Not sure if this is a joke I'm not getting, but the "variants" in the linked study refer to tiny genetic sequences, not phenotypical appearance traits.
There's no reason to believe POLYhj is any lower now than during the Habsburg dynasty. Homozygosity is independent of population-level allele frequency.
