There are just to many factors at play right now to be saying we have any "answers". All we do know that there is a define genetic component.
Most of the current data is skewed in some way. You have to take into account diagnosis methods (which differ worldwide), who is doing the diagnosis's (some countries limit to psychiatrists, others allow psychologists) , those who have been diagnosed later in life since the said definitions were not available then (likely most of the older generations in HFA range before 1980), things like urban vs. rural likelihood of DX, ditto that with terms of things like class, race, gender and socio economic position. You also have about 50% of the DX now being made that are in the HFA / Aspie /PDD-NOS category.
In many ways autism is like celiac disease in the way that while we have some data, we certainly do not have all of it and the data we have is still skewed to some extent. I really do think it is too early to be drawing conclusions yet. Give it 20 years and at least then we will have more reliable data.