Abstract:
A novel two-person "charity game" is used to experimentally investigate whether anticipation of help crowds out incentives to work, and therefore impulses to help. We distinguish two treatments differing in whether the causes of neediness are verifiable or not. Helping behavior does not vary significantly between treatments, but is positively correlated with dictator giving, suggesting idiosyncratic attitudes to help. Needy subjects are unaffected by anticipated help, but react optimally to chance.