The Center for American Progress’s Matt Yglesias emailed to tell me about the latest issue of Science Progress, CAP’s science journal, which includes an in-depth series of articles on patent reform. The article that particularly caught my eye was this piece on patent trolls. In it, Daniel P. McCurdy discusses the controversy over firms whose sole reason for existence is to acquire patents and then use the threat of litigation to extract licensing revenues. The classic patent troll has no products and no employees other than the lawyers required to negotiate licensing deals and file patent lawsuits.

Patent trolls are a real concern. Most famously, in 2006 Research in Motion was forced to pay $612.5 million to a patent-trolling firm called NTP. No one in the case claimed that RIM had directly infringed NTP’s patents. Rather, RIM had independently developed its technology and only found out years after the fact that it might be covered by NTP’s patents. Even more outrageous, the Patent Office had issued “non-final rejections” of the patents at issue in the case, but didn’t move quickly enough to spare RIM from forking over a 9‑figure settlement to NTP.

With that said, it’s important to keep in mind that patent trolls are a symptom of deeper problems with the patent system, not the cause of the patent system’s problems. If we had a well-designed patent system in which only high-quality patents were issued, it would be much harder for patent trolls to engage in the kinds of abusive behaviors McCurdy laments. The reason patent trolling is so profitable is that over the last quarter century the courts have expanded patenting into new areas like software and business methods, and dramatically lowered the bar for receiving a patent. As a result, patents that would have been rejected 30 years ago (like this ridiculous patent on removing white space from database entries, which IBM received earlier this month) are now routinely approved by the Patent Office. As a result, patent trolls are able to buy up low-quality patents by the truckload. Even though the vast majority of the patents won’t survive legal challenges, defendants can’t take the chance that one of them might survive and force the firm into a 8- or 9‑figure settlement.

Patent trolls make good poster children for the patent system’s dysfunctions, but focusing too much on them ignores the fact that abusing the patent system is a game played by large companies as well. For example, Verizon managed to extort tens of millions of dollars from Vonage to settle a lawsuit over an absurdly broad Internet telephony patent. Verizon, of course, isn’t a “patent troll,” but a competitor interested in hobbling an up-and-coming competitor. Any patent reform needs to address the Verizons of the world too, not just the NTPs.

The problems with the patent system have primarily been caused by the United States Court of Appeals for the Federal Circuit, which has jurisdiction over patent appeals and has aggressively expanded patenting over the quarter-century since its creation by Congress in the early 1980s. In the long run, the only solution to what ails the patent system is to undo the mistakes the Federal Circuit made. The Supreme Court has begun to make progress in that direction with recent decisions such as eBay (which weakened patent trolls by making injunctions harder to get) and KSR (which raised the bar on obviousness). The Federal Circuit’s Bilski decision, which placed new restrictions on “abstract” patents, is another step in the right direction. But there’s much more to be done. Most importantly, the courts need to overturn the Federal Circuit’s decisions from the 1990s that ruled (ignoring contrary Supreme Court precedent) that software was eligible for patent protection. It’s not a coincidence that the most prominent examples of abusive patent lawsuits are almost all in the IT sector.

There are also steps Congress could take, but the changes most frequently discussed—switching to a “first to file” system, for example—aren’t likely to have much of an impact on the problems of low-quality patents. A better option would be to follow Jim Bessen and Michael Meurer’s advice and dramatically increase fees for obtaining and renewing patents, which would give patent holders incentives not to waste everyone’s time with low-quality patents.

Finally, given that the Federal Circuit has been responsible for many of the problems with the patent system, Congress should strongly consider phasing out the Federal Circuit and returning jurisdiction over patent issues to the 11 geographically-based circuit courts. As I argued in an issue of TechKnowledge last year, competition among circuit courts is an important part of our decentralized common-law system of justice, and it has served us well in virtually every other area of the law. The experiment with a unified patent appeals court has not gone well, and Congress should consider reversing its mistake.