Hi everyone, I'm Andrea Rapisarda, one of the authors of the paper on the Peter Principle
and I would like to note that we have published a second paper on the same argument, where
we find that our previous results are very robust and do not depend on the size of the
organization or on its specific structure. Moreover even a small amount of randomness
is very useful to improve efficiency. Of course you can continue to maintain incentives for the best
by giving rewards to those who are excellent (like an increase in their salary
or in their periods of vacation ) but you should not change the task where they are excellent!!

Our second paper is the following
"Efficient promotion stategies in hierarchical organizations"
A. Pluchino, A. Rapisarda, C. Garofalo, Physica A 390 (2011) 3496
You can find more info here http://www.dfa.unict.it/home/rapisarda/

Thanks for the attention
Andrea Rapisarda