MightyGem - looking through some old documents, I found a reply from Ray Prouty to an email a colleague had sent about cross-coupling, explaining that the control response from the semi-rigid rotor meant that phase lag was not 90 degrees but more like 80 degrees. He then mentioned that in steady pitch or roll rate (rather than accelerative), that figure might be as low as 60 degrees and doesn't depend on the hub geometry but the Lock number (ratio of aerodynamic parameters of the blade and its flapping inertia).
I don't know whether that would account for the steady state drift into the turn to the right but the Lynx dynamics were rather complex and anything is possible.