First, Richard Dawkins wrote an article on race. Then Steve Sailer commented on it. Jason Malloy of Gene Expression replied, critiquing Sailer's assessment of Dawkins. And then, the discussion started. Much of the discussion revolves around an article by Frank Salter, which proposes that we are as related to our co-ethnics (relative to mankind) as we are to our cousins relative to our co-ethnics. In other words, Frank Salter proposes a genetic basis for ethnic nepotism.
The analogy between the degree of relatedness between a person and their cousin (against an ethnic group genetic background) and a person and members of their ethnic groups (against the genetic background of the species) is accurate, but potentially misleading.
To see why this is the case, imagine that each person has a $-value depending on how close they are genetically to us. Let's say that the average human has a value of $1, the average member of our ethnic group has a value of $10, and our cousin has a value of $100.
So, our cousin is "more valuable" to us (relative to our co-ethnics) and our co-ethnics are "more valuable" to us (relative to our co-humans).
But, we share many more polymorphisms with our cousin than we do with our co-ethnics than we do with our co-humans. So, in relative terms the analogy of me-cousin-coethnic with me-coethnic-cohuman is equivalent (ratio of 1:10), but in absolute terms, i.e., the number of my alleles in polymorphic loci, I am getting a $90 profit by preferring my cousin over my coethnic, whereas I get a $9 profit by preferring my coethnic over my cohuman. Genetic ethnic nepotism predisposition (if it exists) is a much weaker force than kin nepotism predisposition.
Now, suppose that there is an "ethnic nepotism" allele, a genetic predisposition to get into trouble for persons of our ethnic group. Is it plausible that such an allele would be adaptive, especially for race-level groups?
The answer is probably no, due to the fact that the set of different "actions" which would increase such an allele's prevalence is limited. It's only the more "dramatic" actions that benefit many of my coethnics at my expense that would have this effect, but the opportunity for such actions is rare, both because dramatic acts are rare in themselves, and also because for the greater part of human prehistory and history human groups had limited interaction with other racial groups.