Does One Contribution Come at the Expense of Another? Empirical Evidence on Substitution Between Charitable Donations

In this paper I estimate and describe the extent to which an individual's charitable donation to one cause displaces his or her giving to another cause. I use the 2001 and 2003 waves of the Panel Study of Income Dynamics (PSID), in conjunction with the Center on Philanthropy Panel Study (COPPS). This is the first useful major source of panel data on giving to multiple causes. I control for individual-fixed effects and use "college-reunion year" as an instrument for giving to education. I find an economically and statistically significant level of substitution. I also analyze the net effect of shocks to giving to one category on giving to all other categories - testing the extremes of a fixed purse (perfect crowding-out) and zero-crowding-out. While the uninstrumented regressions can generally reject perfect crowding out, the instrumental results (with larger error bounds) do not. I also find a greater level of substitution for "large givers" than for those who make smaller donations. This points to a model with heterogenous motivations for giving: small givers may be driven by shocks and reputation concerns, while for larger givers charities are imperfect substitutes in providing "warm glow" utility.