This paper analyzes the efficacy of the Universal Service Fund's High-Cost Support programs in achieving two of its formal aims—increased availability and comparable or improved service quality. In light of a narrowed rural digital divide, increased facilities-based competition, and significant disparities in support among US states, we separately analyze subprograms active in the 48 contiguous states and Alaska between July 2014 and December 2017. In the 48 contiguous states, we find that Phase II of the Connect America Cost Model (CAF II) support predicts aDSL availability and increased aDSL speeds. Meanwhile, we find evidence that the Alternative Connect America Model (ACAM) may be effective in inducing new fiber deployment, but no evidence that ACAM support affects fiber speeds. With the exception of the Rural Broadband Experiments (RBE), we find that High-Cost Support has no relationship or is negatively related with cable speeds and availability, but we cannot infer if this negative relationship is attributable to “crowding out” by subsidized incumbents or program selection effects. In Alaska, we find that Census blocks targeted by Alaska Plan support have greater availability and faster speeds from aDSL and fiber carriers. Heterogeneity in our results highlight the importance of evaluating High-Cost subprograms separately when analyzing the efficacy of High-Cost programs in achieving universal service objectives.