In IP scholarship, patents are commonly understood as more efficient than other approaches to innovation policy. Their primary ostensible advantage is allocative: as a form of property rights, patents act as a conduit between market signals and potential innovators, ostensibly guiding investment toward inventions with the most social value. Existing accounts recognize that, in practice, signals of social value that patents facilitate may be attenuated because of, for example, transaction costs and limits on the scope and length of patent rights. We show here, however, a different problem with the conventional allocative account. The appropriability mechanism patents rely on, namely excludability, operates in asymmetrical ways for different kinds of information goods. While scholars have noted that patent systems fail to create goods whose value is difficult to appropriate in consumer markets, this fact has not been fully appreciated in the literature, nor have its implications for the standard justification for patents. Through detailed examples in the health context we show that some kinds of information goods will be much more difficult to exclude than others. Importantly, there is no reason to expect that the ease of exclusion will be correlated with social value. The analytic point that emerges is generalizable: patents themselves can have distortive effects, stemming from structural features of exclusion rights. Unlike the problem of attenuation, the problem of asymmetric nonexcludability cannot be resolved by increasing patent scope or length. Because excludability is variable along a continuum, property rights in information, even if formally perfected, and even assuming away conventional transaction costs, will create asymmetrical demand for different kinds of information goods. This argument provides an important new justification for alternatives to patents such as government funding and gives us new insights about how to allocate such funding. It also reinforces the need for a comparative institutional approach to innovation policy, and for incorporating into our debates currently unrecognized implications that patents may have for values such as privacy and free speech.