Data for: Insurance and Solidarity. Experimental Evidence from Cambodia
2020-05-02T06:41:59Z (GMT) by
We conducted a lab-in-the-field experiment in rural villages in Northwestern Cambodia, complemented by a detailed household survey, in order to study whether solidarity is conditioned on the availability of insurance. The game played in the experiment is a modified version of the dictator game. It is one-shot and is played anonymously. In the experiment, subjects are randomly assigned the role of provider or recipient. All receive the same endowment. Recipients can lose a large proportion of their endowment due to a random idiosyncratic shock. Each provider is anonymously matched with one recipient who she can transfer to in case the recipient experiences a loss. We apply a 2x2 design. First, we vary whether recipients have the option to purchase insurance that covers the loss from the shock and thus renders a transfer from the provider unnecessary. This allows us to test whether providers transfer less to a recipient who had the option to purchase insurance. Second, we vary whether recipients are informed that they are matched with a provider who can transfer to them in case of a loss. We can thereby test whether providers reduce their transfers by less when the decision against insurance was made by recipients who were uninformed about the provider compared to when the decision was made by recipients who were informed. The design of our study allows us to relate the subjects' behavior in the experiment to their real-life experience, both in terms of informal support and exposure to insurance. We elicited a rich set of support network characteristics as part of the household survey that preceded the experiment. This allows us to test whether conditional solidarity is more pronounced among subjects who are more engaged in informal support. Furthermore, we selected our sample such that around half of the subjects had access to health insurance, due to their place of residence, while for the other half health insurance was unavailable. We can thereby relate insurance familiarity to a subject's propensity to condition support on the insurance option in the experiment.