NYU Tax Policy Colloquium - Griffith on "Progressive Taxation and Happiness"

The paper at today's Tax Policy Colloquium was Tom Griffith's Progressive Taxation and Happiness. Tom uses the "happiness literature," a recent cottage industry in behavioral economics, to argue mainly that the case for progressive redistribution is stronger than is typically recognized, but more substantially (I thought) for high as opposed to low marginal tax rates.
The happiness literature uses surveys from a variety of countries over time, typically involving some variant of the question "These days are you typically very happy, somewhat happy, or not at all happy?" Recognizing the limitations of this information and its capacity to be influenced by other factors (e.g., people report themselves generally happier if the question is asked on a sunny day), the most significant finding, both in the US over time and across the world, is that increasing wealth isn't reported as bringing much if any extra happiness, either to individuals or societies, once one has gotten above the subsistence level. Claimed implications include (a) economic growth for already-affluent societies has less payoff in terms of human welfare than we thought, (b) progressive redistribution aimed at the people on the bottom is a good thing, because they do gain a lot of welfare from the extra resources whereas those above them don't lose much even if work incentives are hurt, (c) effects on work incentives are possibly good rather than bad up to a point, due to positional externalities (e.g., if I build an extra pool in my front yard it makes everyone else feel worse unless they slave away to build one too, at which point none of us is much better off than when we started).
The empirical evidence is not very strong. E.g., even apart from the obvious objections to how meaningful the data is, suppose that people as they get wealthier raise the happiness level that they define as "very happy." This is a verbal change that would mask actual happiness increases because the survey responses would fail to report them as their terminology changed. Or suppose that wealth increases, whether enjoyed by an individual or a society, improve longevity. This would be a welfare gain, since presumably we want to live longer, that the survey data would miss. But the story has a lot of intuitive plausibility, even if, like me, you really aren't personally into competitive consumption. And the story has if anything more competing credible (and perhaps all true) explanations than it really needs, including habituation to one's current circumstances, rising aspirations that always drive you to seek the next level, and the positional externalities issue (noted above) resulting either from express status competition or more indirectly from rising norms concerning what is, say, an acceptable-sized house.
It's easy to scoff at this stuff, but it could significantly revise how we think about the incentive effects of income and consumption taxes, as well as the affordability of various public goods and the wisdom of the US versus the typical Western European economic system. I am not aware of leading conservative commentators of the Martin Feldstein stripe thinking seriously about or responding to this stuff, but they should.
The economist Robert Frank, who came to our colloquium some years ago, makes similar arguments. The economics profession has tended to respond to him by saying: where's the hard proof? But the same could be said of assuming that the story is not true, and the methodological bias for "hard proof" yields self-delusion insofar as it leads one to accept as a default alternative priors that haven't been given hard proof either. Good Bayesians would not necessarily start from the premises of neoclassical economics, even if those premises are easy and fun to work with.