Where in the United States is median income growing?


That is from 2007 to 2012, the link is here.


Farmers and oil and royalty recipients.

Speaking of North Dakota, Wikipedia shows 2011 median household income as only ~ $52,000 which seems just barely better than the US median.

Does anyone know the latest stats on this? Oil boom may be winning on rates but in terms of absolute levels it seems a far shot: Still $20,000 below Maryland or New Jersey?

Maybe more single earner/person households in ND? That would be my prior when imagining a commodity boom town.

The fracking Great Plains states were historically quite a bit poorer than average before (which was always interesting because they have the highest LE in the U.S. and perhaps the world). I think most used to be in the bottom quarter so "above average" is a huge step up for them.

Someone should make an animated graph of all the states' GDP per capita over time, it would be fun watching DC grow to its current glorious heights.

What's "LE"? Sorry if I'm being dense.

Life expectancy

You can almost say that the energy sector is growing (and hence the income of families affected directly by money created in this sector), the rest goes down.

Quite depressing for a country passed the industrial revolution, no?

Well, clearly we haven't passed it.

Not quite sure what you mean by that. It is precisely because of technological advances that fracking is possible and that many of those green squares appear on the map. The "Industrial Revolution" is never over. The same applies to farming. Technological advances made in Silicon Valley and elsewhere are not just a means unto themselves in their tiny little geographic realms. What you see here is the application of those advances. Where's the anomaly? One could have legitimate environmental concerns about fracking, but other than that, why is this "depressing"?

I mean that growing only in agriculture and energy is not good. Is what you would expect in a different kind of country. Please take into account we are talking about the median here. The top ten % has seen its income grow healthily.

Maybe all that new economy stuff is just marketing bulls**t. A mixture of stock promoters and Ivy League universities selling their wares. It sounds great.

Back in the 90's the NDP government in my province was pushing a high tech agenda, trying to emulate Washington State to the south. Subsidies for high tech firms and lots of blathering on about the wonderful economy where you didn't have to cut trees or catch fish or dig out minerals from the ground. Someone said an rather unfortunate fact; 40% of the current technology spending in the province was done by the forestry industry.

Or median income is generally down, but some segments of the economy- education, restaurants (and maybe hospitality as a whole), chemicals, agriculture, consumer web, healthcare, and yes oil (though note you don't see much booming in Texas, just North Dakota where it just so happened that they started large scale oil production around 2007-2008)- are doing well. Oil and agriculture tend to be concentrated in sparsely populated rural areas (which is by definition in the case of agriculture) so you get green/white counties from those trends. Everywhere else the improvements in these industries are offset or swamped by construction and manufacturing.

Education incomes are down. Not sure why you lumped that in there.

This is a really interesting chart. I would highly recommend going through the other charts in the Census report: http://www.census.gov/did/www/saipe/data/highlights/files/2012highlights.pdf?eml=gd&utm_medium=email&utm_source=govdelivery

This chart covers a really important period ... the largest recession since the Great Depression and the start of a very slow recovery. The imprint of the recession is clearly seen in all the red in areas that experienced housing busts. Large swaths of the country saw little change in median income which sadly has been a pattern for a while. And then there were some pockets (many of which are less affluent in absolute terms, see the other charts) that saw income gains likely related to natural resources ... a very different kind of shock than the housing price collapse. For me the chart underscores the challenge of macro stabilization via interest rates in such a diverse country that may well experience different shocks at the same time and hints at the nontrivial distributional effects of regulation.

Having gone through a major recession, most of the nation showing "no significant change" may not be too bad, eh?

Also, the colors of the large southwestern states tend to get visually overemphasized in an isopleth.

I agree. Especially since the message at the start of the great recession, including and especially this blog, was that the incomes and wealth that we had were illusions. If the current incomes and wealth are more real, then we are actually ahead of then. No?

Agreed. It's been a wrenching experience, like we knew it would be, but I think things are firmer now than in 2007. People are under fewer illusions, so it feels worse, but...

It would be nice if that were true. Expect another crisis in the next 5 years.

All about the counterfactual ... though I did make the mistake too. I claimed that median income had been little changed in recent past too, that's true in inflation-adjusted but not nominal terms, this chart shows nominal income, my bad. But you did get me poking around for the pre-recession data and there is a very cool interactive tool on the Census website: http://www.census.gov/did/www/saipe/data/interactive/#view=Trends&utilBtn=&yLB=8&stLB=0&cLB=0&dLB=0&gLB=0&usSts_cbSelected=true&usTot_cbSelected=true&stateTot_cbSelected=true&pLB=4&multiYearSelected=false&multiYearAlertFlag=false&prStateFlag=false&invalidSDYearsFlag=false If you click over to the "Median Household Income" tab you can see that increases were the norm before the recession. There is no way to sugar coat this recession but its impact was not uniform.

