Let's say there's a 1/1000 chance in infection of the parent if the [adult] child is actively infected during a brief outdoor hug. The chance of the child being currently infected depends on the current base rate but for the napkin's sake let's put it at 1/10. (As an aside, in actuality while there is evidence of pre-symptomatic transmission for SARS-CoV-2 the overwhelming evidence shows that asymptomatic transmission doesn't occur so just a heuristic of "am I actively sick or not" gets you good risk management anyway but let's avoid that for now)

Now let's say that the chance of the parent dying or having a seriously bad outcome given SARS-CoV-2 infection is 1/10.

That brings us to (1/1000) * (1/10) * (1/10) = 1 death per 100,000 of these hug-events. If we stick with my earlier example of someone with 2 years of life expectancy, that would factor out to (365*2*24*60)/(10^5) = 10.512 minutes "lost" per hug, if we assume that hugging someone doesn't have any positive effects on their physical wellbeing whatsoever.