Karl Wärneryd

We consider two-player, perfectly discriminatory, common-value contests (or all-pay auctions), in which one player knows the value of the contested object with certainty, and the other knows only its prior distribution. We show, among other things, that in equilibrium the players win with equal probability. This contrasts with a large class of imperfectly discriminatory contests in which the uninformed player wins with a strictly greater probability than the informed player.

We consider imperfectly discriminating, common-value, all-pay auctions (or contests) in which some players know the value of the prize, others do not. We show that if the prize is always of positive value, then all players are active in equilibrium. If the prize is of value zero with positive probability, then there is some threshold number of informed players such that if there are less, all uninformed players are active, and otherwise all uninformed players are inactive.

Hierarchy can function as an instrument to channel influence activities or power struggles in organizations. Contrary to what has frequently been argued, we show that multi-divisional organizations may involve lower influence costs than single-tier organizations, even though they offer more scope for organizational conflict and have more executives that can be influenced. These benefits derive from two effects. First, part of the conflict in multi-divisional organizations takes place on the division level, where a small number of agents fight over only a fraction of the overall prize. Second, by grouping agents into common divisions, multi-divisional organizations create free-rider problems in rent-seeking. Our model sheds new light on the desirability of divestitures and the transition from the U- to the M-form by US corporations in the 1920s.

We study contests where the set of players is a random variable. If it is known for certain that there will be at least one participant, then aggregate contest expenditure in equilibrium is strictly lower in a contest with population uncertainty than in a non-uncertain contest with the same expected number of players. This suggests an explanation of, for example, why empirical studies show rent-seeking expenditures to be much lower than predicted by other theories.