Census data on income distribution reveal evidence of rising income levels for a rising share of American households
AEIdeas
The top chart above is based on data from the Census Bureau on “Money Income of Households — Percent Distribution by Income Level in Constant (2009) Dollars” (Table 690) for the household income categories: a) $25,000 and under, b) $25,000 to $75,000, and c) $75,000 and over.
I posted and wrote about a similar chart last summer that displayed the distribution of “family income” and concluded in that post that the Census income data may suggest that the reason the middle class might appear to be “disappearing” is because that income group has actually moved into the upper-income category, and not falling into a lower-income group as is typically claimed. That post generated more than 200 comments with some good suggestions, including:
1. The “middle-income” category of $25,000 to $75,000 that I originally selected does not accurately represent “middle-class” Americans, and the lower limit of that category should instead be $35,000.
2. The “upper-income” category of “$75,000 and above” does not accurately reflect “upper income” in some states like California, where you might not be able to buy a house in many areas of the state with that income.
3. The distribution of income in America over time based only on US families (defined as a householder and one or more other people living in the same household who are related to the householder by birth, marriage, or adoption) is not as inclusive as the distribution of income in America based on total US households (defined as all people who occupy a housing unit regardless of relationship, and includes persons living alone). In 2010, there were about 116.7 million total US households and 77.5 million “family households,” according to the US Census.
All three of those suggestions will be addressed in this post.
Let’s start with the household income distribution data in the chart above which shows the following for the three household income groups: a) $25,000 and under, b) $25,000 to $75,000, and c) $75,000 and over, with all household income measured in 2009 constant dollars.
4. In 1967, 55.8% of American households were earning between $25,000 and $75,000 in constant 2009 dollars, one income range that might be used to describe America’s “middle class.” Also in that year, only about one out of seven (14.4%) American households had income above $75,000 (“upper-income”), and slightly less than 30% of American households were earning $25,000 or less, an income category that might be described as “low-income.”
In 1967, there were almost four American households earning a middle-class income ($25,000 to $75,000) for every high-income household earning above $75,000. Further, there were almost two “middle-income” households for every one “low-income household,” so the middle class American households earning between $25,000 and $75,000 clearly represented a significant share of US households in 1967.
Here’s what happened over time:
5. The share of lower-income households fell over time by 4.7 percentage points, from 29.8% of all US households in 1967 to only 25.1% of all US households in 2009 (see blue line in chart), while the share of middle-income households decreased by 12.5 percentage points during that period, from 55.8% in 1969 to 43.3% in 2009 (see brown line in chart). So where did those 17.2% of households go that disappeared from the lower income and middle income categories in the 42-year period between 1967 and 2009? They “disappeared” into the upper-income category of incomes above $75,000, which increased by 17.2 percentage points, from a 14.4% share of American households in 1967 to a 31.6% share in 2009. Whereas “middle class” households were so numerous that they outnumbered “upper class” households by a ratio of almost 4-to-1 in 1967, so many American “middle class” households have moved by the 2000s to the “upper class” by income, that by 2009 the ratio of middle-income to upper-income households had decreased to only 1.37-to-1. Stated differently, in 1969 there were almost 400 “middle-class” households to every 100 “high-income” household; but by 2009 there were only 137 “middle-class” households per 100 “high-income” households.
Now let’s analyze household income distribution over time using the revised income categories: a) $35,000 and under for “low-income” households, b) $35,000 to $75,000 for “middle-income” households, and c) $75,000 and over for “upper-income” (same as before), see second chart above.
6. The share of US households in the revised lower-income category ($35,000 and below) declined from 43.6% in 1969 to 36% in 2009, while the share of households in the revised middle-income category decreased from 42.1% to 32.3% over the four-decade period. Like before, the share of “upper-income” households in the unrevised category of $75,000 and above increased from 14.4% to 31.6%. With these revised income categories for low- and middle-income households, we can say that in 1969, the number of American households in both the low- and middle-income categories outnumbered “upper-income” American households by 3-to-1. But by 2009, the share of American households in the three income categories were almost equal (36% low-income, 32.3% middle-income and 31.6% upper-income).
