A rebate subsidy of rate sr is functionally equivalent to a matching subsidy of rate sm=sr/ (1-sr). Other things equal, an individual should respond identically to the two subsidies. We test the effect on charitable giving of the framing of a subsidy as a rebate or as a match. Subjects make a series of 'dictator' allocation decisions, dividing an endowment between themselves and their chosen charities. Allocation decisions vary by the endowment level, the net price of giving, and the form of the subsidy. We find that contributions are significantly higher with matching subsidies than with rebate subsidies.