It is probably too complicated to be sure, with subtle shifts in air flow easily affecting the total drag.
However, since things (usually) tend towards their lowest energy, and the wheels tend to spin when the car is moving, I would guess that is the lowest energy state, and hence the lowest total drag, compared to clamping the brakes on the bike wheel and stopping it from spinning.
But that statement, lowest energy state, and hence the lowest total drag, is BS. Things have no tendency towards a state of lowest energy dissipation!
So:
Drag on the car * speed = power from the car.
power from the car = power from the spinning wheel + everything else
Increased drag on the car = power the spinning wheel generates / speed
If the wheel does not spin, it does no work.
If the wheel spins freely, it does -very little work-, because it has very little friction. So the spinning wheel adds very little more drag than the fixed one.