In this post, I challenged Alan Reynolds' unconventional view that inequality has not been rising in the U.S. like many have claimed. The post shows that most other conservative economists do not share his view on this topic.
Showing how far Reynolds' views on inequality are from most other observers wasn't satisfactory to some readers, they want a more detailed rebuttal. That was the first post on this topic, I didn't expect that it would be convincing by itself -- I plan on addressing this topic in a variety of posts, so that will come, but for now let me use the little time I have to assemble a few points directed at Reynolds' main objections. As I said, there will be more to follow:
Reynolds' objections are:
1. Piketty-Saez use tax returns, but not everyone files a return, etc., so this is different from total income.
The Census data he uses in rebuttal is a sample too. See below. Tax-returns are used because it provides a homogeneous series over time, Census data do not. Again, see below.
2. The study uses pre-tax income (actually, it is after employer payroll and corporate taxes, but before income taxes). Reynolds argues this is a problem and that after-tax and transfer income that income that includes Social Security and other transfers are a better measure and do not show the same inequality trends. Here's the EPI on this point from a 2000 paper (many of Reynolds' points are not new at all):
The weakest of these critiques, such as the one by Alan Reynolds from the Hudson Institute, ignored a key issue ... and simply argued that there are different ways to measure inequality at a certain point in time (Reynolds 2000). ...[T]hese types of analyses reveal nothing about the key question of whether inequality has increased over time. As we show below, no matter which income definition is applied, the finding of increased inequality stands. ...
So the measurement question is whether differing definitions of income yield different trends in income growth and inequality. We show below that they do not. We agree (and document) that analyses using different income measures will result in different levels of inequality at a specific point in time (e.g., examining a single year). What is important to note, though, is that studies using comparable measures of income inequality demonstrate that income inequality in the U.S. is now historically high, and high relative to other advanced countries.
It's not at all clear to me that we should be using after-tax and after-transfer income. This WSJ report notes the latest data that we have looking at after-tax income inequality (but without transfers):
The share of all income earned by the top 1% of taxpayers rose to 19% in 2004 from 16.8% in 2003, the IRS said. That remains below the 20.8% high hit in 2000, when it was elevated by capital gains related to the stock boom. ... After tax, the share of income of the best-off 1% jumped to 16.5% from 14.4%...
Although dated, the IRS figures are among the best ways to compare the gains of the rich, middle class and poor because they include things that some other reports don't, including capital-gains income and taxes paid. Because capital gains are volatile and mainly reflect swings in the stock market, some experts prefer the Census Bureau data. That showed the richest families' share of total income in 2004 equaled its previous high and rose to a new high in 2005. ...
There's a bit more on the consumption (i.e. including transfers, etc.) versus income inequality measures here in a recent Fedgazette discussion about measuring poverty. It concludes:
Acs and Austin tend to disagree over the utility of consumption-based poverty measures. According to Acs, “Ultimately, consumption is a better measure of well-being than income, but I think it is harder to measure, and income is not a bad proxy for consumption.” But Nichols responded, “I disagree that consumption is a better measure of well-being,” in part because researchers don't know how much consumption is financed by unsustainable borrowing. He added that consumption measures “have just as many problems as income-based measures.”... Said Acs, “I think Austin and I agree that there are pros and cons to all the ... approaches,” both income and consumption.
3. Reynolds says the Piketty-Saez results do not agree with Census data.
Again, see below for the problems with Census data.
4. He notes that "Piketty and Saez measure income per tax unit rather than per family or household." He says that per unit measures show more inequality than per family measures. But this has been addressed. Here's Piketty and Saez:
Because our data are based on tax returns, they do not provide information on the distribution of individual incomes within a tax unit. As a result, all our series are for tax units and not individuals. A tax unit is defined as a married couple living together (with dependents) or a single adult (with dependents), as in the current tax law. The average number of individuals per tax unit decreased over the century but this decrease was roughly uniform across income groups. Therefore, if income were evenly allocated to individuals within tax units, the time series pattern of top shares based on individuals should be very similar to that based on tax units.
And, in a footnote:
Obviously, income is not earned evenly across individuals within tax units, and, because of increasing female labor force participation, the share of income earned by the primary earner has certainly declined over the century. Therefore, inequality series based on income earned at the individual level would be different. Our tax returns statistics are mute on this issue. We come back to that point when we present our wage estimates.
So, that point is well-known and covered.
Much of his rebuttal uses Census data. But there is a problem here:
The first point of call is data from the Census. Census numbers are based on the Current Population Survey, a questionnaire filled out by a sample of Americans, then extrapolated to the nation as a whole. For historical comparisons, go to Historical Income Tables.
Data there are gathered under several categories: households (people living together), families (they have to be related), and individuals. (Formal definitions) As of now, only the household data have been updated to 2005, which is why I recently turned to Table H-13 – Educational Attainment of Householder – to show that most Americans with a college education have lost ground in recent years.
The Census data are the key source for assessing how most Americans are doing. However, they do a poor job of tracking incomes at the very top, for two reasons. First, because Census data are based on a limited sample, not the whole population, they’re unreliable in tracking the income of small groups – and the really rich are a small group, who just happen to bulk large in the economy. Second, the questionnaire is “top-coded”: if the individual interviewed has earnings higher than $999,999, those earnings are recorded simply as $999,999. Since a lot of income growth in the last few decades has taken place among people with multimillion-dollar incomes, the Census data miss an important part of the story. In particular, what you won’t learn from Census data is the extent to which rising inequality is a story, not about the top 20 or even the top five percent of the population, but about the top one and the top 0.1 percent.
Fortunately, there’s another source of information: income tax returns, which aren’t top-coded. Tax return data are especially useful if you want to look at long-term trends going back before 1947, which is when the Current Population Survey data begin; high-income Americans have been paying income taxes since 1913. The I.R.S. does its own analyses of these data, and the Congressional Budget Office produces reports based on a merge of Census and I.R.S. data, but the most convenient and comprehensive analyses come from Thomas Piketty, at the Ecole Normal Superior in Paris, and Emmanuel Saez at UC Berkeley. Their latest data set is at Prof. Saez’s Berkeley home page (Excel file.)
5. Reynolds says much of increase at high income levels due to tax-shifting. Here's Piketty and Saez on that point:
One additional motivation for constructing long series is to be able to separate the trends in inequality that are the consequence of real economic change from those that are due to fiscal manipulation. The issue of fiscal manipulation has recently received much attention. Studies analyzing the effects of the Tax Reform Act of 1986 (TRA86) have emphasized that a large part of the response observable in tax returns was due to income shifting between the corporate sector and the individual sector [Slemrod 1996; Gordon and Slemrod 2000]. We do not deny that fiscal manipulation can have substantial short-run effects, but we argue that most long-run inequality trends are the consequence of real economic change, and that a short-run perspective might lead to attribute improperly some of these trends to fiscal manipulation....
This list is not exhaustive, but it covers his main objections. I expect additional objections will arise, but as I said, this will be followed by other posts so there's no need to take everything on at once, particularly the more obscure points.
Update: Brad DeLong has a second follow-up.