--Summary by Justin Pappas Johnson
Johnson Graduate School of Management, Cornell University

Should governments subsidize free/open-source software? This paper discusses the potential advantages and disadvantages of such public support. It focuses on three distinctive features of software markets: (1) significant economies of scale, (2) innovation, and (3) “network effects.”

Economies of scale arise because once the software is developed, additional copies cost little to produce and distribute, especially if they are distributed over the Internet. The authors argue that from a social standpoint, this feature of the market gives free/open source an advantage over proprietary software. The only cost involved in providing additional copies of free/open-source software is that of distribution, which allows many users to enjoy the benefits. A proprietary firm would probably raise prices to recoup the cost of development. Doing so would raise the firm’s profits but limit access.

However, proprietary developers may have stronger incentives to innovate in important ways. The reason is that they are able to recoup investment costs by charging a profit margin over cost for the product. The profit motive may also lead for-profit developers to design software that better serves the needs of the typical user. This “dynamic efficiency” issue is quite important, and the paper does a fine job of detailing a number of important considerations beyond the basics noted here.

The authors also find public support of free/open-source software problematic when they consider network effects. A “network effect” occurs when the overall benefits of using a piece of software increase as the number of users increases—for example, this may occur when users benefit from having software that is compatible with the software of other users. The authors argue that markets with strong network effects may be easily influenced by a government’s choice to subsidize a piece of software. Such support is tantamount to choosing a winner, they claim, which may not be the proper role for the government to play, not least because it might choose an inferior product.

Generally, then, the authors’ tone about public assistance for free/open source is cautionary. Given the great hype that sometimes surrounds the topic, this critical look at the possible negative consequences of government support is both necessary and refreshing, and the paper succeeds in laying out a number of potential problems associated with such support.

However, the paper is perhaps too negative in its appraisal of free/open-source software. For example, the authors seem to discount the fact that while for-profit firms may indeed have strong incentives to innovate, they also have strong incentives to maintain and expand their market power. Possible tactics here include creating incompatibilities, bundling products to control future markets, and other predatory actions.

The paper also neglects the possibility that free/open source may lead to more innovation in cases where end users are a better source of information on how the software should be modified, or if parallel debugging—that is, debugging done simultaneously by many user/developers—has dramatic effects on the value of the software. This raises the interesting question of why more firms don’t provide their source code to users if so doing would “help the project”—after all, the firm could still charge for access to the code. There are many possible answers to this question, but two obvious ones are that releasing code would threaten a firm’s market power, and that a firm that does not reveal source code may be better able to harvest profits through service and support of software that only it knows how to properly maintain. This argument again suggests that proprietary software developers may be attempting to build market power.

It seems that our current understanding of free/open-source software remains quite limited. As the paper argues, it may be unwise to place too much faith in it, or to provide market-shaping levels of subsidization. Nonetheless, on a case-by-case basis, it seems important to recognize that free/open source may have important social advantages over for-profit software, especially when private attempts to obtain and maintain market power are considered.