Lots of good information in this thread, but I didn't see any reply that
explains why low-e coatings have a direct impact on the overall u-value
of the window. And I saw some incorrect discussion of the radiant heat
transfer impact of low-e coatings. Here's my understanding.

Glass is opaque to infrared radiation (look at the detailed glass layer
properties in most simulation software and the infrared transmittance
will be zero).

The low-e coating is always on an inner face of the window unit, so it
only impacts the radiant exchange between the glass layers (or in the
case of some windows, between the glass and an inner plastic film
layer). Low-e coatings do not participate in the interior room radiant
exchange or the exterior radiant exchange with the surroundings. So you
don't need to be concerned about any double-counting of radiant effects.

U-value is the overall heat transfer through the entire window unit
(with no sun present) divided by the temperature difference. The rated
u-value includes the impact of the inside and outside convection
coefficients. As was said before, the heat transfer through the window
includes conduction through the glass and gas layers, but it also
includes the radiant heat transfer from pane to pane across the gas
layer(s). For a double-pane window, there's conduction through a pane
of glass, then conduction *and radiation* through the gas layer, then
conduction through the second pane of glass. This is why the low-e
coating has a direct impact on the u-value, regardless of how interior
and exterior radiant exchange is modeled.

This FSEC page has a nice diagram and explanation of window coatings.

http://www.fsec.ucf.edu/en/consumer/buildings/basics/windows/how/solar_gain.htm

Mike