Well, we do know that the CFL spectrum is horrible: Very high narrow-band blue, green, and orange peaks. I guess they are using just three phosphors which is good enough to look reasonably good to humans. You can get CFLs with Color Rendering Index values as high as 80 or so. But the plants seem to like it, even if it misses all of the chlorophyll absorption peaks.
You can also look at the CRI of HPS, which everybody says is good for flowering: Usually 25 or so but some grow lights use ceramic radiators to get CRI up to 85. Usually hardly any blue, but plants seem to like it.
Or look at MH: Hardly any red. I suppose you could make a pretty high quality light using a mix of HPS and MH bulbs. Each would fill in parts of the spectrum missing in the other’s light.
All these different light spectra somehow work. Probably because evolution finds solution that are good enough but may not expend the effort to use every frequency of light available. So there is no reason to think that reproducing sunlight perfectly will work any better than some spectrum with holes in it. Look at the PAR spectrum and then look at the sunlight spectrum. VERY different. I think the PAR spectrum should really be the goal.