The difficulty I see with that conclusion is that so many sites with an online history of any decent timeframe have reciprocal links. Also, I know of one case where the -950 ranking fell away after adding a new link - and nothing else was changed or removed.
Establishing cause and effect (rather than just statistical correlation) is a tricky process. When it comes to reciprocal links, there are so very many sites that have them and are not hit by the -950, that noticing this as a common factor may well be correlation but not causation. There are just too many counter examples, in other words.
I'll bet most of the affected sites have a <body> element, too, if you see what I mean by that extreme example.