Point isn't trying to sugarcoat the recession. The critical distinction is whether these maps are just post recession blues or a Tyler Cowenian doomsday scenario of prolonged decades of median income stagnation.

Rahul, why is this an either / or? This chart clearly shows some cyclical forces (the red from a housing bust) and there is arguably an overlay of longer trends of slow growth at the median (very little green) and it shows (even better in the levels) the range of economic outcomes across the country. A chart like this really is worth a thousand words, there is a LOT going on here. As for the "sugarcoating" I see far too little green (in nominal dollars) over a five year period to call this "not too bad."

Using the same chart, compare TX poverty rates to national. Quite interesting.


I suspect we are arguing at cross purposes. But yes, I don't contest any of your points. Though they don't address nor refute what I was trying to get at.

So Rahul what are you trying to get at? Why can't the chart be showing both trend and cycle and the interaction of the two?


Sure it can and does. The hard (and interesting) part is trying to tease out how much is trend and how much is cycle.

One world view (the pessimistic Tyler Cowen-ian) ascribes much greater contribution to trend than other world views. Hence deconvoluting the trend from the cycle matters.

So, how much is trend & how much cycle? In other words you are taking a stand regarding if, say, 15 years down the line successive chloropleths will keep looking dark or not.

How is a housing bubble deliberately engineered to mask the devastating effects of low-IQ immigration "cyclical"?

Rahul - the problem is if you overlay population with the change in median income chart you see that a lot of the areas that are down also have the most people. Much of the white and green space does not contain many people.

Interesting that even San Francisco is in a decline. it seems like, no matter the hype and headlines, the real driver for America's income growth is energy and food, not technological innovation. Could it be that technology is a net negative for laborers?

Yes - modern technological gains are net job destroyers.

You misread the graph. The median is not the mean. The mean may be growing, since the 1% is getting richer, while the median can in theory shrink.

Why should we look at the mean if it is skewed by the 1% - the group that captured more than 90% of the income gains over the past 5 years?

I didn't misread the graph--I don't consider the mean important since it skews to the 1%, many of whom are arguably less labor and more capital anyway.

I suspect the top 15% to 20% is getting richer. There is also a lot of churn, so the people in the top whatever are not the same people as in 2007. Anyway, the mean and median are both important.

I suspect demographics. The people I know are all getting richer. They might not think so, but they're getting bigger houses, better cars, better tech, more entertainment, and more travel. Even the poor people I know are getting richer. But the poor people I know are reproducing more rapidly, which could cause a downward shift in the median even if every individual is getting better off.

I'm curious, does anyone have any evidence that supports or disproves this? It's purely anecdotal.

I think the lastest stat I saw says the middle class is shrinking in income, while the rich and poor are both increasing. But as you allude to, if the poor multiple fast, or if there's lots of shrinking middle class, then the median can decrease even if the mean increases.

Anecdotal for me as well - my peers (all college educated, many with masters and beyond) are doing great but they also cap themselves at 2 kids (many only have one). On the flip side, at least 67% of my peers despise their jobs and want to do something else with their lives.

When I do volunteer work I see many large, poor families. I'm sure there are stats out there that show babies per woman broken out by income.

" the poor people I know are reproducing more rapidly, which could cause a downward shift in the median even if every individual is getting better off."

The poor people you know are reproducing more rapidly? What do you mean? That the gestation period of poor women you know is only 7 months? That they're having multiple births? Overlapping pregnancies?

C'mon, he means that poor people are having more kids than the wealthier classes.

Oh, there's a difference between what he said and what he means. Thanks for pointing that out.

No--what he means is what he said. Are you a non-native English speaker perhaps?

no need to take everything literally. this is a blog , not an courtroom

Babies per year. There now it makes sense tio you as well as everyone else reading this comment.

The children of poor people are not always poor. They can get richer, so their reproduction rates are somewhat inconsequential, unless America is turning into a caste society.

The chart illustrates median family income, which has nothing to do with median family SIZE.

If each household were becoming smaller while population was actually increasing, that could certainly affect mediam income. It could, for example, lead median income to decline even wbile the per capita income of those in the middle increased.

Median *household* income.

The vastly different genetics of the imported poor are turning America into a caste society. Do you deny intergenerational poverty?

