If the reward at this time will be fees for making transactions, then the transaction fee will be proportional to the expense of keeping the ledger, no? So if this expense is too great, the transaction fees will be proportionally too big.
You have that backwards.
You've switched the cause and effect around the wrong way.
The expense of keeping the ledger is proprotional to the block reward.
If the block reward decreases (for instance due to a reduction in transaction fees), then some miners become unprofitable and shut off their equipment. This reduces difficulty, and increases the profitability of the remaining miners. The total worldwide cost of maintaining the ledger automatically adjusts downward to match the total fees being voluntarily paid by the users.
In the end, bitcoin will only really succeed long term if there are usually enough transactions per day to fill most of the blocks. This creates a market force driving most users to voluntarily offer at least a small transaction fee to get their block confirmed quickly, and provides the necessary revenue to keep hashing power high making it difficult for an adversary to attack the network.
If bitcoin has a few thousand users and the majority of them simply sit on their bitcoins, never engaging in transactions, then there won't be enough fees, and hasing power will eventually drop to insecure levels.
If bitcoin has a few hundred million users, and only 0.2% of them engage in transactions on a typical day, then there will be more than enough transaction fees to maintain a significant level of hashing power.
