Proposition: Is Income Inequality a Problem In the United States?

Proposition: Income Inequality Is A Problem In the United States

About Voices of Reason

Apr 1, 2011

Affirming the Proposition: Income Inequality is a Problem in the United States (Rebuttal)

In Mr. Weinberger’s opening statement, he argues that not only is income inequality overstated, but it isn’t even a problem. He asks, “If all income levels are gaining, why does income inequality matter?” As I will illustrate, this is a poor way of framing the argument as it makes magnitude irrelevant. Indeed, following Mr. Weinberger’s logic, the top 1% could be taking home $0.99 of every dollar the entire country earns, essentially turning our society into an oligarchy, yet Mr. Weinberger would ask what the problem is.
Alan Reynolds:
Mr. Weinberger points us to a research paper by Alan Reynolds, a Senior Fellow at the libertarian Cato Institute, who for years has been on a crusade to prove that not only is income inequality not a problem, but it doesn’t even exist. Overall, Mr. Reynolds’ analysis can be accurately described as cherry-picking: he is quick to challenge the data that don’t back his ideological point of view, yet has nothing to say about similar or more severe issues with the data that supports his position. Students of the income distribution will notice that Mr. Reynolds’ work is plagued by either giving an incomplete presentation of the evidence or by making points that are so trivial to the data that one wonders why Mr. Reynolds is even pointing it out. Most of Mr. Reynolds’ analysis has also already been rebutted by other work; in particular, Thomas Piketty and Emmanuel Saez have written a rebuttal to the objections that Mr. Weinberger raises is his last post (shifting tax rates, transfer payments, and capital gains), so I won’t waste my time on it here. 

Consumption Inequality:

Mr. Weinberger argues that because consumption inequality – the purchases of consumer goods – is smaller than income inequality, that there’s no problem. First, the idea that the consumption distribution is more equal than the income distribution is a myth:  a good deal of the academic work that has been done on consumption inequality relies strongly on data from the Consumer Expenditure Survey (CEX), which, among other things, is poorly designed to capture consumption changes at the very top of the distribution.  Studies that adjust for this have found that consumption inequality has actually tracked income inequality closely since 1980. In this context, focusing on consumption inequality instead of income inequality makes little difference.

Second, Mr. Weinberger runs into further problems when he points out that because things like computers, cell phones, automobiles, etc. are widely available, that income inequality doesn’t matter. I like to call this the “But Low Income Households Can Afford Flat Screen TVs” defense. The problem is that focusing on changes in absolute prices for consumer durables like cell phones and TVs is conceptually pointless: what matters is the change in relative prices. If, for example, the price of a cell phone has dropped by nearly $4,000 since the 1980s, and if income and preferences stay constant, then people will obviously consume more cell phones; however, people will also consume less of whatever is now relatively more expensive (in economics, we say that consumers face a budget constraint). So when Mr. Weinberger tells you that cars, TVs, and computers are consumed more today because they are cheaper than they were 30 years ago, all he’s doing is proving the Law of Demand, not commenting on whether the typical American family is better off. Here’s another way to think about it: we know that prices overall are increasing, but if computers and cell phones are getting cheaper, something else must be driving the increase in prices. Health care is an obvious answer; the cost of college tuition is another.


There’s a kernel of truth to what Mr. Weinberger says here: economic evidence does suggest that low-income domestic workers are negatively impacted by immigration from Mexico. However, Mr. Weinberger insinuates that the immigration effect is so large that it is “skewing income data downward.” And based on what evidence? Mr. Weinberger doesn’t say. In any case, for very low-income households, I have no doubt that wages are more depressed today than they would have been absent large migration from Mexico. But those at the bottom of the distribution aren’t the only ones struggling: my last post suggests that the entire bottom 95% of the distribution has seen only minor income growth since 1979. Moreover, to the extent that immigration is having major effects on the income distribution, such effects would have to be depressing incomes enough at the bottom similar to the quadrupling of incomes at the top. And on this point, the economic evidence is not in Mr. Weinberger’s favor. 

Family Size:

It’s true that families are getting smaller, but is it really appropriate to adjust empirical data to account for family size? In principle, this would be important to consider if we were trying to analyze the change in material standard of living: a family with 5 children can’t afford things that a family with 1 child and the same income can. But in making this adjustment, we would be attributing growing incomes to the fact that American families chose to have fewer children. But why should the choice to have a smaller family be considered a form of income growth? If the question is why median household income didn’t rise with productivity, then adjusting for family size tells us nothing: it might tell us why some low- to middle-income households can buy an XBox, but it doesn’t answer the question why they feel worse off than their 1979 counterparts.

Real Compensation:

I’m not sure what Mr. Weinberger’s chart on real compensation is supposed to prove. The chart shows the growth in total rather than average compensation over time. All this chart shows is that total real compensation per hour has increased steadily since 1947, not because real compensation per worker has steadily increased, but because total employment has increased with population growth: If more people are working, then of course total compensation increases. Nevertheless, the growth in average real compensation, that is, average wages plus benefits for an hour of work adjusted for inflation, has actually tracked labor productivity pretty well. But as Mr. Weinberger points out, looking at straight averages can be misleading. Indeed, looking at data on real wages – which constitutes roughly 70% of a typical worker’s compensation – economists Ian Dew-Becker and Robert Gordon find that “over the entire period 1966-2001, as well as over 1997-2001, only the top 10 percent of the income distribution enjoyed a growth rate of real wage and salary income equal to or above the average rate of economy-wide productivity growth.” Similarly on the benefits side, anyone who has looked at data on executive compensation knows that the majority of the gains in benefits have also been disproportionately at the top.    

