I am pretty sure the function of the recurve is to obviate the stress on the string when you release the arrow and it returns to the rest position. In a long bow the string experiences a shock at the limb tips where it loops around the bow at the top and bottom. On a recurve this stress is spread out over the greater area where it lies against the bow. Additionally, as Danger alluded, it allows you to draw the bow a greater distance than an un-recurved bow of the same length.
Compound bows are the ones with the pulleys. I'm sure mgb-phys meant to say composite bows.
Regardless, the advantage of composite bows is not that they're stiffer. Monocoque wooden bows can be plenty stout: difficult to draw. Composite bows obviate the problem of breakage. A bow made only of wood has a limited range beyond which you can't pull it without breaking the wood and this is less than the average archer can pull a bow. A moment of unguarded excitement in battle and you could be holding two pieces. Sinew or rawhide, or whatever, glued to the front of a bow prevents the little initial cracks from which a big break grows.
I have a composite longbow made of wood faced with some unknown plastic. I can pull it as far as I like. However, I hate using it because the string comes to an abrupt, jarring halt when released. I have three recurves all of different lengths, and I prefer any of them to the longbow because the string comes to rest more gently, hitting the limbs of the bow before the tension gets to the limb tips.
Unfortunately, I have never heard that a recurve has a velocity advantage over a longbow. The only reason I can think of in support of it is that the limbs of a recurve are generally shorter and would have less inertia to overcome when the string is released. When they tried to build a working model of Da Vinci's giant crossbow, they found out that the limbs were too massive to spring forward fast enough to give the projectile any meaningful velocity despite the great amount of energy stored in them. So, I don't think it is the recurve itself, but the shorter limbs, that might give a recurve a velocity advantage, if it actually has one.