The biggest advantages of building solar power capabilities on the moon are that you can make if from local materials (eventually), aren't limited in size or mass, and you aren't required to include things like attitute control and propulstion systems. All of these things could eventually lower the costs considerably. You could build a solar energy collector in modules that eventually would cover many square kilometers. Your transmission antenna array could be as big and heavy as needed to transmit megawatts (perhaps gigawatts) to the Earth. That energy might also be useful for other purposes, such as beaming electricity to other spacecraft. You wouldn't be taking up space in the geosynch belt, either.

However, there are some disadvantages to using the moon, too. First, the longer distance would result in higher propagation losses. Second, a power satellite in geosyn orbit stays over the same point on the equator, making it easy to aim the power to a fixed antenna farm on the ground. The same isn't true when beaming from the moon. From the lunar surface, the Earth does rotate, so you'd need multiple antenna farms on the Earth to receive the energy or to stop transmitting for something like 16 hours a day. The first option would make the ground infrastructure more expensive (but you could sell energy to multiple countries) or you'd have a lower return on investment due to reduced utilization. There are a few places on the moon that are constantly in sunlight. That will be some valuable real estate some day. Would we necessarily want to take up all of that real estate with solar collectors? Perhaps not.