If innovations are rapidly made obsolete by subsequent discoveries, firms may have lower ex ante incentives to invest in R&D. This article empirically demonstrates the relevance of this problem and shows that it might be mitigated if the inventing firm reabsorbs its 'spilled' knowledge in its later inventions. Using new data on sequences of patent citations, I estimate the relationship between a firm's stock market value and the citations it receives. Citations on which the firm builds in a future period are positively related to market value, whereas citations on which the firm does not build are negatively related to value. © 2011 The Author(s). The Economic Journal © 2011 Royal Economic Society.