Discrimination as favoritism: The private benefits and social costs of in-group favoritism in an experimental labor market
David Masclet  1, 2, *@  , David Dickinson  3@  , Emmanuel Peterle  4@  
1 : Université Rennes1  (CREM)
Faculté des sciences economiques 7place hoche 35065 Rennes -  France
2 : CREM Université Rennes1
université rennes1
faculté des sciences economiques 7 place hoche 35085 rennes -  France
3 : Appalachian state university
Economics Department Appalachian State University Boone, NC 28608 -  United States
4 : Université de franche comté
Université franche comt
Université de Franche-Comté 45 D Avenue de l'Observatoire 25030 Besançon Cedex -  France
* : Corresponding author

In this paper, we examine labor market favoritism in a unique laboratory experiment design that can shed light on both the private benefits and spillover costs of employer favoritism (or discrimination). Group identity is induced on subjects such that each laboratory « society » consists of eight individuals each belonging to one of two different identity groups. In some treatments randomly assigned employer-subjects give preference rankings of potential worker-subjects who would make effort choices that impact employer payoffs. Though it is common knowledge that group identity in this environment provides no special productivity information and cannot facilitate communication or otherwise lower costs for the employer, employers preferentially rank in-group members. In such instances, the unemployed workers are aware that an intentional preference ranking resulted in their unemployment. Unemployed workers are allowed to destroy resources in a final stage of the game, which is a simple measure of the spillover effects of favoritism in our design. Though we find evidence that favoritism may privately benefit a firm in terms of higher worker effort, the spillover costs that result highlight a reason to combat favoritism/discrimination. This result also identifies one potential micro-foundation of societal unrest that may link back to labor market opportunity.