I seem to recall someone arguing that, in combat between iron age tribes, it was basically a choice between massacre and slavery - if you did neither, they would wreck revenge upon your tribe further down the line.

This wouldn't be charity, as I guess the winners did benefit from having a source of labour that didn't need to be compensated at the market rate, but it would be a case where slavery was beneficial to the victims.

(I think Carlyle was wrong about other supposed cases of slavery proving beneficial for the victims)

I'm not sure about those direct costs. According to my references, a male slave in 10th-century Scandinavia cost about as much as a horse, a female slave about two-thirds as much; that's a pretty good chunk of change but it doesn't seem obviously out of line with the value of labor after externalities. I don't have figures offhand for any other slaveholding cultures, but the impression I get is that the pure exercise of power was not the main determinant of value in most, if not all, of them,