His hypothesis is that winner cities increasingly dominate. Marc Muro and Jacob Whitin write,
growth across communities now tracks exactly with their size. The nation’s bigger communities — powered by well-educated millennial workers and the agglomeration trends brought by digital technology — are now growing notably faster and accounting for more and more of the nation’s growth than before, even as small metros wane and most of the rural hinterland slides into deep decline. In short, fully half of all of the country’s employment growth took place in just 20 metropolitan areas, home to about one-third of Americans, led by the usual suspects of New York, Boston, the Bay Area, Seattle, and Washington, D.C., along with such fast-growing Sun Belt hubs as Dallas, Atlanta, Miami, and Orlando.
Following the link here are some basic issues with the above paragraph.
In terms of output growth since “non-metro” areas are up 19%, far above large cities (14%), the US average (13%), Medium cities (9%) and small (8%). In terms of employment growth large cities are up 15%, medium cities 9%, small 4% and non metro 2%.
If you define economic importance by “% of employment” then big cities are dominating, but if you define it by productivity per employee then non metro areas have crushed it since 2010*. The statement that
“The nation’s bigger communities — powered by well-educated millennial workers and the agglomeration trends brought by digital technology — are now growing notably faster and accounting for more and more of the nation’s growth than before, even as small metros wane and most of the rural hinterland slides into deep decline”
Is a gross misinterpretation of the data provided, while techinically correct it is only so because it drops . The combined evidence implies that the large and medium cities are adding tons of slightly below average jobs in terms of productivity, small cities are adding moderate amounts or above average jobs and that non metro areas are adding a small amount of incredibly valuable jobs.
There are ways this interpretation could be wrong, but none that are presented that I can see.
* there might well be some end point biases going on here so don’t take this to seriously.
The Output by size tier graph contained a footnote citing that the fracking and natural gas boom inflated the rural numbers.
They don’t designate how much, and they don’t note that the increase in productivity for small cities was greater than the employment gains, and lower or roughly the same for medium and large cities than the employment gains.
A footnote without explanation of the size of the effect isn’t sufficient to cover the fact that the productivity gradient is the opposite of the employment gradient, and leads to opposite conclusions of those made.
That nearly all the new employment is in the small number of winner cities is completely consistent with the fact that labor productivity is continuing to rise outside big cities while growing only slowly in those cities on average, and indeed, the former is caused by the latter.
Strictly geography-based output – primarily agriculture, natural resource extraction, and tourism – has become so much more efficient with the use of new capital tools and technologies that these land-using activities can only employ a small share of the labor force, and that trend is continuing. The question is how does the labor force reallocate itself, or as in the early 20th century, where are all those land-based-production displaced workers going to go, and what are they going to do?
Whatever they are going to do, there is no good reason that they would be doing it out there on that non-urban land, because there is no competitive advantage to those locations when an entrepreneur is trying to decide where to site some new, non-land-based enterprise or operation. Land is cheaper in American non-urban areas, but it’s even cheaper than that abroad. Same goes for labor.
The only reason an entrepreneur would site anything in America is because he cannot outsource. You can’t outsource geography-based activities for obvious reasons. You also can’t outsource what I call “proximate services” that have to be done close to other people. And proximate services providers end up being pulled in by anchor-sectors and the economic gravity of lucrative industries which have positive economies of agglomeration. “Because that’s where the money is.”
The thing to keep in mind is that this same story and pattern is playing out everywhere in the world without any apparent exception, which allows us to reasonably infer that it is driven and determined by widely-shared advances in technology which have radically lowered the relative prices or quantity-of-labor required to accomplish certain functions such as transportation, communication, information processing, and mechanical manipulation.
Meanwhile, those proximate services are precisely those functions that are left after all the “easy, quick, and cheap to automate” tasks have already been automated. That means they are done by humans at human speed and can’t easily be sped up, which strongly limits labor productivity growth.
So what the future looks like – especially in developed countries – is more centralization of production and employment into a few big winner cities with the few winner sectors, and gradual emptying out of increasingly gloomy “hinterlands,” which will become “creamed”, that is, especially depopulated of anyone with the human capital to work in a winner sector.
I wonder how much of the large metro employment/ output represents government employment. My impression is that there is more government employment in cities, and especially larger cities. For example, government is the largest employer in New York City, with well over half a million employees (https://www.labor.ny.gov/stats/nyc/), weighted heavily towards municipal employees. I’m no economist, but I understand that the output of government is simply measured as its cost.
Why do we like congested cities?
Because goods queue up, we can see how long the lines are and that is how we price. It is the principle of agglomeration. Way stations in commerce become small towns because goods congest and townspeople get a measure of relative scarcity, they can price. The congestion balances up the chain (delivery cost measures at the same error level) and total cost of production is implied in the local price.