We know no such thing. We know that wealth is distributed unequally (see, for example, the Gini coefficient). Whether or not an unequal distribution of wealth is unfair is a decidedly normative question - and beyond the purview of economists.
One of the many problems that I see is they assume that various attributes are independently normally distributed - which many studies have shown not to be the case. An individual's performance in a wide variety of seemingly unrelated disciplines tends to be highly correlated (http://en.wikipedia.org/wiki/General_intelligence_factor). Simply put, a person who is a great at one thing is also more likely to be good at other things.
Furthermore, there is no reason to expect the output elasticity of intellect to be anywhere near 1. Someone who is just the tiniest bit smarter than me may see a shortcut that they I don't that saves them a tremendous quantity of resources.
Study after study has shown that 'overpaid' people (sports stars, actors, etc) get paid a wage that is almost exactly equal to the marginal revenue they bring in - that is they are paid perfectly appropriately.