| People | Open Source Law & Economics | News | Antitrust Links | Recruiting | Contact Information |
A large market failure exists in the market for most types of software, including all consumer desktop software. This is because the Kaldor-Hicks efficient price and terms of consumer software is $0 with no license restrictions, but private companies will almost never have the incentive to produce software at this price and under such terms. This first graph illustrates the failure of such markets.
Note that marginal cost curve quickly declines to zero, because the cost of producing additional copies of software is effectively zero. The large grey area is the deadweight loss caused by the failure of the consumer software market.
Currently United States public policy does almost nothing to rectify this market failure. The best way to rectify this market failure is through increased funding of open source software projects. Malaysia and Brazil are doing this. The United States, the world's largest economy and the worldwide center of software development, needs to take the lead.
Fortunately many individuals, companies, universities, and non-profit organizations have stepped forward where both the market and US policy have failed and are funding and producing free, high-quality, open-source software. Links to many such programs are at the bottom of this page. Below is an illustration of the gain to the national economy when software is released free of cost and licensing restrictions.
The graph illustrates two effects of this same consumer software market moving from a monopoly model to an open-source model. First, the relatively small level of monopoly profit is converted to consumer surplus. Second, and much more importantly, the large deadweight loss is corrected and converted to consumer surplus, a pure gain to the national economy. Moving from a market failure to a Kaldor-Hicks efficient market is the closest thing in economics to a free lunch.
Many of the most important antitrust class actions are brought on behalf of classes whose members are virtually all computer owners. The various class actions against Microsoft are one example, but the example I'd like to work with is In re Graphics Processing Units Antitrust Litigation ("GPU Antitrust"), which is the private class action that was brought in tandem with the government investigations of collusion between the two giants that dominate the industry for graphics hardware ("GPUs" or "Graphics Cards"), Nvidia and ATI.
Funding Open Source Software is a Better Way to Compensate Victims of Price Fixing
The largest antitrust class action settlement involving computer hardware in recent memory was against the companies that fixed prices on DRAM, where settlements thus far are about $325 million. Assuming that the GPU class manages to achieve the same result, and that there are 50 million class members, that is about $6.50 per class member.
This is not very much money on a per person basis. In fact, the main benefit to class members of the settlement will not be the dollar value of the settlement if it is distributed, but the deterrence effect that the settlement in scaring the defendants and other computer companies from future violations of America's antitrust laws, thus keeping the prices class members pay down, and encouraging the fair and open competition between companies that is the pillar of our free market economy.
Superiority #1 - An Open-Source Settlement Allows More Class Members to be Compensated
Sending out tens of millions of checks for under $10 is an economically wasteful activity. The average hourly wage in the United States in December 2007 was $17.71. Optimistically, the process of receiving class notice, filing a claim form, and cashing the claim check would take 15 minutes total. A 30-minute estimate is more realistic in GPU Antitrust because not all GPUs are the same. Some graphics cards cost $50, others $800. Thirty minutes multiplied by the average hourly American wage gives us an average time cost for class members of $8.86. Further, most potential GPU class members don't actually know they are class members. Indeed, most class members only have a foggy idea of what a GPU is, much less if their computer has one. (Most PCs sold in the United States do have GPUs, but many laptops and lower-end desktops instead have what is called "integrated graphics processing" instead of a separate GPU.)
Thus the precentage of class members who will file claims is likely to be low. Most class members will receive no other benefit from the lawsuit other than its deterrent effect on future computer industry price-fixing.
Superiority #2 - Free Open Source Software and Cheaper Commercial Software Provides a Greater Economic Benefit to the Class Than Cash Checks
The development of open source software benefits class members directly in the form of free open-source software and cheaper commercial software, and does not require class members to spend any time on the claims process. As the quality and quantity of free software increases, commercial software vendors will either have to cut prices, or keep prices steady and rapidly improve quality in order to compete. Thus even consumers who continue to buy its commercial software will benefit from the competition that a well-funded open source project provides commercial software.
Superiority #3 - Selection of Open Source Projects Can Easily Be Weighted To Benefit Class Members in Proportion to the Damages Each Suffered
Some types of buyers, namely those who purchased more expensive high-end graphics cards, suffered more damages than the average computer user. Weighting the distribution of a cash GPU settlement by amount of damages consumer class members suffered, however, would require class members to figure out and prove, or at minimum estimate, how much they paid for their graphics cards. For most class members this would be a difficult and time-consuming process, if not an impossible task.
By contrast an open-source software settlement could fund specific software projects according to whatever general demographic information is available about the class. The majority of class members who bought a basic desktop personal computer or laptop have a fairly cheap graphics cards and spent an average of $50, thus suffered a relatively small amount of damages. Those that purchased "media center" computers to use as the hub of their home entertainment system might have spent an average of $200 on their graphics cards. Graphic designers, engineers, and digital video producers might have spent an average of $1000. Finally there is a subculture of networked video game enthusiasts who purchase and frequently upgrade their graphics cards, allowing them to competitively play the most advanced networked video games. This submarket is so large than both defendants produce expensive, specialized graphics cards with this demographic in mind. They might have spent an average of $700.
