There is some correlation, but it's not perfect.
I've thought a lot about this and I think beauty is both shorthand for "that looks like someone healthy and very functional" and also is something that can be faked, so it's not a reliable indicator.
For this reason, I think it will always be both valued and controversial. To whatever degree it genuinely signals something of underlying value, like health and competence, it's useful shorthand. To whatever degree it is a signal that is hackable, it's shorthand with a poor signal-to-noise ratio.
I don't think we will ever completely solve it. We will always be caught between those two facts.