Boards still script succession as a single heroic handoff, then act surprised when the crown fits nobody for long. The work beneath the title has changed. It is more interdependent and complex, and the science says performance follows when leadership is treated as a team property, not a personality trait. That has hard implications for how firms groom successors and share the top job.
What the evidence actually says
Two independent meta-analyses converge on a clear result. Shared leadership, where influence and decision rights are distributed across team members, is associated with better performance. In a synthesis of 50 effect sizes from 3,198 teams, researchers found a positive relationship between shared leadership and team performance, with larger effects in real organizations than in classrooms or labs (D'Innocenzo, Mathieu and Kukenberger, 2016). How you measure sharing matters too. Studies that captured leadership as a social network, such as influence density or decentralization, showed stronger effects than those that simply aggregated individual ratings (D'Innocenzo, Mathieu and Kukenberger, 2016).
A separate meta-analysis of 42 independent samples reported an overall correlation of 0.34 between shared leadership and team effectiveness, and the gains were larger when the work was more complex, the condition that now defines much of knowledge work (Wang, Waldman and Zhang, 2014). The type of leadership shared matters. Distributing newer forms like transformational or charismatic leadership correlated more strongly with team effectiveness than sharing traditional behaviors like initiating structure or consideration, and shared leadership showed unique, incremental effects on performance beyond other factors (Wang, Waldman and Zhang, 2014).
The implication is not subtle. When influence is broadly held and practiced as a networked capability, teams do better, and the advantage grows when tasks are knotty. That is the setting in which most succession failures are born: a single leader trying to carry complexity that is better handled by a system.
Turn succession into a system
If performance follows shared leadership, then the bench you need is not a shortlist of heirs. It is a web of people who can take the lead on consequential work and hand it back. That suggests four design moves.
- Rotate real authority, not titles. Put rotating leads on core missions and time-box them. Think product lines, strategic projects, customer segments. Give each lead the budget levers and the accountability that match the scope. Do not hoard approvals in a central office.
- Treat influence as a network to be cultivated. Map who others seek for advice, decisions, and escalation. The meta-analytic record suggests network measures track real shared leadership better than aggregate ratings, and they relate more strongly to performance in the field (D'Innocenzo, Mathieu and Kukenberger, 2016). Use that map to grow density where it is thin.
- Share the newer forms of leadership intentionally. If transformational behaviors travel better than task structuring, build rituals that spread vision-setting, stakeholder storytelling, and sense-making across the team. That aligns with the evidence that sharing these “new-genre” behaviors is more strongly tied to effectiveness (Wang, Waldman and Zhang, 2014).
- Make complexity the training ground. Place emerging leaders in cross-functional, ambiguous work where the shared-leadership advantage is strongest. Let them practice passing and receiving the baton without ceremony, because the meta-analytic link to effectiveness strengthens as complexity rises (Wang, Waldman and Zhang, 2014).
This reframes succession planning. You are not only picking a next-in-line. You are building a resilient system in which the next-in-line is one of several people already leading in plain sight.
The top job is not exempt
Shared leadership is not just a middle-management trick. A review of 87 companies that used co-CEOs argues the arrangement is more viable than many assume, even if it remains rare and imperfect (Feigen, Jenkins and Warendh, 2022). Among roughly 2,200 firms in major indices from 1996 to 2020, fewer than 100 adopted co-chief executives, and there are visible failures, including Chipotle Mexican Grill, SAP, and Research In Motion/BlackBerry, especially under stress (Feigen, Jenkins and Warendh, 2022). Those caveats matter. They point to context and design, not to an automatic win.
What the co-CEO record does establish is possibility. The existence of dozens of firms that have sustained a split at the very top weakens the reflex that leadership must be indivisible. For boards, that opens a practical path. You can pilot shared top leadership explicitly, for a defined scope and time, while watching the same metrics you would for any executive transition. Treat it as an experiment that tests whether your operating model can handle distributed authority where it is most visible. The underlying bet is consistent with the team-level science. If complex work benefits from shared influence, then the most complex portfolio of all, the enterprise, can sometimes benefit too. That is an interpretation, not a guarantee, and it demands careful governance.
Start in 90 days
You do not need to rewrite the org chart to begin.
- Name three mission-critical domains and assign rotating executive sponsors with full decision rights for 6 to 12 weeks. Publish the rotation calendar and the exit criteria for each sponsor’s tour.
- Instrument the network. Run a confidential, lightweight advice-and-influence survey each quarter to see who colleagues turn to on strategy, customers, and people. Use the results to broaden participation and to verify whether shared leadership is actually growing, a method consistent with evidence that network-based measures track performance better in the field (D'Innocenzo, Mathieu and Kukenberger, 2016).
- Shift development to complexity. Move your succession candidates into cross-functional, ambiguous work and train them to share transformational behaviors. The evidence ties that sharing to higher team effectiveness, especially when the work is complex (Wang, Waldman and Zhang, 2014).
- Consider a bounded co-leadership trial at the top. If you must navigate a major transition, study the co-CEO record, including its failures and rarity, and define a narrow, time-limited co-leadership mandate with explicit review points (Feigen, Jenkins and Warendh, 2022).
The practical read is straightforward. Succession risk falls when leadership is redundant on purpose. The science backs distributing influence in the places where work is hard, and even at the top there is precedent for sharing power. Firms that still search for a single savior are betting against both the evidence and the way modern work gets done.
Sources
- D'Innocenzo, Mathieu and Kukenberger (2016). A Meta-Analysis of Different Forms of Shared Leadership–Team Performance Relations. Finds a positive shared-leadership to performance link, stronger in field settings and when measured as a social network. https://journals.sagepub.com/doi/10.1177/0149206314525205
- Wang, Waldman and Zhang (2014). A meta-analysis of shared leadership and team effectiveness. Reports an overall rho of .34, stronger effects in complex work and for shared transformational leadership, with unique incremental effects on performance. https://pubmed.ncbi.nlm.nih.gov/24188392/
- Feigen, Jenkins and Warendh (2022). Is It Time to Consider Co-CEOs? Reviews 87 co-CEO companies and argues the model is viable though rare, with notable failures such as Chipotle, SAP, and BlackBerry. https://hbr.org/2022/07/is-it-time-to-consider-co-ceos