When you are discussing cultivars you are discussing trees with subtle genetic differences from the species tree. They arise from poor chromosomal division during the develop of sexual organs in flower development, plant cultivars are genetic abnormalities that are not fatal to that particular plant. (Witches brooms are similar, but the abnormality occurs in a branch meristem). The extent of these abnormalities can be difficult to predict. Some abnormalities will not allow seeds to develop and grow, while many abnormal plants may not show any difference to the species plant, but may be a little slow to develop or are more disease prone.
The abnormalities may prevent proper root development, or stop adventitious roots from from cuttings, or change leaf color etc. (There are species of plants that will not flower as cuttings, but will as grafts)
The genetic blend for each cultivar is unique, because it arose from a mistake. To propagate a cultivar by seed ruins the genetic code of the cultivar, so it is now different. So if you have a kashima maple and propagate it by seed, the seedlings cannot be called kashima, even though they look similar.
So cultivars to remain true have to be propagated by cutting, layering or bioengineering, so that they are exact genetic copies of the original. For ease, grafting wins hands down as it easy, quick and has a high success rate. Also some cultivars are just too abnormal to produce roots or survive on them for long, so grafting is a good means of keeping that cultivar alive.
Sometimes grafting will dwarf a plant, usually when cross grafting species. Sometimes it speeds up development, Brent Walston thinks that grafting speeds up shishigashira growth.
Many horticultural and agricultural plants are grafted because the cultivar is not resistant to particular diseases, eg wine grapes are grafted to a root stock that is resistant to phylloxera, a subterranean aphid. Even though the cultivar does well on it's own roots, it has low resistant to these aphids.
So there are many reasons why plants are grafted, though usually it's for specific reasons. For Japanese maple these plants are desired for their interesting and pretty foliage, many of these cultivars may not form roots as cutting or layers, or they are weak on them, so grafting is the best method of propagating. Since these are predominantly garden plants, it doesn't matter if there is a graft.
There are some JM cultivars that grow very well on their own roots, such as shishigashira, kashima, kiyohime, deshojo, shideshojo.
A bonsai maple without an obvious graft on the trunk is more desirable then one that has, hence the interest in cutting grown maples.
For a longtime the standard response to growing JM on their own roots was it can't be done, their not strong, etc, but this is not necessarily true, it depends on the cultivar The more people who attempt to propagate different cultivars by cuttings or layers, and talk about it, the broader the propagating base becomes for all of us.
In regards to the opening post about Kagiri nishiki cultivar. I would think it unlikely to be successful on its own roots as its is just too different from the original species plant. But you don't know until you try, so try a layer on a small sub-branch, one that doesn't matter too much, and let us know how it went.
Paul