I agree with you. There have been lengthy arguments about "hashrate follows price" or "price follows hashrate" for years now. I'm in the former camp.
Also: supply is not a function of price and cost in the case of bitcoin mining. Supply is pretty much constant, miners just fight among each others over the fixed-sized cake. So only way in which mining influences price is by one variable: the percentage of coins sold by miners into the market.
I think this percentage is increasing with falling price, up to some point... some miners will reach > 100% and switch off. If they want BTC they'll buy instead. Mining efficiency (of the remaining miners) increases and cost decreases. Their selling-percentage decreases. All this has positive effect on price, further decreasing the percentage that has to be sold to cover cost. I'm not sure we've reached low enough price (or competitive enough mining landscape) for this to play out to a meaningful extent.
If the price drops so that the increased portion of the cake they earn is worth less in fiat terms, then would they not need to sell a larger percentage of coins to cover costs?
Yes, at first. What I meant was when inefficient miners start to drop out, the collective selling-percentage would decrease again.