An open source software settlement of GPU Antitrust thus ought to fund both general consumer software benefiting the average class member who owns a cheaper graphics card, as well as software that would benefit class members who fall into the groups who purchased more expensive cards. Examples of general consumer applications that the GPU Antitrust settlement fund should fund are the Firefox web browser, and the Open Office software suite, which like the commercial Microsoft Office suite includes a word processor, spreadsheet, and PowerPoint-style slideshow software. For video gamer class members a worthy recipient might be BZFlag, a "first person shooter" video game with advanced graphics that already has been downloaded and enjoyed by millions of people. Video production class members would directly benefit from further funding of Jahshaka, a multi-featured video editing software. Even if particular class members don't like and will never use these particular programs, funding one open source project benefits all other free software projects by increasing the amount of free software code available for these other free software projects. Further, more free software increases competition in software markets, resulting in lower commercial software prices.
Superiority #4 - Faster and Larger Settlements
Many antitrust defendants, in particular computer hardware defendants like Nvidia and ATI, would receive benefits from an open source settlement that would partially offset the cost.
Computer software and computer hardware are classic examples of what economists refer to as complementary goods, that is two products with a negative cross-elasticity of demand. This means the subsidy of one complementary good via a settlement will increase the demand good for the other complementary good. Because defendants would receive benefits as well costs from a software-funding settlement, they would be willing to agree to a larger settlement, and do so sooner, again to the benefit of the class, as well as to the economy as a whole.
Turning back to the concrete example of GPU Antitrust, if more computer owners download a bigger and better-funded free Jahshaka video editing software because of a software-funding class action settlement, their desire for a higher-end graphics card will increase, to the benefit of ATI and Nvidia. Assume for each extra dollar in open-source development the defendants would increase their profits over time by 20 cents. Thus a $400 million software-funding settlement, which would cost them $400 million but increase their profits over time by $80 million, costing them a net $320 million. These companies therefore would agree to a $400 million software-funding settlement faster and more readily than a $350 million cash settlement.
Superiority #5 - All Consumers Harmed by Price-Fixing Benefit
A price fixing conspiracy such as the in my GPU Antitrust example harms two groups of consumers, represented in the graph below as Consumer Group 1 and Consumer Group 2. Only Consumer Group 1, however, can possibly be compensated by a cash settlement. The following graph shows the two groups of consumers harmed by collusion in the graphics card market.
The point on the demand curve marked "Cournot Equilibrium" shows the price and quantity produced that would have been sold in a market free of illegal collusion, but still dominated by the Nvidia/ATI duopoly. However, because of the conspiracy, prices have increased and production and sales have decreased to the point on the demand curve marked "Collusive Price & Quantity." Consumers responded to the collusion in two different ways. Consumer Group 1 paid the higher prices that resulted from the collusion. The higher prices that these consumers paid is the illegal profit earned by the conspirators, and is represented graphically by the light blue box. Most of Group 1 will have standing to sue the companies engaged in the illegal behavior for their losses.
Consumer Group 2 also suffered economic harm as a result of the collusive price raising, not because they paid a higher price for their GPUs, but because they didn't purchase the them at all. Their losses can are shown by the grey triangle to the right of the light blue box. A class action settlement compensating only Consumer Group 1 thus does not compensate all the victims of the conspiracy, but compensating Group 2 with cash for their antitrust injury is impossible because indentifying them individually to send them checks would require knowing what individuals would have purchased in the past had prices been lower.
While payments to Group 1 does nothing for Group 2, an open-source software settlement benefits both groups because they are both participants in the market for computer hardware and therefore also in the market for computer software.
Also, as an economic matter, while Group 2's injuries are smaller in magnitude than Group 1's, they are worse from an economic perspective. Group 1's injuries result in a transfer of wealth from the buyers to the sellers, which in the short-run do not otherwise harm the economy, though collusion does a lot of long-run harm to the economy. Group 2's injuries, however, are the result of the size of the American economy shrinking, because the supply of graphics cards is reduced by the colluding companies in order to create enough scarcity of the good to support their collusive price.
Conclusion
Normally when a portion of a class action settlement is not used for either payments to class members or their attorneys, the funding is considered cy pres, or "as close as possible" to the impractical ideal of a cash distribution. Here I have presented strong evidence that most, if not all, of class action settlements involving computer users should be used to fund open source software because class members are likely to receive a vastly greater direct economic benefit in the form of lower prices and improved software quality than if some small portion of them were simply cut a check.
At minimum, every computer class action should reserve some part of the settlement for open source software, since absent an open source software many if not most class members will receive no direct economic benefit from the settlement.
SourceForge.net 2007 Community Choice Award Winners
SourceForge.net 2006 Community Choice Award Winners
Damicon's List of Open Software
Google's Open Source Software Page
http://www.thewestonfirm.com/opensource.html - A Law and Economics Analysis of Open Source Software and Computer Industry Antitrust Class Action Settlements - Version 4 - 4/22/08