Finally, the bottom chart above displays income distributions over time for US households based on the three new income categories: a) $50,000 and under for “low-income” households, b) $50,000 to $100,000 for “middle-income” households, and c) $100,000 and over for “upper-income” households.
Once again, I think the results are pretty striking:
7. The share of “middle-income” Americans (defined here as US households earning between $50,000 and $100,000 in constant dollars) has remained relatively stable over the last 40 years at about 30%.
8. At the same time, the share of low-income households earning $50,000 or less has decreased by 13.50 percentage points from 63.6% to 50.1% from 1969 to 2009, while the share of high-income households earning more than $100,000 has increased by 14 percentage points from 6.1% to 20.1% over that period.
According to this last analysis, middle-class households earning $50,000 to $100,000 have remained a fairly constant share of all US households over time, while there has been a significant increase in the share of high-income households that has offset the decline in the share of low-income households. Stated differently, in 1969 there were more than ten low-income American households making less than $50,000 for every high-income household making more than $100,000; by 2009, that ratio had fallen to only 2.5 low-income households per high-income household.
Bottom Line: The top two charts above show that the share of American “middle class” households making either: a) $25,000 to $75,000 or b) $35,000 to $75,000, did decline starting in the 1970s, but it was because a greater share of American households were moving up to a higher-income category ($75,000 and above), not down into a lower-income category (which were declining as shares of all households, though at a lower rate than the decline in the share of middle-income households). And that movement of the middle-class (and the lower-income group) was so significant that between 1967 and 2009, the share of American households earning incomes above $75,000 more than doubled, from 14.4% to 31.6%. Further, the bottom chart above shows that the share of US households earning $100,000 or above has more than tripled from 6.1% in 1969 to 20.1% in 2009.
On the previous CD post, Ken commented that although “Many prominent people like Paul Krugman claim that the middle class has been in decline since the 1970s, that assertion is incredibly and verifiably wrong.” According to the percent distribution of household income in Table 690 from the Census Bureau, I think Ken is exactly right. Despite all of the reports on stagnating household income, decreased mobility for the middle-class, the top 1% reaping all of the benefits of income/wealth gains over time, increasing income inequality, current generations doing worse than their parents, the general decline of the middle class, etc., the Census Bureau data and the charts above tells a different story of an America with documented evidence of rising income levels for a rising share of American households.
One aside, not entirely relevant to the post, but prompted by it: the term “middle class” is increasingly used to refer to people whose income is near the median. This is not at all what it historically meant.
Historically, there were two classes: wealthy landowners, whose income came from tenants and who did not need to work; and workers, whose income came from labor. Landowners lived well; workers, not so much. The median income, of course, was a blue collar, or “working class” income.
The “middle class” was a term invented to describe workers whose skills generated enough income to allow them to live “well” – much better than blue-collar workers and often as well as at least some landowners. Thus, “middle class” was a term invented to describe people who lived like the moderately wealthy but who, although they might have some income from capital, primarily derived their income from working.
Today, these are the people increasingly targeted for high taxation. The great irony of the policies that “tax the rich”, justified by the need to “help the middle class” is that they are, in fact, high taxes on… what was once called the middle class.
Thomas Boyle is “right on!” History and Evolution of the term “middle class”
Historically, the term “middle class” is first attested in James Bradshaw’s 1745 pamphlet Scheme to prevent running Irish Wools to France.
The term has had several, sometimes contradictory, meanings. It was once defined an intermediate social class between the landed nobility and the peasantry of Europe. While the nobility owned the countryside, and the peasantry worked the countryside, with the development of population groupings into cities, a new class, the bourgeoisie (literally “town-dwellers”), arose around mercantile activities which took place mainly in the cities. Another definition equated the middle class to the original meaning of capitalist: someone with so much capital that they could rival nobles. In fact, during the industrial revolution, to be a capital-owning millionaire was the essential criterion of the middle class.
