Evaluating Dilemmas

I recently read a large number of letters of reference for candidates applying for the position of Assistant Professor of Economics at Smith College. Most of the reference letters were very good. Indeed, they seemed, on average, a little too good. A significant number of reference writers said that the job candidate was one of the best he had ever worked with. I suspect that reference writers exaggerate the quality of the person about whom they are writing. Actually, I have written many letters of recommendation for students, and I know that I exaggerate their good qualities. Letters of reference play an important part in our economy by helping employers match positions with people, so great harm is caused when reference writers are not completely candid. I don’t feel bad, however, about writing letters that are a little more glowing than they should be, because prisoners’ dilemma makes me do it.

When you are writing a letter of recommendation you probably have two goals: You want to give an honest evaluation of the candidate to a potential employer, and you desire to help the candidate. Since one of these goals is best filled by being truthful and the other by exaggeration, on average, recommendation writers should embellish their recommendations at least a little. Of course, a problem arises when everyone plays this game. If everybody exaggerates, then when a reference writer indicates that his candidate is average, it will be taken to mean that the candidate is below average.

Now imagine that you have an average candidate, and you want to write her a letter that causes employers to think a little bit more highly of her than they should. Because everyone exaggerates, it’s now not enough to say she is above average; you might have to label her as, say, extremely good. Alas, if all reference writers follow this trend, then even saying someone is extremely good does not signal quality. To help a candidate who truly is above average a reference writer is almost forced to say he is one of the best ever. Of course, all of this exaggeration causes recommendations to be taken less seriously. This is where the prisoners’ dilemma manifests since all reference writers are worse off because they all exaggerate, even though it is in each reference writer’s self-interest to exaggerate. Since we all exaggerate, the effects of exaggeration cancel out. Unfortunately, this still reduces the value of recommendations, so all reference writers would benefit from some binding agreement forbidding any embellishment. If, however, everyone else stopped exaggerating, then I could greatly benefit by being the only one to deceive, since such deception would then not be suspected.

This evaluating dilemma increases the importance of networking and personal connections. A letter of recommendation I write to a friend should be given more weight than one I sent to a stranger—not, obviously, because my friends know me to be particularly truthful, but rather because I have a greater incentive to be honest with my friends. I will probably never need anything from the strangers who see my recommendations, so I most likely wouldn’t be hurt if they find out I had deceived them. A friend would probably be annoyed with me if I recommended a student whose performance was below what I promised. Since I don’t want my friends to dislike me, I tend not to lie to them when I know that such lies will be detected.

When companies require internal reviews of employees by supervisors, they can also face evaluating dilemmas. The supervisors will want to be perceived as honest and effective evaluators. They will also, however, probably want to be considered nice people to work for. Thus, supervisors have some motivation to be honest and some incentives to exaggerate, causing them on average to be a little too kind when evaluating. To stop this evaluating dilemma, companies often require supervisors to rank employees. By mandating ranking, companies prevent evaluators from labeling everyone as above average.