You say that this paper argues that the top earners should be taxed at a rate that maximizes the amount of revenue collected from them. And this is because this money collected from them is better used for other societal purposes (government services, welfare, entitlements, etc.) that the poor and middle class rely on. I.e. the money gives much more value to society as a whole than the marginal extra value it would give to the rich guy if he kept it. That's fine, but to me this seems to be taking a too narrow view and suffers from a kind of local maxima problem. Don't you really want to be maximizing, in the aggregate, the total revenue you collect from everyone; i.e. set tax rates and deductions/credits so as to maximize the total revenue collected. Don't just set rates, etc. so as to maximize what you collect from a small subpopulation (the very rich), but rather set rates, etc. to maximize total revenue collected from everyone.
Taking this view one might be able to argue that it is actually better to tax the rich at a lower rate. For example, with the extra money that they would get to keep with lower rates, the rich might invest this money in new businesses which would hire a lot of people that otherwise would not have been hired --- and these people would then be paying taxes on their earnings and spending their money on goods and services in the community, which demand would further increase demand and revenue, etc. So it is possible that the revenue lost by taxing the rich at lower rates could be more than made up by the economic effects of how the rich end up deploying this money that they get to keep (and otherwise would have had to give up). I'm not saying this is definitely true, but just showing a plausible argument. And maybe optimizing total revenue in this way would still effectively tax the rich at a 70% rate, but it seems you're solving the wrong problem when you just consider tax rates on the rich; you should consider the whole revenue picture.
