I recall a study into the new manager/honeymoon period a few years ago - basically it examined whether results improved after installing a new manager.
Obviously I can't be assed to search for it on the net now, but surprisingly they concluded No - on average there is no honeymoon period where things pick up results wise.
I'm sure encyclopaedia smeetonnia will remember the said study

Edit - OK, I can be arsed. Wow this study is 10 years old, how come I remember this but barely know what day it is!
http://www.le.ac.uk/sp/sf-review/97-98/97article7.html
Using this procedure, we find that the average improvement recorded by the teams which did not terminate after the point of comparison is, in fact greater than that recorded by the teams which did terminate the manager, no matter whether the comparisons are made over 6, 12 or 18 matches before and after. In other words, the research shows that the average short term recovery in results produced by termination is not as great as that which would be expected in the normal course of events from a club experiencing comparable results which retains its manager.