Public Subsidies for Open Source? Some Economic Policy Issues of the Software Market
Schmidt & Schnitzer (2002)
This paper contains a critique of the assumption that government intervention in the form of subsidies or policies in favour of OSS is necessarily a good thing. It applies “modern economic theory” to the software market to allow systematic discussion of the implications for government support for OSS projects, looking at potential market failures in the software market.
The core of the argument for OSS is that in a static economy, the market for a given software product would be a natural monopoly with massive economies of scale; since the almost all of the costs are incurred in development, sales cannot be priced at marginal cost by a proprietary developer.
On the other hand
Open source software is more efficient from a static point of view because it is priced at marginal cost (ie sold at distribution cost or given away for free)
However, in a dynamic market, it is argued that proprietary ownership allows developers to turn into profits at least some of the consumer benefits created by improvements to the code. They refer to Schumpeter’s (1942) argument that the profit motive is a key driver of innovation, the “creative destruction” inherent in the capitalist system, a phrase much misused during the dot.com boom at the end of the last decade, but useful in this context.
Unlike OSS developers,
… a proprietary software developer has strong incentives to engage in market research to identify [customers needs] and to adopt and develop the software as closely possible to what customers want.
OSS developers have little inclination or incentives to work with non-IT users (eg accountants for a finance application), or to make an application that works for non-expert users. Which explains why most successful OSS applications that are used by the developers themselves.
Network effects (direct/indirect and strong/weak) and switching costs in the marketplace are also discussed; they can work to make the market sticky and tippy (ie stable only where there is a monopoly supplier). They can favour both OSS and proprietary products – apache and MS Office are obvious examples in each camp.
Many governments around the world have mandated or subsidised the use of OSS, but that misgivings have also been expressed about the wisdom of this, not least by the libertarian part of the OSS movement.
Possible government responses
In the concluding part of their paper, models for government support of OSS are reviewed in turn.
Direct subsidies for OSS development are criticised, not least because it encourages rent-seeking activities, “there are many examples of … projects that got captured by large companies who then managed to acquire vast amounts of public subsidies…”. Governments do not have a good record of selecting and nurturing winners, and there is no reason to believe that they will do any better with software than motor vehicles.
If OSS is to be subsidised, it is argued that the BSD (not viral GPL) model should be used, to minimise the welfare costs to consumers.
A policy of public sector adoption of OSS is not seen as justifiable in general, being seen a distortion of the market, leading to increased prices for the remaining proprietary application.
The final option of a subsidy for coordination of OSS development is seen as least harmful if it encourages general compatibility and compliance with open standards.
In summary, government support of OSS is seen as unlikely to be of general social benefit in a dynamic market, removing as it does the linkage between developers and society that the marketplace and competition provides.
The best critical overview of OSS and government policy I have seen so far, raising many valid points. However, it does not crop up significantly in CiteSeer, so I suspect there may be others.
At the moment, there are two areas where I can see that this paper could be criticised.
Firstly, the market analysis does not as far as I can see take into account the value that can be added by private companies using OSS applications as the basis for the service they provide. That is the service, not the application, can be the differentiator – providing the linkage between OSS and the market that would otherwise be lacking. Drupal is one example: most of the development is driven by consultants who are incentivised by their clients. In short, the paper does not consider whether
Software is a service industry operating under the illusion that it is a manufacturing undustry (E Raymond)
Secondly, the argument does not take into account the specific needs of Public Authorities (PAs), and the possibility that the market they provide is too small and/or diverse (particulatly in fragmented, multilingual Europe) for packaged proprietary applications to be viable. The are (surely?) many successful models of PAs forming consortia to share costs and meet their unique needs.
Two books published in 2002 appear to cover similar ground from the US perspective:
- Government Policy toward Open Source Software Hahn (2002)
- Government Preferences for Promoting OSS: A Solution in search of a Problem Evens and Reddy (2002)
And on the other side is (are?) a miriad of EU funded projects FLOSS, GPOSS, COSPA and more.