Stephen Rose:

Finally, Mr. Weinberger points to an article by Stephen Rose who claims that the middle class is shrinking because people are getting richer. At first glance, it sounds like Rose is making a unique argument: the percentage of people in households that bring in more than $100,000 has doubled since 1979 from 12% to 24%. But if you think about it for a minute, you’ll notice that Rose’s point is actually consistent with the conclusion that the U.S. has rapidly growing inequality. Census data show that the income of the household at the 80th percentile was $79,851 (in 2009 dollars) in 1979 and $103, 448 in 2007, which is pretty close to the 24% of households Rose claims now make above $100,000. But by focusing on income levels rather than growth rates, Rose totally misses the boat: that’s only about a 1% annual income growth rate, the same anemic growth rate I presented for the 80th percentile in Figure 1 in my last post, which doesn’t even come close to the 2.8% growth rate our parents and grandparents experienced from 1947-1973. The 1% annual growth rate at the 80th percentile is also about double the annual income growth rate of the typical, or median, American household, but still far less than the 15% growth rate at the very top. Is a 1% annual income growth rate supposed to be the income surge Rose is implying? Don’t let Rose’s semantics fool you: following his own logic, if income growth rates continue their trend since 1979, then in the year 2497 Stephen Rose the Fifth could claim that 80 percent of American households make over $100,000 per year. Feel the surge! 

In summary, it’s obvious that denying that income inequality exists has clearly become a part of the conservative movement, presumably because the implications of such a situation require policy prescriptions that aren’t in alignment with the right’s philosophical goals. But there are substantive things to say about the income distribution. Economists don’t actually know all the reasons why the top 1 percent’s incomes have soared while the typical American family has been left in the dust, nor do economists agree about what we should do about it. But if Mr. Weinberger’s last post is a representative sample of how conservatives feel about income inequality, all I can glean is that conservatives aren’t interested in discussing substance: denying that income inequality is a problem might fill an ideological void, but it doesn’t make the problem go away.

Read Mr. Mitchell's Closing Statement

Read Mr. Weinberger's Closing Statement

Read Mr. Weinberger's Rebuttal

Read Mr. Mitchell's Opening Statement

Read Mr. Weinberger's Opening Statement


  1. How do you plan on responding to Mr. Weinberger's point on the distribution of wealth?

  2. Mr. Mitchell-

    Until you can respond to Mr. Weinberger's point on transfer payments my vote is with him.

  3. Uh, if consumption inequality and income inequality are not different, then the liberal argument that tax cuts don't stimulate the economy because the rich don't spend their money is false.

  4. "I’m not sure what Mr. Weinberger’s chart on real compensation is supposed to prove. The chart shows the growth in total rather than average compensation over time."

    No, the chart shows exactly what it says that it shows: Real Compensation PER HOUR. For example, if Real Compensation Per Hour is $20/hr, and it increases 100%, then it will be $40/hr. Changes is population have nothing to do with it.

  5. "Indeed, looking at data on real wages – which constitutes roughly 70% of a typical worker’s compensation – economists Ian Dew-Becker and Robert Gordon find that 'over the entire period 1966-2001, as well as over 1997-2001, only the top 10 percent of the income distribution enjoyed a growth rate of real wage and salary income equal to or above the average rate of economy-wide productivity growth.'"

    You've missed the point entirely. I'm sure it was an honest mistake ;).

    Wages are irrelevant because businesses are today compensate with benefit packages instead of wages. This is thanks to tax incentives given to businesses by the left because they don't believe people can be responsible with their own money, so they use the tax system to alter the way it's spent. It makes no difference, then, that 70% of compensation is wages, because in earlier decades, that number was much higher!

    This is so obvious that the only conclusion one can come to is that you're deliberately evading the point.

  6. Who is Anonymous and why are they so incoherent? Consumption vs income inequality has nothing to do with whether tax cuts are stimulative or not, Mr. Weinberger's chart shows the rise in population (and hence the rise in total compensation), and I'm not sure how you deduced that wages aren't important. Is this what flies for conservative conventional wisdom these days?

  7. I don't know how Mr. Mitchell is losing in the poll. Mr. Mitchell's dominance in this debate is so one-sided I'm surprised Mr. Weinberger has any votes at all.

    Mr. Weinberger: Next time, forget the Alan Reynolds stuff. It's revealing that none of his work is actually published in an academic journal, while Mr. Mitchell continuously cites professional articles to back his position. If I have to choose between an article printed on a biased think-tank's website and a journal article discussing empirical findings, I think the choice is pretty clear.

  8. I can understand Anonymous' point--adding benefits compensation to wage numbers would provide a more accurate picture, no?