Slack

Many years ago, I read a great book by Tom DeMarco called Slack. It's been a while since I read it and I've read lots of other things in the interim, but if I remember correctly, the basic thesis of the book is that a completely "efficient" organization (in terms of resource allocation) can actually be less effective. This is especially true in "knowledge work" -- situations where the work is not about producing a tangible widget.

Making sure that people are always busy is a sure-fire way to ensure that the organization will not be able to cope with sudden fluctuations in workload without disruption. In other words, there needs to be a certain amount of slack in the system to allow it to absorb variety. Employees will be more productive in the long run if they have time to think, investigate new ideas, and learn new skills. 

If workers are already at 100% capacity, there is no way to take advantage of new ideas or new opportunities as they arise, much less to handle a sudden, temporary increase in demand. Trying to add new work to a system already at capacity means that something has to give (in terms of some work not getting done or, if people are forced to work extensive overtime, a decline in morale). On the other hand, if workers have some slack in their schedule, they can quickly ramp up for those occasions which demand higher output (assuming that it is only temporary). 

A great analogy (I can't remember if this was from the book itself) is that of a highway. Building a highway that no one uses is obviously not a good use of resources (i.e., too much slack), but filling every square inch of the highway with cars, though it may be an "efficient" use of space, will be counterproductive -- everyone will be at a standstill.