Knowing the gap - intermediate information in tournaments

Abstract

Intermediate information is often available to competitors in dynamic tournaments.
\nWe develop two simple tournament models with two stages: one with intermediate information
\non subjects’ relative positions after the first stage, one without. In our
\nmodels, equilibrium behavior in both stages is not changed by intermediate information.
\nWe test our formal analysis using data from laboratory experiments. We find no
\ndifference between average first and second stage efforts. With intermediate information,
\nhowever, subjects adjust their effort to a higher extent. Subjects who lead tend
\nto lower their second stage effort, subjects who lag still try to win the tournament.
\nOverall, intermediate information does not endanger the effectiveness of rank-order
\ntournaments: incentives do neither break down nor does a rat race arise. We also
\nbriefly investigate costly intermediate information.