content on page A can have a negative effect on the rankings of content on page B with Panda
I noticed the same thing right at Panda 1.0. The negative effect seems to be related to "click distance" in some way. I really can't pin the "decay rate" down so far - and after all it's more important to fix the most hurting pages and regain some traffic.
However, I also agree with this thread - Panda is a "score" or an algorithm factor, and not a penalty. Using the word penalty to describe any ranking loss can lead into confused thinking.