Property Rights and Loss Aversion in Contests

Research output: Contribution to journalArticle

Open Access permissions





  • Subhasish M. Chowdhury
  • Joo Young Jeon
  • Abhijit Ramalingam

Organisational units


We analyze the effects of property rights and the resulting loss aversion on contest outcomes. We study three situations: in ‘gain’ two players start with no prize and make sunk bids to win a prize; in ‘loss’ both the players start with prizes and whoever loses the contest loses their prize; and in ‘mixed’ only one player starts with a prize that stays with him if he wins, but is transferred to the rival otherwise. Since the differences among the treatments arise only from framing, the expected utility and the standard loss aversion models predict no difference in bids across treatments. We introduce a loss aversion model in which the property rights are made salient, and as a result the reference point varies across treatments. This model predicts average bids in descending order in the loss, the mixed, and the gain treatment; and higher bids by the player with property rights in the mixed treatment. The results from a laboratory experiment broadly support these predictions. There is no significant difference in bids in the loss (gain) treatment and bids by property rights holder (non-holder) in the mixed treatment. A model incorporating both loss aversion and social preferences explains this result.


Original languageEnglish
Pages (from-to)1492-1511
JournalEconomic Inquiry
Issue number3
Early online date17 Oct 2017
Publication statusPublished - Jul 2018


    Research areas

  • Contest, Experiment, Framing, Property-rights, Loss Aversion

View graph of relations

ID: 115730065

Related by journal
  1. Quantifying the deterrent effect of anticartel enforcement

    Research output: Contribution to journalArticle

  2. Forecasting with Social Media: Evidence from Tweets on Soccer Matches

    Research output: Contribution to journalArticle

  3. Fiscal rules and government borrowing costs: international evidence

    Research output: Contribution to journalArticle