Boards are again betting the company on a single successor. The context is harder than the last cycle: hybrid work is contested, AI is rewriting roles, and public markets punish stumbles in weeks. The risk is not only picking the wrong person. It is overconcentrating power in a way that slows learning and leaves the firm brittle between coronations.
There is a better path. Build rotation into the fabric of leadership before the crisis. The best field evidence on how leadership structure relates to outcomes comes from teams, not just the C-suite. It points in a clear direction.
Why circulation beats concentration
Two independent meta-analyses synthesize decades of team research. Across 3,198 teams and 50 effect sizes, shared leadership is positively associated with team performance, with stronger effects in real organizational settings than in classrooms or labs. The association is also stronger when leadership is measured as a social network, meaning who is actually influencing whom across the group, rather than as a simple average of individual traits (D'Innocenzo, Mathieu and Kukenberger, 2016). A second meta-analysis of 42 independent samples reports an overall association of rho = .34 between shared leadership and team effectiveness, and it finds the relationship is stronger when the work is more complex. It also notes that newer forms of leadership such as transformational behaviors have a stronger shared effect than traditional initiating-structure routines (Wang, Waldman and Zhang, 2014).
These are team-level results, not CEO experiments. They do not prove that rotating the top job will raise corporate performance. They do, however, establish a mechanism that matters in knowledge work. When leadership circulates through the network rather than pooling at a single point, teams tend to perform better, especially on complex tasks. If the bulk of an enterprise's value creation happens in such teams, the logic of distributing and rotating authority up the line is not a stretch. It is a disciplined extrapolation with a clear caveat.
What top jobs can borrow, carefully
Sharing authority at the summit remains rare. Among roughly 2,200 companies in the S&P 1200 and Russell 1000 between 1996 and 2020, fewer than 100 used co-chief executives. The model has notable failures, including at Chipotle Mexican Grill, SAP, and Research In Motion/BlackBerry. Yet a study of 87 firms with co-CEOs argues the arrangement is more workable than common wisdom admits, provided roles are explicit and the partners bring complementary strengths (Feigen, Jenkins and Warendh, 2022). The sample is small relative to the market, and the evidence is observational. It still matters because it shows that shared power at the top can be designed to function, not only to fail.
The practical borrowing is not to default to permanent co-CEOs. The lesson is to normalize shared and rotating authority in the run-up to succession and in discrete moments when the job’s demands diversify. That includes time-boxed co-leads during transitions, shared mandates for cross-functional growth bets, and a rotating chair for the executive committee. The rotation is not symbolism. It creates repeated, lower-stakes exposures to the full span of decisions, and it forces the system to document who decides what, on what information, and with what cadence. That is the same social network clarity the stronger meta-analytic effects depend on in teams (D'Innocenzo, Mathieu and Kukenberger, 2016).
How to run rotations without chaos
Firms avoid rotation because they fear muddled accountability. The team evidence points to design choices that counter that risk. First, design for complexity. Shared leadership’s relationship with effectiveness is stronger when the work is complex (Wang, Waldman and Zhang, 2014). Rotate authority where the firm faces complex, interdependent problems such as platform strategy, AI product integration, or supply resilience. Keep routine operations under clear single-point ownership.
Second, measure the network, not just the org chart. In the meta-analysis, social-network-based measures of leadership dispersion correlated more strongly with performance than simple aggregates of individual leadership scores (D'Innocenzo, Mathieu and Kukenberger, 2016). In practice, that means instrumenting decision forums. Track who speaks, who proposes, who synthesizes, and who gets followed. Use those data to ensure rotation does not become cosmetic and to identify emergent leaders.
Third, use explicit splits when sharing the very top. The co-CEO evidence highlights both viability and failure. The failures are reminders to define domains, escalation paths, and an external face for investors in advance, not in the first earnings call under pressure (Feigen, Jenkins and Warendh, 2022). When a board authorizes a time-bound co-lead during a handover, write the roles and the sunset clause. Treat the pair as a single performance unit for incentives, then separate rewards to match the next roles once the rotation ends.
Finally, treat rotation as a rehearsal loop. The team meta-analyses emphasize that leadership is a set of behaviors distributed across a network, not a fixed trait lodged in a single individual (Wang, Waldman and Zhang, 2014; D'Innocenzo, Mathieu and Kukenberger, 2016). Rotations give more people repeated practice with those behaviors at scope. Over twelve months, run three to four rotations of the executive chair for key strategy reviews, assign co-leads for one or two company-critical initiatives, and publish a forward calendar of who holds which mandate when. The calendar is a control system. It signals that rotation is planned, not improvised.
The practical read
Succession without rotation is a cliff. It depends on a single bet and a fragile pipeline. Succession with rotation is a system. It converts leadership from a title to a repeatable pattern of influence across the network, which is where the meta-analytic associations with performance are strongest. The co-CEO record is mixed, and rare, but it offers a proof of possibility that shared authority can be made to work at the top when roles are explicit.
Start before the chair is empty. Map the firm’s complex decisions and schedule rotating ownership. Instrument the forums to see the leadership network you actually have. Use time-boxed co-leads for transitions with hard edges and clear exits. When the board finally names the next CEO, that person steps into a system that already shares and rotates power. The company inherits resilience, not just a ruler.
Sources
- D'Innocenzo, Mathieu and Kukenberger (2016). A Meta-Analysis of Different Forms of Shared Leadership–Team Performance Relations. Finds a positive association between shared leadership and team performance, stronger with social-network measures and in field settings, supporting rotation as a performance-relevant design. 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 = .34 association between shared leadership and team effectiveness, stronger in complex work and with newer leadership behaviors, grounding rotation in knowledge-intensive contexts. https://pubmed.ncbi.nlm.nih.gov/24188392/
- Feigen, Jenkins and Warendh (2022). Is It Time to Consider Co-CEOs? Reviews 87 co-CEO companies, notes the rarity and notable failures, yet argues the model can work with clear role design, informing how to structure shared authority at the top. https://hbr.org/2022/07/is-it-time-to-consider-co-ceos