One factor is the lens' maximum resolution. The less sharp a lens is at its peak resolution, the easier I think it is to label it parfocal. If a lens is razor sharp when focused properly, even a little movement one way or the other will be quite noticeable. Not quite so much if the lens is a little soft to begin with.
Another factor that may make it easier to consider a kit lens parfocal is the lens' maximum aperture. At f/3.5-5.6 the depth of field is a lot deeper than at f/2.8, so even if the focus moves a little between 18mm and 55mm there is a lot of overlap in the DoF at the two focal lengths of the EF-S 18-55mm f/3.5-5.6 IS II. There is nowhere near as much overlap between 17mm and 55mm at f/2.8 using the EF-S 17-55mm f/2.8 IS. At 55mm with a subject distance of 10' the DoF at f/5.6 is twice the DoF at f/2.8: 2.15' vs. 1.07'.
I'm not convinced the cheaper kit lenses are truly parfocal as much as that they are effectively parfocal due to the narrower maximum apertures and lower peak resolution. In most cases you can focus at the longest focal length and then as you zoom out the increase in DoF is enough to compensate for any small movement in the true center of focus. With a true parfocal lens you could also focus at 18mm and zoom in without needing to refocus.
For a more in-depth look at these differences between true parfocal lenses and those considered by many to be effectively parfocal, though they actually are not, please see this 2016 blog entry (written about three years after the answer above was written) from Roger Cicala of lensrentals: Mythbusting: Parfocal Photo Zooms