I’m always happy to see prominent publications calling for patent reform, but I found this article from the Economist a little frustrating. They seem to clearly understand that the experiment with allowing patents on software and “business methods” are a big part of what’s wrong with the patent system. To me, the obvious conclusion is that the courts made a mistake in allowing those types of patents. But the Economist disagrees. In addition to making the bar for getting such patents “much higher” (in unspecified ways), it calls for the creation of a new, shorter patent category for software. This would be an improvement, to be sure, but only because the optimal length for a software patent is 0 years.

But shorter terms for software patents would at least be a step in the right direction. My stronger objection is to the Economist’s final recommendation:

Patent cases should be heard by specialized courts (as happens in other areas of law), rather than non-expert juries in advantageous jurisdictions in Texas. That would make life harder for trolls.

This gets it wrong in several ways. For starters, those “advantageous jurisdictions in Texas” have become de facto specialist patent courts, since they have hear a disproportionate share of the nation’s patent cases. So it’s not clear why further specialization at the trial court level would improve matters.

But the more important point is that we already have a specialist patent court at the appellate level. The creation of this court, called the US Court of Appeals for the Federal Circuit, has been a disaster.

When properly limited, patents promote innovation by rewarding inventors for their efforts. But too much patent protection can be counterproductive. Broad or obvious patents work as an innovation tax, forcing genuine innovators to pay tolls to people who made only trivial improvements before them.

The courts are charged with preserving this balance, ensuring that the system rewards past innovation without unduly burdening future innovators. Congress establishes the system’s broad parameters, but the details are fleshed out in common-law fashion by the courts. Until 1982, responsibility for developing this body of law was divided among 13 appeals courts, 12 of which had generalist judges who spent most of their time ruling on other types of cases.

In 1982, Congress became concerned that patent law was too complex for generalist judges. So it created a new court, the United States Court of Appeals for the Federal Circuit, and gave it exclusive jurisdiction to hear appeals in patent cases.

This had the unintended consequence of dramatically increasing the influence of the patent bar over patent law. Not only do Federal Circuit judges spend all their time hearing arguments from patent attorneys, but some of them are former patent attorneys themselves. In its first two decades, the Federal Circuit gradually shifted patent law in the pro-patent direction favored by most patent attorneys. Patents became easier to get and harder to invalidate. The courts allowed tougher punishments against infringers. And the Federal Circuit unilaterally eliminated traditional limits on patenting software and “business methods.”

The creation of the Federal Circuit had another unintended consequence, too. The Supreme Court relies on disagreements among appeals courts — known as “circuit splits” — to help it figure out which issues require its attention. And when the Supreme Court takes a case, the existence of multiple, conflicting precedents gives the justices more raw material from which to fashion their own decisions.

But when the Federal Circuit became the only court ruling on patent cases, there were no more circuit splits and no more competing legal precedents. That might be why the Supreme Court seems to have barely noticed that the Federal Circuit was dramatically reshaping patent law in the 1990s. The high court reviewed only about a dozen Federal Circuit decisions between 1982 and 2004, and the ones it did review tended to be on narrow, technical issues. The Supreme Court finally began to give the Federal Circuit’s handiwork some serious scrutiny when Chief Justice John Roberts took the bench. And the justices did not like what they saw. In the Chief Justice’s first three terms, the high court heard five different patent cases, and all of them resulted in unanimous or near-unanimous reversals of pro-patent decisions by the Federal Circuit.

But a lot of the damage had already been done. Hundreds of thousands of low-quality patents had been approved under the permissive rules the Federal Circuit had developed during the 1990s. Those patents may be technically invalid under recent Supreme Court decisions, but that’s of little help to a small company that can’t afford to litigate the question.

So contra the Economist, a key element of any serious patent reform agenda should be to decentralize authority over patent appeals. The regular appeals court handle all manner of complex subjects — antitrust, copyright, commercial litigation, and so forth. Patent law is complicated, but it’s not that complicated. The other appeals courts don’t have the Federal Circuit’s in-depth legal expertise, but it also doesn’t have the Federal Circuit’s incestuous relationship with the patent bar. And inter-circuit competition is essential for our common-law system to work well.