Let's try using a few more advanced stats-- offensive rating and offensive win shares. These statistics are based on field goals made, field goals missed, assists, offensive rebounds, and turnovers and are meant to serve as a catch-all offensive statistic that factors in efficiency. I've also included PER and TS%. (Note that 2011 numbers for Perkins are small sample size and should probably be ignored.)
Kendrick PerkinsCareer offensive rating: 103
Offensive rating, last three years (2010-11 first): 97, 105, 105
Offensive Win Shares, last three years: -0.1, 1.4, 1.1
Career PER: 12.9
PER (last three years): 10.2, 15.0, 13.2
TS% (career): .584
TS% (last three): .568, .613, .591
Nenad KrsticCareer offensive rating: 105
Offensive rating, last three years: 113, 111, 104
Offensive Win Shares, last three years: 1.7, 2.0, 0.6
Career PER: 13.6
PER (last three years): 13.0, 13.7, 13.6
TS% (career): .531
TS% (last three): .546, .527, .504
Based on offensive rating/offensive WS (which takes into account turnovers-- Perk's biggest weakness), it certainly looks like Nenad has him beat by a fair bit (Perk maxed out at 105, Nenad is at 113 and 111 the last two years). PER slightly favors Perkins (but the difference is almost entirely due to defensive rebounding/blocks) and while TS% favors Perkins, it doesn't take into account aspects like floor-stretching, volume (Nenad has averaged a higher pts/36 min the last few years), and turnovers. Nenad turns the ball over roughly half as often as Perkins does-- and that's a pretty [dang] important weakness for the Celtics. That alone more than makes up for any shooting efficiency differences between them.
It's fair to question whether Nenad's offensive improvement makes up for what he gives away on defense and on the boards, but to me it's clear that Nenad is a fair bit better of an offensive player.
------------
Glossary
http://www.basketball-reference.com/about/glossary.htmlhttp://www.basketball-reference.com/about/ws.htmlhttp://en.wikipedia.org/wiki/Offensive_rating