To Renata Hesse
The following six (6) pages of this facsimile are a comment on the Microsoft Settlement in the Microsoft antitrust case. This comment has been simultaneously submitted by email.
As a professional working in the technology sector, I often have occasion to use Microsoft software and competing products. I am therefore concerned that the Revised Proposed Final Judgment in the Microsoft antitrust case has a number of deficiencies that prevent the Judgment from providing certain and effective relief for Microsoft's violations of the Sherman Act. Unless these flaws are corrected, the Revised Proposed Final Judgment is clearly against the public interest and will positively harm third parties.
This Comment addresses five serious deficiencies of the Revised Proposed Final Judgment, The deficiencies are discussed in the order they appear in the Judgment, not necessarily in their relative order of impact on injunctive relief. The deficiencies are:
Although it is unreasonable to expect a truly optimal Judgment that best serves the public interest, the existence of any one of the above deficiencies--and certainly the coexistence of several of them--will not end Microsoft's unlawful conduct nor avoid a recurrence of violations of the Sherman Act, and is thus outside the reaches of the public interest.
1. Judgment provides no remedies for past unlawful conduct
Although the Revised Proposed Final Judgment provides limited remedies "to halt continuance and prevent recurrence of the violations of the Sherman Act by Microsoft" (Competitive Impact Statement, Section I.), it does not in any way "undo its anticompetitive consequences" (Competitive Impact Statement Section IV.B.) , There is no provision in the Judgment to remedy any past anticompetitive actions by Microsoft: all provisions in the Judgment attempt to alter the current and future behavior of Microsoft. As such, the Judgment does not effectively restore the competitive conditions experienced by Microsoft prior to its violations of the Sherman Act.
An effective remedy for Microsoft's past illegal actions requires a careful balance to empower injured competitors while not unduly damaging Microsoft. A simple but fair remedy would create a pool of Microsoft's money based on a percentage of sales of Microsoft Operating System Products since the filing of the antitrust complaint till the time of the Final Judgment entered by the Court. The parties damaged by Microsoft's anticompetitive behavior (e.g., Sun Microsystems, Netscape Communications Corp., etc.) would be payed from this pool. The size of the pool and the relative payment terms to competitors are details that require careful consideration.
2. Allowing volume discounts anticompetitively maintains Microsoft's monopoly
Allowing volume discounts serves no procompetitive interest and is in fact very much against the public interest as it serves to illegally maintain Microsoft's monopoly. Section III.A. of the revised proposed final judgment stipulates that "Nothing in this provision shall prohibit Microsoft from providing Consideration...commensurate with the absolute level or amount of that OEM's development, distribution, promotion, or licensing of that Microsoft product or service." Section III.B.2 provides for a licensing fee schedule that "may specify reasonable volume discounts based upon the actual volume of licenses of any Windows Operating System Product..." These provisions allow Microsoft to continue to leverage its monopoly position to illegally maintain that monopoly. The Competitive Impact Statement entirely ignores the anticompetitive ramifications of these terms.
Unlike traditional manufacturing, where the production or distribution of a large quantity of a product can generate "economies of scale" and thereby procompetitively justify non-uniform pricing (e.g., volume discounts), the licensing of software has no significant economies of scale. A comparison with traditional manufacturing is useful. For a car dealership selling hundreds of cars per month, there is economic justification for the car manufacturer to provide a volume discount to the dealership: the distribution costs (shipping) per car are lower than for a dealership selling only ten cars per month, with software however, the only economy of scale obtained is slightly cheaper production materials: compact disks for distribution and paper for documentation and product boxes. OEMs typically only include a compact disk with a new computer purchase, for which the volume production cost is under one dollar (US$1.00). Hence the economies of scale afforded by large scale OEMs to Microsoft are less than one percent (1%) of the retail value of typical Windows Operating System Products. Hence there is no significant procompetitive reason to allow volume discounts to large OEMs.
Allowing Microsoft to offer volume discounts will further entrench its monopoly position. With volume discounts, Microsoft would retain the ability to price its Windows Operating System Product licenses at an artificially low cost to the largest OEM vendors. These vendors would thus have a strong incentive to continue to offer exclusively or predominantly the Microsoft Operating System Product on new Personal Computers, The largest OEM Personal Computer suppliers would have a free market incentive to choose alternate Operating System Products if Microsoft's Operating System Product were instead priced at an open market value. Avoiding volume discounts increases competition while preventing Microsoft from leveraging its monopoly to stifle competition.
This deficiency of the revised proposed final judgment is remedied by deleting the words "distribution" and "licensing" from the last paragraph of Section III.A. and by modifying Section III.B.2 to read "the schedule may not specify volume discounts based upon the actual volume of licenses of any Windows operating System Product or any group of such products." These modifications will still allow Microsoft to compete in the marketplace based on the merits of the windows Operating System Products, but prevent Microsoft from anticompetitively erecting barriers to competitive products.
3. Restrictions on disclosure of communications protocols maintains barrier to competition
The Revised Proposed Final Judgment maintains a significant barrier to competing Non-Microsoft Middleware Products by restricting the disclosure of Communications Protocols. Section III.E. of the Judgment provides that Microsoft shall disclose Communications Protocols "on reasonable and non-discriminatory terms." Such terms, however, prevent a large number of established and nascent competitors from obtaining the Communication Protocols. "Reasonable and non-discriminatory" license terms act as an anticompetitive barrier to potential Microsoft competitors, while providing no procompetitive advantage for Microsoft.
