Return-to-office fights and AI anxiety have drawn attention to a quieter structural shift: workforces are aging, and firms need older and younger employees to learn from each other under strain. Training participation falls sharply with age, yet demographic pressure is rising. Across the OECD the old-age dependency ratio, the number of people 65 and over relative to those of working age, climbed from 19 percent in 1980 to 31 percent in 2023 and is projected to hit 52 percent by 2060. Only about a 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). No amount of office policy resolves this math. Firms need mechanisms that transfer power, not just skills.
Mentoring is often sold as the answer. The research record supports mentoring as a positive practice, but it is more modest than many program brochures imply. The good news for leaders is that modest does not mean marginal. It means you should design mentoring with clear aims and tight feedback loops.
What the strongest evidence actually says
Two large meta-analyses set the baseline. A Journal of Applied Psychology synthesis found that mentoring is generally associated with positive career benefits for protégés, across both objective outcomes and subjective attitudes (Allen, Eby, Poteet, Lentz and Lima, 2004). The headline needs context. The gains on hard outcomes were small in magnitude. The meta-analysis reported a small correlation with compensation, around 0.12, and a modest association with promotions around 0.31. The bigger lifts sat in the softer domain such as career satisfaction. The authors’ implication is plain: mentoring helps, but it is not a guaranteed pay bump.
A broader cross-domain meta-analysis that pooled 112 studies reached a similar conclusion. Mentoring was linked to favorable outcomes across behavioral, attitudinal, health, relational, motivational and career domains, yet the typical effects were small. In workplace settings, corrected correlations clustered in the low teens, roughly 0.14 for career attitudes and 0.16 for situational satisfaction (Eby, Allen, Evans, Ng and DuBois, 2008). The strength of effects differed by context. Academic mentoring showed the strongest associations, while workplace mentoring sat in the middle of the pack. That pattern suggests structure matters. When there is a curriculum and clear milestones, mentoring tends to do more.
These two syntheses are not a license to dismiss mentoring. They are a call to calibrate expectations. If a program is marketed as a fast track to higher pay, the average participant will be disappointed. If it is positioned to support integration, learning momentum and career clarity, the evidence says it can deliver.
The 2020s problem mentoring can actually help solve
Aging reshapes the talent equation. The OECD projects a steep rise in the old-age dependency ratio through 2060, and warns that without productivity gains per capita GDP growth could slow by about 40 percent across the bloc. At the same time, older workers receive less training than their mid-career peers (OECD, 2025). That is a double bind: more experience on the payroll, less investment in updating it.
Mentoring is one of the few scalable ways to bridge this gap in both directions. The formal research base has not yet tested many age-diverse or reverse-mentoring designs with the same rigor as traditional arrangements, which the 2008 meta-analysis notes as an area needing study (Eby, Allen, Evans, Ng and DuBois, 2008). Still, the demographic pressure is not going away. When roles and tools change quickly, pairing people for focused knowledge exchange is a practical hedge. The key is to be honest about what mentoring can and cannot do on its own. It will not offset a training deficit for older workers, and it will not replace performance management for younger ones. It can, however, make both training and management stickier by giving people a human guide and an expectation to apply what they learn.
Design for small effects, then earn bigger ones
If you want mentoring to help across generations, design it as if effects are small and need compounding.
- Tie mentoring to a concrete capability. The meta-analytic pattern that academic contexts outperform workplace ones implies that structure increases value (Eby, Allen, Evans, Ng and DuBois, 2008). Borrow that logic. Define the skill or transition you care about, set a simple syllabus, and add checkpoints.
- Measure what the science says moves. Expect modest shifts on objective outcomes. Track them, but put more weight on near-term subjective gains such as career clarity and situational satisfaction, where effects are reliably positive, if small (Allen et al., 2004; Eby et al., 2008). Use pre-post surveys and a brief follow-up after three to six months.
- Protect time. Older employees are less likely to be in formal training. If mentoring is to carry any load here, give pairs scheduled time and access to learning materials. Otherwise the program competes with production and loses (OECD, 2025).
- Match for complementarity, not myth. Avoid stereotypes about what each generation brings. Match on specific gaps and assets, for example, regulatory know-how with data tooling, or stakeholder management with new product workflows. The evidence base does not yet validate one-age-direction fits all, so treat age direction as a design choice to test, not an article of faith (Eby et al., 2008).
- Train the mentors. Even light-touch guidance increases structure, which the evidence associates with stronger results. A one-page playbook with questions, a goal-setting template and advice on feedback is often enough to lift consistency (Eby et al., 2008).
The practical read for leaders
Leaders face a simple trade-off. Aging will strain growth and skills unless knowledge moves faster than roles change. The OECD’s projections make the macro case to act, and the mentoring literature makes the micro case to act with humility. The meta-analyses show consistent but small benefits on the outcomes you can measure quarterly, and larger perceived benefits on the ones that keep people engaged and oriented (Allen et al., 2004; Eby et al., 2008; OECD, 2025).
Treat mentoring as an operating mechanism, not a morale perk. Budget for it, specify the capability at stake, measure what it moves, and feed those results back into matching and content. Do that and you will get the only return that matters in a multigenerational firm: knowledge handed on in time for someone else to use it.
Sources
- Allen, Eby, Poteet, Lentz and Lima (2004). Career benefits associated with mentoring for protégés: a meta-analysis, Journal of Applied Psychology. Finds mentoring is associated with positive career outcomes but with small effects on hard measures like compensation, larger on satisfaction. 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, Journal of Vocational Behavior. Synthesizes 112 studies and shows mentoring has generally small positive effects, stronger in structured academic contexts than in workplace settings. https://pmc.ncbi.nlm.nih.gov/articles/PMC2352144/
- OECD (2025). OECD Employment Outlook 2025. Documents aging-driven pressure on workforces and lower training participation among older adults, framing why multigenerational mentoring matters. https://www.oecd.org/en/publications/oecd-employment-outlook-2025_194a947b-en.html