I can see how this could be an issue. That said, this strategy could only be successful if miners get the minFee down so low that blocks are frequently more than 100% full, encouraging higher fee premiums, so it would require many miners to collude to have an effect. If fees are only depressed slightly by a few miners pushing minFee down so blocks are on average, say, 60% full, then the effect would be relatively small. Also note that even in the current system, it’s the case that 50% of miners could collude to push the gas limit to below 8 million and try to earn more revenue from supply constriction.
That said, if this ends up bring a problem, there is a way to do a default strategy change (not even a soft fork!) after the fact to make it robust up to 50% collusion: instead of asking each miner to vote the adjustment for one block, we ask each miner to locally compute what the minFee “should be” based on eg. the last 256 blocks, and the default strategy would be to push the fee as close to that value as possible. This way if a small coalition tries pushing it up or down, as soon as they are even slightly successful the majority would start maximally pushing back in the opposite direction.