Be semantic history what it may, the modern usage of the term “middle class” dates to the 1913 UK Registrar-General’s report, in which the statistician T.H.C. Stevenson identified the middle class as that group of people falling between the upper class and the working class. Included as belonging to the middle class are professionals, managers, and senior civil servants. The chief defining characteristic of membership in the middle class is possession of significant human capital, i.e., skills as opposed to wealth and property as the basis of their financial status.
Within capitalism, “middle class” initially referred to the bourgeoisie (“upper middle class” or owners of capital and the means of production) and the petite bourgeoisie (“lower middle class” or managers of capital and the means of production). However, with much of the petite bourgeois world fracturing into, on the one hand, poverty, and on the other, proletarianism (wage-earning employment), and with the rise of finance capitalism, “middle class” came to refer to the combination of the labor aristocracy, the professionals, and the white- collar workers.
Census data — mapping poverty in America, an interactive map.
http://www.nytimes.com/newsgraphics/2014/01/05/poverty-map/?nl=todaysheadlines&emc=edit_th_20140106
Not poverty. The less rich. Poverty is an absolute concept; however, the Census Bureau redefined poverty to be a percentage of the median income.
Could you please define income?
Is just pre-tax, pre-benefit cash paid as wages?
Or is it post-tax, post-benefit, total compensation?
Does it include interest, dividends and realized gains?
the one possible fly in this ointment seems to be the <$35k cohort which has been growing since 2000.
that might well be an immigration issue though.
“…<$35k cohort which has been growing since 2000."
Besides the illegal immigrant and Puerto Rican(poverty rate of 50%) population…
could much of it be the 75% of social security recipients who choose "early retirement" at age 62. This rapidly growing population receives annual benefits of ~$15,000?
It might be immigration as you suggest. Or early retirees, as Buddy suggests.
My suggestion is that the low point around 2000 represents the baby bust of the 1970s. Younger households are more likely to be single-earner households. The baby bust generation reached their early 20s in about 1992 through 2002. We should have seen a declining share of one income households during this period.
Sorry, but I have an appointment which prevents me from researching the data.
i’m trying to make sense of what you are saying here.
mark’s argument is that people have gotten richer.
i suppose we could track each decile and see how they gained or lost, but even that can be misleading as it misses the movement of individuals and mistakes starting wages remaining the same for economic immobility.
if, in real terms, more people earn more money, and we can show this using monetary breakpoints, how is that “torturing data in the extreme”?
i really am not seeing what you are getting at.
I think what he means is that you are not “torturing data to the extreme” in the way I like it. 😉 Presumably he thinks that the percentage share of total income going to each quintile/quartile/decile has stayed the same, but because total income has been rising, each group’s income shows up as increasing in real dollars in the graph above, but not when shown as a percentage share of total income. I’m not sure if that’s actually the case- I haven’t seen the data- or why people like him are so obsessed with percentage share and why they think it has to change.
When pundits talk about “income inequality”, they try to paint a picture of a “hollowed out” middle class descending into poverty, and the rich are becoming mega-rich at the expense of everyone else. It is important to note that real income (in constant dollars) is increasing across the board as people move up into new buckets – the standard doom-and-gloom story is simple not true. In fact, “income inequality”, where income is earned honestly (that is, not through government rent-seeking), ought to be celebrated, as it is proof that the lives of innumerable people are being improved. Income inequality is proof that society is continuing to evolve to meet the needs of it’s members, and far more effectively than government central planning can (which tends to make people’s lives worse)
but that is a deeply misleading way to look at the situation.
income growth across quintiles is only a measure of inequality if people remain in the same quintiles, which they don’t.
let’s take a very simple example of 5 jobs that pay 1, 2, 3, 4, 5.
now let’s look at 5 people who get them.
that looks very unequal.
but if people all start at job one and work up to 5, then the system is PERFECTLY equal.
thus, if the gap between starting wage and peak wage in an economy widens, even if every single person in the economy starts and ends at the same wage, it looks like more inequality (by quintile) even though there is absolutely none by individual.
the methodology you are trying to use is inadequate to measure actual opportunity or equality.
