Can tax reform really fix the income gap?

Based on a spate of recent posts (see here and links therein), a commenter (HT: Greg) asks a good, tough question of yours truly: on the one hand, I’ve argued long and hard that while we definitely need more progressive tax policies, the fact that the growth of inequality is largely a pretax phenomenon implies that tax changes alone won’t reverse the trend.

Yet, in this post reviewing the recent CRS inequality report, I point out that a) the increase in capital gains plays a large role in driving inequality trends, and b) if we taxed such gains as regular income (instead of at much reduced rates), that would help to reduce inequality.

So how can I argue on the one hand that tax policy is inherently limited as a tool against rising inequality, and on the other, that we should employ tax policy to push back on inequality?

I could invoke Walt Whitman: “Do I contradict myself? Very well, then I contradict myself, I am large, I contain multitudes”—and leave it at that.

But better yet, let me explain.  First, the evidence as shown in the chart here shows that increased inequality is a pretax story.  Whether it’s the increase in earnings or wealth inequality, the latter including outsized gains from assets, that’s all occurred in the so-called primary distribution of income, before taxes and transfers get into the mix.

Now, that doesn’t mean that more progressive taxes and transfers can’t help offset higher inequality.  They can, they should, and they do.  Unfortunately, as I stress here, they’ve become less effective in that regard over time.

But there’s another policy constraint here—it’s the “building-a-dam-against-an-ever-rising-river” problem.  As long as market outcomes become increasingly unequal almost every year, whatever redistribution we’re accomplishing through the tax code will have to be constantly ratcheted up.  That’s tough even in a rational political environment—it’s impossible in the current one.

Then there’s the question of how much of a direct difference we could make in the growth of inequality by just depending on taxes (the word “direct” is important, as I’ll  show in a moment).  Using table B-1 from the CRS report, I can simulate what the Gini index might have been in 2006 if the tax system hadn’t become less progressive.  And the answer is: it wouldn’t have directly changed measured inequality much at all—it lowered the Gini by only 0.005 compared to its actual 2006 level.*  More progressive taxation will help, but at the end of the day, you can’t bring a knife to a gun fight.

There are, however, very good indirect reasons to think that taxes matter a lot more in terms of moving inequality than the above rap would suggest, and I should be more careful to reflect this insight.  Important work on this question by Saez et al (see here; paper here), for example, shows strong correlations across time and place between higher marginal tax rates and reduced income concentration.

With higher taxation, they argue, there’s less “rent seeking” (economese for rich people figuring out ways to claim more riches—ways that don’t lead to better overall economic outcomes).  As Saez et al put it, “Lower top tax rates induce top earners to bargain more aggressively for higher pay” and bargaining here mean using their clout, power, friends on the board, etc., to claim pay packages that go well beyond their productive contributions to the firm’s output.  This is a classic zero-sum outcome–the execs’ gain is someone else’s loss.

Of course, the notion that there’s separable independence between the primary distribution (market outcomes) and the secondary (post-tax and transfer) is wrong.  Tax policy itself affects market outcomes, for better or worse.

The trickle-downers have way overdone this for decades, arguing, against evidence to the contrary, that tax cuts unleash torrents of growth.  What Saez et al are identifying is a new pattern that looks like it has a lot to do with inequality: trickle-up economics.

*I did this by substituting the 1996 terms for taxes into the 2006 table and recalculating the Gini index.  One wrinkle here is that you have to rescale the income shares so that they sum to one.