Is the lone county west of the Rockies that experienced an increase Sherman County, Oregon? (or Hood River County?) What explains the anomaly there?

(and if I'm distinguishing colors properly, I count five counties east of the Mississippi River, three in Pennsylvania and two in SE Alabama: the PA counties are all home to fracking operations? what of the Alabama counties?)

Google setup a server farm in Sherman County a few years ago. Not sure it has anything to do with the gains, but were I inclined to search for an answer, I'd explore that connection first.

It was a data center actually. Huge investment in a small populous area, may've had a impact. http://www.google.com/about/datacenters/inside/locations/the-dalles/index.html

Wrong County.

Google's data center is in Wasco County, not Sherman county. Given where the data center is located (The Dalles) and where the population centers -- by which I mean towns of a couple hundred people -- in Sherman are, very unlikely it's Google employees commuting. First, employment at the data center is really not very large, and Google actually rotates Bay Area-based engineers and managers in and out for the more technical jobs given the tiny local workforce. Second, it'd be a 30 - 50 mile commute when there is plenty of just-as-cheap housing in The Dalles.

It's Sherman County, Oregon. I know that area somewhat, and Sherman doesn't stand out from the surrounding areas at all, so it is a bit surprising. My thoughts

1. A lot of wind farm development, that'd be my #1 guess. And fits nicely with the nationwide story.

2. Very small population -- under 2000 total -- so it could also be a very idiosyncratic thing, like one local small business makes good (e.g., Cowboy Bill's Sagebrush Breath Mints became a surprise national hit)

Mostly the county is wheat farming and wheat prices are up over that period, but that factor would affect lots of other similar counties so that's probably not it. Also worth noting incomes were pretty low to begin with, true many other places, but does make it easier for the wind farms or whatever local factor to pull them up.

Where can I get Cowboy Bill's Sagebrush Breath Mints?!

"and two in SE Alabama:"

Those look to be Pike and Coffee counties. They are rural counties with populations in the 35-50K range.

Where I live people are getting richer. 2008 is a distant memory with home values at historic highs.

The map shows income, not wealth. It doesn't reflect property values.

Every time somebody sold one of those homes for a record value they had income. Often sharp jumps in income and combined with the high amount of house turnover during that period it's certainly going to be reflected in the income stats. Hmmm, I wonder what income minus "capital gains" looks like over the last 10 years?

One of the things that the map indicates is net increase in population. Areas that have shown a decrease in median household income, like much of Arizona, are also places where the jobless have gone looking for work. Some areas that have shown an increase in median household income are places where the population has declined due to loss of jobs.

When people say things as wrong as this I cannot dismiss it as simple ignorance--it's got to be politically motivated lying or astroturfing at best.

North Dakota's net increase in population is massive. Likewise, New England has negative net migration rates. You're so wrong I cannot help but wonder why you'd even say that.

You must be saying that Arizona has had a big decrease in population between 2007 and 2012. The state's population in 2000 was a little over 5 million, in 2012 it was about 6 1/2 million, an increase of 30%. Do you think that all these new Zonies arrived with jobs? And did all those that eventually got hired were in the upper income strata? Maybe most of the new residents became doctors, lawyers and cops, but I doubt it. A lot of them went to work at Walmart and Fry's.

They're Mexicans, not displaced jobless Americans

North Dakota's "massive" increase in population is oil field workers and truck drivers living in man camps. Their income is figured in where they actually live, probably Texas, Louisiana or Oklahoma.

They actually do live in the man camps.

What is the source of your claim that the income is reported elsewhere?

During the great Alaska pipeline extravaganza there were people from all over the world working in Alaska. Are you saying that there were Brazilians walking around with the address, "General Delivery, Deadhorse, Alaska" on their driver's license? That Israelis had their kids going to school in Cordova? None of these people "lived" there then, they worked there at temporary jobs, just like oil field trash are doing now in the Williston Basin.

Think about your use of the term "oil field trash" next time you fill up your tank at less than 4 dollars per gallon

Oil field trash refer to themselves as oil field trash. Why would anybody think about the fact that they showed up for work? They are being paid for their labor, are they not? And their appearance at the rig for each shift doesn't have any effect on the price of gasoline.

Much of the data on income comes from payroll data.

In other words it is based on who and what the oil companies report are on their payroll in North Dakota, etc..

I ask you again do your have anything but your theory to support your claim that oil workers income is recorded in another state?

