By 0.2. Given the many flaws of PER (I hope you can use Google to find this information, but I'm doubtful), I would be hesitant to rely entirely on that and ignore other stats, but you seem lazy.
So not only do you rely heavily (and seemingly exclusively, given the frequency you mention it) on PER, but you think (1) MJ's Wizards seasons dragged down his career PER by 4 (his Bulls' PER was 29.1, not 31.9), and (2) you ignore that MJ had three years in college to develop while LeBron had the worst season of his career coming in straight out of HS (where his PER was worse than any MJ Wizards season). That's not even bad logic; it's a failure to use logic.
Also isn't PER normalized to the league each year, meaning that when the league is worse outliers tend to be more extreme? The mid to late 1990s were unquestionably a down time for the league due to expansion and a series of terrible drafts. Normally it's not a nit I would pick, but when you're hanging your hat on a 0.2 difference maybe it's something you should think about. Thanks, I'll take my answer off the air.