Every L&D leader has watched this movie before. The manager who worked through six weeks of coaching, showed up differently in every session, and left the program a visibly better leader, until the quarter got busy, the coaching conversations slowly disappeared from the calendar, and eighteen months later you couldn’t point to what had changed. Or the leader whose 360 scores climbed right after the program, then slid back down the following cycle, as if the improvement had been rented rather than owned.
You approved the budget for that program, you saw the change happen, and then you watched it go.
This describes what happens to leadership under pressure, and there’s a name for the pattern underneath it.
The Weight of the Modern Workplace
Pressure on leaders has become the operating condition rather than the occasional spike. AI is rewriting what a manager is expected to know and decide in real time. Spans of control keep widening as organizations flatten. Restructuring has become close to constant. Hybrid work adds a layer of complexity to every interaction that used to be simple, and decision velocity has outpaced the time leaders have to think clearly before they act.
The data bears this out at the top of organizations, where you’d expect the most support and the most experience managing it. According to the Brunswick Leadership Stress Index, more than 50% of C-suite leaders describe themselves as very stressed, and roughly 66% say that stress has risen year over year.
None of this means leaders are fragile. It means pressure is now a permanent design constraint, and leadership development can no longer be built for ideal conditions and hope the real world cooperates. It has to survive the conditions leaders are actually operating in.
The Reversion Effect
Here’s the pattern we’ve named after watching it show up again and again:
The Reversion Effect is what happens when a leader who has demonstrably changed their behavior in a development program returns to their pre-program behavior the moment real organizational pressure arrives.
Under pressure, managers lose access to their new leadership model and default to the identity that made them successful before they became leaders.
This is exactly why the highest performers are the most at risk. Their identity as an individual contributor, the one who got the promotion in the first place, was the most rewarded version of themselves they’ve ever had. That makes it the most deeply held, and the first thing pressure reaches for.
Why It Happens
Start with the biological floor. Under cognitive load, the brain prioritizes speed and protection, reaching for well-practiced pathways over recently learned ones. That’s well established, and it’s why so much writing on stress and leadership stops right there.
It’s not the whole story, though, and stopping there misses the part that really matters to an L&D buyer. Under pressure, the brain reaches for a specific pattern: that person’s pattern, not just any old habit. Reversion is identity-consistent, not generically regressive. Two managers under identical pressure, with identical training, will revert to different behaviors, each to the version of themselves that was rewarded before they ever led anyone. That specificity is what a purely neurological explanation can’t account for.
The peer-reviewed evidence supports this at scale. In a study of 304 employees rating their leaders separately under routine and stressful conditions, Klebe, Klug & Felfe (2022) found a real, measurable gap in how the same leader showed up, and that gap widened when the leader was under more strain. The inconsistency itself hurt the people being led, independent of whether the leadership was good or bad in either state, which reframes reversion from a disappointing training outcome into an organizational risk in its own right.
We’ve seen the same thing from the practitioner side. Across our coaching team, 30 coaches working in dozens of different organizations, different industries, systems, and cultures, have independently reported the same pattern. That kind of convergence, across genuinely different environments, is hard to explain any other way.
What reversion looks like before you can name it:
- Coaching conversations go first. They’re the least identity-native behavior a leader has, and the easiest thing to defer when the calendar tightens.
- 360 scores improve, then decline. The measurement window usually closes before the real pressure test arrives.
- Delegation reverses. The leader starts taking work back without ever deciding to.
Why the Common Fixes Don’t Hold
Three explanations dominate the conversation about why leadership development doesn’t stick, and each one is held by serious people for good reason. Each is also partly right, which is exactly why they’re worth taking seriously before showing their limits.
More skills training.
This is the default assumption, and it’s the quickest one to rule out. It fails because the skills were already there. The manager who reverted could still run the coaching conversation; they simply chose not to. Knowing how was never the constraint. (We made a version of this argument in our post on why knowing your DiSC style isn’t the same as changing how you lead.)
System alignment.
This is the most credible of the three, and it’s the position held by the field’s leading professional body: leaders don’t return from a program to a blank slate, they return to a system, and if that system contradicts the new behavior, the behavior fades. This is largely true, and it genuinely helps. But the limit is precise: aligned systems reduce reversion without eliminating it, and the variance that remains tracks the person, not the environment. That’s exactly what the 30-coach observation tells us: the pattern shows up across very different systems, which means it isn’t a property of any single one of them.
Reinforcement.
Handle this one carefully, because reinforcement genuinely works, and the decay data supports using it. The problem is what reinforcement is aimed at. Reinforcement aimed at behavior, layered on top of an unchanged self-concept, builds scaffolding. Scaffolding holds, right up until the pressure exceeds it or the scaffolding comes down.
