Monday, August 01, 2011

Income Mobility for All Income Groups is Significant


From a 2010 St. Louis Fed report "U.S. Income Inequality: It’s Not So Bad":

"One big problem with inferring income inequality from the census income statistics is that the census statistics provide only a snapshot of income distribution in the U.S., at a single point in time. The statistics do not reflect the reality that income for many households changes over time—i.e., incomes are mobile. For most people, income increases over time as they move from their first, low-paying job in high school to a better-paying job later in their lives. Also, some people lose income over time because of business-cycle contractions, demotions, career changes, retirement, etc. The implication of changing individual incomes is that individual households do not remain in the same income quintiles over time. Thus, comparing different income quintiles over time is like comparing apples to oranges, because it means comparing incomes of different people at different stages in their earnings profile.

The U.S. Treasury released a study in November 2007 that examined income mobility in the U.S. from 1996 to 2005. Using data from individual tax returns, the study documented the movement of households along the distribution of real income over the 10-year period. The study found that nearly 58 percent of the households that were in the lowest income quintile (the lowest 20 percent) in 1996 moved to a higher income quintile by 2005 (see top chart above). Similarly, nearly 50 percent of the households in the second-lowest quintile in 1996 moved to a higher income quintile by 2005. Even a significant number of households in the third- and fourth-lowest income quintiles in 1996 moved to a higher quintile in 2005.

The Treasury study also documented falls in household income between 1996 and 2005. This is most interesting when considering the richest households. As shown in the bottom chart above, more than 57 percent of the richest 1 percent of households in 1996 fell out of that category by 2005. Similarly, more than 45 percent of the households that ranked in the top 5 percent of income in 1996 fell out of that category by 2005.

Thus it is clear that over time, a significant number of households move to higher positions along the income distribution, and a significant number move to lower positions along the income distribution. Common reference to “classes” of people (e.g., the lowest 20 percent or the richest 10 percent) is quite misleading because income classes do not contain the same households and people over time."

Thomas Sowell sums up the issue of income inequality and income mobility this way:

"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."

14 Comments:

At 8/01/2011 3:14 PM, Blogger Benjamin Cole said...

Interesting study. What is unclear is if there are large numbers of families or households camping out on the "borders" between quintiles.

For example, I have a slightly better year, my net goes from $110k to $120k, and I migrate into a higher quintile, or vice-versa. What looks like mobility is an artifact arbitrary lines across income groups.

Additionally, I might be incredibly wealthy and have lots of income, but like Dodgers owners McCourts, declare losses year after year.

Nevertheless, I suspect there is some mobility, as the young of wealthy families often seem dissolute.

I did know a scion of one wealthy L.A. family, and all he did was smoke pot all day. Then, he borrowed $1 million from his Dad, bought a lot in L.A., got a construction loan, built a house, and sold it for $4 million, netted $2 million. Went back to smoking pot all day.

He would bounce up and down in the quintiles, I imagine.

 
At 8/01/2011 3:58 PM, Blogger Hydra said...

looks like regression to the mean. Has nothing to do with economics. Could be the same for people gaining or losing weight as it is for gaining or losing money.

 
At 8/01/2011 4:07 PM, Blogger Benjamin Cole said...

BTW--
Martin Feldstein, Reagan's top-gun on the economy, has a must-read op-ed in today's WSJ.

Now, Feldstein is a Harvard professor so he is not as smart as Morganovich, or Vange, but he is really smart, and he likes a cheap dollar.

 
At 8/01/2011 6:12 PM, Blogger Che is dead said...

This comment has been removed by the author.

 
At 8/01/2011 6:35 PM, Blogger George said...

To answer Benjamin's questions, one would need a longer serial study, or more information. One way would be to take a really large number of people earning 10,000, or a hundred thousand, or a million dollars a year, and generate a graph of the distribution of incomes they had ten years later, twenty years later, thirty years later,... and compare with the overall distribution at that later time.

 
At 8/01/2011 7:56 PM, Blogger kleht said...

If we are going to show the bottom income groups moving up, it seems that we should also be showing the top income groups moving in the same direction - up. Or show the two groups moving up as well as moving down. I would think this would give a better picture of what's actually happening.

