Reverse Contests

Working Paper: CEPR ID: DP14411

Authors: Aner Sela

Abstract: We study two reverse contests, A and B, with two agents, each of whom has both a linear reward function that increases in the agent's effort and an effort constraint. However, since the effort (output) of the agents has a negative effect on society, if the agents' effort constraints are relatively high, the designer in reverse contest A imposes a punishment such that the agent with the highest effort who caused the greatest damage is punished. Conversely, if the agents' effort constraints are relatively low, in reverse contest B, the designer awards a prize to the agent with the lowest effort who caused the smallest damage. We analyze the behavior of both symmetric and asymmetric agents in both contests A and B. In equilibrium, independent of the levels of the agents' effort constraints, both agents are active and they have positive expected payoffs. Furthermore, the agents might have the same expected payoff regardless of their asymmetric values of the prize/punishment or their asymmetric effort constraints.

Keywords: No keywords provided

JEL Codes: D44; D72; D82


Causal Claims Network Graph

Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.


Causal Claims

CauseEffect
High effort constraints (D20)Overall effort levels (D29)
Punishment on agent with highest effort (C73)Overall effort levels (D29)
Reward for lowest effort (J33)Lower effort levels (D29)
Reward for lower efforts (J33)Positive expected payoffs in contest B (C72)
Asymmetry in agents' reward functions and effort constraints (D82)Expected payoffs (G19)
Different contest designs (C90)Agent behavior (L85)

Back to index