Dr. PatentLove

or: How I learned to Stop Worrying and Love Proposed Patent Reform.

The blawgosphere is alive with discussion of proposed patent reform.  Patent Troll Tracker today mocked Dennis Crouch for his support of the status quo.  Here at Patent Prospector, we maintain a healthy amount of skepticism about the proposed legislation.  Patent Hawk earlier pointed out how the new damages apportionment rules could potentially raise the cost of litigation for accused infringers by requiring the disclosure of more (highly proprietary) information about the cost, revenue, and profit of accused products and processes.  For my part, in an earlier post I called the new damages apportionment rules "the crudest and most error-prone way to handle bad patents."  The worry is that judges and juries will tend to "split the baby" by lowering some damages awards, which in may cases are properly justified by the economic value of a patented technology.  The political reality is that a Congressional amendment to the law suggests the inference that current reasonable royalty awards have been too favorable to patentees.  (See, for example, how many injunctions have been awarded to non-practicing entities since Mercexchange v. eBay.)  Information asymmetries between judge or jury and the expert witnesses and parties should be part of the equation in designing the rules, as some recent antitrust law scholarship suggests.

Despite this skepticism, I cannot agree with Dennis that "the proposed patent reforms now being debated by the Senate do virtually nothing to address these serious problems and instead potentially cause harm to the current regime."  Specifically, Dennis points out "massively overlapping claims" as a problem for the current patent system.  Since earlier posting strong criticism of the proposed damages apportionment rules, I've come around to seeing how the proposed rules could actually improve the efficiency of patent markets afflicted with the transactions costs associated with "massively overlapping claims."  Here's how:

First, to define the problem, a patent thicket can arise when multiple patents cover a product, and each of the multiple patentees (rationally) tries to encompass as much of the profit as possible.  The multiple patentees may behave this way even without fraudulent intent because each patentee tends to have more information about the value of their own invention than does any other patentee or the accused infringer.  All the accused infringer can do is try to pay as little as possible for each license.  But in some (perhaps many) cases the result is either underproduction of the patented product or a feast for the earlier licensors and famine for the later.

More generally, this problem can be identified in markets in which complementary inputs are provided by multiple firms to a producer selling in a competitive market.  French mathematician and economist A.A. Cournot pointed out that the producer will tend to underproduce in these circumstances because each complementary input supplier will try to encompass all of the profit.  Profs. Lemley and Shapiro have argued (with empirical support) that "stacked" patent royalties can resemble Cournot complements in their economic effect on production.

How do proposed damages apportionment rules help untangle patent thickets? The new rules explicitly provide that patentees seeking a "reasonable royalty" (i.e., the proposed rules don't change how damages for "lost profits" are calculated) may receive only a portion of the profits from a patented invention when neither the "entire market value" nor the "marketplace licensing" measures of damages appear to provide an accurate measure of the economic value a claimed invention has contributed to an infringing product or process.

By definition, where a patent thicket exists, the "entire market value" measure for "reasonable royalties" does not apply.  The problem of Cournot complements remains, however, even when existing "marketplace licenses" are available as a measure of reasonable royalties.  Thus, the "apportionment" provision of the proposed rules is useful to the extent that it tells judges and juries to think about how big or small the slices of the pie need to be to ensure that everyone who contributed to its baking gets a slice.

Here's another way to look at it:  From the point of view of everyone in the market for ideas (except, perhaps, the opportunistic few  who help themselves to the pie first), patent thickets are bad.  Because no participant in the market has complete information; nobody is going to agree about how the pie gets sliced.  Cournot knew that markets could solve his problem: owners of complementary inputs can integrate.  In patent markets, this can be achieved through patent pooling.  But patent pools often cannot be formed; transaction costs go up as fast as N(N-1)/2 where N is the number of patentees in the market.  And when patent pools are not formed, the difficult job of deciding how the pie will be sliced up must be done by other institutions in society, including, as a last resort, the judicial system.

To be absolutely clear: in my view, the proposed rules regarding damages apportionment could be improved.  In particular, I would much prefer to see the magic words "availability of non-infringing alternatives" somewhere in the statute.  (There is still no better way to define the economic value of an invention than by comparison to its alternatives.  Why not make that explicit in the statute?)  Nonetheless, I felt a desire to clarify some of my earlier overstatements.  The proposed damages apportionment rules may not, in fact, be disastrous for the emerging market for ideas.  In instances where a patent thicket has developed, apportionment could conceivably lead to more efficient production, and more efficient allocations of profit among all patentees who have contributed to the product or process sold.