Master of Public Policy
Master of Public Affairs
Master of Development Practice
Doctoral Program
Global and Executive Programs
Undergraduate Programs
Diversity, Equity, Inclusion
Student Groups
Alumni
Directories
Faculty
Publications
Working Papers
Centers
News Center
Policy Initiatives
For Students
For Employers
Career Events
Employment Statistics
Client-based Projects
About GSPP
Events
Make a Gift
Find a Program
Find a Faculty Member
Find a Center
The New (Commercial) Open Source: Does it Really Improve Social Welfare?
Working Paper: GSPP10-001 (January 2010)
Abstract
The number of open source (“OS”) software projects has grown exponentially for at least a decade. Unlike early open source projects, much of this growth has been funded by commercial firms that expect to earn a profit on their investment. Typically, firms do this by selling bundles that contain both OS software and proprietary goods (e.g. cell phones, applications programs) and services (custom software). We present a general two-stage Cournot model in which arbitrary numbers of competing OS and closed source (“CS”) firms decide how much software to create in Stage 1 and how many bundles to supply in Stage 2. We find that the amount of OS software delivered depends on (a) the degree of substitutability between proprietary products, (b) the number of OS and CS firms competing in the market, and (c) the savings available to OS firms from cost-sharing. However, code-sharing also guarantees that no OS firm can offer better software than any other OS firm. This suppresses quality competition between OS firms and restricts their output much as an agreement to suppress competition on quality would.
Competition from CS firms weakens this quality-cartel effect, thus mixed industries often offer higher welfare. We find that Pure-OS (Pure-CS) markets are sometimes stable against CS (OS) entry so that the required OS/CS state never occurs. Even where mixed OS/CS industries do exist, moreover, the proportion of OS firms needed to stabilize the market against entry is almost always much larger than the target ratio required to optimize welfare. We examine various policy options for addressing this imbalance with tax policy, funding of OS development, and procurement preferences. We find that the first-best solution in our model is to tax OS firms and grant tax breaks to CS firms. Conversely, government policies that fund OS development or establish procurement preferences for OS software actually increase the gap between desired and actual OS/CS ratios still further. Despite this, funding OS development can still improve welfare by boosting total (private government) OS investment above the levels that a private cartel would deliver.