I’m putting together some classes to handle common tasks like mouse click and drag, navigation, etc. For the sake of modularity I’ve given each its own collision traverser. Is that a bad thing? Meaning, will that incur a significant performance hit versus having everything use one traverser through e.g. base.cTrav?
There is some performance benefit to combining multiple collision passes into a single collision traverser. The amount of benefit depends on the commonality between the different traversers; if the different traversers are colliding with the same objects in the scene graph, there is more benefit than if they are unrelated objects.
For the precise cost/benefit in your specific situation, nothing beats making an actual measurement. 
David