Boards know the risk. One resignation, one health scare, one scandal, and a company discovers whether it has a leadership system or a single point of failure. The debate tends to lurch between hero worship and panic hiring. There is a better way hiding in plain sight: rotate leadership by design and measure how well authority travels before you need it.
The strongest signal is at the team level
Two independent meta-analyses point in the same direction. A review of 42 samples found a positive relationship between shared leadership and team effectiveness of rho = .34. The effect was stronger when the work was more complex, which maps directly to knowledge-intensive contexts where succession risk bites hardest. It also mattered what was shared. New-genre leadership, such as transformational behaviors, was more strongly associated with effectiveness than traditional task-structuring behaviors (Wang, Waldman and Zhang, 2014).
A second meta-analysis pooling 50 effect sizes from 3,198 teams found that shared leadership consistently related to higher team performance. The relationship was stronger in real organizational settings than in classrooms and labs, strengthening external relevance. How researchers measured shared leadership mattered. Network-based measures that captured density and (de)centralization showed larger performance links than simple aggregates of individual ratings, a signal that the pattern of who leads whom, and how often, is the mechanism to watch (D'Innocenzo, Mathieu and Kukenberger, 2016).
These studies live at the team level. Applying them to CEO rotation is a reasoned inference, not a causal proof. Still, the spine of the argument is clear. Where work is complex and interdependent, distributing leadership roles and making the network healthier are associated with better results. Succession, in this light, is the ultimate stress test of whether leadership is a property of a few people or a property of the system.
Measure the network, not the nameplate
If network shape tracks performance, then boards should ask for network health metrics, not just a list of high potentials. The research suggests two design implications. First, build density. More ties for advice and decision input tend to coincide with better outcomes at the team level. Second, reduce over-centralization that traps all paths through a single node, because those structures correlate less with performance than more distributed ones in field studies (D'Innocenzo, Mathieu and Kukenberger, 2016).
Translate this into practice. Map the decision network quarterly for the top one or two leadership tiers, not as a social exercise but tied to real decisions taken in the past 90 days. Look for bottlenecks where one executive is the only gateway for multiple critical flows. Rotate who runs the monthly operating review and who presents external updates to investors or regulators. Document decisions and their owners so that accountability follows the work, not the title.
Wang, Waldman and Zhang (2014) found that sharing transformational behaviors carries more weight than simply dividing task oversight. That implies you should rotate not just who signs off, but who sets direction, frames trade-offs, and communicates meaning. Include those lenses in your rotation design and in performance conversations so the shared elements are the ones that matter for effectiveness.
Co-leads as a rehearsal, not a religion
What about the very top? The co-CEO model remains rare. Among roughly 2,200 companies in the S&P 1200 and Russell 1000 over 1996 to 2020, fewer than 100 were led by co-chief executives (Feigen, Jenkins and Warendh, 2022). The record is mixed. The article names notable failures, including Chipotle Mexican Grill, SAP, and Research In Motion/BlackBerry, especially under stress. Yet the authors’ review of 87 companies with co-CEOs argues the model is more viable than common wisdom allows when roles are clearly carved and trust is high (Feigen, Jenkins and Warendh, 2022).
Treat co-leads as a bounded experiment. Use them to rehearse a future in which authority is shared because the work demands it, not because a board cannot decide. Set explicit domains, decision rights, and escalation paths. Tie shared goals to shared incentives. Insist on a single board-facing joint report, even if operating lines are divided. Above all, set a review horizon. The point is to prove that the company can route power through more than one capable node, not to enshrine a structure that the culture will not sustain.
A rotation playbook boards can adopt now
You do not need a co-CEO to build a bench. You need a cadence. Four moves help.
-
Run real rotations on real work. Assign time-boxed, overlapping stints where executives trade P&L ownership for adjacent units. Do not rotate only in back-office roles. Include external exposure such as investor days or regulatory hearings so successors practice direction-setting behaviors that matter for effectiveness in complex settings (Wang, Waldman and Zhang, 2014).
-
Publish a decision ledger. Record the top 20 decisions each quarter, who led, who advised, and the cross-functional ties involved. Over time, the ledger becomes a network map you can interrogate. You want to see more people setting direction and fewer single-threaded approvals, consistent with the network-based measurement effects observed in field settings (D'Innocenzo, Mathieu and Kukenberger, 2016).
-
Rehearse the handoff. Pick two or three plausible shock scenarios and run tabletop exercises where a different executive acts as incident lead in each. The point is less the scenario and more the muscle memory of shifting authority while keeping tempo. Capture lessons and fold them into the rotation calendar.
-
Try co-leads where the work is naturally dual. Joint ventures, multi-sided platforms, or regions with distinct regulatory regimes can justify co-ownership. The arrangement is rare and has failed in prominent cases, so treat it as a proving ground under tight governance rather than a default structure (Feigen, Jenkins and Warendh, 2022).
The practical bet is simple. Succession is not an event to be survived. It is a continuous audit of how widely leadership behaviors are distributed and how gracefully they shift hands. The team-level evidence links shared leadership with better outcomes in complex work. The co-CEO record shows the top seat can be shared, though it is uncommon and requires discipline. Boards that design rotations, measure the network, and rehearse handoffs will stop betting the firm on a single name. They will build a bench, and the company will be ready when it counts.
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 leadership network, guiding how to design and measure rotation. https://journals.sagepub.com/doi/10.1177/0149206314525205
- Wang, Waldman and Zhang (2014). A meta-analysis of shared leadership and team effectiveness. Shows an overall rho = .34 link between shared leadership and effectiveness, stronger in complex work and for transformational behaviors, informing where rotation pays off. https://pubmed.ncbi.nlm.nih.gov/24188392/
- Feigen, Jenkins and Warendh (2022). Is It Time to Consider Co-CEOs? Reviews 87 co-CEO cases, notes rarity and notable failures, yet argues the model is viable with clear rules, shaping how to use co-leads as a bounded experiment. https://hbr.org/2022/07/is-it-time-to-consider-co-ceos