Jon Henke
Bruce "McQ" McQuain
Dale Franks
Bryan Pick
Billy Hollis
Lance Paddock


Recent Posts
The Ayers Resurrection Tour
Special Friends Get Special Breaks
One Hour
The Hope and Change Express - stalled in the slow lane
Michael Steele New RNC Chairman
Things that make you go "hmmmm"...
Oh yeah, that "rule of law" thing ...
Putting Dollar Signs in Front Of The AGW Hoax
Moving toward a 60 vote majority?
Do As I Say ....
QandO Newsroom

Newsroom Home Page

US News

US National News

International News

Top World New
Iraq News
Mideast Conflict


Blogpulse Daily Highlights
Daypop Top 40 Links


Regional News


News Publications

The American Dream
Posted by: Dale Franks on Thursday, May 19, 2005

I wasn't able to get to this yesterday due to time constraints, but Alan Reynolds' Wall Street Journal piece requires a closer look.

This last week, the subject of income distribution has been featured in the press, and the conclusions being made are...questionble.
Major newspapers are in the throes of Mobility Mania: who "makes it" in America, and why; who doesn't, and why not. The Wall Street Journal began a series last week titled "Challenges to the American Dream." The New York Times followed suit with a multiparter on "Class in America," which aims to disparage the notion that the U.S. is a land of opportunity by claiming that "new research on mobility, the movement of families up and down the economic ladder, shows there is far less of it than economists once thought and less than most people believe."

Yet the scholarship commonly cited in support of such assertions—new research by Gary Solon of the University of Michigan, David I. Levine of Berkeley, and Bhashkar Mazumder of the Chicago Fed, among others—says no such thing. A paper last fall by Mr. Solon observed that several of the newest estimates, including two from Messrs. Levine and Mazumder, suggest that it has become substantially easier to move from one economic class to another (as a 1997 Urban Institute study also concluded). Those new results were statistically weak, however, and an alternative estimate from Messrs. Levine and Mazumder pointed in the opposite direction—implying family background might have grown more important between the early 1980s and early 1990s. But they described the latter result as merely "suggestive," and Mr. Solon now suspects the data were distorted. As for the latter's own research, he concluded that "our estimates are still too imprecise to rule out modest trends in either direction."

The discovery that something has not changed, or might have moved imperceptibly in either direction, would not normally be considered front-page news.
Ah, but income distribution is far too important a political issue to be left to the dry statistics of economists. It is, after all, usually one of the prime exhibits for raising taxes on "The Rich". Besides, the data are so susceptible to...ah...subtle interpretation. In fact, if you don't know some of the background behind the statistics, you're awfully likely to get a story from the press that is "fake, but true". Consider the following:
• Households with two full-time workers earn five times as much as households in which nobody works. Median income for households with two full-time earners was $85,517 in 2003 compared with $15,661 for households in which nobody worked. Median income for households with one worker who worked full-time all year was $60,852, compared with $28,704 for those who worked part-time for 26 weeks or less.

Alan Blinder of Princeton emphasized this point in a 1980 study: "The richest fifth of families supplied over 30% of the total weeks worked in the economy," he wrote, "while the poorest fifth supplied only 7.5%. Thus, on a per-week-of-work basis, the income ratio between rich and poor was only 2-to-1. This certainly does not seem like an unreasonable degree of inequality."

• Experienced supervisors earn twice as much as young trainees. Median income for households headed by someone age 45 to 54 was $60,242 in 2003, compared with $27,053 for those younger than 24. When we define people as poor or rich at any moment in time, we are often describing the same people at earlier and later stages of life. Lifetime income is a moving picture, not a snapshot.

• Those with four or more years of college earn three times as much as high school dropouts. Median income for college grads was $68,728 in 2003, compared with $22,718 for those without a high school diploma.
These are three factors that should be blindingly obvious: People who work more, earn more income. Older people who have lots of career experience earn much more money than entry-level 20-somethings. College grads earn more than high school dropouts. Income, in other words, is most strongly correlated with work, age, and education.

For instance, in the top income quintile there is an average of two workers per household. In the bottom quintile, there is less than 1 worker per household. In other words, in about half the households in the lowest income quintile no one works at all. That fact alone puts quite a different spin on income disparities. Moreover, the little known fact is that the number of households in the lowest quintile with no workers has been rising for the last 40 years or so.

And, the factor of time is important, too. In general, people start off life in the lower quintiles, then move to higher ones because they get pay raises over the years. They get promotions, and move to new, better paying jobs. So, age and experience alone causes quite significant movement between income quintiles for most people.
To deny progress, the Times series claims that "for most workers, the only time in the last three decades when the rise in hourly pay beat inflation was during the speculative bubble of the 90's." Could anyone really believe most workers have rarely had a real raise in three decades? Real income per household member rose to $22,966 in 2003 from $16,420 in 1983 (in 2003 dollars)—a 40% gain.
Per-capita real income has skyrocketed over the last 20 years. Yet, still, the idea persists that workers have been cheated out of pay rises. It's hard to reconcile that idea with the numbers.

