My orbital mechanics is pretty rusty, but I think the cars would
actually have a significant velocity relative to the geosynch station
(having climbed all the way up from Earth), and similarly would need a
big kick on leaving the station so that they would drop out of geosync
orbit all the way down to a perigee of 1 earth radius. This is quite
a big drop. My very rough calculations came out with a value of about
10,000 m/s for the velocity differential between the circular orbit and
the elliptical one at apogee. Maybe that is tolerable with your EM rail,
if it is long enough. A similar velocity differential exists at the
Earth side as well. (Obviously you need greater than LEO speed there,
which is about 7000 m/s.)
Hal