EVs could coast if a manufacturer chose to make one that allowed that without shifting into neutral. In practice, when letting off the accelerator, existing EVs will instead regen brake.
Modern EVs have easy adjustment for this. The Hyundai/Kia EVs for example have shift style paddles for adjusting this on the fly which includes a mode for regen only when depressing the break pedal.
It's true though that using this mode will extend the life of your tires.
Next time you do this keep an eye on the actual power readout. See if it's actually zero or if it's reporting ~3kW of braking or accelerating.
Whether that is more or less efficient than a zero-power coast followed by some kind of braking exactly at the end... I assume the difference is so tiny that it makes no difference.
How so?
For the purposes of tire wear, applying regen braking in a car is the same as applying brake pads. Generating 5kW of electricity of 10 seconds vs generating 5kW of heat for 10 seconds, same same.
Let's say you're on the highway driving in an EV. You have cruise control on. You go down a hill. The EV's cruise control applies regen braking down the hill, using the tires to slow you to your desired speed.
Let's say you do the same in an ICE vehicle. You will coast down the hill, gathering speed. Cruise control in an ICE vehicle generally will not brake for you. So more of your energy from the hill gets removed as air resistance. When you slow due to air resistance it does not wear the tires.
The same logic applies each time you push the gas pedal slightly harder than you needed to and then back off.
That's an assumption I disagree with. Brake pads will always be less smooth than engine braking. For the same braking action, I assume more brake dust and slightly higher tire wear due to brakes not able to provide fine speed adjustment.
The down-the-hill scenario is interesting, it brings new comparisons: is there more tire wear from maintaining a chosen speed, vs letting the car overspeed and then braking? How does air resistance contribute in each case?
I maintain my earlier opinion that the differences between all these scenarios are minimal and can be ignored. But if you have some physical model that helps calculate these scenarios, it could be fun to play around with.