The most important finding in this research, confirming a major theme in the literature, is that leadership and value are critical to the success of open-source collaborations in the public sector. Collaborations with a strong leadership structure, and more importantly a single leader who is persistent, passionate and willing to spend a great deal of time maintaining and improving the organization are much more likely to succeed. Value is also a critical component, and requires that efforts meet the wants and needs of members and clients, whether they be in the form of software, documentation, research or even policy advocacy. (24)
Just like in any vibrant, efficient organization. The difference here, however, is government involvement in open source tends to have "multiplying" effects; that is, the more government body X contributes to a project, the more likely it will be that other governments will chip in directly and indirectly by starting their own projects that government body X can then borrow. In short, a little leadership can go a long way.
Peter Quinn, former CIO of the Commonwealth of Massachusetts, was fond of saying that all state governments needed to be able to issue license plates, collect parking ticket fees, etc. There were few good reasons to not collaborate on such common bureaucratic infrastructure or, rather, the code that powers it.
This is why governments should get involved in things like the Government Open Code Collaborative. Governments don't need to rely on vendors for much of the code they use - they can write their own and share it amongst themselves. This strikes me as a much smarter, more efficient way to expend my tax dollars than on proprietary licenses.