One often hears in this context that household income has declined sharply over the last forty years or so. That also gives the impression of workers being worse off. But it ignores the fact that the number of people her household has declined from more than four people per household to about two per household today. Real household income has declined. But, income per household member has increased sharply.
It is statistically dubious to compare long-term growth of average income in any top income group with growth below. Only the top group has no income ceiling, and the lower income limit defining membership in that top group rises whenever incomes are rising. In 2003, a household needed an income above $86,867 to make it into the top 20%, but an income above $68,154 (in 2003 dollars) would suffice in 1983. When the Census Bureau averaged all the income above $86,867 in 2003, they were sure to come up with a larger figure than in 1983, when the average was diluted by including incomes nearly $20,000 lower.
This is also an important consideration. As real incomes rise across the board, getting into the top quintile means that the entry point to that quintile requires a much higher income. And the fact that there's no top end to the highest quintile skews the results.

You see, income quintiles divide the population by income levels, not by actual numbers of people. The number of people in the top income quintile is about the same as the number of people in the lowest income quintile, namely, about 14 million in each. The population is distributed in a bell curve, with the vast majority of people in one of the middle quintiles. Ah, it appears that I have a made a bit of a mistake here. I was going from memory instead of checking the numbers before posting, and my memory failed me. It is the number of very rich and very poor that are similar. After rechecking, though, I see that the top quintile and bottom quintile contain vastly different numbers of people. As it happens, the top quintile contains slightly more than 24% of the population, while the bottom quintile contains a bit less than 15% of the population. But this makes the income distribution figures even more skewed because there are 65% more people in the top quintile than in the bottom quintile.

I think it might be interesting to divide the population into 5 equal population quintiles by income, and then look at the income data. I don't know of anyone who does that though. And, now I do. The Heritage Foundation appears to have done this, although I offer the link with the caveat that this comes from a conservative think tank, and not a peer-reviewed academic source. With that in mind, they present the graph below:

Income Distribution in equal population quintiles.

It's clear, though, that discussions of income distribution are far more complicated than they appear upon casual inspection.
Return to Main Blog Page

Previous Comments to this Post 

An interesting analysis. I have to wonder if the households that have "no workers" counts those who have already retired.. In which case the increase in households where noone works could largely be explained away by an aging population...
Written By: Jamie Rosensteel
I’m hoping retirees are the reason that households with no workers actually earn any $ at all. Otherwise I’m quitting tomorrow, after I get my check.
Written By: Sharp as a Marble
You see, income quintiles divide the population by income levels, not by actual numbers of people....I think it might be interesting to divide the population into 5 equal quintiles, and then look at the income data.
I’m a little confused. The Effective Federal Tax Rates document you linked to a while back says that quintiles "contain equal numbers of people, but because households vary in size, quintiles generally contain unequal numbers of households." Did you confuse "number of people" and "households" in your post?
Written By: JWG
URL: http://
I’m a little confused. The Effective Federal Tax Rates document you linked to a while back says that quintiles "contain equal numbers of people, but because households vary in size, quintiles generally contain unequal numbers of households." Did you confuse "number of people" and "households" in your post?
No. It’s an entirely different series of reports, so the quintiles are calculated differently. Household and personal income use different types of quintile. Personal income quintiles are created by the Census Bureau, and, in fact, the different quintiles contain different numbers of persons.
Written By: Dale Franks
OK, thanks...I get it now. From the CBO:
CBO defines income quintiles as having equal numbers of people, but it allows the number of households to vary. In contrast, the Census Bureau defines quintiles as containing equal numbers of households and differing numbers of people. Person-based quintiles give equal weight to all people; household-based quintiles give less weight to people in large households than to people in smaller households....Basing quintiles on people results in more households in the lower quintiles; the share of income in the lower quintiles is thus greater, relative to household-based quintiles, at the expense of the upper quintiles. Moving from household-based to person-based quintiles increases the income shares of the lower and second quintiles and lowers the shares of the upper three quintiles. However, the distribution of income shares does not differ significantly for the two definitions of income quintiles.
Written By: JWG
At the risk of displaying my ignornace, a couple of thoughts:

I think a quintile is, by definition, always a fifth of elements in a given set. Whether it’s households or individuals that are the elements in the set, every quntile should, theoretically, have an equal number of them.

Dale, you also state, "Real household income has declined. But, income per household member has increased sharply." Has real household income declined?

Looking at the report Income in the United States 2002 (specifically at table A1 on page 23), it looks a lot like median household income has increased from just over $30,000 a year in 1967 to just over $40,000 in 2002. These figures are in 2002 CPI-U-RS adjusted dollars, which should more or less account for inflation, no?

Your statement is so clear and concise that I’m second-guessing myself, but I can’t see how the median household income levels reported through the census can be considered a decline.

Written By: JMD

Add Your Comment
  NOTICE: While we don't wish to censor your thoughts, we do blacklist certain terms of profanity or obscenity. This is not to muzzle you, but to ensure that the blog remains work-safe for our readers. If you wish to use profanity, simply insert asterisks (*) where the vowels usually go. Your meaning will still be clear, but our readers will be able to view the blog without worrying that content monitoring will get them in trouble when reading it.
Comments for this entry are closed.
HTML Tools:
Bold Italic Blockquote Hyperlink
Vicious Capitalism


Buy Dale's Book!
Slackernomics by Dale Franks