An aging workforce turns a familiar management tool into an economic necessity. The old-age dependency ratio across the OECD has already climbed from 19 percent in 1980 to 31 percent in 2023, and is projected to reach 52 percent by 2060 (OECD, 2025). With fewer younger workers supporting more older citizens, how skills, norms, and tacit knowledge move across age cohorts will shape capacity far more than seating charts or office mandates. Mentoring is often put forward as the glue. The best evidence says it can help, but not in the way executives often assume.
What the research actually shows
The most cited synthesis inside organizations, a meta-analysis in the Journal of Applied Psychology, reported that mentoring is generally associated with positive career benefits for protégés, across both objective outcomes and subjective ones such as career satisfaction (Allen, Eby, Poteet, Lentz and Lima, 2004). The magnitude matters. Effects on hard outcomes were positive yet small in size. The authors reported modest associations with compensation, approximately r = .12, and a stronger but still limited link with promotions, approximately r = .31. Benefits were larger for attitudinal outcomes like career satisfaction than for pay. That is an important calibration: mentoring correlates with feeling better about one’s career more reliably than with a bigger paycheck.
A broader multidisciplinary meta-analysis that pooled 112 studies and 116 independent samples reached a similar conclusion, and extended it (Eby, Allen, Evans, Ng and DuBois, 2008). Across behavioral, attitudinal, health, relational, motivational, and career domains, mentoring was associated with favorable outcomes, but the effects were generally small. The corrected correlations for career attitudes were around rc = .14 and for situational satisfaction around rc = .16. The study also showed that effects differ by setting. Academic mentoring displayed the strongest associations, roughly rc .11 to .36. Workplace mentoring sat in the middle, roughly rc .03 to .19. The authors cautioned against overselling what mentoring can do, a warning worth heeding when programs are justified with promises of rapid pay growth or dramatic performance gains.
These syntheses are correlational and integrate diverse methods. They do not license strong causal claims. What they do offer is a map of where mentoring seems to matter most. If leaders want a blunt instrument for compensation, this is not it. If they want to move attitudes, career clarity, and satisfaction, mentoring is more likely to pay off.
The multigenerational frame managers cannot ignore
Demography makes the mentoring question sharper. The OECD projects the old-age dependency ratio rising to 52 percent by 2060, a structural shift that tightens labor supply and increases the value of retaining and developing workers at all ages (OECD, 2025). Yet participation in training is uneven by age. Only about one third of adults aged 60 to 65 participated in training in 2023, compared with more than half of those aged 25 to 44 (OECD, 2025). That gap is a signal. If older workers participate less in formal training on average, then on-the-job vehicles for knowledge exchange become more important in practice.
Mentoring is a plausible bridge because it can carry both tacit knowledge upward and new practices downward. The evidence base, however, is thinner on age-diverse and reverse mentoring configurations than on traditional pairings, and the multidisciplinary meta-analysis underscores heterogeneity across contexts while warning against inflated claims (Eby et al., 2008). Taken together with the workplace effect sizes that sit in the low to moderate range (Eby et al., 2008) and the small objective career associations reported in the classic meta-analysis (Allen et al., 2004), the design implication is straightforward. Use mentoring to strengthen cross-generational learning and belonging, and do not present it as a shortcut to higher pay.
How to design mentoring for real, if modest, gains
Start by aligning aims with what the evidence supports. If the goal is to improve protégés’ career attitudes and situational satisfaction, the two meta-analyses suggest that mentoring is well suited to the task, with consistently positive, albeit small, associations (Allen et al., 2004; Eby et al., 2008). Make those aims explicit. Promise clearer navigation of roles, faster sensemaking in new domains, and a stronger network. Do not promise higher compensation. The meta-analytic link to pay is small in magnitude, and the broader literature cautions against overselling (Allen et al., 2004; Eby et al., 2008).
Second, treat mentoring as a system, not a match. The workplace effects reported by Eby et al. (2008) sit between weak and moderate depending on outcome. That means small design errors can wash out already modest benefits. Build rhythm. Provide light scaffolding for conversations, including goal-setting and reflection prompts, so the relationship stays focused on growth and sensemaking rather than drifting into unstructured chats that go nowhere. Keep the administrative burden low so the program is easy to sustain. Pair mentoring with access to formal learning, particularly for age groups that participate less in training on average (OECD, 2025). The point is not to replace courses with conversations. It is to make learning continuous by turning everyday work into an arena for guided practice.
Third, make it cross-generational by design. The OECD figures on aging and training gaps make a practical case for pairing older and younger colleagues so skills, norms, and tacit know-how flow in both directions (OECD, 2025). Reverse mentoring is a reasonable component here, especially where younger employees can share emerging tools and practices while drawing on experienced colleagues for context and judgment. The evidence synthesised to date gives fewer direct estimates for reverse or explicitly age-diverse mentoring structures than for traditional dyads, so treat this as a design choice informed by logic and demographics rather than as a proven high-yield variant (Eby et al., 2008).
Finally, measure what mentoring is likely to move. Track career satisfaction, role clarity, and perceived support, the domains where meta-analytic associations are strongest (Allen et al., 2004; Eby et al., 2008). Use brief, repeated measures so you can adjust. If you choose to monitor objective career markers, interpret small shifts soberly given the small effect sizes reported in the literature for compensation and the setting-specific ranges for workplace outcomes (Allen et al., 2004; Eby et al., 2008). If scores stagnate, fix the system before you expand it.
The practical read for a graying economy
The demographic arithmetic in OECD countries tightens over the coming decades. That makes skill transmission across age cohorts a strategic question, not a human resources nicety (OECD, 2025). The two major meta-analyses give leaders a clear brief. Mentoring is associated with positive outcomes, particularly for attitudes and satisfaction, and the typical effect sizes are small in workplace settings (Allen et al., 2004; Eby et al., 2008). Set expectations accordingly. Use mentoring as connective tissue for a multigenerational workforce and as a complement to formal training where participation differs by age. Promise clarity, connection, and learning. Deliver those, and you will have used mentoring for what it does best.
Sources
- Allen, Eby, Poteet, Lentz and Lima (2004). Career benefits associated with mentoring for protégés: a meta-analysis. Mentoring is positively associated with career outcomes, with small effects on compensation and stronger links to subjective satisfaction, which calibrates realistic goals for programs. https://www.semanticscholar.org/paper/Career-benefits-associated-with-mentoring-for-a-Allen-Eby/a13c80a7f263f4378c07b474a0628105b91c6d28
- Eby, Allen, Evans, Ng and DuBois (2008). Does Mentoring Matter? A Multidisciplinary Meta-Analysis. Across 112 studies, mentoring relates to favorable outcomes but generally with small effect sizes, varying by context, and the authors caution against overselling. https://pmc.ncbi.nlm.nih.gov/articles/PMC2352144/
- OECD (2025). OECD Employment Outlook 2025. Ageing will lift the old-age dependency ratio to a projected 52 percent by 2060 and older adults train less than mid-career peers, sharpening the case for cross-generational knowledge transfer. https://www.oecd.org/en/publications/oecd-employment-outlook-2025_194a947b-en.html