Mega: I don't think your inverse square calculations directly apply to MH's in hoods, especially when parabolic type reflectors are involved. The inverse square law is for a point source in 3d space (i.e. a star). Suppose you have the entire insides of your hood made of a "prefect" reflective material. If you put a certain amount of light into the hood, that light will have no place to go but into the tank below. In this case you could make your hood as tall as you want and it wouldn't affect how much light goes into the tank. This would be like fiber optic cables, that carry light for miles with little reduction in intensity (certainly not an inverse square law situation).
I do think raising the lamps will affect distribution of the light. Suppose you have a bulb very close to the surface. The light directly under it will be high, and the light to the sides of it will be relatively low. Raising the light will re-distribute the light somewhat, and the previously high region will be lower, and the previously low region will be higher. But I think the total light into the tank will be about the same (the extra air space will have negligible effect on intensity, unless you live in Houston). Losses out the ends of the reflector will increase as the light hits the sides of the hood and not the water, but this is still not an inverse square phenomenon. You could maybe approximate it geometrically though.
At night, take a strong flashlight and focus the beam on a coral with the flashlight right up to the side glass. That one coral is getting say 100lux (or par, or uE, or whatever), and say the coral next to it is getting 0. The tank as a whole is getting 100lux. Now move the flashlight back a ways. The cone of light from the flashlight will spread and now that first coral may only be getting 50lux, but the one next to it is now getting some light too. But the total light going into the tank will still be 100, as long as the cone of light from the flashlight is all hitting the tank. Are you seeing what I'm trying to say? Look at it from an energy balance perspective. If you raise the lamp and you loose all your light, where did it go? If you are flooding the whole room with light, then yes, your tank has less, but if 80-90% is pointed at the tank, then it's not going to magically vanish.
I would like to see someone with a PAR meter do some experiments on this. However, I'm certain that with the geometries and reflectors involved in typical hood installations, the inverse square equation is an inadequate model at best.