Dynamics of Open Source Movements

• Published date: 01 June 2014
• Author: Glenn Ellison,Susan Athey
• Date: 01 June 2014
• DOI: http://doi.org/10.1111/jems.12053

Dynamics of Open Source Movements

SUSAN ATHEY

Graduate School of Business,

Stanford University/Knight Management Center,

Stanford, CA 94305-7298

athey@gsb.stanford.edu

GLENN ELLISON

Department of Economics,

Massachusetts Institute of Technology,

50 Memorial Drive, Cambridge, MA 02142-1347

gellison@mit.edu

This paper considers a dynamic model of the evolution of open-source software projects, focusing

on the evolution of quality, contributing programmers, and users who contribute customer

support to other users. Programmers who have used open-source software (OSS) are motivated

by reciprocalaltruism to publish their own improvements. The evolution of the open-source project

depends on the form of the altruistic benefits: in a base case the project grows to a steady-state

size from any initial condition; whereas adding a need for customer support makes zero-quality

a locally absorbing state. We also analyze competition by commercial firms with OSS projects.

Optimal pricing policies again vary: in some cases the commercial firm will set low prices when

the open-source project is small; in other cases it mostly waits until the open-source project has

matured.

1. Introduction

Open source software (OSS) has many varieties, but common features are that code is

freely available and contributions are made by a diffuse set of programmers often work-

ing as volunteers. Well-known success stories include Linux, Apache, which dominates

the market for web servers, and PERL and PHP, which are leaders in scripting software.

But OSS is a much broader phenomenon: as of July 2012, SourceForge.net hosted 324,000

OSS projects developed by 3.4 million programmers. There is a great deal of diversity

in project characteristics and outcomes: projects aim to serve very different user bases;

have different internal organizations; and some have thrived while others have risen

and fallen (or never risen at all). There is also diversity in the relationships between com-

mercial firms and OSS. Some software companies compete directly with open source

projects in critical areas, for example, Windows Server competes with Apache, but it is

also common for firms to actively support open source projects.1

Weare grateful to the Toulouse Network for Information Technology and to the National Science Foundation

(grants SES-0550897 and SES-0351500) for financial support. We thank participants in the 2005 Toulouse

Network conference, in particular Josh Lerner and Jean Tirole, for helpful comments.

1. Lerner et al. (2006) report that the fraction of corporate contributors to open source projects ranges from

22% for the smallest projects to 44% for the largest. IBM is a particularly noted supporter of open source and

claims to have invested billions. A sample description (in an interview of James Stallings by Linuxplanet.com)

is:

C2014 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume23, Number 2, Summer 2014, 294–316

Dynamics of Open Source Movements 295

Though recent, the literature on OSS has developed rapidly. It now contains both

enlightening theoretical papers and convincing empirical analyses that present findings

derived from diverse methodologies. The largest part of this literature focuses on why

individual programmers join open-source projects and contains evidence for the im-

portance of both economic career-concern and noneconomic intrinsic motivations.2A

smaller literature presents insights on competition between open-source and traditional

software products.3

Our paper departs from most of the literature in a few ways. First, we conceptualize

an OSS project as a community of programmers within which reciprocally altruistic

preferences are a driving force. Second, our analysis differs in our focus on the dynamic

life cycle as projects take off, grow, and/or decline. Third, some of the specific questions

we examine differ: why programmers go to the effort to commit code to the project;

how projects are affected by competition with for-profit firms; and how development is

affected by the importance of customer support.

Our formal model assumes that at each point in time the open-source product

can meet a fraction of the “needs” of a community. We refer to the fraction of needs

that a product meets as its “quality” and assume that quality naturally depreciates

over time (perhaps because new needs such as compatibility with new hardware and

software become relevant). Quality increases if these declines are more than offset by

improvements made by a community of programmers. We assume that programmers

are motivated both by their own needs and by a form of reciprocal altruism akin to that

in Akerlof (1982) and Rabin (1993): altruistic feelings are activated when a programmer

benefits from using the product (which is more likely when the quality is high), and

then disappear at some random future time. The model is therefore at its core a dynamic

system with two important state variables: the current quality of the software product;

and the size of the set of programmers who currently have altruistic feelings toward the

community.

Our choice to focus on reciprocal altruism is influenced by several descriptive

papers. Shah (2006), for example, reports that individuals typically first become involved

with an open source community when they have a need that they meet by using the

software. She reports that the most commonly cited reason for remaining involved for

at least a brief period is a sense of reciprocity, for example, “Others helped me, so

I should help them,” and that participation becomes a hobby for a a much smaller

number of deeply involved participants. The relevance of such altruistic preferences is

also supported by surveys which note that programmers cite such motivations along

with other intrinsic (enjoyment of intellectual stimulation) and extrinsic (career concerns)

motivations.4We do not wish to deny the importance of these other factors. Instead, as

LP: I’d like to ask you a few things about IBM’s Linux Technology Center. You have about 250

people working there. Can you tell me exactly what they’re doing with their time? Are they

helping customers or developing the kernel, or doing other work?

Stallings: They’re making contributions. Their full time job is making contributions to the kernel.

That’s it. They don’t have another job sweeping the floor or working on Websphereor anything

like that.

2. See, among many others Lerner and Tirole(2002) and Johnson (2002) for theoretical analyses, Fershtman

and Gandal (2007) for empirical evidence on extrinsic motivations andd Lakhani and Wolf (2003) for survey

evidence.

3. Two noteworthy papers here are Casadesus-Masanell and Ghemawat (2006) and Economides and

Katsamakas (2006).

4. See Ghosh et al. (2002), Hertel et al. (2003), and Lakhani and Wolf (2005).