Let's simplify the case to the basics:
Will a towed ship be more efficient than a propeller driven ship at equal speed compared to the shore? So disregard all other factors.
First, let's split the efficiency of the propeller in mechanical efficiency and hydrodynamical efficiency. Mechanical efficiency is the efficiency from energy content of the fuel to the torque + rpm on the propeller, the hydrodynamical efficiency is the efficiency from torque + rpm on the propeller to thrust and speed.
A hydrodynamical efficiency of 80% of an inland ship is actually rather high. Somewhere between 60-65% is not uncommon actually. This is because of the draught requirement forcing them to have smaller than optimal propellers, also the propeller is surrounded by tunnels and many rudders because maneuvering is very important for them (often two rudders per propeller, sometimes rudders in front of the propeller for backing performance).
Then there is another effect, called thrust deduction. The propeller is situated aft of the ship which means that it is generating a low pressure area on the ship aft body. This is an extra drag an can account for 10% or more extra required thrust. In other words: the towing force on a ship can be more than 10% lower than the thrust a propeller needs to provide.
Let's also assume that you can take off all appendages of the ship regarding propulsion: less rudder area (you need a bit to keep course) no tunnels, no shafts, tunnel thrusters etc. This reduces the resistance even further.
So, let's now also assume that the mechanical efficiency of the tower (mule?) is in the same order of magnitude of the engine of the propeller driven ship (why wouldn't it be?). I would also assume that the efficiency of torque + rpm on the wheels to the track is much higher than the hydrodynamic efficiency of a propeller.
So, this all points into one direction. Using a mule would indeed be significantly more efficient (in terms of dollars per mile) than a ship propeller, comparing at equal speed along the shore.
Should the mule be underwater or on the shore? Underwater has many practical issues, including it's own resistance that is not negligible anymore. But a mule on the shore is always towing with an angle, meaning a transversal force is applied on the ship which needs to be counteracted. This will cost efficiency. However, this is many orders of magnitude more practical than underwater.
So, why aren't we doing this? Probably because of the many practical issues and high costs surrounding building this track. Using the costs of building and maintaining this track you can make up for a lot of propeller propulsive efficiency.