CEO tenure has collapsed and nobody knows why

The thing that's getting lost in this framing is the difference between *measuring what changed* and *understanding what the change means*. You've got the measurement problem pinned down pretty well—tenure fell, but we don't actually know if boards got pickier or if the job itself became less stable or if CEOs started jumping ship voluntarily. Fair enough. But there's a deeper identification problem you're skirting: we don't have shared language for what "performance" even means in a CEO context, and that matters way more than it should. When a board forces out a CEO after three years, the official story is almost always some version of "strategic realignment" or "market conditions." The actual reasons—missed a specific analyst target, board member's pet project tanked, activist investor threatened a proxy fight, the CEO wouldn't play ball on some governance thing—get buried or contradicted depending on who's telling the story later. I've sat in depositions where the defendant's version of why they got fired bore almost no resemblance to what HR had documented, and nobody was technically lying. They just had different reference frames for what counted as "performance." So when you ask if shorter tenure is good or bad, you're working with data that's already been filtered through whoever won the interpretation fight about what happened. The actual mechanism—impatience, higher standards, voluntary movement, measurement inflation—might be resolvable with better record-keeping. But I'd want to see someone actually look at what boards *documented* as the reason for departure, not retrospective analysis of stock price or innovation metrics. Until then the "why" staying open is kind of structural to the problem.

The thing that's getting lost here is the distinction between "tenure got shorter" and "survivorship got harder." Those aren't the same problem and they'd have totally different causes, but the data on median tenure collapses them together. Say you have a cohort of 100 CEOs hired in 1990. Ninety of them make it five years, seventy make it ten years. Now run the same experiment in 2010. Maybe eighty make it five years, but only thirty make it ten years. Your median tenure drops—looks like impatience. But what actually happened? The distribution of *who gets fired when* changed. You can have faster churn at the top without boards getting trigger-happy; you just need more people flaming out earlier for reasons that are orthogonal to patience—worse onboarding, steeper expectations on day-one performance, boards that are actually enforcing their own strategic plans instead of letting the CEO freeform for five years. I ran scheduling at a mid-sized system and we had this exact problem with operator retention. The numbers said we were hemorrhaging people. But when you looked at the actual curves, we weren't firing more people at three years in. We were losing *different* people—ones who couldn't hack the new (worse) dispatch software, ones who had family stuff come up because we'd stopped doing flexible shifts. Same median tenure number, completely different root cause, completely wrong fix if you'd just looked at the summary stat. The CEO research probably needs to bucket people by how they exited and when, then ask whether the *conditional* probability of survival changed for each bucket. Otherwise you're just watching the shadow move and guessing at what cast it.

I'd push back gently on the "identification problem is hard" framing, because I think there's a simpler mechanism hiding in plain sight that the CEO literature mostly ignores: what happened to the rest of the labor market at the same time. Tenure collapsed everywhere, not just in the executive suite. Median job tenure for all workers dropped from about 4.2 years in the early 1990s to around 3.5 by the 2010s. The Bureau of Labor Statistics has been tracking this for decades. The difference is we have a story for why that happened—globalization, offshoring, automation, the shift away from long-term employment contracts as a norm. But somehow when we talk about CEOs we treat it like a separate phenomenon caused by activist investors or governance rules, when it might just be that the entire employment relationship fundamentally changed. Boards stopped thinking in terms of "grow this person into the role over a decade" and started thinking in terms of "hire for immediate fit, replace if conditions shift." That's not activist pressure; that's just how labor markets work now. The tricky part—and maybe this matters for your "is it bad?" question—is that the incentives for regular workers and CEOs diverged sharply even as tenure converged. Shorter tenure is fine if you're a software engineer with transferable skills and a hot market. For a CEO, especially in capital-intensive industries, it probably means less appetite for long-term bets. But I haven't seen anyone actually study whether CEO turnover explains changes in capital expenditure or R&D spending, which would matter more than tenure itself.

The thing that keeps me from buying most of these stories is that people keep diagnosing the *symptom* as if it's the cause. Boards got more independent, say-on-pay votes started, activist investors got louder—yes, all true. But none of that actually explains why a board would suddenly decide their perfectly functional CEO needed to go. You need a *reason*, and "we have better governance now" doesn't supply one. What I'd actually want to know is whether the job itself changed in ways that made it harder to stay. Not metrics getting harsher in theory, but like—did the actual demands on a CEO shift? More fragmented media attention means your mistakes blow up faster. More activist attention means you can't ignore a 300-basis-point miss. Faster technology cycles mean your five-year strategic plan is worth less than it used to be. I spent enough years watching dispatch schedules get shredded on day one to know that systems designed for stability fail when the environment stops being stable. Maybe the CEO role just got more volatile, and tenure collapsed because people couldn't actually succeed at the same tenure lengths anymore, so boards churned them sooner because they *had to*. That'd flip the "is it bad?" question. If CEOs are leaving because conditions won't let them stay, shorter tenure isn't a governance victory—it's a sign that the role got genuinely harder in some structural way. Whether that's actually true is worth checking.