Last week I attended an event on software patents at GW Law School. The event made me uncomfortable because it was—as one would expect at a law school event—dominated by lawyers. The concerns of the legal academics, practitioners, and lobbyists participating in the round table discussion were very different from those one would expect for a policy audience. For example, the participants agreed that there is no elegant way to partition software patents from other patents under current law and that current Supreme Court jurisprudence is unsophisticated, relying on the wrong sections of the U.S. Code.
Missing from the discussion was the single most important fact about patents: that they are negatively correlated with economic growth.
Patent grants have exploded in the past two decades or so, and real GDP per capita growth has declined over the same period. Now, patent proponents can argue (rightly) that correlation is not causation—growth could have been even worse over the past few decades had we not had strong patent protection. But correlation is correlated with causation, so proponents of strong patent laws should have to explicitly make that argument using real evidence.
In addition to U.S. time-series data, we can examine the international cross-sectional evidence. As Petra Moser concludes in her recent JEP article:
Overall, the weight of the existing historical evidence suggests that patent policies, which grant strong intellectual property rights to early generations of inventors, may discourage innovation. On the contrary, policies that encourage the diffusion of ideas and modify patent laws to facilitate entry and encourage competition may be an effective mechanism to encourage innovation.
Taken together, absent some additional evidence from patent proponents, this time-series and cross-sectional evidence suggests we are on the wrong side of the Tabarrok Curve.
If there is any evidence that software patents in particular have a positive effect on innovation or growth, I have yet to see it. Here’s hoping proponents of the current system will take up the challenge and respond with such evidence. But if they do not, then we should abolish software patents even if it means adopting some relatively bizarre legal formulations as the lawyers fear.