Tracking the same households over time shows significant income mobility

Thanks to James Pethokoukis for his ongoing efforts to cite empirical evidence that exposes many of the myths about “exploding income inequality” and “wage stagnation” in America (see recent posts here, here, here, and here).

In an editorial in Investor’s Business Daily last year, Thomas Sowell made this insightful comment about the issue (emphasis added):

Only by focusing on the income brackets, instead of the actual people moving between those brackets, have the intelligentsia been able to verbally create a “problem” for which a “solution” is necessary. They have created a powerful vision of “classes” with “disparities” and “inequities” in income, caused by “barriers” created by “society.” But the routine rise of millions of people out of the lowest quintile over time makes a mockery of the “barriers” assumed by many, if not most, of the intelligentsia.

As I reported on the Enterprise Blog last March, the dynamism of the U.S. labor market that Sowell describes is generally underappreciated, and has received almost no attention from those complaining about income inequality and stagnant household income. Contrary to prevailing public opinion that households get stuck at a given income level for decades or generations, there is strong empirical evidence that households move up and down the economic ladder over even very short periods of time.

For example, recent research from the Federal Reserve Bank of Minneapolis is summarized in the table below, based on income data from the Panel Study of Income Dynamics that followed the same households from 2001 to 2007.  The empirical results answer the question: For households that started in a given earnings quintile (20 percent group) in 2001, what percentage of those households moved to a different income quintile over the next six years?

Below are some of the main findings about the income mobility of American households over a short, six-year period from 2001 to 2007 based on the Minneapolis Fed study:

1. Looking across the first row of data in the table from left to right, we can see that for those U.S. households that were in the lowest earnings quintile (bottom 20 percent) in 2001, only 56 percent of those household remained in that same income quintile six years later in 2007, and almost half—44 percent—had moved up to one of the four higher income quintiles by 2007. Five percent of the lowest-income households in 2001 had moved up to one of the top two quintiles by 2007.

2. Looking across the fifth row of data from right to left reveals that of those households in the highest earnings quintile (top 20 percent) in 2001, only 66 percent remained there six years later and 34 percent had moved down to one of the four lower income quintile by 2007. Five percent of those highest-income households had moved all the way to the bottom income quintile in only six years.

3. For those households in the middle-income quintile in 2001, 42 percent remained in the same quintile in 20007, about one-third (32 percent) moved to a higher-income quintile, and slightly more than one-fourth (27 percent) moved to a lower-income quintile.

4. The bottom row in the chart shows that for households in the second, middle, and fourth income quintiles in 2001, more than half of each group moved to a different earnings quintile by 2007 (61 percent, 58 percent, and 55 percent, respectively).

Bottom Line: In the discussions on income inequality and wage stagnation, we frequently hear about the “top 1%” or the “top 10%” or the “bottom 99%” and the public has started to believe that those groups operate like closed private clubs that contain the exact same people or households every year.  But the empirical evidence displayed above tells a much different story of dynamic change in the labor market—people and households move up and down the earnings quintiles throughout their careers and lives. Many of today’s low-income households will rise to become tomorrow’s high-income households, and some will even eventually be in the “top 10% or “top 1%.”  And many of today’s “top 1%” or top income quintile members are tomorrow’s middle or lower class households, reflecting the significant upward and downward mobility in the dynamic U.S labor market – an important point in any discussion about “exploding income inequality.”

13 thoughts on “Tracking the same households over time shows significant income mobility

  1. This truth, that households move up and down the economic ladder over time, needs to be broadcast again and again, until everyone is aware.

  2. Thanks for this. Fantastic original work and summary.

    The only thing that saddens me as a patriotic American more than the fact that there are some who would work with such conviction to trick folks into believing that the “land of opportunity” should somehow “evolve” into the “land of shared success” is how many people abandon so much to buy into it.

    Think about the number of things a person must lose in order to believe that “the 1%” owe the 99% something special.

    Hope: I can live my dreams if earn them.
    Determination: I relentlessly do what it takes to earn my dreams.
    Pride: My contribution to society has the value that society assigns.
    Confidence: No one is in charge of my destiny but me.
    Self-respect: No one owes me anything. I get what I earn.
    Character: When I make a fair deal I live with the consequences, good and bad.


    • DanFarfan – I am forwarding your post verbatim to my friends. Dan Adams (Dilbert) wrote that people are abandoning the middle class dream of achieving success for having the rewards shifted to them from a successful person out of “Fairness”. They may have abandoned Hope, Determination, Pride, (Self)Confidence, Self-Respect, and Character or , most likely, they never had it developed in them to begin with. Even more likely, they have had developed in them the concept of sharing the reward out of fairness without understanding the fairness of sharing the work, risk, and sacrifice in order to achieve success.

  3. When I read the RSS in the headline, Sowell was the first thing to come to mind! I haven’t gotten around to reading any of his books yet, but I have heard interviews when he talks about such things. I was wondering if he would be mentioned in the article hehe Thanks for bringing this into the light. The PDF is only 30 pages too, so I may actually be able to read this one!

