Tuesday, November 13, 2007

Despite Myths, Income Mobility is Alive and Well

From today's WSJ editorial "Movin' On Up," which is based on a Treasury Department study to be released today:

One of the notable, and reassuring, findings is that nearly 58% of filers who were in the poorest quintile in 1996 had moved into a higher income category by 2005. Nearly 25% jumped into the middle or upper-middle income groups, and 5.3% made it all the way to the highest quintile.

Of those in the second lowest income quintile, nearly 50% moved into the middle quintile or higher, and only 17% moved down. This is a stunning show of upward mobility, meaning that more than half of all lower-income Americans in 1996 had moved up the income scale in only 10 years.

Also encouraging is the fact that the after-inflation median income of all tax filers increased by an impressive 24% over the same period (see chart above). Two of every three workers had a real income gain -- which contradicts the Huckabee-Edwards-Lou Dobbs spin about stagnant incomes.

Those who start at the bottom but hold full-time jobs nonetheless enjoyed steady income gains. The Treasury study found that those tax filers who were in the poorest income quintile in 1996 saw a near doubling of their incomes (90.5%) over the subsequent decade. Those in the highest quintile, on the other hand, saw only modest income gains (10%). The chart above tells the story, which is that the poorer an individual or household was in 1996 the greater the percentage income gain after 10 years.

Bottom Line: A prevailing mythology persists that: a) those in the poorest income quintile are doomed to a life of poverty, with no hope of ever rising to a higher quintile, b) those in the lowest income quintile are doomed to a life of falling real income, c) those who end up in the top quintile become lifelong members of an exclusive private club that is closed to new members, and d) the richest income group gets richer and richer over time, at the expense of lower income quintiles.

The new Treasury data support previous research, which overwhelmingly shows that the prevailing myth is pure, populist hokum. The truth is that the rich get richer and the poor get richer, and today's poor are often tomorrow's rich. Income mobility is alive and well today, as it has been for decades. No, make that centuries.


At 11/13/2007 9:03 AM, Anonymous Anonymous said...

This treasury department study you quote fails to take into account or even acknowledge the differences in income and expense people live with in different parts of the country. By being so broad of a study it fails to have meaning.

I think you are in Flint, Michigan. Well, that house you can buy in Flint for $1500 would cost $500,000 where I live. So I can understand how you might think that being in the top quintile would be a good thing.

In the county where I live a single person qualifies for low income housing with an annual income under $80,000.

If a person is in the low part of the highest quintile in my county and surrounding counties they are considered "poor" and I guarantee you that they feel and look "poor."

The numbers in the Treasury Dept. study are so broad that they have no meaning.

I would like to see what the difference in costs were for these people in 1996 vs 2005. Show me that and the numbers might have more meaning.

Many regions that were doing well in 2005 were not so good in 1996. Things have changed but have things gotten better for people or just moved them into a higher tax bracket?

At 11/13/2007 9:31 AM, Anonymous Anonymous said...

Anonymous said:
"I would like to see what the difference in costs were for these people in 1996 vs 2005. Show me that and the numbers might have more meaning."

The U.S. Bureau of Labor Statistics (http://www.bls.gov/) has data that you seek. You’ll have to spend a few minutes building custom tables for what you want, but it’s all there for the taking.

At 11/13/2007 9:57 AM, Anonymous Anonymous said...

In dollar terms, a 90% increase in income ranging from $0 to $18,000 is less than a 10% increase in incomes ranging from $88,000 to $100 million.

Expressing the increase in percentage terms is highly misleading when there is such variation in the base amount. Of course, I've come to expect as much around here.

At 11/13/2007 10:12 AM, Anonymous Anonymous said...

By the way, if you want to see what the change in income in each quintile REALLY looks like in dollar terms, instead of meaningless percents -- look here:


Mark Perry -- if you're up to it, why don't you post the graph on your front page and make a comment.

At 11/13/2007 2:56 PM, Anonymous Anonymous said...

Most people view salary increases in terms of a percentage increase in relation to what they make now. And that's how a corporation (in the real world) would give an equitable raise to many people; when employees are evaluated for a raise, they might get 3, 4, or 5 percent, etc, based on their performance. It doesn't make any sense for a low-income earner to get the same dollar amount raise as a person in the upper income.

Looking at these increases in relation to cost of living would be illustrative -- yes, it would illustrate how the higher income earners have lost far more buying power than the poor.

Also, raises don't move your entire income into a new tax bracket, that's not how tax brackets work.

At 11/13/2007 6:37 PM, Blogger juandos said...

