Tuesday, 22 July 2014

Smoking gun found as drug deals go down

I've been a bit exercised by these "pay for delay" drug deals where patent holders buy off generic drug competition - see here and here - and I know I'll have to let them go, and get on with other things, but one final post on the topic, as I've been pointed towards a brand new piece of empirical research on them that finds that they are very likely anticompetitive.

Up front, I should say that I'm normally minded, when I see some business behaviour that doesn't look like it fits with what the economics theory would suggest a firm would do, to look for a benign, rational explanation. Businesses will often have logical and proper reasons for what they do, even if on first inspection an economist can't see what those reasons are. And I've even changed my mind on some things that I would once have regarded as out of hand anticompetitive (retail price maintenance, for example) and I can now see why a company might have a legitimate reason to do it, and why consumers might not be harmed or could even benefit.

But at first blush these "pay for delay" delays looked suss to me, and although there are respectable and even heavyweight competition economists (Willig, for example) who believe they are, or can be, above board, I'm at a minimum still of the view that most "pay for delay" deals are rorts on the consumer.

So I was intrigued when a blogging colleague put me on the trail of "Do "Reverse Payment" Settlements of Brand-Generic Patent Disputes in the Pharmaceutical Industry Constitute an Anticompetitive Pay for Delay?", by Keith M. Drake, Martha A. Starr, and Thomas McGuire, NBER Working Paper No. 20292,July 2014, © 2014 by Keith M. Drake, Martha A. Starr, and Thomas McGuire (the copyright thingy is there because the NBER papers say third parties citing them have to put it in).

You can see the abstract here and you can read the whole thing if your organisation has a sub to the NBER or if you fork out US$5 online. If you're in the teaching trades, and in particular if you're in the competition teaching trades, you might want to share it with your students, as the topic is interesting, the writing's accessible, and the maths and stats are fairly easy.

The researchers looked at settlements of patent drug litigation between incumbent patentholder companies and generic competitors, and split them into two buckets - those where there were "reverse payments" from the patentholder to the generic, and those where there weren't. "If, in settlement, the brand manufacturer in effect buys a longer period of monopoly sale by “paying for delay” with a “reverse payment,” expected profits to the brand go up", they reasoned (p12) - "longer" here being longer than the average outcome that would have been expected from fighting on in the litigation. So they looked for the stock price impact of news of settlements, which should quickly reflect that rise in expected profits by way of higher share prices. And, since it's at least possible that settlements might have been struck for good reasons that weren't anticompetitive, they were especially interested in whether the settlements with reverse payments had a bigger effect on stock prices than the ones that didn't.

They were pretty careful, too, to isolate the impact of the news of the settlement, by comparing the actual share price movement with three different measures of what might have happened to the share price in the absence of the settlement. So they looked for "abnormal" or "excess" returns over and above what the company or the share market might have delivered in any case.

This is their key result (pp26-7): "For multiday event windows" - that's where the share price impact is measured over a few days - "cumulative abnormal returns for the reverse payment settlements are 5.5% to 6.0% higher than those for the other settlements and in all cases the difference is significantly different from zero...the incremental stock price jump of approximately 6% upon announcement of a settlement with indication of a reverse payment compared to one without is consistent with the hypothesis that reverse payments buy an anticompetitive delay in generic entry".

Incidentally, they also looked at trading volumes (again compared with the volume that might have been expected in any event), and again the same pattern came through. Investors were much more interested in news that the money had changed hands than in news that it hadn't.

Maybe I'll have my mind changed by some new evidence, but on this showing, if there are strong incentives to wreak a rort (check), and it looks like a rort (check), and it's carrying a large sack of non-consecutively-numbered dollar bills under one arm (check), it's a rort.

1 comment: