Yes, exactly. Basically I’m approaching this externality problem in a fact that the fee market is kind of a small market placed in a single block space whose total revenue is called block gas limit. That’s why the miner uses a strategy for profit maximization in a condition of volatile P and Q.
Well it’s hard to understand this point. The 8 MM hard limit block, just a status quo of each block, doesn’t mean the fixed Q of commodity(computation resources). Since the total revenue of miner is estimated like this. (If wrong, plz tell me 
)
In this regard, the Q of the commodity is not exactly the same one with the hard limited resources, so it’s not the fixed constant.
I mean, provided the fixed revenue of block gas limit, the Q of commodity could be volatile following the smart contract each tx sender writes. So, even in a hard limit, miner’s supply would not be the vertical. Since the Q is always going to be volatile. But if we regulates the Q(the gas used), as I’ve already talked with @4000D , there would be no more sophisticated contracts like Bancor, MakerDAO, etc. It’s not that good way .
So, if tx fee is decreased x %, there will be no difference of miner’s total revenue since there is always tx sender’s demand to be in a block.
Whatever the level of revenue(block gas limit) or tx fee is, miners will always try to fully fill their block gas limit to maximize their profit in general. But, if the demand elasticity is too high, when the tx fee is go upper, the total revenue would be decreased.
In this point, just using the theoretical approach, miner(seller)'s total revenue can be maximized in two conditions.
i) Elasticity of demand equals 1.
Using the demand elasticity, we can check the profit maximization condition. That means, in our world, the elasticity of tx senders has to be 1. That makes the miner’s total revenue maximized. But as you already estimated, the elasticity of tx senders was over 1.ii) “Marginal Cost = Marginal Revenue” condition of supplier.
This is also a profit maximization condition of each supplier in every market whatever the type of market is. (I’m still researching on these…)
As I mentioned above, whatever the hard limit is, the Q and P are simultaneously the volatile variables. And the maximization strategy itself is not influenced either.
So, I wonder how the fixed fee(fixed gas price right?) scenario goes on. First it is possible for the miners to have horizontal supply.
But in this case, the pigouvian tax burden would substantially go to the tx senders. Since the elasticity of miner is infinite.