Perhaps an inertia only dyno might be affected by something such as this, but I can't see how horsepower across an rpm range would be negatively affected -at- the motor. With an eddy current dyno I've had my previous car dynoed many of times with several different tire combinations (taller:28", and shorter 26") which would mimic a gear change, and with the same motor combination horsepower was not affected. The eddy current also took in consideration tire height which I'm not certain if an inertia only such as the dynojet does?
The inertia dyno accounts for anything that affects the ratio of engine RPM to roller RPM as it "knows" both.
Think of it like this, the engine accelerates the rollers, plus everything in between the crankshaft and the roller. With a higher numerical gear, the engine has more mechanical advantage over the roller so it accelerates the mass quicker. The HP calculation accounts for all that, BUT, the dyno does not account for the momebt of inertia of all the driveline components. They were all accelerated quicker too because the drum was accelerated quicker. So the driveline sucked out more power because even though it had the same moment of inertia, it was accelerated quicker.