Document Type
Article
Publication Date
1-1-2015
Abstract
One of the puzzling aspects of open-source software (OSS) development is its public good nature. Individual developers contribute to developing the software, but do not hold the copyright to appropriate its value. This raises questions regarding motives behind such effort. We provide an integrated model of developers' incentives to describe OSS development and compare restrictive OSS licenses that force all modifications to be kept open with non-restrictive OSS licenses that allow proprietary ownership of modified works. Different incentives govern effort provision at different stages of the software development process. We show that open-source licenses can provide socially valuable software when a proprietary license fails to do so. We also show that restrictive OSS licenses generate greater effort provision in the design stage of software development relative to non-restrictive licenses. Endogenizing licensing choice, we find that a project leader chooses a non-restrictive OSS license if reputational concerns drive developers' incentives, a proprietary license when there is a large population of users in the market and a restrictive OSS license if user population is small but reputational benefit is high. Our results resonate well with empirical findings and suggest additional testable implications about the relationship between licensing and software project characteristics. Finally, we also find that the market under-provides restrictive OSS licenses relative to the efficient level, suggesting the need for subsidizing restrictive licenses in some cases.
DOI
10.1515/bejeap-2014-0007
MSU Digital Commons Citation
Atal, Vidya and Shankar, Kameshwari, "Developers' Incentives and Open-Source Software Licensing: GPL vs BSD" (2015). Department of Economics Faculty Scholarship and Creative Works. 13.
https://digitalcommons.montclair.edu/economics-facpubs/13
Published Citation
Atal, V., & Shankar, K. (2015). Developers’ incentives and open-source software licensing: GPL vs BSD. The BE Journal of Economic Analysis & Policy, 15(3), 1381-1416.