The U.S. government provides large subsidies for investment in renewable energy. I consider the differential impact of these subsidies based on the form in which the government awards them.  Most federal subsidies for renewable energy are in the form of non-refundable tax credits.  Wind project developers frequently receive more in subsidies than they owe in taxes, and this prevents them from using the tax credits themselves.  Instead, they use special financing arrangements to effectively sell the credits to investors who can use them. I study what fraction of the subsidy is transferred to outside investors or lost to transaction costs in these arrangements.  From 2009-2012, wind developers chose between a non-refundable tax credit subsidy and a different subsidy in the form of a grant.  The size of the tax credit subsidy was based on production, and the size of the grant was based on project cost.  I exploit geographic variation in productivity and cost across projects to estimate a model of subsidy choice, and infer that wind developers value these tax credits at $0.86 on the dollar. Using a back-of-the-envelope calculation, I show that moving to subsidies in the form of grants would have increased investment by up to ten percent.  [back]