What is your source of this graph? It looks like something that was "drawn" with excel. I would disagree that the max block size would affect the amount that people are willing to pay for their TX to get confirmed.
Block size does affect how much users will pay for a transaction because of the economics. If there are more transactions than will fit in a block, then users will have to compete for space in the block, and miners will accept the transactions that maximize their fee revenue. If the block size is increased such that there is enough room for all transactions, then there will be no competition for space to drive up fees.
Assuming there is no competition for space, then miners will include any transaction with a fee (assuming that the cost of including the transaction is 0) because not including a transaction with a fee is just throwing away money. Fees could drop to 1 satoshi per transaction because there is no reason for miners to turn away 1 satoshi transaction fees (other than pride).
You also need to remember that there is no "set" fee for a TX fee, but rather the fee is based on how big the TX is (among other things). Also there are many "free" TX that are confirmed in every block (or almost every block)
You are describing the current policy that is implemented in the reference code. There is nothing preventing miners from adopting other fee structures that will increase their revenue, and there is nothing preventing users from adopting other fee structures that will lower their costs. In the end, economics will determine the fees, and not the policies of the Bitcoin Foundation.