It is shown that spillovers can enhance private returns to innovation if they feed back into the dynamic research of the original inventor (Internalized spillovers), but will always reduce private returns, if the original inventor does not benefit from the advancements other inventors build into the “spilled” knowledge (Externalized spillovers). I empirically identify unique patterns of knowledge flows (based on patent citations), which provide information about whether “spilled” knowledge is reabsorbed by its inventor. A simple model of sequential innovation with dynamic spillovers is developed, which predicts that market value and R&D expenditures should rise with Internalized spillovers and fall with Externalized spillovers. These predications are confirmed using panel data on U.S. firms between 1981 and 2001. To the extent that firms internalize some of the spillovers they create, the classical underinvestment problem in R&D will be mitigated and the central role of spillovers in promoting economic growth will be enhanced.