Istvan makes a number of good points and Matt Cutts has certainly alluded to a loss of some link juice when using 301's, although Google's official line is that there is no loss. I'd not seen the 15% number before, which is certainly high enough to be discernible above the 'noise'.
I support his contention that, in terms of getting existing inbound links repointed, it's best to focus on the few high value links and then look for new links driven by quality content. This has the double benefit of cleansing some of the now-devalued link types, whilst appealing to Google's measurable preference for 'fresh' links.
