Income Inequality Follow-Up

I want to follow up on the most recent inequality post. Recall that the WSJ editorial page says that:

[N]ew data show that the bottom 50% of Americans in income ... paid a smaller share of total income taxes in 2004 (3.3%) than in Bill Clinton's last year in office (3.9%). That 3.3% is the lowest share of total income taxes paid by the bottom half of earners in at least 30 years, and probably ever. ...

By contrast, ... the top 5% and 10% of earners saw an increase in their tax share over that same period, with the top 5%'s share rising to 57.1% in 2004 from 56.5% in 2000. If this isn't the definition of a highly "progressive," aka redistributionist, tax code, we don't know what is.

Though I don't agree with defining progressivity according to share of taxes paid or by percent of income, I'd rather see it defined according to marginal tax rates, here's an example of how taxes can become more progressive according to the progressivity definitions used by the WSJ editors, and by others:

Suppose that the tax rate is 10% on the first 100 in income, and 30% after that.

Let person A have an income of 125. Taxes are 17.50.
Let person B have an income of 225. Taxes are 47.50.

Then the average tax paid as a percentage of income is 17.50/125 = 14% for person A. Similarly, person B's average tax rate is 47.50/225 = 21%. The shares of total taxes paid are 27% and 73% (e.g. 47.50/(17.50+47.50) = 73%)

Now move 50 from A to B, i.e. move 50 up the income ladder. This could be from government policy that redistributes income, productivity changes, or other reasons.

Person A now has an income of 75 and a tax burden of 10%. a smaller amount than before, and person B has an income of 275 and a burden of 23%, a larger amount than before. The shares of taxes paid have increased from 73% to 89% for person B, and have fallen from 27% to 11% for person A. Thus, taxes have become more progressive in the sense that average tax burdens have tilted upward, and in the sense that the share of taxes paid has increased for higher income people and decreased for lower income people.

Nobody should be surprised that the share of taxes paid by a particular group goes up when their share of income goes up. Nor should we be surprised that if a group's share of income falls to historic lows, the share of taxes will fall to historic lows as well. Share of taxes paid is a poor metric for assessing progressivity for this reason.

What is important, and where the debate should be focused, is why the shift in income shares has taken place, i.e. what is causing the more unequal distribution of income. Is it due to productivity or some other factor? If it's due to changing market power relationships, allocative inefficiencies, or government policies that transfer income upward, the fact that the higher income people now pay a larger share of total taxes may be of little consolation. Do conservatives really want to argue that increased progressivity is a good thing if it arises from redistributing income up the income ladder? Obviously not, and neither would liberals. So before we begin pointing to changes in the shares of taxes paid as evidence of increased fairness, we should be certain we know why the changes have occurred.

Currently, we don't know for certain, or even nearly so, and this is the subject of debate. The split is between those who think growing inequality is due to productivity changes, skill-based technological change in particular, and those who believe it arises from other factors such as changes in public policy (e.g. taxes, union busting) and globalization.

I believe that public policy plays a large role, but even if it is skill-based to some notable degree, public policy can still be a factor. For example, are we certain that the opportunity to acquire the skills needed to realize the skill-based premium are fair and equitable, i.e. that everyone has an equal chance to compete in the global economy? If the opportunity to acquire skills is not equal, has public policy (and doing nothing is a policy) played a role? If so, the role of public policy in bringing about inequality may be understated because econometric investigations designed to determine the source of income inequality will have trouble distinguishing the role of public policy in the skill acquisition process.