Wednesday, March 2, 2016
Interesting study of patent litigation in China: is the conventional wisdom about protectionism wrong?
Here’s a very interesting new paper on patent litigation in China based on an analysis of 471 suits for patent infringement (Love, Helmers & Eberhardt, Patent Litigation in China: Protecting Rights or the Local Economy?). The authors aim to subject to empirical analysis (which they say, to the best of my knowledge correctly [UPDATE Mar. 10, 2016: I'm wrong. See Mark Cohen's blog post on this.], is virtually non-existent to date) the conventional wisdom that the system is biased, probably deliberately, against foreign patent-holders and they can’t get a fair shake. They conclude that the conventional wisdom is wrong:
Though many suggest China set out to create a system that would benefit domestic industry at the expense of foreign firms, our findings suggest that the system has accomplished the opposite. Contrary to conventional wisdom and high-profile anecdotes, foreign litigants in Chinese patent suits play the role of patentee more often than defendant and fare just as well in their suits as privately owned Chinese firms. Moreover, state-owned monopolies—parties the Chinese government presumably has the greatest incentive to protect—rarely sue and, when sued, lose a significant share of their cases.
On the whole, our findings suggest that the Western technology community may have been too quick to write off the Chinese patent system as a rigged game. To the extent that Chinese authorities sought to establish a protectionist system, they appear to be failing. Rather, they seem to have opened the door for foreign innovators to seek redress against local copyists. Industries that have long accused Chinese firms of idea theft may be well advised to take a peek inside.
I think this is a valuable paper that deserves wide circulation—we should always welcome careful empirical work that challenges conventional wisdom—but I think the authors don’t adequately explore the implications of a key methodological problem: that the population of cases they study is not representative of the population of patent disputes and is subject to selection bias. They look only at cases that go all the way to judgment; they don’t have access to cases that settle, they don’t have access to cases where a plaintiff brings suit but the court refuses to docket the case, and of course they don’t have access to cases that never become cases at all because the plaintiff is convinced (rightly or wrongly) that it will lose and so doesn’t bother suing, or the defendant is convinced that it will lose, and so folds immediately upon receiving a threatening letter from the potential plaintiff. For those interested, the selection bias involved in studying reported cases is modeled and analyzed in a classic paper by Priest and Klein, The Selection of Disputes for Litigation (1984).
They acknowledge this problem in a footnote at the end, but seem to overlook it in the main text when stating their conclusions about what their study shows. For example, they state, “when foreign companies sue, they win relatively frequently . . .” One simply cannot say even that, based on this data, since it’s possible that 99% of foreign suits are never accepted by courts, or are ultimately dismissed before judgment for some reason. (Of course, that’s unlikely, but the point is that there is no way to know from this data what the percentage is, and it’s an important number.)
They also say, “the case-level data suggests that patent suits are rarely litigated in smaller inland cities where, conventional wisdom holds, protectionism is most often encountered.” But this result is consistent with the conventional wisdom being absolutely right, and known to be right by non-local patent holders, who therefore don’t bother bringing suit in such cities, or if they do bring suit, find such suits rejected before judgment by local courts.
There are some areas where this methodological problem might be less acute. For example, the authors find that “successful foreign patentees received a median damages award of 100,000 RMB in suits against private Chinese firms, exactly the same amount that private Chinese patentees received when they sued private domestic parties. Interestingly, Chinese patentees received 20 percent less in suits against foreign companies and 60 percent more in suits against state monopolies.” Those with more mathematical sophistication than me can think about whether my intuition is correct that award size might be less susceptible to selection bias than win rates.
In any case, to the extent one can say anything from a study of reported judgments, this study says something new and interesting and adds something valuable to our knowledge.