professionals in the Financial and Corporate Sectors. American Economic Journal: Applied Economics, 2(3), 228-255. https://doi.org/10.1257/app.2.3.228 (verified 22 November 2026)
  3. Clark, B., Joubert, C., & Maurel, A. (2017). The career prospects of overeducated Americans. IZA Journal of Labor Economics, 6(1), 3. https://doi.org/10.1186/s40172-017-0053-4 (verified 22 November 2026)
  4. Cortés, P., & Pan, J. (2023). Children and the Remaining Gender Gaps in the Labor Market. Journal of Economic Literature, 61(4), 1359-1409. https://doi.org/10.1257/jel.20221549 (verified 22 November 2026)
  5. Dale, S. B., & Krueger, A. B. (2014). Estimating the Effects of College Characteristics over the Career Using Administrative Earnings Data. Journal of Human Resources, 49(2), 323-358. https://doi.org/10.3368/jhr.49.2.323 (verified 22 November 2026)
  6. Deming, D. J., & Noray, K. (2020). Earnings Dynamics, Changing Job Skills, and STEM Careers. The Quarterly Journal of Economics, 135(4), 1965-2005. https://doi.org/10.1093/qje/qjaa021 (verified 22 November 2026)
  7. Spurk, D., Hirschi, A., & Dries, N. (2019). Antecedents and Outcomes of Objective Versus Subjective Career Success: Competing Perspectives and Future Directions. Journal of Management, 45(1), 35-69. https://doi.org/10.1177/0149206318786563 (verified 22 November 2026)
  8. von Wachter, T. (2020). The Persistent Effects of Initial Labor Market Conditions for Young Adults and Their Sources. Journal of Economic Perspectives, 34(4), 168-194. https://doi.org/10.1257/jep.34.4.168 (verified 22 November 2026)

Sources

Share this article