It's a Rawlsian question of "how can we structure incentives such that everybody is better off?"

Now, the metric for "everybody is better off" is still an open problem, but there seem to be near pareto optimal steps we can take, like a high threshold wealth tax.