Example - A college can pay $75k per year for an Angel or Blackboard license, and host it locally (or contract the hosting out to Angel/BB). Or, they can adopt a F/OSS solution like Sakai, and instead of paying $75k/yr to a corporation outside of the area they can pay $75k/yr for a programmer to maintain and enhance Sakai for their needs. Dollar costs are the same. However, by hiring the programmer, that improves the local economy and keeps that money local, as opposed to sending it out of area/out of state/etc.
Except that not both options are equally efficient. The 75k spent on the proprietary product will pay support from developers that are experts in the product (after all, they built it and should be actively maintaining it), while the 75k spent on the FOSS solution will go to a local programmer who is most likely not as expert as product's authors.
There are certain cases where 'paying' for the FOSS solution could be more efficient: when the product is sufficiently popular that it is easy to find an expert to support it, or there is a for-profit company with sufficient expertise (maybe one that employs the main contributors of the project) that does provide support for the FOSS product. In these cases the open source product has a bit of an advantage in that it would be slightly more difficult for the maintainer/vendor to pull lock-in tricks on you.
Of course this assumes that the commercial and the open source products are of equal quality.