Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Charitable giving increasingly requires recipients to undertake costly prior action. A common justification is that willingness to undertake costly actions signals greater productivity from transfers. We demonstrate that, if the distribution of recipient types is unknown, recipient costs indivisible and productivity unobservable, conditional charity, once instituted, may not yield information adequate to refute its efficiency claim. Consequently, donors who inefficiently provide conditional charity will not correct themselves. Donors who wrongly provide unconditional charity may however subsequently correct themselves. This offers grounds for scepticism regarding efficiency claims for conditional charity. We also provide reasons for encouraging donor competition and indicator targeting. Copyright 2007 , Oxford University Press.