How about eye-witness testimony? Or does evidence have to come from some academic in a Boston suburb or Princeton, NJ in a doctoral thesis? When every possible hole has been made in the Williston Basin the population of western North Dakota and eastern Montana will drop precipitously, though maybe not to the level it had been. Not too long ago there were less than 35,000 students in public schools in the state.

If you pick the right years, you can show anything you want. Would look very different if you chose 2009-2013 or any other combination of years

This. This exactly. But one apt comparison might be 2003-2013. Even better would be to avoid the median income, in favor of GDP per working age person.

Build your own diagram then

In 2012 natural resources & mining accounted for 0.64% of employment, up from 0.55 in 2007.

If you dropped all the oil workers in North Dakota in New York city or Los Angles no one would ever notice it.

From 2007 to 2012 natural resource & mining employment grew 17% and was one of only two sectors where employment was higher in 2012 than in 2007. .The other was education and health services that was up 11%. Actual employment in resources and mining rose 127,000 while employment in education and health rose 1,999,000.

According to the Federal Reserve oil & gas drilling capacity is now growing at an 8% rate.

"If you dropped all the oil workers in North Dakota in New York city or Los Angles no one would ever notice it."

Bartenders would.

Of course, Alaska isn't even included on the map. It doesn't count, apparently.

Bartenders in North Dakota would notice it but I bet that bartenders in NYC & LA wouldn't.

You can identify the regions that weren't as rich as they thought they were in 2007.

I think an even better interactive map, which also uses census data and other data as well, is at Patchwork Nation: http://www.patchworknation.org/ It follows the format of the book by the same name that was, I believe, underwritten by the Pew Foundation. I take the app with me when we travel

It looks like having a low population density helps in our brave new world, just like Ben Franklin said, while the rapidly growing immigrant destination states of California, Nevada, Arizona, and Florida ... not so much.

Alternatively, low population may simply make it harder for the same income change to reach statistical significance.

You'd see the same thing on a global scale: the big winners in the last half decade have been lightly populated countries with a lot of land or sea and thus natural resources per capita, like Australia and Norway.

The current trend won't last forever, but we ought to admit its existence rather than just subscribe to Julian Simonism uber alles.

Ranked by median household income or by per capita income, three of the top five states were Maryland, Connecticut & New Jersey. Now, fascinatingly, they are also among five most densely populated states in the US. What say Ben Franklin? Where's your low property density premium?

OTOH perhaps you are one of those enamoured by relative rates alone & not absolute levels? But then you ought to be migrating to some place like Sierra Leone to enjoy catch-up growth. Or perhaps Mongolia or Turkmenistan?

And the south side of Central Park South has even higher density and average income than Connecticut, so all we need to do to raise income is to build 1,000' apartment towers.

You were the one proposing silly correlations with population density not me.

Obviously geographically small "suburb-states" are going to have high density and high income. It's an extremely spurious correlation.

Looks like having a bunch of natural resources in the ground helps.

The counties with low population densities, let's say Greenlee County, AZ, St. Louis County, MN, Iron County, WI and Big Horn County, MT, may have had statistically insignificant changes in median household income in the years 2007-2012 but that doesn't mean that average was the same as the US median household income of $51,371. Areas like these have been in an economic downturn for decades. It's pretty likely that the median household income in these and similar places is far less than $50,000 annually.

Still says nothing about livability...Migrant labor - I got it.

Median income for individual earners would be a far more relevant and important stat, but for some reason, the govt, the media, and the sycophants who slavishly adhere to the range of talking points from CorpGovMedia never seem to care much for the Median income for individual earners stat.

Maybe because focusing on mean income of households make american workers think income is higher than it really is.

The median income for americans over 25 years of age is about $27K.

But I bet that particular stat never got published on this particular blog....

I would have agreed until I got married.

So, the entire nation is flat or going backwards overall, except a handful of highly rural and formerly poor counties that are enjoying a temporary RELATIVE boom thanks to prostituting themselves out to fracking.

It really does trickle down too: cops, social service workers, strippers and hookers, meth cookers... They are all getting a piece of the action.

While it lasts.

Dear Prof Cowen:
I respectfully urge you to NOT publish maps of US with area on the screen proportional to area on the ground; instead maps should be proportional to population
when our visual system sees a map like the one you show, we process the information, and one thing we do is weigh area X color.
Thus, in the graphic above, our visual system pays a lot of attention to sparsely populated areas like NV, ND, WY, etc, whilst the aprox 20 million people in LA and CHI and NYC are virtually invisible.

The graph nanny

I would like to see a map that compares the change in median income to the change in median household expense. Median incomes can decline and people may still be better off if median expense has declined more.

Comments for this post are closed