In this paper, we examine the propriety of using patents as a proxy for firm innovative performance. We find that in the pharmaceutical industry during 1985-2001 they are not. Replicating the methodology in a study by Comanor and Scherer (1969) upon which many current studies base their use of patents as proxy, we demonstrate simple correlations of 0.1911 between lagged-patent filings and actual new product introductions, and partial correlations of 0.1785 when holding firm size constant. Our results raise serious questions about the usefulness of employing patent measures in contemporary studies that examine the innovative performance of firms.