Well I have been puzzling over this problem. I have come to the conclusion that indeed you don't need to pump huge volume of coolant. The problem is mostly about airflow.

I can't seem to find radiator efficiency data, but I was able to find some good data on oil coolers (really the same problem)

http://www.oilcoolers.com/LCHX%20Specifications.pdf
If you plow through this one thing you will see is that heat rejection is not strongly dependent on oil flow. For example if you look at the 8406R radiator you see that for an air flow of 20lbs/min the heat rejection varies from 200 BTU/min to 320 BTU/min

Over a 5 to one variation in the oil flow rate. So pump the oil as fast as you want , it doesn't make much difference. The air flow is what does the magic.

I suppose another way of thinking about it would be if we pump super fast so that the water is absolutely flying through the radiator then there would be almost no loss of temp of the water. You would have a situation where a lot of water lost a little bit of temp in order to shed the required heat. In this situation the entire radiator would be a one temp since the water is only dropping a tiny bit in temp. Since the radiator is at a higher average temp it can shed more heat.

The other situation is when you are pumping slowly so that the coolant loses a lot of temperature through the radiator. Somewhat less heat is shed, since the back end of the radiator is cooler it is somewhat less effective a dumping heat so in total less heat is removed.

So the question is in the automotive situation where are you with respect to this point of diminishing returns?