When did Wikipedia become a credible source?

Then again its holymoly...

What part of the following confuses you holymoly: "The study tracks what happened to these tax filers over this 10-year period. One of the notable, and reassuring, findings is that nearly 58% of filers who were in the poorest income group in 1996 had moved into a higher income category by 2005"?...

Regardless of base income or percentages the fact remains that according to the Treasury Dept. poor people didn't ALL remain as poor as they were...

holymoly if you got something that is at least as credible as the Treasury Dept. figures why not roll them out for us?

At 11/13/2007 6:38 PM, Blogger David Foster said...

These are *family* incomes, right? If that's correct, I wonder how the study would treat 2 people who were single 10 years ago but are now married to each other.

Surely no one would do something as dumb as taking 2 people making $50,000/year each, and later making $100,000/year as a couple, and conclude that their family income has doubled...But I wonder how this is handled, especially given the age of the people in the sample.

At 11/14/2007 8:21 AM, Blogger doclawson said...

Regression to the mean is a beautiful thing to behold.

At 11/14/2007 9:09 AM, Blogger Ironman said...


If you really want to see income distribution changes, first here's a post showing the number, percentage and percentile changes for all income earners between the ages of 15 and 74 from 1995 to 2005 for incomes from $0 to $95,000 in constant 2004 dollars.

If you want to see how the distribution of income changed as these income earners got 10 years older, see this post.

And if you want to compare the income distribution between a given age group in 1995 with its counterparts in 2005, see this post.

At 11/14/2007 9:25 AM, Anonymous Anonymous said...


If you had bothered to follow the link, you'd have seen that the Wikipedia entry was using data from the Census bureau.

"Distribution of income in the United States of America from 1967 through 2003. Numbers are normalized to 2003 United States dollars. The data source is DeNavas-Walt, Carmen; Bernadette D. Proctor, Robert J. Mills (08 2004). Table A-3: Selected Measures of Household Income Dispersion: 1967 to 2003 (PDF) (English). Income, Poverty, and Health Insurance Coverage in the United States: 2003 36-37. U.S. Census Bureau. Retrieved on 2007-06-16."

At 11/14/2007 9:34 AM, Anonymous Anonymous said...

ironman -

That is interesting data at political calculations, but I'd like to see more careful analysis. The age groupings are very strange. 15-24? That mixes teen and adult workers. Not to mention that 10 year windows are subject to demographic shifts within each window. Since incomes change so rapidly during that timeframe for those individuals in the 25-34 year age group who are on a *decent* age-earnings trajectory, you would want to make sure there wasn't a change in, say, the number of 25 year-olds, say, versus 34 year-olds over the 10 years. A 10 year window is enough for a "baby boomlet" to have major impact within such broad age categories. I'd like to see a true regression-based wage adjustment.

At 11/14/2007 7:07 PM, Blogger Ironman said...


I don't disagree with you regarding the data - the limiting factor is that the U.S. Census doesn't break down the Age 15-24 group into Age 15-19 and Age 20-24 groups as it does with each of the older age groups.

As it happens though, for the years of the analysis, each age grouping corresponds to people born in say the 1940s, or the 1950s, etc., which is why I selected them!

Your point about the number of people in each age group is a good one, and one I addressed elsewhere in the series.

As things are still in development, I'll just drop a hint and suggest that there are more entertaining ways than regression-based wage analysis to chart the impact of the baby-boom generation in the distribution of income by age.

At 11/14/2007 7:47 PM, Blogger juandos said...

"If you had bothered to follow the link, you'd have seen that the Wikipedia entry was using data from the Census bureau"...

Ahhh, the claim that Wikipedia was supposedly using Census bureau data... Oh yeah! That makes them credible to you?!?!

Why didn't you just link the Census Bureau data instead?

At 11/14/2007 8:23 PM, Anonymous Anonymous said...

Cool. I didn't draw the connection that that was your page. Looking forward to seeing more analysis from you in the future. The bump up in the middle range of the distribution in the 25-34 year old is very interesting.

You need to get a new shtick, cuz your wahmbulance is running out of gas. I didn't feel like making my own graph when someone else already did. If you think the graph is bogus, do your own analysis of the same data.

At 11/14/2007 8:39 PM, Anonymous Anonymous said...

Taking David's comment about these being *family* incomes (which the neo-liberals on this site conveniently ignored), I don't believe the study takes into account how many of these incomes rose as a result of someone getting a second job, or another member of the household getting a job. Can anyone confirm or deny this? Obviously, such information has an impact on what the study actually shows about wages.


Post a Comment

<< Home