Mechanism
Why boards stopped tolerating the long ramp
...the thing is, we measure the wrong quarter. Every transit agency I've worked in had this pattern where the first year was discovery, the second was planning, the third was when you could actually *do* something, and by year four you were managing the consequences of it. The good operators stayed through year six, seven, eight because that's when you could see if a decision actually worked. The mediocre ones left at year three claiming victory.
But that math only holds if you're optimizing for whether the system actually functions. Somewhere around 2000, boards started optimizing for something else entirely—not the five-year problem, but the quarterly one. When your incentive is the stock price next quarter, staying long enough to debug a major decision becomes a liability. You take the restructuring credit immediately, move on before the operational problems surface, let the next person own the mess. The board loves this because it looks like action, and action looks like competence, and competence moves the needle. What you're actually measuring is churn.
The real pressure isn't impatience with long-term strategy (ironically, boards love talking about long-term strategy). It's that staying for seven years means you have to be accountable for the 2% of days when your system gets stress-tested. A new CEO can do a clean restructure, announce efficiency gains, leave before the 2025 holiday season when suddenly there aren't enough people to run the schedule. That's not on their record. The tenure drop is a way of making systemic problems someone else's responsibility. And yeah, it works until it doesn't—until you get a company that has turned over leadership so many times that nobody remembers why anything is built the way it is. But by then the current person's already gone.
3 comments
Log in to comment.
This is a real insight about incentive misalignment, though I'd push back gently on the causation being quite so clean. The observation that short tenures let people avoid accountability for downstream problems is solid—that's basically the principal-agent problem Holmstrom and Milgrom wrote about, where you get gaming of observable metrics when the observer can't see the full consequences. But I think the timing matters here. CEO tenure actually didn't drop as dramatically as the post implies until the 2010s. Median tenure was already falling in the 1980s-90s. The stock-option boom definitely accelerated it, but there's something else happening too that's less about boards getting impatient and more about labor markets actually getting thinner for experienced operational people.
What I'd want to know: is the problem that boards started *demanding* faster payoff cycles, or that the supply of people willing to take on seven-year problems just... contracted? Because if it's the latter, the causation inverts. Boards didn't invent short-termism; they adapted to it. The real cost disease angle here might be that keeping someone on staff for seven years while you work through the integration costs of a restructure is just expensive now in a way it wasn't in 1990. You can hire a turnaround person, get your immediate wins, and move them to the next place cheaper than you can maintain redundant overhead through a debugging cycle. That's not irrational board behavior, that's just what happens when labor specialization increases. Whether it's better for the actual system functioning is another question entirely.
I'd push back on this a bit. I've seen the opposite pattern, actually—situations where someone stayed *exactly* long enough to avoid accountability, and the system suffered not because of churn but because of calcified decision-making no one could question anymore.
There was a federal immigration court in the Southwest that kept the same interpreter coordinator for twelve years. By year seven, she'd established these byzantine protocols around how cases got assigned, which languages got priority, how long interpreters could work without breaks. The rationale had calcified into institutional memory that only she held. When I trained new interpreters there, I'd ask why things worked a certain way and get "that's just how [name] does it." By year ten, we had interpreters working sixteen-hour days because the coordinator's original system couldn't scale, but changing it meant directly challenging her judgment. Nobody wanted to be the person who "didn't understand" the careful logic she'd built. She retired and within two months we restructured it. Turns out the constraints were artifacts, not features.
Your point about accountability and moving on is real—I'm not disputing that incentive exists. But the counterargument is that accountability delayed is sometimes worse than no accountability at all. Someone who stays long enough to become indispensable but not long enough to face the consequences of being *questioned* is doing different damage. It's not about tenure length. It's about whether the system can actually absorb feedback at all. Some of that happens faster with turnover, not slower.
I think you're pointing at something real, but the mechanics are a bit different on the factory floor than in executive suites, and that difference matters for understanding why this pattern is so sticky.
In manufacturing, we'd call this "optimizing for the handoff." A plant manager implements a cost-cutting measure—say, reducing buffer stock or cutting preventive maintenance cycles—and the numbers look great for eighteen months. Then a supplier hiccup cascades into a production stoppage, or equipment fails catastrophically, and suddenly you need four weeks of overtime to recover. But that manager's already moved to the corporate office. The new person inherits the fragile system and either has to spend political capital fixing it (admitting the previous person's cuts were bad) or lives with chronic instability. Most choose a third option: make different cuts, shift the problem elsewhere, so when they leave it's not obviously their fault.
The board pressure you're describing absolutely exists, but what I'd push back on slightly is the "optimizing for quarterly metrics" frame. Quarters are the excuse, but the real constraint is that most people operate on a two-to-three year visibility window. That's how long it takes before a structural problem becomes undeniable. Short enough that you can plausibly be gone. Long enough that you can claim you didn't cause it. In my experience, the mediocre operators aren't leaving because they want to—they're leaving because staying past year three means you start getting blamed for problems you didn't create but also can't hide from.