Abstract
The intuition behind this study is that those who have a history of building a welfare state (henceforth called natives) may form subjective entitlements to this welfare state. This perception would be observable in their reluctance to share their welfare with those who have no history of building it (henceforth called immigrants) even when not sharing is costly (i.e., decreases efficiency).
To test this intuition, we design an experiment in which subjects are paired and proceeds symbolizing the welfare are to be divided between them. To create a situation analogous to building a welfare state, one member of a pair is randomly selected as the one responsible for creating proceeds to be divided, holding productivity constant between parties. That is, credit for having created the proceeds is determined through a random assignment, which is public knowledge for everyone.
Orthogonally, we also systematically pair subjects based on whether or not they have real-life migration backgrounds.
That is, in half of the pairs (called homo pairs), partners are either both migrants or were both born in the UK, while in the other half of the treatments (called hetero pairs), pairs are mixed with respect to their migration backgrounds (i.e., a migrant is paired with a UK born
subject).
Additionally, we collect unbiased third parties' views about the fair division of resources between parties.
This setup allows us to test whether beliefs about the fair division of the proceeds diverge more between parties in the homo than in the hetero pairs, and also whether those named responsible for creating the to-be-divided resources would believe they deserve and also claim a higher share than those who are not responsible for creating it. Additionally, we test whether these two channels interact. Finally, we are able to explore unbiased third parties' views about the fair solution.
Finally, we also control for the heterogeneity of subjects’ pre-treatment attitudes towards (1) immigration to and immigrants in the UK, (2) Brexit votes,(3) redistribution in general, and (4) efficiency concerns. This allows us to investigate whether treatment effects vary according to subjects’ characteristics along these aforementioned dimensions.