I agree with Will Wilkinson's point that real social inequality has (mostly) been falling for some time in the United States. Today many an upper middle class person is plausibly happier than many a billionaire. Yet most self-made billionaires work very hard to get to that position, which creates a possible tension between cardinal and "observed choice" or "ordinal" metrics of welfare. Why work so hard for so little? Presumably many of these billionaires really want to "be there," even if they are only marginally better off or in some cases worse off.
Here are a few possible implications, not all of which are (or can be) true:
1. Higher marginal tax rates aren't very unjust, because lower incomes don't make wealthy people much less worse off.
2. Higher marginal tax rates are very unjust, because they undo results that the wealthy have worked very hard for and cared very deeply about.
3. Work is fun for the (self-made) wealthy, so higher marginal tax rates won't much discourage their work effort.
4. Greater wealth is barely worth it for the wealthy, so higher marginal tax rates will very much discourage their work effort.
Will's paper convinces me that the distinction between ordinal and cardinal measures of human welfare is more important than ever. Conservatives often cite #2 and #4, or in other words they have an ordinal normative theory and a cardinal predictive theory. Liberals are more likely to cite #1 and #3, giving them a cardinal normative theory and an ordinal predictive theory. In neither case is there an outright contradiction, but arguably both groups end up holding an odd mix of positions.
It would be interesting to take each group aside and present them with the abstract questions of cardinal vs. ordinal understandings of well-being, yet without explaining to them the possible policy implications of their answers.