selection doesn't even apply directly to the individual, but to the survival of individual genes through individuals.
if a gene establishes itself within the genepool of a group, that is because it has been beneficial to every individual (and the genes that survive through that individual), by it's own merit - not by merit of being widespread in the group already, along the way.
Some form of "Group Selection" works, with stuff like the Green Beard Effect. There can be subgroups where sexual attraction is influenced by something like that. If group selection can work with a pair of traits in two halves of a population, it seems like it could work for a single trait in both halves of a population.
If you treat "race" as a word meaning "extended families that reproduce to some extent," then you have a great evolutionary argument for in-group altruism. By being slightly sensitive to race, you allocate slightly more resources to your own set of genes, meaning that people who share your genes, in turn, have more resources they can allocate back to you. In conditions of scarcity, that really helps. However, I don't know if it helps enough to deal with defectors (i.e. if there's a group that allocates lots of resources to group members, someone born with a "mooching" gene could afford lots of kids. It is basically the same math that wrecks cartels, just on a longer timeline).
That doesn't really matter for my point, which was that group selection is a very wrongheaded way to look at the issue. Abstracting an organism's genotype and phenotype as "an individual" is accurate shorthand, while group selection just doesn't happen.
Well, a gene can spread in a small group through drift (even if detrimental). If it allows that group to out-compete the next tribe over, it will spread farther.
Of course, no one has ever seen this in action, so it's pretty unlikely that group selection is very important in evolutionary history.
if a gene establishes itself within the genepool of a group, that is because it has been beneficial to every individual (and the genes that survive through that individual), by it's own merit - not by merit of being widespread in the group already, along the way.