Having read the paper, and taking the time to think about the whole 'signaling' explanation for open source, I'd say that in addition to ego and joy, signaling is a major motivator for contributing to open source software. While many programmers may contribute to any arbitrary project because of their interest or dependence on the software, the most effective way to signal your skill as a developer would be to contribute to software that is either a) used by the prospective employer to support their business, or b) developed by the employer himself (likely to support their business).

However, this theory doesn't take into account that one may contribute to an OSS project, because an employer is paying them to do this in the first place. Both employers and employees participate in OSS. Employees want to demonstrate their skill so that they would get employed, while employers want to demonstrate their organization's engineering talent to attract even more talent. The paper seems to assume that OSS signaling is simplex, whereas it is actually duplex. But beyond signalling, an organization open sourcing software can be seen as a way to encourage standardization of that software (Android, WebKit, V8, and in the future maybe Dart) instead of a competitor's potentially closed alternative (which might be more expensive to deploy/port/debug and so on).

Sun Microsystems' open sourcing of Solars and JVM can be seen as an attempt attract more talent to the company (which at that time was fairly unpopular, especially when compared to Google which was an open source darling).