Thursday, September 12, 2013

Day 2 of University of Florida Workshop on Standard Essential Patents and FRAND: Hylton on "A Unified Framework for Competition Policy and Innovation Policy"

Professor Keith Hylton:  good that Harrison raised the issue of the purpose of the IP Clause.  This paper is not centrally about FRAND but more about antitrust and innovation. 

Hylton starts off with the premise that IP and antitrust are dealing with the same topic in many cases.  Many antitrust defendants got their market power from innovation.  Takes us back to Schumpeter.  Firms innovate, gain temporary monopoly, others copy what they do, profits back to competitive level.  Innovation and antitrust policy should have a lot in common.

Paper with David Evans in 2008 presents an informal discussion of the law:  section 2 seems more lenient than one might expect.  Why?  Because of the need to foster innovation.

Model here:  enforcement authority sets the penalty knowing that if sets it too high, new products won't be created. 

Williamson's diagram of monopolization with cost reduction.  After monopolization, firm gets some market power and can raise prices.  But efficiency gain (E) may outweigh deadweight loss (D).

Becker and Landes on internalizing consumer harm:  static penalty internalizes consumer harm.  Firm will engage in conduct only when efficiency gain outweighs deadweight loss.  Encourage some monopolization with cost reduction.

Hylton posits that maybe the internalizing penalty is too large, if the firm must earn some return from creating the market in the first place.  What would the ideal enforcement authority do?  Get the firm to invest and create the market.  One possibility:  tell the firm you won't punish the firm at all, then hit them with a penalty ex post, but that won't work after the first time.  Hylton refers to this as "snookering."  Another:  if the firm can perfectly price discriminate, don't penalize the firm at all, if worried about distributive justice can redistribute through lump sum tax.  But perfect price discrimination not possible either.

The objective function has the profits from incentive (sum of T + E - cost of investment).  Another component:  what consumers get if firm invests and monopolizes.  Consumers get W.  But then there is the enforcement action, so there is an enforcement cost, so consumers get W - PC where P is probability of enforcement and C is cost of enforcement.  So we might want to moderate the penalty for monopolization, or even give the firm a reward.

Optimal penalty then is a dynamic penalty:  weighted penalty of Becker penalty plus innovation subsidy.   Static penalty is always too high.  Could give monopolist an innovation subsidy, or patentee a patent and a reward on top of the patent.       

My question:  IP is often both an input and an output.  Dynamic efficiency is not only a function of rewarding the first inventor but also of allocating rewards among a series of inventors.  IP rights that are too strong risk undermining dynamic efficiency by reducing the rewards of subsequent inventors who build upon the initial invention. I don't see where your model takes that into account.

Hylton:  that would be a more complicated model.  One way to look at it is where to draw the line between IP and antitrust.  IP takes my consideration into account by limiting patentable subject matter, etc.  IP already takes that into account.

Sokol:  You assume there are no shifts in policies of enforcement.  Could you have repeat "snookering" with changes in administration, etc.?

Yoo:  Only if credible.
Page:  You're suggesting that the monopolizing act is the same thing as as the cost-reducing act.  In many antitrust contexts, that's not the case.

Hylton:  That's a critique of the Becker/Landes model as well. 

Yoo:  I'll give you an example of snookering.  Nationalization of industry after foreign investment.  Another is compulsory licensing of patents.  But I'm not sure you can even snooker them once.  They'll discount innovative activity on the possibility this will happen.

Page:  Thurman Arnold would bring suit and then in the consent decree require royalty-free licensing.

Hylton:  There is mission creep in the IP/antitrust interface.  Investment takes place under assumption that the IP laws permit something, then you find out there is potential antitrust liability.  I assume that will work itself out in the long run.

Carlton:  I think your analysis would suggest that any extension of antitrust law is snookering.  Aspen Skiing.

Yoo:  The bridge between this paper and FRAND is my paper.  We tend to think of commercialization as a separate act, but how do we tweak that; the public goods analysis of my paper gives you a basis for allocating the surplus of each consumer to the innovator.

No comments:

Post a Comment