First things first: A hearty welcome to my new co-blogger Art Carden.
To the task at hand: A while ago I wrote on how "Capital should be untaxed" should be the default view in economics: Zero capital taxation should be the starting point for future discussions among serious people, it should be taught in principles courses. Interfluidity wrote a valuable critique, read the whole thing. I'm here to emphasize two claims he discusses:
1. If the real world has increasing returns to scale--if doubling inputs more than doubles outputs--then rational workers would want to subsidize capital rather than tax it (Short wonkish PDF, page 4).
2. If in the real world some workers can become productive from accumulating human capital, then perhaps capitalists (and workers who aren't good at building up human capital) would rationally want to treat human capital as, well, capital, and leave it untaxed. In practice this would probably mean taxing high wage-earners at lower rates than low wage-earners: If high wages are mostly caused by human capital we want to ensure that there are good incentives to accumulate human capital.
Both of these ideas have the same implication: Given a few basic assumptions, if there are workers who are unable to accumulate human and physical capital themselves, and if these workers are wise, these workers would gladly tax themselves to pay the cost of government, and they might even gladly tax themselves to subsidize the growth of capital.
If I'm a worker who can't accumulate human or physical capital to make myself more productive, I crave opportunities to at least use such capital in my job. And one rational way to satisfy that craving is to make a lot of sacrifices so that everyone around me has an incentive to accumulate physical and human capital.
So once we make the Chamley-Judd zero capital taxation result a bit more realistic, once we make it sound more like the real world, we find that regressive taxation looks a bit better than before. It's not that all signs point in the direction of regressive taxation--as Interfluidity and I both note, some tweaks of the model support progressive taxation, some support regressive taxation. But if the reality-based community cares about the well-being of people who don't have great skill at accumulating human and physicial capital then they should use zero capital taxation as their starting point.
There are sound arguments for positive capital taxation, sound arguments for negative capital taxation. Let's start with zero and go from there.
Coda: To quote Interfluidity:
...those of us who support redistributive taxation don't believe that the world works in the way that Chamley and Judd assume...
Indeed. And once we look at the way the world works, there are at least some reasons for having high tax rates on those who don't have the ability to accumulate capital and low tax rates on those who do.