A case for outsourcing (and its slippery slope)

I think there is a case for outsourcing software development in a software development company: when you have so much work in a temporary basis that you can do it yourself. I think I can't strength temporary enough: if it isn't temporary, you are outsourcing your supposedly main value-creation activity, and getting yourself in a problem (read the previous posts once again).
However, as stated in the title, there is a slippery slope. You may presume that it is temporary, and suddenly, it becomes a permanent assignment. The best people to do the job are the ones who are already doing it, so you keep outsourcing to them. Voilà, you are outsourcing a permanent position, depending on your contractor, etc.
How often this slippery slope happens? Only 90% of succesful outsourcings; if it is succesful, you don't want to change, do you?
On the bright side, if your outsourcing failed miserably, as they usually do, you don't have to worry about this