Going easy on a misstep isn't automatically safe. Psychological safety peaked at moderate leniency.

One of your people has missed a commitment, or cut a corner, in a way the whole team can see. You have a choice about how much weight to put on it, and everyone is watching which way you lean. Leniency feels like the generous answer and, for many leaders, the safe one. A recent paper asks whether it is.

The paper is "Is leniency always divine? Nonlinear consequences of leader leniency for follower psychological safety and job performance," by Jiang Xu (Beijing Sport University) and Jingzhen Cao (China University of Petroleum, Beijing), published online October 15, 2026 in the Journal of Managerial Psychology. Here is how far I got: the publisher's page returned a 403 to me, so everything below comes from the published abstract as carried by Crossref. I have not read the full paper.

The abstract describes a time-lagged, multi-source field survey, analyzed with multilevel path modeling. It is a survey of people at work, not an experiment, so it can show patterns and not prove cause. The abstract I read does not give a sample size, the number of teams, or the industry. Those are the first things I would want before leaning on the result, and their absence from the abstract says nothing about whether the paper reports them.

The central finding, in the authors' words:

The results show an inverted U-shaped relationship between leader leniency and follower psychological safety, which is positively associated with job performance.Published abstract, via Crossref

Read plainly: followers felt safest under moderate leniency, and safety fell off when leniency went past that point. Psychological safety, in turn, was positively associated with job performance. The authors frame this as a "too-much-of-a-good-thing" effect, drawing on fairness theory.

The abstract adds a condition. The relationship varied with the severity of a third party's wrongdoing: "nonlinear at higher levels of severity but approximately positive and linear at lower levels." So when the misstep was minor, more leniency tracked with more safety. When it was serious, there was a point past which leniency stopped helping. The abstract's practical note is that leaders may benefit from aligning leniency with wrongdoing severity when responding to misconduct.

Why this lands differently at the top. Everyone in the building reads how the person at the top handles a lapse, and reads it as a statement about the standard. A manager who goes easy on one person is making a local call. The most senior person going easy is setting a rule for everyone who is watching, including the people who did the work correctly. That is where fairness theory has traction: the people who were not involved still form a view about whether the response matched what happened. The paper's setting is one where a third party's wrongdoing is the trigger, which is the position most observers are in.

The part the research doesn't say, but that I'll say plainly: I would not read this as a case for being harder. A survey finding that safety peaks at moderate leniency does not tell you where your own team's peak is, and I can't tell from the abstract how leniency was measured or whether it was rated by followers or by someone else. My reading is that the useful variable is proportion. The mistake the curve warns about is a response that is the same size no matter what happened. Those cautions are mine, not the authors'.

One thing to do this week. Think of the last two times someone on your team visibly fell short, and write down, for each, how serious it was and how you responded. Then ask whether a person who watched both could have worked out the rule you were following. If they could not, say the rule out loud at your next team meeting, before the next lapse and not after it.

How you calibrate a response is a pattern in how you lead, and it is hard to see from the inside. The Leadership Profile is built to make those patterns visible; you can start at /wiring-map.