Let's take an analogy. Let us say that you make a fully flexible reservation for a hotel room 10 months in advance.
At that stage, the hotel does not expect occupancy to be very high and you can get that room for £80. One week before you are due to stay at the hotel, your plans change and you cancel. At that stage, the hotel expects a high level of occupancy and therefore the price for the same room as the one you booked has gone up to £250. Does the hotel owe you £170 when you cancel? Of course they don't. Yet this is precisely what you argue BA should do, viz. value the reservation that you cancel at the current price it could obtain for the seat (and therefore only charge you the difference between the current lowest price on the reservation that you release and the current lowest price on the new reservation you make).

Secondly, the price that you paid in the first place was premised on the current system, whereby you pay the fare difference between the fare for the inventory you were booked in and the fare for the inventory you need for the new reservation. This is, incidentally, standard practice in the airline industry. If the airline industry were to move to the system that you suggest, then the prices for semi-flexible advance reservations in lower fare buckets would have to go up to produce roughly the same amount of revenue per flight.

Thirdly, your argument assumes that the seat that you release will definitely be occupied. This assumes a 100% load factor, which is rather unrealistic even in the current system but would be even more unrealistic under the system that you propose.