A study by BU Law professors James Bessen and Michael Meurer (via a good Popular Mechanics article via @normative) analyzes stock market events associated with patent troll litigation, and comes up with a $500 billion estimate for lost wealth. But less than 10% of this reflects wealth transfers to the patent trolls themselves. The authors theorize that the difference reflects lost innovation incentives, but it's hard to see why that wouldn't be reflected in companies' stock prices already: are investors that unaware of the patent thicket ex ante, before the lawsuit is filed? And if they are, why do we believe the market is efficient in predicting the effects of patent litigation on market value? If one smartphone manufacturer is sued by a patent troll claiming a blocking patent, shouldn't we expect the rest of the smartphone market to face similar problems?
Certainly the existence of a patent lawsuit means costs to a defendant beyond the eventual damages or license fees; there are the costs of attorneys, there's lost executive and engineer time sitting in depositions and engaging in other litigation tasks instead of productive activities. There's increased cost of capital from the overhang of the threat of litigation. And, yes, at the margin, there's decreased innovation incentives. But much of this should have been captured in the existing stock prices (and, for that, reason, any study estimating the costs of litigation based on stock market events will systematically fail to count existing depression of stock prices; the authors themselves note that their study doesn't capture the losses to privately-traded firms). If market values are really declining ten times as much as what is eventually paid out to the trolls themselves suggests to me what I've said in other circumstances: markets underestimate litigation risk before litigation is filed and overestimate litigation risk once litigation is filed and plaintiffs' lawyers and their agents start inserting self-serving information into the marketplace. This is a great data set: I'd love to see if companies facing stock shocks from being sued by patent trolls then outperform peers in the long run because the market overreacts to the realized threat of litigation.
It would not surprise me if patent trolls do cost the economy as much as $500 billion or more. I don't think this paper dispositively demonstrates that. Much of the loss from patent litigation and the threat of patent litigation (which would include resources diverted to defensive patent filings) is unmeasurable, mostly because we cannot observe the but-for world where incentives are better aligned to see what consumer surplus we're missing from depressed innovation.