but the relationship between ex ante winning chances and effort incentives is qualitatively
similar in rank-order tournaments where observed performances rather than perceived
abilities determine payoffs (Lazear and Rosen (1981), Meyer (1992), Baik (1994)), or if a
single agent competed against a fixed standard.
The main difference with respect to tournaments based on observed performances is
that the accuracy of the nomination committee’s information about a candidate’s ability
at the beginning of the selection process may also affect her incentives now. As we
show, the predicted relation between a candidate’s optimal effort and the precision of
the information the decision-maker has about her is non-monotonic. If the committee
already has a very clear idea of the manager’s qualification for the post, incentive are
weak because any change in her performance is unlikely to affect the final decision.5
Moreover, a candidate’s incentives will also be weak when only little is know about her
ability: due to the high variance of beliefs in this case, it is unlikely that the perceived
ability of the candidate will lie close to that of one of the other candidates at the end of
the nomination period. Thus, an increase in effort is unlikely to affect the final decision
even if the agent’s observed current performance could have a strong effect on perceptions
about her ability.
Similar incentives arise in other situations where several applicants compete for a
limited number of positions, and selection is based on perceived skill or talent, in particular
other hiring or promotion procedures,6 or nominations of election candidates by political
parties.
This paper provides a theory of agents’ incentives in such situations and tests the
predictions of this theory versus alternative theories by means of a quasi-experiment in
professional soccer. When a nation qualifies for a major international tournament, such as
the Soccer Euro Cup, the national coach is responsible for selecting the players who will
be on the national team that plays in the Cup.7 Our theory then predicts that players who
decision-maker’s information about all agents is identical.
5The intuition for this result goes back to the career concerns literature, building on Holmstrom (1982),
where uncertainty about the agent’s skill is necessary for him or her to find it profitable to exert effort.
6 In firms, an interesting tension may arise between two possible goals of hiring and promotion systems:
motivating internal agents by the prospect of a future promotion, and selecting the most promising
candidate out of a pool of both internal and external candidates. Chan (1996) analyzes this conflict. In
the example we consider no such tension arises, since the principal cannot contract with the agents prior
to the nomination decision.
7 National team composition is flexible in friendly matches between nations or qualification matches
for international Cups, but not for Cups.