A couple things need to be addressed here. First, the flow rate is a product of the sum of the total restrictions in a system. If you have a massive restriction in your system, say a Whitewater or Cascade, then adding more smaller restrictions in the cooling loop will lower your flowrate even more.
Fluid in a closed loop actually does accelerate and decelerate depending on the cross sectional area of the tube it flows through. Take a garden hose, for instance. The cross sectional area of the hose is constant throughout its length, so the water travels at the same velocity through it. But when you place a nozzle (or just your finger) over the end of it, the water leaves the hose with a much higher velocity. The rest of the water in the hose still travels at the same velocity it did before, though. What remains constant is the flow rate, the volume of fluid per second.
The advantage of larger diameter tubing comes in when you take into account the friction that is caused by the interaction between the water and the walls of the tube. Larger diameter tubes have less surface area exposed per unit volume of water, which results in less restriction. Less restriction in the cooling loop means that the centrifugal pump can operate more efficiently, which means a higher flowrate. The water in a 1/2" diameter system can actually have a lower velocity than a comparable 3/8" system while still having a much higher flowrate. Since flowrate is far more important than water velocity, especially with modern die impingement blocks, 1/2" tubing will inevitably outperform 3/8" tubing.
This is pretty much the CliffNotes version of things. If you really want to get into the nitty gritty aspects of one-dimensional flow, a good place to start is cpufan's post here:
http://www.ocforums.com/showthread.php?p=2815392#post2815392 . If you read that and still want more, then you're definitely beyond the scope of these forums - I'd recommend a fluid dynamics textbook.
Hope this helps!
