Reverse Settlements, Part 4: What Is the Baseline?

In my final post on reverse settlements I want to offer three thoughts that are more directly related to the legal question of how to treat reverse settlements under antitrust law.

First, it strikes me as odd that we scrutinize reverse settlements of Paragraph IV challenges differently than settlements of patent suits of non-drug, even non-health products.  As Einer acknowledges in his Texas Law Review piece, nearly all patent litigation affects market structure and thus both the level of competition and the amount of consumer welfare (Elhauge and Krueger 2012).  In each of those cases, because the public is not party to the litigation, settlements between patent holders and alleged infringers will – in theory and perhaps in practice – tend to hurt consumers.  (The monopoly-duopoly wedge that gives rise to the problem of reverse settlements is by no means unique to the drug market.)  Yet my patent law colleagues tell me there is no systematic review of non-drug patent settlements as is being urged of drug settlements in the FTC v. Watson case.  It seems that under the FTC view, drug patents would be treated more harshly than other patents. I am not sure why that should be the case under antitrust law.

Second, the critical question in the antitrust litigation is the baseline against which reverse settlements are judged.  Reverse settlements are only problematic under antitrust law if they extend patent duration or scope beyond some baseline.  Should that baseline be expected duration with full litigation and no settlement – as critics of reverse settlements urge – or something else?  For expected litigation to be the baseline, one has to assume that Hatch-Waxman modifies patent law and that patent duration after litigation is what is now required.  I am not sure these assumptions are appropriate.

For one thing, Hatch-Waxman does not require litigation until judgment rather than settlement.  Indeed, it does not even require suit.  It tries only to encourage suit by offering the first-filing generic 180 days of exclusivity with the pioneer.  Nor does Hatch-Waxman prohibit settlement.  That Hatch-Waxman does not achieve the expected patent length with litigation until judgment is a function of litigants’ behaviors.  The behavior that folks focus on is reverse settlement after suit.  But expected patent duration is also affected by the fact that generics do not file suit against all drugs with weak patents – only those with high sales (Hemphill and Sampat 2011).  Those other non-litigated, branded drugs also enjoy “extended” patent life.  Why is that not an antitrust violation?  There is inconsistent treatment of different branded drugs when reverse settlements are not allowed but non-suits are.

A simpler, more consistent approach to interpreting Hatch-Waxman is just to suppose that Hatch Waxman tried to encourage generic entry via, among other thing, patent litigation against branded drugs, but failing to anticipate settlement, it was not super-effective.  Antitrust law does not demand it be more effective.  And nothing stops Congress from amending it and making it more effective, say by banning settlement of Paragraph IV challenges.  That Congress failed to do so does not imply that reverse settlements violate antitrust laws.  It just means there is perhaps more work for Congress to do.

Nor am I sure that Hatch Waxman modified patent law for the purposes of antitrust law. It is not clear that Hatch-Waxman determines “legal” patent duration any more than a statute that affects incentives to sue – say one that reduces court fees or one that reallocates attorneys fees – should be thought to modify substantive patent law.  Indeed, under this logic, the Food Drug and Cosmetic Act, by reducing the time during which drug companies can market drugs could also be thought formally to be reducing patent life.  But that is not how it is treated.  If a drug company a drug company was found ex post to not have met all the requirements of the FD&CA but have been approved anyway, e.g., if it hides negative test results from the FDA, it is unlikely to be treated as having violated antitrust law.  It simply violated the FD&CA.

After four long posts on reverse settlements, it might be helpful if I summarize my position. I believe reverse settlement probably extend patent duration relative to expected duration given litigation without settlement.  I am not sure, however, that this fact implies they should be per se illegal.  There are a number of reasons I think extending patent life relative to the litigation-without-settlement baseline could be welfare enhancing (post 1) or may not imply higher producer surplus (post 2).  Post 3 explains why focusing on a single drug patent makes little economic sense when trying to calculate antitrust-relevant outcomes.  Finally, Post 4 questions the litigation-without-settlement baseline.  All these point certainly counsel agains a per se rule that reverse settlements are invalid. They probably also counsel against a rule that evaluates patents as the 11th Circuit has done.  I am probably more comfortable with the 2nd circuits more accomodating approach.  Congress is certainly within its rights to modify Hatch Waxman to ban settlements [1].  But I am not sure that would enchance social welfare.  To answer that question we need to know the return to R&D spending by drug companies – a notoriously difficult question on which economists are divided.

[1] Banning settlement could also have the counterproductive effect of lowering the number of generic suits.  Presumably reverse settlements offer higher returns than litigation, otherwise first-filing generics would not settle.  Eliminating settlement thus lowers expected returns to filing Paragraph IV challenges.