mine run at a power factor of .62, and pull 20VA. for a 15W bulb I'll happily take a watt or two penalty (it can't be in the ballast because it's plastic and not hot like the magnetic one was) to stop that horrible noise.

It's costing you 5 watts.
Which isn't bad. It's just when some one has a ton of lights like that.

I'm not quite sure how you come to that conclusion. real power isn't at 20W. The surface temp of the housing is lower with the electronic ballasts vs the magnetic ballasts with the same room temp, and light output appears the same.

Apparent power includes the power supply inefficiency.
That's why all my lumens per unit of power ratings usually includes apparent power not real power, when the power factor is less than ideal.

I don't get billed on apparent power though, I get billed on kWh. maybe if they took power factor into account it would matter?

But they do charge for apparent power.
The old electro mechanical meters will spin the same speed with a 140w 500va and a 500w 500va resistance load.
Some electronic meters I have seen collect "80% VA" and "50% VA" kwhs.
So yeah they do.
I had to do the experiments my self with an electromechanical meter because when I searched all I could find was people arguing over it and not actually testing it.

The power company should bill for apparent power as it loads the generating stations and the grid the same if not worse than true power.

Yes, they can measure power factor, but my poco doesn't bill residential customers for having a low power factor. it's been a long time since I've seen a mechanical meter around these parts.
for commercial customers:

With a lot of the junk coming out of China it's real easy to run well under a power factor of .85 since a lot of the gadgets are running power factors around .5 to .6 and those items include CFL, cheap LED replacements, cell phone charger, cheap florescent ballasts, laptop power supplies, lots of things.

Plus I would like to potentially go off grid some day. So I would be penalized all the time for low power factor.

I'd also like to go off grid (not a likelyhood for a while here) - but an inverter shouldn't draw more real power in DC than the real power output + 10% (or whatever efficiency the inverter is rated at) - it just won't be able to supply more VA than it's rated for.

This sums up Osram Sylvania's position nicely:
http://ceolas.net/Docs/Sylvania_on_CFL_Power_Factor.pdf
apparently .6 for a cheap electronic is much better than your average magnetic ballast with .25. not everything is worth the spend for high power factor.

This shows some power factor correction with capacitors:
http://www.ecospecifier.com/media/72...t%20Lights.pdf