Open Source Software, Competition and Innovation

Abstract

The entry and success of open source software (OSS), for example, Linux's entry into the operating systems market, has fundamentally changed industry structures in the software business. In this paper we explore the process of OSS innovation and highlight the impact of increased competition and different cost structures on innovative activity in the industry, which has been neglected in the literature thus far. In a simple model, we formalize the innovation impact of OSS entry by examining a change in market structure from monopoly to duopoly under the assumption that software producers compete in technology rather than price or quantities. The model takes into account development costs and total cost of ownership, whereby the latter captures items such as network externalities. The paper identifies a pro‐innovative effect of both intra‐OSS and extra‐OSS competition.

Keywords:

• Open source software
• innovation
• strategic interaction

Acknowledgements

The authors thank Roger Sherman and Zhentang Zhang for valuable comments. Previous versions of this paper have been presented at the International Industrial Organization Conference (IIOC) at Northwestern University, Chicago, the Annual Meeting of the Canadian Economic Association (CEA), Toronto and the annual conference of the European Association for Research in Industrial Economics (EARIE) at Humboldt University, Berlin. The usual disclaimer applies.

Notes

1. For a review of the literature see Rossi (2006).

2. Within the formal model we do not examine if the OSS business model as such is pro‐innovative, see also the discussion in Section 2.

3. An example of a fork that emerged because a sub‐group of programmers felt that new contributions and patches had not been released quickly enough is “The Community OpenORB Project” (cf. http://openorb.sourceforge.net/).

4. In fact, this phenomenon of user‐developers is by no means restricted to the realm of OSS. See, for example, Franke and Shah (2003), who analyse cases of user communities that support innovation activities for sports‐related products.

5. Famous examples of software which were forked are: Berkeley Software Distribution (NetBSD), forked into OpenBSD and FreeBSD, and GNU Compiler Collection (GCC) forked into EGCS.

6. In particular, the emergence of no‐cost competitors in the software market has altered the type of competition. In such market segments neither price nor quantity are important competition axes but technology became the crucial determinant (see Bresnahan and Greenstein, 1999; Shy, 2001).

7. In fact, also the price of most proprietary software is often inessential from a consumer's point of view. The majority of software is sold as pre‐installed, thus its ability, reliability, compatibility, etc.—in short its technological content—drives the consumer's decision, while the price is a matter between the soft‐ and hardware producer.

8. It is always possible to overcome network externalities by corresponding investments, for example, paying someone to program a missing piece of application software. Of course, the required investment costs might be exorbitantly high. However, network externalities can be interpreted as a component of the TCO.

9. Shy (1996) similarly argues that not every newly developed technology is adopted leading to the same conclusion.

10. As no single software product is at the technological frontier in terms of every one of its technological dimensions, the technological level of any specific piece of software must always be lower than the global technological level.

11. Even though we use the term “firm” here, it is applied in the sense of agency, in that it may either refer to traditional proprietary software firms or OSS developer communities. Thus we are not referring to for‐profit firms that base their business on OSS.

12. For example, adaptation of the software in terms of programming interfaces or adaptation to users, or training of users is usually more expensive if the software has a small user base.

13. Network effects determine the substitutability as it might be difficult to change from a software with a large user base to one with a small one. Because one might loose positive network effects, for example, changing from MS Office to KOffice might prevent or hamper the exchange of documents with outsiders.

14. Or put differently, if both firms opt for technological stand‐still, global developments will eventually reduce the dissemination of the two products to zero.

15. Obviously all other properties laid out in Lemma 1 directly extend to the asymmetric case.

16. The intuition for the latter effect is, of course, that a higher k_b forces firm b to offer a higher technology level such as to compensate users for the TCO, and this in turn triggers firm a via the strategic interaction to respond with a higher technology level itself.

17. Notice that Proposition 1 is purposefully derived under the assumptions c_a  =  c_b  =  c and k_a  =  k_b  =  k, such that our results stem solely from the strategic interaction of firms, that is, competition, rather than from the possible cost advantages that an OSS competitor may or may not have.