What a bunch of bozos. Chart 3 is GIGO.
Median income, you know, the 50th percentile, you know, smack dab in the middle of the income distribution, you know, the heart of the middle income bucket was $49,455 in 2010. Chart 3 supposes that the median is actually in the lower income bucket.
You might try reading the actual post, where Mark pointed out that some previous commenters thought that $75k didn’t really reflect upper income on the coasts, so he raised the thresholds for them. I’d like to see all 9 income groupings from the original data charted, rather than aggregating some of them as he did.
Chart 2 is the most promising chart as a reflection of the 2010 three income buckets.
Incomes across all the quintiles have risen since 1967 but not at the same rate* so there will be absolute movement in households from the lower to middle to upper in the Chart 2 fixed income buckets by arithmetic.
I like how you just wave away the amazing rise of incomes across all quintiles and subsequent movement up the ladder for so many families as just “arithmetic.” 😉 Who’s the bozo here? Why does it matter to you so much that the rich got richer faster when everybody got richer? Care to point out any other country in the world where everybody got richer for 30 years, so that even their “poor” make around $10k/year, which is considered rich for the vast majority of people on this planet?
wait, now we are supposed to look at median income as lower income?
i have absolutely no idea what you are trying to get at here.
You are aware, aren’t you, of the subtleties of statistics? You are aware, that it is possible for everyone’s income to rise, but the median income fall, right? In fact, this is exactly what’s happened.
A statistic isn’t created in a vacuum and isn’t really all that informative all by itself or even compared to the same statistic computed in the past. If you don’t understand how the statistic is computed, nor the mathematical artifacts of that statistic, as well as the context in which that statistic is computed, then you might as well not even mention it.
So you’re saying that there’s always going to be a bottom 20%? I’m shocked, SHOCKED, I tell ya!!
The problem with the myopic thinking in terms of percents, deciles, and other fixed distribution, is that it allows for dishonest discussion about how the bottom 20%, as if the very existence matters, because it always will. The thrust of Mark’s post is that if you want to discuss the state of material well being of these groups, then let’s do just that.
The string of graphs very clearly contradicts the left’s assumption that wages have stagnated for all but the very top percent. In fact, the set of graphs above absolutely destroys that assumption by showing that the lower deciles in income distribution are much better off than just a couple decades ago. And they show that the middle class has been “shrinking” because people have been getting wealthier, i.e., moving up the income ladder, not down as the left simply assumes with no basis in reality at all.
What’s interesting is that the percentage share of the lower- and middle-income groups in all the charts were declining and the upper-income group’s share increasing for 30 years until 2000, then all of a sudden everything flatlined for the last 13 years, as if hit by a freeze ray. 🙂 Boosh did it! 😉 It would be fascinating to dig deeper into the data to disaggregate why that happened.
I’m guessing this is what has given rise to the squeezed middle class meme, flacked by opportunistic politicians and other propagandists. It’s not actually that the middle class has been squeezed recently: it was actually being squeezed for 30 years, upwards. It’s just that that upward rise has stopped for more than a decade now, so that lack of continued movement up feels like a move down to those conditioned to think that way.
The real question is what were the forces driving those trends for 30 years and why did they stop with the dot.com boom.
To answer my own question, my guess would be increased globalization and outsourcing, increased automation through technology, and the failure of the tech sector to create new sectors and jobs to fill those two gaps since 2000. I see the first two continuing in the years ahead, but the third finally kicking into gear soon.
More data in income inequality — source Census & CBO:
http://www.brookings.edu/research/opinions/2014/01/06-income-gains-and-inequality-burtless?utm_campaign=Economic%2520Studies&utm_source=hs_email&utm_medium=email&utm_content=11575522&_hsenc=p2ANqtz-8_GlMdxsO3EvHhQI51jpd7h4l9F4P1bd-KDC_v3smIQH725SK-tqM-8uLVsUs0W9JVpNbq-1xm6RlbxdDP9DrMXQKmOg&_hsmi=11575522
More data on income inequality — from Pew Research
http://www.pewresearch.org/fact-tank/2014/01/07/5-facts-about-economic-inequality/
More data on income inequality — from Pew Research
http://www.pewresearch.org/fact-tank/2014/01/07/5-facts-about-economic-inequality/
“More data on income inequality — from Pew Research…
Geez! I thought I was reading something from the Daily Kos..
