One thing that becomes clear from the discussion [comparing Wikipedia to OSS] is how dangerous it is to use “open source” as a metaphor in describing other forms of participative production. Although common, the metaphor almost always ends up reducing the complex open-source model to a simplistic caricature. The discussion also sheds light on a topic that I’ve been covering recently: Yochai Benkler’s contention that we are today seeing the emergence of sustainable large-scale production projects that don’t rely on either the pricing system or management structure. Benkler’s primary example is open source software. But panelist Siobhan O’Mahony’s description of the evolution of open source projects reveals that they have become increasingly interwoven with the pricing system and increasingly dependent on formal management structure.
Libertarians have long recognized that the firm and the market can be mixed and matched in complex ways. Firms obviously rely on markets to obtain raw materials and to sell their finished products. Markets are also organized by firms, as in the case of the New York Stock Exchange. Firms also sometimes use markets internally, as with Koch industries, which uses “market based management” to give individual divisions within their firm greater incentives to productivity. Moreover, as Coase described, firms constantly face decisions about which of its goals they should accomplish internally (i.e. using the methods of the firm) and which they should outsource (i.e. use the methods of the market). No one would claim that any significant industry could be run purely as “markets” or purely as “firms.”
Benkler’s claim is that peer production is a new model for organizing economic activity that’s on par with, and often interdependent upon, firms and markets, in precisely the same way that firms and markets are intertwined with each other. He describes how peer production is often combined with firms and markets in a variety of innovative ways. For example, he describes one firm that uses peer-production-like processes to maintain their internal documentation. Benkler has never claimed that peer production renders markets and firms irrelevant, simply that they are being supplanted in particular cases where peer production works better.
Hence, the fact that some of the largest open source projects are organizing themselves in firm-like structures is an example of Benkler’s thesis. So is the fact that firms are contributing to open source projects. In the former case, an effort that had initially been organized on a “pure” peer production model discovered that some of their goals could be achieved more efficiently using the methods of the firm. In the latter case, a firm discovered that some of its goals could be more efficiently accomplished using the methods of peer production.
Carr then quotes a report arguing that whereas open source projects were once “self-governing, accepting volunteer contributions with most participants motivated by the cause,” most are now supported by a “majority of volunteers are sponsored by vendors, well-supported by in-kind donations of hardware, marketing and legal services.” This is silly; it’s akin to saying that now that Google and Microsoft are billion-dollar companies, there are no more software startups. Of course as open source projects get larger, they’ll find the methods of firms and markets useful to accomplish some of their goals. Of course the highest-profile projects have generated significant attention and support from mainstream software companies. There are more small, decentralized, “pure” open source projects than ever, and some of those will be tomorrow’s Apaches and Firefoxes, just as there are lots of small startups today, a few of which will be tomorrow’s Googles and Microsoft’s.
Despite their adopting some of the methods of firms and markets, projects like Firefox and Apache remain fundamentally oriented toward peer production. The crucial difference is what Benkler calls indirect appropriation. In the traditional proprietary software model, an individual or firm produces a piece of software so that he can gain exclusive rights to it and sell those rights to others. This is the “market” method of organization. Not surprisingly, this is the dominant organizational scheme for the production of tangible goods because making copies of tangible goods is expensive. GM wouldn’t be able to give away cars and make it up on service contracts. Benkler calls this market-oriented production scheme “direct appropriation.”
The commercial employees who contribute to Linux, Mozilla, Apache, and other open source projects, in contrast, are rely on indirect appropriation. They give away their efforts for free and depend on a variety of ancillary benefits to make it worth their trouble. They sell service contracts, integrate the software into their own products, use their participation as a marketing opportunity, and use their influence over the evolution of the process to help strengthen their firm’s market position. What they don’t attempt to do is assert proprietary rights over the software they produce, the way that virtually all firms making tangible widgets do.
The fundamental issue here is that IBM’s relationship to other open source developers is intrinsically different from its relationship to commercial vendors with whom they buy and sell proprietary rights in software they’ve produced. The fact that IBM pays its employees for both activities, or that they make a profit from doing so, in no way changes the fact that peer production is a new and important way to organize the creation of intangible products.