If you read the thread, you'd see that I posted a two-sentence "proof" (<-- I used the scare quotes to indicate that it probably wasn't a proof), and then asked people to poke holes in it.
I agree with @molecular. The economic pressure can also be relieved, for example, by people voluntarily leaving the economic system. This would drop the supply curve such that it meets the demand curve at a point near the quota (Qmax).
What is interesting, is that either way (by fork or by people leaving the system), somehow the result is that Q* ends up to the left of Qmax! If this simple result is true, it would imply that it is not possible to use a block size limit to drive up fees.
Of course if bitcoin becomes too much expensive to use other cheaper systems will get their shares of the market reducing the usage of bitcoin.
I don't see how this could be good for bitcoin in any possible way.If the economic pressure is relieved by people leaving (or never entering) the economic system, then I think this would result in a decline (or a levelling out) of Bitcoin's market cap.
So again, if the "economic pressure from deadweight loss gets relieved somehow" theory is true, then not only does it mean that we can't use the protocol to artificially increase fees, but it also means that any attempt to do so will instead have either no effect (the code will fork around it) or a negative effect (growth within the Bitcoin economy will stall or reverse [rather than fees increasing]).
Food for thought!