Hsu's 0.6 estimate is not implausible. As long as heritability of IQ is significantly less than one -- and all researchers have found that to be the case -- there will be some regression to the mean in IQ, just as there is in other human traits, such as height and athletic ability.
Hsu doesn't present his estimate as an estimate, just asserts it as a fact. Elsewhere he admits there may be confounding factors. I'm simply not convinced, though I would never argue that intelligence cannot be inherited or that it is completely based on the intelligence of the parents.
ETA: In thinking of ColinsMum's objection regarding which mean to use, I'm thinking now that the mean that makes the most sense is that of biological children of the same parents, raised in the same environment. I'd like to see a plot of the average IQs for children of high-IQ people, if anyone can find some.