"Shareware" software developers typically provide software products (including middleware) free of charge for end users to evaluate, and only demand payment if the end user decides to continue using the software product. Such developers would be unable to comply with "reasonable and non-discriminatory" licensing terms unless a very large percentage of end users payed for the software product. similarly, the entire "open source" class of software would be unable to meet "reasonable and non-discriminatory" terms as the "open source" licenses allow virtually unlimited duplication and derivation rights. Several important Non-Microsoft Middleware Products are "open source", notably the Samba program (http://www.samba.org), that provides file transfer and print Services through the Microsoft SMB Communications Protocol. The Samba program is a well-established and widely used alternative to Microsoft Middleware Products, but it would be effectively prevented from competing with Microsoft through the adoption of "reasonable and non-discriminatory" licensing terms for future changes in the SMB protocol.
This deficiency of the Revised Proposed Final Judgment can be remedied by a simple wording change. The phrase "reasonable and non-discriminatory" in Section III.E. of the Judgment should be changed to "royalty free". Since Microsoft's ability to hide Communication Protocols serves only to prevent competitors from effectively interoperating with Microsoft products and does not in any way increase competition, a mandatory royalty free license would serve to allow both large and small competitors to interoperate with Microsoft products.
4. Arbitrary five year term of Judgment harms the public interest
The Competitive Impact Statement in Section IV.C. claims that a five year time frame for the Judgment "provides sufficient time for the conduct remedies contained in the Proposed Final Judgment to take effect...and to restore competitive conditions to the greatest extent possible." The Competitive Impact Statement provides neither evidence, nor precedence, nor logic to support this claim.
In fact, a five year term may well be too long. The provisions of the Revised Proposed Final Judgment may turn out to be so effective at restoring competition that Microsoft loses its dominance in less than two years in the Operating System market for Personal Computers and becomes unnecessarily hobbled by the restrictions of the Judgment. In such a case, Microsoft would be unfairly restricted from competing in the market for another three years, possibly causing great economic damage to Microsoft and depriving consumers of the fruits of a vibrant competition in the Operating System market.
Alternatively, the provisions of the Revised Proposed Final Judgment might not be sufficient to hinder Microsoft's anticompetitive actions, and Microsoft could continue to violate the Sherman Act through an extended seven-year Judgment period. Clearly such a situation would severely harm the public interest, again depriving consumers of the benefits of a competitive market and stilting the entire Operating System and Middleware market. The arbitrary five year Judgment term length would only be beneficial in the most serendipitous of circumstances, and the arbitrary two-year extension does not mitigate this fault.
The overriding concern of this Judgment is to prevent Microsoft's anticompetitive actions and to restore competitive conditions to the market, and it is that principle that should guide the term length of the Judgment. The most straightforward application of this principle would be to terminate the Judgment when Microsoft no longer enjoys monopoly status. This could be achieved with the following replacement for Section V. (Termination) of the Revised Proposed Final Judgment:
With this revised termination clause, the Judgment will stand exactly as long as necessary for the public interest. An alternate definition of monopoly status (i.e., instead of "fifty percent market share") may also be acceptable, provided it is logically and legally defensible, and maintains the intent of the Judgment.
This new termination clause will ensure the return of healthy competition to the Operating System market without unduly burdening--or harming--Microsoft. At the point that Microsoft's windows Operating System Products have less than fifty percent share of the Personal Computer Operating System market, there is clearly healthy competition in that market, with at least one other dominant competitor to Microsoft. There is then no further reason to impose the conditions of the Judgment. However, Microsoft is not prevented from maintaining its monopoly on the technical merits of its products. The ongoing terms of the Judgment would not be onerous to Microsoft should it maintain a monopoly position without resorting to anticompetitive actions.
5. Definition of "Non-Microsoft Middleware Product" maintains barrier to competition
Although the Revised Proposed Final Judgment seeks to "restore the competitive threat that middleware products posed prior to Microsoft's unlawful conduct (Competitive Impact Statement, Section IV), the proposed definition of "Non-Microsoft Middleware Product" serves instead to maintain barriers to competition. Section VI.N. of the Revised Proposed Final Judgment stipulates that a software product, among other requirements, can only be considered a "Non-Microsoft Middleware Product" if "at least one million copies were distributed in the United States within the previous year." This requirement is explained in the Competitive Impact Statement, Section IV.A. as being "intended to avoid Microsoft's affirmative obligations...being triggered by minor, or even nonexistent, products that have not established a competitive potential in the market..." As the Competitive Impact Statement makes clear, the definition of "Non Microsoft Middleware Product" intentionally limits the possible competitive impact of nascent middleware products. Such a limitation is antithetical to the desired goals of the Judgment.
This deficiency of the Revised Proposed Final Judgment can be easily remedied by deleting Section VI.N.(ii) and thus removing the restriction on number of copies distributed. The Competitive Impact Statement in Section IV.A. states that the restriction on number of copies distributed "is intended to avoid Microsoft's affirmative obligations--including the API disclosure required by Section III.D. and the creation of the mechanisms required by Section III.H.--being triggered by minor, or even nonexistent, products..." In other words, Microsoft should not endure an onerous burden in its obligations. However, deleting Section VI.N.(ii) would not create such a burden. Since Section III.D. already specifies that APIs and related Documentation shall be disclosed via the Microsoft Developer Network or similar mechanisms, Microsoft will not require any further effort to make the APIS and Documentation available to ISVs or other middleware developers that have not established a competitive potential in the market--but that nevertheless have the potential to become competitors with Microsoft, Furthermore, the mechanisms required in Section III.H. (such as the creation of Add/Remove icons) are sufficiently generic that they will only need to be created once--and likely already exist--to accommodate all Microsoft and Non-Microsoft Middleware, and hence the expansion of the number and kind of possible middleware competitors to Microsoft again does not create an undue burden on the company.
This Comment has been submitted through both e-mail and facsimile copy.