4.5 MM is what teams are paying per win, that's not what a win is worth, there's a pretty big difference there. And because of that difference, what teams as a whole are paying per win doesn't mean shit, because since that 4.5 MM figure was divined only by looking at free agents, we can be sure that almost each and every one of them is overpaid, or at the very least over valued. This is what's called the "winner's curse" in common value auctions, with MLB Free Agency being a rather nice analogue.
Here you go again with this divisional approach to changing environments, and it hasn't gotten any less stupid. Look, I know that you think because teams play an unbalanced schedule with regards to intra-divisional and intra-league games, it is wise to approach environment changes not on a league-to-league basis, but on a division-to-division one. But look, there's a reason why people aren't doing this on a grandiose scale: it's overstating the situation, by like, a lot, because who is to say that the games a team plays against a really good divisional opponent would not be played or at least equaled by a different series or set of series against opponents throughout the rest of the league? Regardless of division, teams will play roughly the equivalent of their respective league averages throughout the season, because while the schedule may be unbalanced as far as divisional play goes, there are still a ton of games played.
And even if some tiny shred of a reliable (this last bit is highly unlikely, but I'll toss you this bone just to show how incredibly asinine your overall point is) difference were to be found with regards to "toughness" of competition played against, there is absolutely no reason to believe that said difference would be any greater (if it is at all) than the change in opponent ability seen when changing leagues. And, for the fourth time this month: the AL and NL have gotten extremely close to one another as far as ability goes, with almost no difference arising in the 2010 season.
Oh yeah, and Carlos Pena is going to be now four years removed from what was by far the best season of his career, has been trending downward ever since, and that one fantastic season also occurred at age 29, right around his peak. He's a big, lumbering, power-hitting first baseman, and those don't tend to age all that well, so while we can reasonably expect his BABIP to increase next season, and perhaps his other statistics in lock-step, it's an awful lot to hope for another well-above-average season from him, and his career averages (which are helped out immensely by those two seasons that came completely out of nowhere) are probably the best one can reasonably hope for. Oh yeah, and for the record, his 162-game WAR average is 1.9, or to put it differently, decidedly not worth 9 MM a year by your own asinine judging system.