You haven't proven that ASIC Miners withold coins at a higher rate than GPU miners did. I mined with GPU for 18 months and sold less than 10% of my coins to cover electrical costs. Still while daily volume may be lower than total supply it is much larger than the rate of NEW production. New ASIC miners don't "own" the complete supply they only "own" the newly created coins which is a tiny portion of the market now. It was different in 2010 when the daily production was sometimes more than daily exchange volume but those days are gone.
Well, that's true, I didn't spend most of the coins I mined with GPUs either, but I meant more of that being a possible example of how the spending of coins mined might affect price.
Since we have actual data, I suppose an interesting thing to do would be to track how coins are spent after being mined in a block, and then see if we can use this to predict historical BTC prices. I suspect this might be difficult because of the way pools are setup, but it's a neat idea.