Essays on Lottery contests : Theory and Experiments

This thesis presents four studies on a variety of lottery contests. A contest is a game in which players incur irreversible expenditure of costly resources to win some valuable reward. The contest success function, which maps the resource investments into corresponding success probabilities, is the mathematical representation of the selection mechanism in a contest. This thesis focuses on lottery contests where chance plays a crucial role in determining the outcome. The first chapter, called “That’s the ticket: Explicit lottery randomisation and learning in Tullock contests” looks at how mere difference in representation can affect behaviour in an experimental lottery contest. A simple lottery contest with a single winner prize is implemented using two different frameworks one more abstract, following convention in majority of existing literature in this field, and one more operational, closely resembling the familiar institution of a lottery. Behaviour in the operational framework is closer to the theoretical prediction and reflects quicker adjustment towards best-responses. The findings have been replicated in two university locations in two continents. The second chapter, “A combinatorial multi-winner contest with package-designer preferences” provides a general framework for modelling a special type of contest that has been discretely studied in the existing literature. This is a highly stylized model of a multi-winner contest characterized by a group lottery contest among overlapping groups with perfect substitution of effort within a group. Best-responses are characterized for any connected network, and equilibrium properties have been discussed for certain types of networks. The third chapter, titled “Equivalent multi-winner contests: An experiment”, shows the strategic redundancy of three multi-winner contest mechanisms and proves outcome equivalence through laboratory implementation. This chapter also argues that such lottery-based multi-winner mechanisms generate similar contest investment as in a single-winner contest for suitable parameter values. The fourth chapter, “In-group and out-group motives in group conflicts: An experimental study”, employs a novel design to distinguish between different group-related attitudes in an inter-group conflict. Depending on the treatment, financial as well psychological consequences for in-group or out-group members are removed by matching current experimental subjects with the pre-recorded decisions made by previously participating subjects. This has been crossed with minimal identity conditions. No significant evidence of group bias is found. Observed behaviour shows higher support for conformity with others in decision-making. Overall, the findings from the three experiments indicate that decision-making patterns in lottery contests are largely unaffected by strategy-preserving procedural variations, but are responsive to representational differences