What all three fixes share is an assumption: that the leader is still selecting a response. Better skills give them more to select from. Better systems give them a better environment to select within. Better reinforcement strengthens what they’ve already selected. All three require conscious selection at precisely the moment conscious selection is least available.
What Really Interrupts It
If skills, systems, and reinforcement aren’t sufficient on their own, what is?
Identity change requires action under real conditions, sustained over time, not reflection alone. Herminia Ibarra’s research on leadership transitions supports this directly: her outsight principle holds that leaders think their way into a new identity only after acting their way into it, and the final stage of that transition is internalization. The Reversion Effect is what shows up when pressure arrives before internalization has finished.
Ibarra’s data also explains why so few leaders manage this transition on their own. In her 2013 INSEAD survey, only 47% of leaders facing a major change in what was expected of them had actually been promoted into a new title; the rest were expected to lead at a bigger scale with no new title and no formal support, a shift she calls the do-it-yourself transition. Most leaders can’t generate the right conditions for that kind of change alone. That’s precisely why development has to build those conditions deliberately.
What are those conditions? Sustained duration instead of a single event. Real working conditions instead of a classroom. Peer witness. External reflection from someone who won’t let the drift pass unnoticed, the kind of structure you also see in adaptive leadership approaches built for conditions leaders haven’t faced before.
What that structure actually looks like, and how it’s built, is a bigger question than this post can answer. It’s the whole reason the next one exists.
What This Means If You Develop Leaders
If you’re responsible for a development budget, the Reversion Effect gives you a sharper set of questions for your next program review:
- Are we measuring change at program completion, or under pressure?
- Does the design put leaders in conditions where the old identity is likely to resurface?
- What happens between sessions, and for how long after the program ends?
- Are we developing the identity, or just decorating it?
This distinction is measurable. In our own practitioner data, programs built around this approach have shown 82% sustained behavior change measured under pressure, and 90% three-year retention against an industry baseline closer to 65%, a 25-point gap. (This is preliminary practitioner research and carries a self-selection caveat, but it points at what’s possible when development is built for real conditions instead of ideal ones.) It’s a pattern worth watching alongside what tends to happen to high performers who stall out and start looking elsewhere.
The Handoff
The manager in that first vignette didn’t forget the training. Somewhere along the way, they stopped being the person who used it.
Ibarra’s answer to stepping into a bigger leadership identity has three parts: redefine your job, redefine your network, redefine your leadership identity. The first two are well understood and widely practiced. The third is the one almost nobody knows how to build deliberately.
That’s the subject of the next post.
Does your leadership development survive the pressure test?
FAQ: Leadership Under Pressure
What is the Reversion Effect in leadership?
The Reversion Effect is the pattern in which leaders who have demonstrably changed their behavior during a development program return to their pre-program behavior when they encounter real organizational pressure. It happens because development that changes behavior without changing identity leaves the new behavior dependent on conditions. When pressure removes the capacity required to sustain it, the leader acts from the identity underneath, which, for a leader promoted for individual excellence, is usually the high performer they were before they led anyone.
Why do leaders revert to old habits under pressure?
Under pressure, a leader stops selecting a response and starts acting from self-concept. The common explanations (a skills gap, misaligned organizational systems, insufficient reinforcement) each describe a real contributing factor, but all three assume the leader is still choosing. Peer-reviewed research on leadership inconsistency found that leadership behavior varies measurably within the same person between routine and stressful conditions, and that the variation widens when the leader is strained.
What are the early signs a leader is reverting?
Three signals tend to appear before the pattern is obvious. Coaching conversations disappear first, because they’re the least identity-native behavior and the easiest to defer when the calendar tightens. Delegation quietly reverses, with the leader taking work back without consciously deciding to. And 360 feedback scores improve immediately after a program, then decline in the following cycle, often because the measurement window closes before real pressure arrives.
Does leadership training actually change behavior?
Yes, and that’s part of the problem. Leadership training reliably produces measurable behavior change, often quickly. What it doesn’t reliably produce is behavior that survives sustained organizational pressure. For anyone buying development, the distinction that matters is whether that change was measured at program completion or under the conditions that actually test it.
Can better systems or reinforcement prevent leaders from reverting?
They help, and both are worth doing. Aligned organizational systems reduce how often reversion happens, and reinforcement slows the decay of newly learned behavior. Neither eliminates the pattern. The evidence for that comes from what practitioners observe across organizations: the same reversion appears in companies with very different systems and very different levels of follow-up support, and the variance that remains tracks the individual leader rather than the environment.
Enjoyed this post? You might also like…
