Zubrin's panning of space solar power in Entering Space

I just finished reading Zubrin's "Entering Space", and was somewhat
disturbed by his economic analysis of space solar power. Not that I
found it off the mark for what he analyzed - but that it seemed to
ignore what seems like an obvious alternative. Too obvious - there
must be something I'm missing here.

Can someone explain why, given Zubrin's arguments about launch costs
for the commonly conceived solar to microwave power sat, it isn't far
more effective to simply put large mirrors up at GEO to light up solar
power farms on Earth at night? That should about double the power
output of a solar power farm, without greatly increasing operating
costs. The mirrors could also be used during early and late daylight
hours, to augment lighting of the solar farm - eliminating most of the
value of expensive sun-tracking hardware.

My back of the envelope estimate (including LEO launch costs coming
down to $2000/kg due to frequent launches of solar mirrors - a side
benefit) comes in well under 2 cents per KW-hr for the power added by
the space mirrors. I assumed that the mirror could either solar sail
up to GEO, or that there'll be enough LEO to GEO traffic to justify an
inexpensive solar powered tug.