Since 1979 more than thirty states have adopted "performance funding" for public institutions of higher education. Under performance funding, a portion of the state appropriations for each institution is determined by the institution's achievement of performance goals on such metrics as retention and graduation. We argue that several characteristics of higher education institutions are likely to weaken the effect of performance incentives on graduation rates. To test these expectations, we develop a comprehensive database that identifies the institutions subject to graduation performance metrics. While most previous researchers have coded each state with a simple binary measure, indicating whether performance funding exists or does not exist in each year, we have determined for which exact institutions and in which years graduation metrics have been used to allocate state appropriations. We combine this detailed performance-funding data with institution-level data on graduation rates and other important factors from the Integrated Postsecondary Education Data System from 1997-98 through 2015-16. We estimate a difference-in-differences model that reveals no significant impact of performance funding on the graduation rates at 4-year institutions, but we find that performance funding is associated with a significant increase in graduation rates at 2-year institutions under certain conditions.