  4. Not sure if any of you looked at the underlying Fed data, but I had a strong suspicion about what this movement means… I was right. You needn’t look much further than Table 1 or Figure 1 of the paper by the Fed to see that movement between lower quantiles can easily be achieved, because the differences between income levels at the bottom, and even the middle, are small and earnings are still low. Income is ZERO up until the 20th quantile. Figure 1 displaying income distribution explains why mobility is so easy for lower groups, because the variance within each is to tiny due to the fact the distribution of income is anything but normal, rendering any analysis based on quantiles statistically insignificant. At least the “intelligentsia” know how to read this data instead of imputing it with rhetoric to fit their own story. Nice try though :)

    • Nicole,

      Had you looked and read past table 1, you would have noted that the earnings, not the income, up to about 20% are zero. When you actually read the article, you will learn that the bottom 20% with zero earnings contains retired and disabled people living on incomes from SSI, SDI, etc.

      The fact that about 5 to 6% of the top 20% are leaving each year says a great deal. Remember, for everyone leaving the top 20, someone is added. That high turnover makes it hard to have a defined class. As a member of the “intelligentsia”, with whom you seem to identify, do you understand the nature of this mathematics.

      Any analysis that assumes constancy in the membership of various categories over decades is obvious mathematical nonsense. This does not stop the “intelligentsia” from making such assumptions to support their belief systems, the pillars of which are that the “rich are getting richer” and the “rich get rich on the backs of the poor” along with the belief that we must take from the rich and give to the poor. We are currently doing an extraordinary job of taking from the many to give to the retired, which also happen to be earnings poor.

      If the “intelligentsia” figures a way to “read” this data as supporting their beliefs, perhaps they should provide their adroit “data reading” services to the global warming deniers and the anti-evolutionists who have require similar fact/data spinning to justify their preconceived notions. The “intelligentsia” are missing their real “calling”, spinning data.

    • Nice try Nicole, except your looking at “Earnings” and the graph above is based on “Income”. You may want to revisit your next to last sentence.

  5. So there is a lot of mobility but almost half of it is down, even before the crash. The middle group had 32% go up and 28% go down so on balance there was 4% movement up over 6 years. That’s nothing to crow about.

    • I copied down the wrong number. 27% went down, so 5% went up, though the total is 101% so something got rounded the wrong way. So 27% went down during the booming bubble? Some always go down because of sickness and such but 27% during good times? That’s terrible. That’s something to look into. I hate to think what these numbers would be now after the crash.

  6. I applaud AEI for trying to bring a very important variable into the mix. It is undoubtedly true that many people view quantile boundaries as impervious barriers, divisions between discrete classes. This is simply false. However, it strikes me that commenters responding to Nicole have missed the forest for the trees. Yes, she mixed up earnings and income, but the point is that the probability of moving vertically among quantiles declines as incomes increase (simply as a function of the space between boundaries). No reasonable analysts are going to argue that there should be some sort of grand convergence across the spectrum (indeed that would be counterproductive). However, to argue that movement up and down the ladder is free from structural friction is just incorrect. Where people start matters a great deal. Some of the commenters appear to implicitly believe that, ceterus paribus, achievement of any of the quantiles is equally likely. If you accept that this is not the case, which the data will bear out, then it is tough to make the argument that we should not do what we can to facilitate the *capacity* to move by ensuring that those at lower quantiles have the same human capital development and affiliated services needed to compete with those who have a far more fortunate starting position.

  7. Mr.Perry,

    I agree mobility is more important than a snapshot, but I wanted to point out some housekeeping items and ‘theoretical’ concepts. In terms of housekeeping, it appears your data is off (I am asusming your table above is in reference to Table 27 in the paper; if I am mistaken, my apologies and can you clarify which table you are looking at?).

    I would also suggest focusing on Table 29 as this can be used to compare to the authors’ previous studies. Most notably, their 2001 study analyzed mobility from 1989 to 1994 and a brief comparisson between the two would seem to suggest mobility then is virtually the same in from 2001 to 2007. In the end, it isn’t whether mobility does or does not exist so much as whether it is increasing, decreasing, or in this case, staying the same.

    Finally, it is always potentially troublesome to take a point and data from a paper in which the greater amount of said paper states something akin to the opposite of your point (the authors conclude inequality increased from 2001 to 2007; I do recognize, however, this isn’t necessarily the same thing as mobility. Perhaps, inequality correlates to the volatility within mobility, but I digress.).

    Which brings me to my ‘theoreticals.’ I call them theoreticals because the data is either not yet fully known or was not neatly available in my limited browsing. Specifically, in terms of the 99% it is possible the 1% have a significantly different mobility than the top 20%, so this research does not necessarily negate the argument of unfair options for the 99%. The second point is the data referenced only goes to 2007. It is arguable mobility has changed since then, particularly amongst financial firms which received government bailouts (thus preventing downward mobility). While time will tell, I would caution the mobility argument does not prevent the 99% argument today.

    Overall I agree with the concept of mobility, but this very same concept is what concerns me most today in that government intervention may have stopped it. Overall, nice post (I can’t say I agree with Mr. Pethokoukis, but that argument is for another comment on a different post). Just my two cents.

  8. Okay how about we actually look at these numbers. The people who have the highest chance of not moving up or down are those in the top 20%. The next ones with little chance of moving up or down are those in the lowest 20%. The chance of going from the middle 20% to the highest 20% is only 8%. So what we see here is variation but very small variation with a slight step up or down. And since when we get to the lower brackets it the difference in salaries are less. So obviously a slight increase in salary will have a higher chance of pushing you up to the next bracket.

    If you notice it becomes more and more difficult to go up to the next bracket the higher you are. So if you were in the 4th 20% and wanted to get into the top 20% you had only a 21% chance. Saying that todays low income households have a decent chance of getting into the top 10% is obviously pretty wrong.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>