To support the aerodynamic explanation:
While preventing blade sailing during shutdown, the rotor is controllable down to very low RRPM. Until about 10% RRPM, my inputs to maintain a level disc do not change, only more cyclic into wind is required.
I would find it hard to believe that a rotor at such low RPM is still behaving as a gyro.
This seems to support an aerodynamic cause for the phase lag.
------------
To really analyze this though we will need someone who knows the math behind gyroscopic forces.