The Leader in the Glass Office: Why Promoting Your Best Worker Can Break Team Cooperation

A short summary of Hattori, K. (2026) “Who Disciplines the Leader? Leadership as Exposure”, ZBW - Leibniz Information Centre for Economics, Kiel, Hamburg.

Background

The word leadership brings to mind a leader guiding followers, or a leader monitoring how hard they work. Leaders set direction, motivate members, communicate information, supervise, and hold the team together. Leadership, on this understanding, is a device through which a leader influences other people.

This paper turns that view around. Is the leader really only on the disciplining side?

Being a leader is rather like working in a glass office. Your conduct is in plain view of the team, and the followers decide what to do after seeing it. If so, taking the leader’s role may expose the leader’s own conduct to the followers’ observation and immediate response, and thereby discipline the leader herself.

Consider a long-running joint project. The leader works first, and the followers decide whether to work after watching her. If the leader shirks, the followers can see it and respond at once: then I won’t work either. The leader’s shirking is recognized immediately and punished by a refusal to reciprocate. The followers, on the other hand, can shirk after seeing that the leader has already worked, so they can free ride on the leader’s sunk effort. Punishment from the leader, that is, the end of cooperation, can only be delivered in the future.

What the paper focuses on is this asymmetry in the timing of punishment. The leader’s action is behind glass before the followers move, and shirking is punished on the spot. The followers can free ride on effort the leader has already put in, and their shirking is deterred only by future punishment. Leadership is not merely a position from which one moves other people. It is an institutional position in which one’s own conduct is exposed to others’ immediate response, a glass office. That is where its work as a disciplinary device lies.

What the paper does

Consider a repeated game played by a two-member team. In each period each member chooses whether to work or to shirk. Working raises team output but carries a private cost. Members differ along two dimensions. One is ability, how large a contribution the member generates when she works. The other is patience, how much weight she places on the future of the cooperative relationship, or her long-term orientation.

Two team structures are compared. In a horizontal team, the two members choose effort simultaneously. In a vertical team, the leader moves first and the follower moves after observing her.

In a horizontal team, each member is tempted to shirk while the partner works, and that temptation can be held in check only by the threat of future punishment, namely the collapse of cooperation. In a vertical team, if the leader shirks, the follower immediately stops working, so the leader’s temptation is checked on the spot. What remains is the follower’s temptation to shirk after seeing that the leader has worked.

The question of who should lead therefore ceases to be a question of who should be senior. It becomes a question of whose incentive constraint to remove, that is, of whom to put in the glass office. Leadership is a mechanism for allocating enforceability inside the team.

What it finds

Make the myopic member the leader

The first result is counterintuitive: cooperation is easier to sustain when the less patient member leads.

A member with a long horizon values the future of the relationship and is therefore disciplined reasonably well by future punishment. Making her the leader spends the disciplinary power of the role on someone who was already easy to discipline. It is better to place in the leader’s role the member on whom future punishment has the weakest grip, and thereby expose her conduct to the follower’s immediate response. That is the basic logic of leadership as discipline.

The able member works better as a follower

The second result is that the more able member should often be the follower.

This is not to say that able people make poor leaders. On the contrary, it is precisely because she is able that placing her in the follower’s role has a strong disciplinary effect. The follower carries the only incentive constraint left in the team, and an able follower who shirks forfeits her own large contribution, so her temptation to shirk is smaller.

Conversely, when the able member leads, the follower can free ride on a large contribution that has already been made, which strengthens the follower’s temptation to shirk. Isolating the disciplinary effect of leadership, then, an able member can be worth more as a follower than as a leader.

“Able but myopic” depends on the task

In reality, ability and long-term orientation can pull in opposite directions. Abler people often have more outside offers and better job opportunities, so they are less likely to remain in the current team for long, which effectively makes them myopic. High ability then calls for placing this person in the follower’s role, while a short horizon calls for placing the same person in the leader’s role.

What settles this tradeoff is the nature of the task. When the two members’ efforts mesh together, that is, under complementary tasks, the value of keeping the able member as a follower is large, because her working after the leader has worked produces a strong joint effect. When one member’s effort substitutes for the other’s, that advantage weakens, and it becomes more attractive to place the able but impatient member in the leader’s role and discipline her there.

But if leaders change people, the conclusion changes

Everything so far isolates one aspect of leadership: that the leader moves first and is observed. Real leaders, however, influence their followers. How they work, how they think about the future, and how committed they are to the goal can transmit from leader to follower.

When that influence is strong enough, the conclusion reverses. Making the able, future-oriented member the leader lets those traits spill over to the follower, so the conventional assignment, in which the able and patient member leads, becomes the better one. The paper is therefore not arguing that able people should not be made leaders. Where a leader’s power to change her followers is weak, or where leader and follower interact too little to influence one another, leadership as discipline matters; where it is strong, leadership as influence matters.

Why it matters

Implications for promotion

Organizations often promote their most able member into the leader’s role. This is a natural rule, but the paper shows that it can undermine cooperation. The problem is not that the promoted person lacks managerial skill. The problem is that promoting the able member removes her from the follower’s role, the one position where discipline still operates.

This gives a different reading of the Peter Principle. Promotions of strong performers may fail not because those people are incompetent as leaders, but because the incentive structure inside the team changes and the arrangement that was sustaining cooperation falls apart.

Implications for team design

The disciplinary effect of leadership does not come from a title. What matters is how clear the glass around the leader is. Where conduct is fully visible, through clear glass, discipline is strongest; and even through frosted glass, where only the outline shows, discipline still works, since perfect observation is not required and a certain degree of transparency suffices. But if the glass becomes a wall, so that conduct cannot be seen at all, discipline is lost however formal the hierarchy, and the team reverts to being horizontal in substance.

Supervision and direction are also substitutes for the disciplinary effect of leadership. Where followers will not cooperate voluntarily, an organization can pay the cost of supervision to draw out cooperation. But when supervising is difficult and expensive, placing the hard-to-discipline member in the leader’s role can solve the same problem more cheaply.

In closing

People often observe that someone becomes noticeably more responsible on becoming a leader, and this is usually read in terms of a sense of responsibility or of personal growth. This paper offers another interpretation. A leader may change not because holding the role reshapes her character from within, but because her conduct is exposed to other people’s immediate response. Leadership is a position from which one moves others, and at the same time a position in which one’s own conduct is made enforceable.

This research was supported by JSPS KAKENHI Grant Number 24K04912.