Stalin stayed in Moscow during WWII to lead resistance against the approaching German armed forces.  He was putting his own life at risk.  Why do leaders have to take costly actions to persuade followers to follow?

A seminal paper Leading by Example (jstor link) by Ben Hermalin (AER 1998) has an answer based on a classical signaling model.  Suppose a leader wants followers to exert costly effort.  The ideal level effort depends on some factor known only to the leader.  Stalin might have a better idea as to the chances of success against the Germans; or a C.E.O. might have better information about the state of demand.  What if the leader simply tells the followers his information?  Of course, this is Cheap Talk. If the leader wants the followers to exert high effort whatever the true state of affairs, the followers cannot believe anything the leader says.  For example, Stalin might tell his troops to fight hard whether they have a good chance of beating the Germans or not. Communication breaks down and the followers’ effort cannot be a function of the truth.

The picture changes if the leader can himself exert costly effort and lead by example.  Then, he might only be willing to work hard if and only if he thinks the chances of his effort paying off are good.  Stalin’s decision to stay in Moscow signals that he believes that his life is likely to be safe as the chances for Soviet success are good.  If the Germans have the advantage, he does truly risk his life and would prefer to leave Moscow.  As the leader’s incentives to send the signal or not depend on the true state of affairs, his “message” is credible.  The followers can then fight hard or work hard if and only if the leader leads by example and works hard himself.

“Actions speak louder than words” and Costly Signaling is more informative than Cheap Talk.  Hence leaders must lead by example because words cannot be believed while actions can.  This is the heart of Hermalin’s idea.

Once you buy into the framework, it is easy to think of variations.  Hermalin does not talk explicitly about competition but let’s add that in.  If things are going well for the organization in the product market, then the competition does not matter – you have a good product so you can slack off.  More importantly the leader slacks off and the followers do too.  If the organization is defunct if the crew does not pull the oars hard, the captain has a good incentive to pitch in and row.  So, this is similar to the first Hermalin story except the leader works hard when the going is bad and effort is very important and not when the going is good.  The followers do the same.

Another variation: Middle managers looking to get promoted have weird incentives.  They want to signal how hard-working they are to their superiors.  They want to work hard even when the going is bad.  So, the followers cannot filter out the true state from the middle manager’s effort.  The followers will slack off.  The middle manager cannot be a credible leader by example as signaling to his superiors destroys the credibility of the signal to his juniors.  This is a good reason to appoint someone who does not want to be senior manager into the middle position.  Hence, looking at the leading by example issue alone, academic department should appoint Chairs who do not have any ambition to rise further in the hierarchy.  Many other examples can be given…