W3C home > Mailing lists > Public > www-patentpolicy-comment@w3.org > October 2001

Anti-Trust Issues of RAND

From: Daniel Probst <daniel.probst@vwl.uni-mannheim.de>
Date: Tue, 9 Oct 2001 16:26:56 +0200
Message-ID: <000e01c150ce$737643a0$e0379b86@vwl.unimannheim.de>
To: <www-patentpolicy-comment@w3.org>
                Comments on the W3C Patent Policy Framework

                            Dr. Daniel A. Probst
                       Department of Economic Theory
                            Mannheim University

                        preliminary notes: 9:10:2001

1  Introduction

   In their Patent Policy Framework the W3C has suggested incorporating
   patented subject matter into standards and allowing the collection of
   "Reasonable and Non-Discriminatory" (RAND) license fees by members. I
   would like to point out that the adoption of such a policy instead of
   the current Royalty Free (RF) model leads to anti-trust issues due to
   the specific nature of the software market.

   Before arguing that the suggested policy change is problematic, I
   would like to draw in doubt the W3C's claims that a policy change is
   even necessary. There seems little indication that, in the past,
   development of the Web has been severely held back because either 1)
   the holders of intellectual property rights (IPR) did not grant
   royalty free access to their IP, or 2) knowledge was not created
   because royalty fees needed to pay for the investment were
   unavailable. The current state of academic economic research does not
   allow the conclusion that patents in the software industry are
   necessary to recoup R&D investments.[Footnote 1] The claim that without
   royalty paying patents, necessary technologies would not be available
   for future standards is unfounded.

   The W3C has argued that so-called submarine patents pose a threat to
   the efficiency of the standardization process (and this could be
   mitigated by RAND licenses). At a superficial level the welfare losses
   due to such strategic behavior seems orders of magnitude less
   important than the anti-competitive effects of RAND licenses described
   in the next two sections. Furthermore, the W3C has 1) presented only
   one example of problematic behavior (i.e., in the context of P3P), 2)
   not argued why this should be a recurring event, 3) not explained why
   RAND is the only/best way of mitigating such problems.

   Having argued that the necessity of RAND licenses in software
   standards is less than compelling, the following two paragraphs make
   the point that software standards containing patentable subject matter
   are more likely to run afoul with anti-trust law than is the case in
   other sectors.

2  Anti-Trust Issues Concerning Market Entry

   Many standard setting bodies in other industries accept patentable
   components of standards. They try to mitigate the ensuing
   anti-competitive barriers to entry by requiring non-discriminatory
   licenses. However, in the software sector this is more problematic due
   to the specific nature of software.

   The W3C proposal of RAND licenses strongly discriminates against a
   significant and increasingly important sector of the software
   industry, namely Open Source Software (OSS). Most OSS-licenses like
   the GPL are incompatible with royalty payments as required by RAND
   licenses. Market entry would therefore be blocked for a growing and
   extremely competitive segment of software suppliers. The ensuing lack
   of competition would be detrimental to the welfare of both consumers
   and the economy as a whole.

3  Anti-Trust Issues Concerning Royalty Payments

   The current draft of the W3C Patent Policy Framework does not
   acknowledge that software patents are available in only a limited
   number of countries (and given the strong opposition towards a
   strengthening of IPR in this sector in the EU, there is a reasonable
   probability that this will remain so).

   Imagine, for the sake of the argument, two US software companies
   producing two competing software products based on some future W3C
   standard. Assume furthermore that both companies also own patents on
   different parts of the standard. If the software prices in Europe were
   to reflect the RAND license fees the companies pay to each other, then
   this pricing strategy would run afoul with European competition law.
   Given that there no legal basis necessitating such royalty payments in
   Europe, the pricing behavior would be interpreted as coordinating on
   artificially high prices.

4  Conclusion

   In light of the arguments presented above, the current W3C Patent
   Policy Framework makes the impression of being:

    1. a coordinated anti-competitive attempt by a number of large US
       based software companies to restrict market entry by OSS into
       future markets by designing standards which are known to be
       incompatible/non-implementable with current major OSS licenses.
    2. a coordinated attempt by a number of large US based software
       companies to fix artificially high prices of software products
       through royalty payment agreements even in those countries in
       which the royalty payments have no legal foundation because the
       underlying claims are not patentable subject matter.

   If the W3C Patent Policy Framework goes through as it is currently
   drafted, I would expect every successful RAND standard to be closely
   scrutinized and, if necessary, blocked by anti-trust authorities (at
   least in the EU).


   [1] While patents certainly have strategic value for the holder
   (i.e. to block competitors, or as a bargaining chip), current
   empirical research show that patents are not widely used for the
   purpose of recouping R&D expenditures via license fees. Indeed
   empirical research has great difficulties in identifying the surge in
   research expenditure which should have been induced by past
   strengthenings of IPR in the US.

Dr. Daniel Probst
Economic Theory
Mannheim University
Received on Tuesday, 9 October 2001 10:24:53 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:06:44 UTC