No its not.
Oversteer is not simply large angle stuff, any time the rear tyre yaw rate is greater than the front tyre yaw rate a car is oversteering (feel free to disagree - but keep in mind that's Skip Barbers' definition not just mine). As such if you lift and the front turns in, the car is oversteering.
No it doesn't, see above.
Why? Its more than possible in reality.
Its not even remotely that simply. A mid point CoG will give a car a greater Polar Moment, but that alone is not enough to automatically mean its a fairly serious issue. Suspension type (swing axle = bad, multi link = good), layout (roll centre location, roll bar set-up), wheelbase (long = good, short = bad), track width (wider = good, narrow = bad), track width offset (the wider the rear in comparison the better), tyre width (front and rear and the relationship between the two), CoG height (low = good, high = bad) and many other factors all affect how serious the degree of lift off oversteer is and all can (to a degree) be changed to affect how a car behaves in relation to lift off oversteer.
Now PCars doesn't get it perfectly right (AC is better in this regard), but its not as bad as claimed and SMS were quite open from the beginning about the fact that the default set-up on all cars was as neutral as possible to act as a kind of 'assist' for pad users.
While I would agree that Pcars is not perfect (yet to find the sim that is) I wouldn't say 'fails miserably' and certainly I have had engines blow due to overheating, brakes overheat and increase braking distance (road cars suffer from this far more) and tyres overheat and become very, very tricky as a result.
Exactly. No such thing as a 'correct' set-up exists for a race car and I don't really blame SMS for going for a neutral/understeer bias as a starting point. A team will try and set a car up to suit (as much as possible) the style and preference of the driver.
Or hold a constant throttle.