Boards still behave as though the fate of a company rests on finding the right individual to sit at the top. The research on leadership has spent forty years quietly dismantling that belief, and the picture it leaves is more useful than the myth it replaces.
Shared leadership is not a slogan, it is a measured effect
The strongest evidence does not concern the lone leader at all. It concerns leadership distributed across a team. A meta-analysis by D'Innocenzo, Mathieu and Kukenberger (2016), covering 3,198 teams, found a positive relationship between shared leadership and team performance, stronger when leadership was measured as a genuine network property rather than a single figure. Wang, Waldman and Zhang (2014), aggregating 42 samples, put the overall relationship at about 0.34, and found it was largest for the transformational, new-genre forms of leadership rather than the old command-and-structure kind. A third meta-analysis reached the same conclusion and identified the mechanism: shared leadership works partly by building trust and consensus inside the team, not by magic (Nicolaides et al., 2014). Authority spread across capable peers, the arrangement that a company like Semco institutionalized, has real and repeatable performance behind it.
Changing the person at the top is expensive and rarely decisive
Set the hero belief against the data on succession itself. In a meta-analysis of 60 samples covering 13,578 CEO successions from 1972 to 2013, Schepker and colleagues (2017) found that CEO succession hurts performance in the short term and has no significant direct effect in the long term. The change is a shock with a bill attached, and the long-run outcome is decided less by the identity of the new chief than by the strategic change that follows and by whether the successor came from inside or outside. Internal successors tended to improve long-term performance and pursue less disruptive change. Bringing in an outside savior produced more strategic upheaval, which was associated with worse long-term results.
How the handover is planned matters more than the charisma of the person handed to. Studying 204 successions, Zhang and Rajagopalan (2004) found that relay succession, in which a designated heir is groomed over time, led to better post-succession performance, especially when prior performance was weak and the environment turned unstable. The unglamorous practice of preparing a successor beats the drama of parachuting in a star.
The hero is partly in the eye of the beholder
Part of the reason boards overpay for saviors is a perceptual bias with a name. In their classic work on the romance of leadership, Meindl, Ehrlich and Dukerich (1985) showed that observers over-attribute organizational outcomes to the leader, and that the same executive is judged far more charismatic when the company is doing well. Charisma, in other words, is largely something followers project onto results, not a fixed trait that produced them. The heroic-CEO narrative is partly a story we assemble after the fact to explain performance we cannot otherwise account for.
Two at the top can work, with a caveat
Even the sharing of the very top job is more viable than the unity-of-command orthodoxy admits, within limits. Feigen, Jenkins and Warendh (2022) studied 87 companies led by co-CEOs and argued the model is workable, while noting it remains rare and has visible failures. The nuance comes from Krause, Priem and Love (2015), who studied 71 co-CEO pairs and found that a moderate power gap between the two was associated with better firm performance, but that the benefit reversed once the gap grew too large. Shared command at the summit is not a free lunch. It works when the division of authority is clear enough to avoid paralysis and balanced enough to avoid a figurehead. One of the first systematic studies of the arrangement, by Arena, Ferris and Unlu (2011), found co-CEO structures cluster around mergers, family firms, and transitions, contexts where complementary skills and mutual checking outweigh the coordination cost, which is a more precise prescription than either the unity-of-command orthodoxy or the co-CEO fashion.
The practical read
Stop designing the organization around the search for one indispensable person. The evidence favors building leadership capacity across the team, where the performance effect is real and repeatable (D'Innocenzo, Mathieu and Kukenberger, 2016; Wang, Waldman and Zhang, 2014). Treat succession as a shock to be managed with a groomed internal successor rather than a lottery to be won with an outside star (Schepker et al., 2017; Zhang and Rajagopalan, 2004). And remember that some of the hero's shine is a trick of attribution (Meindl, Ehrlich and Dukerich, 1985). A company that distributes leadership and plans its handovers is not settling for less. It is refusing to bet its future on a story.
Sources
- D'Innocenzo, Mathieu and Kukenberger (2016). A Meta-Analysis of Different Forms of Shared Leadership. Across 3,198 teams, shared leadership is positively related to team performance, stronger as a network property. https://journals.sagepub.com/doi/10.1177/0149206314525205
- Wang, Waldman and Zhang (2014). A meta-analysis of shared leadership and team effectiveness. Overall relationship about 0.34, largest for new-genre transformational leadership. https://pubmed.ncbi.nlm.nih.gov/24188392/
- Nicolaides and colleagues (2014). The shared leadership of teams: a meta-analysis. Shared leadership lifts performance partly by building trust and consensus. https://doi.org/10.1016/j.leaqua.2014.06.006
- Schepker, Kim, Patel, Thatcher and Campion (2017). CEO Succession, Strategic Change, and Post-Succession Performance. Across 13,578 successions, succession hurts short-term performance with no significant direct long-term effect. https://www.sciencedirect.com/science/article/abs/pii/S1048984317301984
- Zhang and Rajagopalan (2004). When the Known Devil Is Better Than an Unknown God. Relay succession with a groomed heir produced better post-succession performance, especially under instability. https://journals.aom.org/doi/10.5465/20159598
- Meindl, Ehrlich and Dukerich (1985). The Romance of Leadership. Observers over-attribute organizational outcomes to leaders; charisma is largely projected onto results. https://doi.org/10.2307/2392813
- Feigen, Jenkins and Warendh (2022). Is It Time to Consider Co-CEOs? A study of 87 co-CEO companies arguing the model is workable but rare. https://hbr.org/2022/07/is-it-time-to-consider-co-ceos
- Krause, Priem and Love (2015). Who's in charge here? Co-CEOs, power gaps, and firm performance. A moderate power gap between co-CEOs helps performance, but a large gap reverses the benefit. https://doi.org/10.1002/smj.2325
- Arena, Ferris and Unlu (2011). It Takes Two: The Incidence and Effectiveness of Co-CEOs. Co-CEO structures cluster around mergers, family firms, and transitions where complementary skills outweigh coordination costs. https://doi.org/10.1111/j.1540-6288.2011.00305.x