Walter Williams wrote in this 2008 commentary: THE POVERTY HYPE
‘The psychology of victimhood and the politics of envy are powerful political tools and we see them being exploited this political season. Politicians telling Americans how bad off we are reminds me of one of Aesop’s Fables where a dog was carrying a piece of meat across a bridge. Looking down into the river, he saw his shadow, which appeared to him as another dog carrying a larger piece of meat. Attacking the “other” dog, he dropped his piece of meat into the river and it was gone for good. Aesop’s lesson is something to keep in mind as politicians offer their solutions to income inequality.‘…
vic-
who cares?
that data misses most of the story.
young people earn less. you earn more through a career. then you retire and earn less.
you may start and end in the bottom quintile, but hit the 4th one in the middle.
measuring quintiles and gini coefficients misses the whole game if people are mobile among income brackets, and, in the US, they are.
this is the same sort of bogus statistical mistake that mark has debunked a dozen times.
why trot out these disproven tropes?
further, there is absolutely no evidence that wealth redistribution makes the poor any better off in real terms over time.
it just makes the rich (and the society as a whole) poorer.
the poor in the Eu are as poor as those in the US.
we just have more wealthy people.
http://www.forbes.com/sites/warrenmeyer/2013/12/10/do-we-care-about-income-inequality-or-absolute-well-being/
More data — Income inequality and social mobility (the Gatsby Curve, etc.)
http://pubs.aeaweb.org/doi/pdfplus/10.1257/jep.27.3.79
Also read Coming Apart, Charles Murray, for the effects of social stagnation and the effects on the American value system.
Your data citation doesn’t really match what’s in the given source. You say it’s from table 690 and give a link. I open that, download the pdf, look at table 690 and it only has a fraction of the data; years 1990, 2000, 2008, 2009, while the graph clearly implies yearly data going back to 1970. Where did the rest of it come from?
Download the excel spreadsheet instead, it’s all there. 🙂
What about rampant inflation? Because 6 grand today will most definitely not buy me a brand spankin new Corvette today. FYI, back in the 60’s the Vette was less than 6 grand. 97% of the dollars purchasing power has been lost since 1913. You want to kill poverty, then we need to kill the FED.
This is addressed by looking at constant (inflation adjusted) dollars. But one thing these graphs don’t capture is the massive improvements in background technology and quality of life. In 1975, $700 non-adjusted dollars could get you a 25″ color TV. Now, that’s $3000 or so of purchasing power, which I could use to buy 4 flat screen 60″ 1080p TVs. It would be interesting to figure out a way to incorporate these benefits into income statistics
Ben
“It would be interesting to figure out a way to incorporate these benefits into income statistics.”
There is a way to *attempt* to adjust for those benefits called CPI. , It is far from perfect, and has little relevance for most individuals, as everyone’s situation is different.
It’s interesting to consider that most things that depend mainly on consumer preference such as the TVs you mentioned, have become far better and cheaper , while those areas of our lives with heavy government involvement, like healthcare, education, and defense have become more costly *despite* improvements in technology.
No. I think you are missing the real reason why the rich account for a growing percentage of households. It is, for the most part, because the rich are eating the poor and middle class relatively (compared to the rich) faster than the poor and middle class can reproduce themselves.
Of course, I’m exaggerating a bit. It is not literally true that ALL of the rich dine on the poor with fava beans and a fine Chianti. I concede that many of the poor and middle class have also been starving to death, committing suicide, disappearing into the wilderness, becoming undocumented household slaves to the rich, or escaping to Cuba, North Korea, Venezuela and other such worker’s paradises. So, please, no right-wing nitpicking about how I aggregate away important details.
Yep, my Acme KTC (Krugmanian Thinking Cap) still appears to be working.