The graphs as shown appear to be comparing apples to oranges here. I may be missing something, of course. But my understanding is that while the bottom income groups' income did indeed rise from 1996 through 2005, the top groups' income rose at a substantially higher percentage rate.

 
At 8/01/2011 8:34 PM, Blogger Sam said...

I'm not sure about the conclusions here. The quintiles represent the relative distribution, not the absolute income levels. I'm trying to understand what it means if 58% of the people in a quantile move up. I think that for that to happen a large number of people must move down into the lower quintile, otherwise it won't be a quintile anymore. Put another way: if a lot of people start making more money, the quintile their in doesn't change, just the boundaries move.

 
At 8/01/2011 8:35 PM, Blogger Sam said...

I mean "...the quintile they're in..."

 
At 8/02/2011 3:24 AM, Blogger Ron H. said...

George

"To answer Benjamin's questions, one would need a longer serial study, or more information. One way would be to take a really large number of people earning 10,000, or a hundred thousand, or a million dollars a year, and generate a graph of the distribution of incomes they had ten years later, twenty years later, thirty years later,... and compare with the overall distribution at that later time."

Aahh...Isn't that exactly what this study does?

Although the number of households isn't indicated, since the study is based on census data, one might safely assume that it's a very large number.

Benji's questions are meaningless.

 
At 8/02/2011 3:39 AM, Blogger Ron H. said...

"If we are going to show the bottom income groups moving up, it seems that we should also be showing the top income groups moving in the same direction - up."

The actual dollar values that define each quintile isn't important, and would be different numbers between starting and ending dates.

While some who were in any given quintile in 1996 may still be in the same quintile in 2005, their income could have increased significantly without changing their relative position.

Those in the top quintile can't move into a higher quintile.

 
At 8/02/2011 3:54 AM, Blogger Ron H. said...

Sam: "I'm not sure about the conclusions here. The quintiles represent the relative distribution, not the absolute income levels. I'm trying to understand what it means if 58% of the people in a quantile move up. I think that for that to happen a large number of people must move down into the lower quintile, otherwise it won't be a quintile anymore."

You are absolutely correct. For 58% to move out of the lowest quintile, 58% must move down from other quintiles. We just don't know what percentage came from each one, based on only the information we have.

 
At 8/02/2011 11:31 AM, Blogger morganovich said...

"For example, I have a slightly better year, my net goes from $110k to $120k, and I migrate into a higher quintile, or vice-versa. What looks like mobility is an artifact arbitrary lines across income groups."

58% of people crossing the line is unlikely to be an artifact.

and you appeal to authority is as ridiculous as it is self defeating.

you have nothing like the education of professional training that i do, so by your own logic, you should shut up and listen.

 
At 8/02/2011 11:41 AM, Blogger morganovich said...

"I think that for that to happen a large number of people must move down into the lower quintile, otherwise it won't be a quintile anymore. Put another way: if a lot of people start making more money, the quintile their in doesn't change, just the boundaries move."

sam-

this is easily explained by 3 factors:

entry to the workforce.

exit from the workforce.

immigration.

new entrants and immigrants tend to start at a low level. they are constantly refilling the bottom quintiles. as you progress in your career, you move up. when you retire, most people drop back down again.

it's a life cycle thing.

this is a large potential flaw in the "only the top groups are gaining" argument.

if most people wind up moving up to the top groups late in their career, they are still participating in that.

income over about $55k puts you in the second to highest quintile. lots of people can expect to hit that in a career.

also note:

this is household data. to examine actual well being, you need to adjust ti for household size, which has been dropping and for mean number of earners in the household, which has been going up in the top brackets but not as much in the lower ones.

the 15% of american households earning over $100k have a mean number of earners of 2.

the national median is 1.35 with an income of 45k.

adjust that for earners and you get $67k if both worked.

this would seem to imply that you could get significant fluctuation in incomes around pregnancy as well.

 
At 8/02/2011 11:47 AM, Blogger morganovich said...

here is some data to play with:

http://en.wikipedia.org/wiki/Household_income_in_the_United_States

the average household earning 24,999 has .79 workers and 2.14 members

the average at 99,000 is 2.12 and 3.33.

normalize that for workers and dependents and you get 14,787 vs 13,588.

the higher bracket is actually lower per worker per dependent.

you need to be very careful of household data's comparability.

 

Post a Comment

<< Home