Dollar per mW it is, but green is far brighter than blue, mW to mW.
1) Are we using input wattage into the diode for comparison? If so, then efficacy may be a big explanation for a green advantage over blue.
Speaking of this, I guessed that the Arctic blue laser might be in the range of about 150 lumens per watt. This value is important in calculating apparent magnitude of the light from space, of course, Is this value a fair guess?
2) I suspect that our eyes ability to see color is best modeled using photon counts rather than spectral energy levels. Blue photons require more energy to produce than do the other colors of the spectrum, ignoring violet. [E =hf.] This would give green -- red would have an even greater advantage here -- an advantage. But, I suspect, the lumens per watt efficacy rating would reflect this issue.
3) Is there something special about 445nm that reduces our reception to it that is not seen in the spectral sensitivity distributions for our color cones? This seems unlikely.
4) Color contrast? This can cause us to see one color better given certain background colors. Yet, given a black background of a dark sky, this shouldn't matter, though a green light might stand out nicer than a blue one if shone into a blue sky.
5) Scattering. Though I'm about to appear to contradict something I just posted in a prior post, a green beam may, perhaps, appear brighter due to the better receptivity to green light, not just the cones but the rods also. [I think I've read that the rods do contribute, though only slightly, to color determination even at photopic levels of brightness, and the rods like green best, as do the cones, but with separate peaks.] In other words, what few photons we gain by Rayleigh scattering of "blue" photons we more than make up for with the improved receptiveness of "green" cones. An analogy might be in comparing the weight disadvantage of a bigger racing engine to the h.p. advantage, both must be taken into consideration, else most racing engines would be huge.
That's the old way of doing it, with an AOM (AcoustoOptic Modulator). I've built a voice-over-laser transceiver that uses direct amplitude modulation of the current going to the laser diode. It's only 12mW but could easily be scaled up to a full 2W quiescent output level with funds thrown at it. You could easy send AFSK (audio freq. shift keying) data through it. I've also built from scratch a CW (morse) over laser transceiver as well. It's only at 2.5mW output but the transmitter side could control any diode based laser of any power.
For a "pointer" without safety features, 5mW is the limit. If you add FDA required safety features there is no power limit. If you build it yourself, there is also no power limit, instead the legality falls on application and intent.
I would like to learn more about any requirements for class IV, or greater, laser safety. Do you have a thread on this or other reference?