Sir Stanley Hooker was placed in charge of supercharger design at Roll Royce and wrote many articles and memos on supercharger design and testing. One such article is here:
AC Eng Perf Analysis at R-R . This is from Enginehistory.org, as you can see in the link.
This is a 1940 article on defining a new term, called shaft horsepower from the engine tests. He defined shaft horsepower as Brake Horsepower + Compressor Horsepower. Brake horsepower is what drives the propeller and compressor horsepower is the power absorbed by the supercharger. They HAVE to account for it because the crankshaft must produce shaft horsepower when the engine runs ... at LEAST in mechanical stress. Since the supercharger TAKES power from the Shaft Horsepower, Brake horsepower is what is left to drive the propeller.
To make a long story shorter, the engine they tested, was a Merlin XX. They used 90% efficiency, used the same friction and pumping characteristics as the Allison V-1710 at the time, as indicated just above table 4. The engine made 1,234 shaft hp at sea level and 1,298 shaft horsepower at 20,000 feet. At seal level, the supercharger used 216 horsepower and at 20,000 feet it used 225 hp. That leaves 1,018 Brake horsepower at sea level and 1,073 horsepower at 20,000 feet for the Merlin XX used in test.
So, at seal level the supercharger consumed 17.5% of the shaft horsepower while, at 20,000 feet, it consumed 17.33 % of the shaft horsepower. This was AFTER Hooker improved the basic single-stage engine to produce the Merlin 45, and the Merin XX used his better-matched impeller and diffuser parts.
Just for clarity, the Merlin XX used a single-stage, 2-speed supercharger.
I assume that when the 2-stage engine was developed, the second stage used a bit less horsepower since the first compressor usually had the larger impeller diameter. But, how MUCH less power it might use is up for "guestimate." I'll assume it uses something like 90% of the power used by the first compressor stage as a second-order estimate. To be conservative, as a first-order estimate, I'd assume the same power consumption for the second stage as the first stage, and further assume it to use about the same percent as this test. So far, I haven't found a supercharger analysis for a 2-stage Merlin to confirm anything, but I'd assume both stages use about 17.5% of the Shaft Horsepower and go from there.
The Merin 61 made 1,565 hp Brake hp @ 3,000 rpm, +15 psi boost, at 11,250 feet. That means shaft horsepower was about 2,407 with a 35% loss to superchargers as a 1st-order approximtion.