I'm a houdini-boy (Of course I do test its evaluation and play out many moves to see if it doesn't suddenly find something, but I find myself agreeing with it here), so bear with me, but it can't find much advantage at all for white in the QGA surprisingly. Same for many lines of the QGD, the main exception being the exchange variation because the static pawn structure makes it more difficult for black to make freeing exchanges.
What's funny is that black will even play moves like ...a6 (even after white has played Bxc4 for instance), which when I first saw it, seemed ridiculous and cheeky to the extreme. In lines where white plays e3, white has an extra center pawn, but his queen's bishop is blocked in, and the pawns simply don't make that many threats, to the point where black even has time to move pawns on the queenside with ...a6, ...b5, then bring the bishop to the long diagonal. He will generally get in ...c5 very quickly if he desires, always giving him the option of trading bishop pawn for center pawn, like what black did in reverse when he played ...dxc4. White in my opinion should fight for an advantage in more modest ways, such as thinking of a pawn breakthrough with d5, or e4 followed by d5. But it's not so easy to achieve and it often just leads to lots of exchanges that don't really net white much progress.
White can also be more aggressive with an e2-e4 push early on, rather than e2-e3, but even in these lines, white's pawns on e4 and d4 are just pawns -- pawns in themselves are not going to create threats against the position, meanwhile they can be attacked.
So I think the QGA is in effect as good as the QGD, even if it may look less natural. Black's apparent concession of the center is surprisingly not bad at all -- it is rather difficult for white to make anything out of his central majority. It should be noted that in the QGA, due to the open lines, black's pieces are actually less bottled up than in the QGD; for instance, the c8 bishop usually finds a good home on b7 pretty easily. Black just has to be careful about possible central pushes for white like d4-d5 or e4-e5; I think that if black plays accurately, those ideas will not give him any problems.
The last question then for you is taste. You can try it out, see if you like it, because it is totally sound and arguably not worse than black's main defenses against d4. It often does tend to be more of an equalizing opening, but I had always been fascinated about how black was able to play those slow ...a6 moves and totally get away with them. But yes, it's generally more for solid players.
Note that when I say QGA I mainly mean it as a strategic formation, not the declaration that black is going to take and hold the pawn. There are some strange lines where black really can hold the pawn (e.g., 1 d4 d5 2 c4 dxc4 3 Nf3 c6 followed by ...Be6 might hold it), but the moves he makes are so awkward that when I face it with white I just don't care that he keeps it. If it's that important to him I'll let him lose coordination and he can have his ill-gotten gains.
That doesn't mean it's necessarily bad though! I think white is better, but black's position is still solid there, and it might freak out white players who are told that black automatically loses when he tries to hold the pawn. There is still a game to be played.
I really enjoy QBA as White... but I avoid it all together with Kings Indian in QP game as black.