I am not a rocket expert but this is not a question to rocket experts, this is an economics question. Taking up fuel instead of cargo and using fuel to bring back the launch vehicle safely as opposed to parachuting the launch vehicle (or its engines, the most complex part) and launching more useful cargo into orbit.
The question is about cost of production of a rocket and cost of fuel and cost of unit of cargo per launch, not about difficulty of a controlled descent. Assuming it's perfectly easy to do a controlled descent (which is probably much more difficult to do from a much higher altitude than a few dozen meters above Earth, given that the rocket also needs to reorient itself and maneuver a much longer distance down to the surface) then the question is: how much cargo are you not taking up because you are taking up a huge amount of fuel to do this controlled descent, so it's not just a question of cost of fuel, it's a calculation of cost of launch of unit of cargo and cost of launch in total, etc.
One thing I don't have to be a rocket scientist to know is that if you do this, you are not being very efficient with your rocket, you are using huge amount of fuel just to haul fuel, you are not launching as much cargo as you can. You are launching and landing fuel. Then why not launch more cargo and a parachute to land some of the launch vehicle and instead work on manufacturing the rocket cheaper, more efficiently with a manufacturing conveyor line, with robots, with fewer people?
Why not use the money to improve efficiency of manufacturing of the rocket?
Musk wants to be in this business end to end, he wants to manufacture rockets and launch them and do everything. I don't need to be a rocket scientist to know for sure that he will not be as efficient at all of these as he can be at just some of it, that's why we have complex supply chains and specialisation.
Maybe he wants to be a total solution, but that would be Apple of space, not Linux of space so to speak. How about concentrating on the manufacturing side, creating the production process that is the most efficient and the cheapest and then mass producing launch vehicles for all who want to launch and operate them?
No, you don't have to be a rocket scientist to look at the overall business model.
Lets say a disposable rocket costs 150 million, thats 150 million per launch.
If a reusable rocket costs 300 million and gets 10 launches thats 30 million per launch, if it costs 600 million thats 60 million over 10 launches, if it costs 1 billion, thats 100 million over 10 launches. A reuseable rocket can potentially save a lot of money even if it is vastly more expensive.
Considering he is figuring out how to make his current rockets reuseable this will make launches dirt cheap. As fuel is only 2.5 % of launch costs even if you double it relative to cargo orbited to make the rocket reuseable, you still save a lot of money. If a rocket costs 200 million, and fuel costs 5 million, if you only launch half as much cargo but get to reuse the rocket 10 times you spend 50 million on fuel to save 800 million on the 10 replacement rockets you would have destroyed if you launched at full capacity on a disposable rocket. Thats 750 million dollars saved if you only launch at 50% capacity. As a rocket is much lighter landing than taking off launch capacity will likely be reduced by less than 20% which means 1.45 billion in savings over the 8 disposable rockets for the same mass orbited. Thats 250 million for 10 uses of a reuseable rocket vs 1.64 billion for 8 disposable rockets and fuel for the same mass orbited. Thats around an 85% reduction in cost per unit of mass orbited.
The only cost between launches will be recertification and refueling, If the rocket is designed right, recertification will be cheap (10-20% of manufacturing cost or less). Reuseable rockets will also increase manufacturing efficiency, as every man-hour spent on the manufacturing line will enable 8 times as much mass orbited as a disposable rocket, which means reuseable rockets increase manufacturing capacity by a factor equal to the number of times a rocket can be reused times the percentage of capacity lost to recovery. A rocket that can be used 10 times to launch 80% of the mass of a disposable will increase manufacturing capacity by a factor of 8 as 1/8 as many rockets will have to be made per unit of mass orbited. If the rockets can be reused more than 10 times it gets even cheaper.
You don't have to be a rocket scientist to look at the overall business model, but having a grasp of basic mathematics really helps.