In the current study, Kanazawa argues that humans are evolutionarily designed to be conservative, caring mostly about their family and friends, and being liberal, caring about an indefinite number of genetically unrelated strangers they never meet or interact with, is evolutionarily novel. So more intelligent children may be more likely to grow up to be liberals.
That's a rather unconventional definition of "conservative" and "liberal" they're using there. One might almost suspect an agenda.
I'm also unclear on why things which are "evolutionarily novel" should be correlated with higher intelligence.