We talk quite a bit on this blog about the importance and value of agglomeration, which is the additional level of productivity which comes from locating activities close to each other. Agglomeration is why Auckland growing will be good for all New Zealanders, and at a smaller scale why central parts of Auckland growing will be good for all of Auckland. While agglomeration economies are well studied in terms of observing that they most definitely do actually exist, it has been a little less certain exactly why they exist.
A recent article in The Atlantic Cities reports on research done by MIT (Boston version, not Manukau version) on this question – with some perhaps unsurprising but at the same time interesting results:
There have been plenty of theories. Adam Smith famously figured that people become more productive when we’re able to specialize, each of us honing a separate area of expertise. And when lots of us elbow into cities, we’re able to specialize in ways that we can’t when a rural farmer must also double as his own butcher, accountant and milkmaid. Other economists have suggested that cities become great agglomerators of industry when factories cluster together around economies of scale and communal access to transportation.
“We think there’s an underlying completely different way of thinking here, which is very different from the economist’s way of thinking,” says Pan, a doctoral candidate in computational social science in the MIT Media Laboratory’s Human Dynamics Lab. Previous work by researchers at the Santa Fe Institute has proven the math behind the power of cities: As they grow in population, all kinds of positive outcomes like patents and GDP and innovation (and negative ones like STDs and crime) grow at an exponential factor of 1.1 to 1.3.
This means that all the benefits (and downsides) that come from cities don’t just grow linearly; they grow super-linearly, and at roughly the exact same scale, with a growth rate that looks on a graph something like this blue line (a linear relationship is shown in red):
As for why this happens, though, Pan pushes aside theories about the location of manufacturing or the specialty of trade. “It’s more fundamental than that,” he says. “Cities are about people. It’s just that simple.”
We have discussed the work of the Santa Fe Institute in before in this post with a video presentation by theoretical physicist Dr Geoffrey West.
The full paper in which the study is published highlights that fundamentally it’s the density that we’re able to form social ties which generates the super-linear increase in outcomes. As noted, “the larger your city, in other words, the more people (using this same super-linear scale) you’re likely to come into contact with.”
However, unlocking the potential of cities is dependent upon ensuring that sufficient transport infrastructure is available for the city to effectively function as a large connected entity, rather than just a series of smaller disconnected placed which just happen to be next to each other. This is detailed further below:
“What really happens when you move to a big city is you get to know a lot of different people, although they are not necessarily your ‘friends,'” Pan says. “These are the people who bring different ideas, bring different opportunities, and meetings with other great people that may help you.”
Maybe this doesn’t sound like a novel discovery – that the inherent power of cities lies in our individual connections to each other. Other researchers have nipped at a similar idea, calculating, for example, the “social interaction potential” of place. But until recently, most economic thinking sidestepped the sheer value of human interaction in favor of explanations about the proximity of manufacturing, or the processes of production.
This explanation for the power of cities also raises some curious questions about those places that remain an exception to the model. In some African cities and Eastern mega-cities, innovation and productivity don’t grow super-linearly. Populations grow, but the benefits don’t accrue with them as we would expect. This is likely because transportation infrastructure in those places is so poor that people aren’t able to connect across town to each other. “To live on the west side of Beijing,” Pan says, you never go to the east side.”
What I find fascinating is to analyse whether the internet has the ability to change this fundamental relationship – because it’s now so easy to be in instant contact with a huge number of people no matter where you are. Yet none of the recent research seems to suggest that technological advances make agglomeration less important – in fact it seems like as employment specialises more and more, agglomeration actually becomes more critical than ever.
Processing...
“Gerry Carlino of the Philadelphia Fed has found that the number of patents per capita increases by roughly 20 percent to 30 percent for every doubling of employment density.”
http://www.nytimes.com/2013/06/07/opinion/brooks-the-power-inversion.html?_r=0
so that’s a sub-linear growth of a negative externality, which I guess is the same as a super-linear growth of its positive version?
Per capita.
yup. so if employment doubles then you get 2*1.3 = 2.6 times as many patents as the original scenario. Pretty bloody superlinear,
No, that is linear. If you graph it it makes a line. That would be the red line in the graph above, the blue is superlinear.
linear on a per capita basis yes, but super-linear across a population.
If you take a city of 10 people and double the population to 20 – but do so at the same density then you will get twice as many patents as you had initially (i.e. linear returns) – because you have doubled the population.
Consider another scenario where you take the same city of 10 people and double the population, while also doubling the density. In this case you will get 2*(1+0.3) = 2.6 times as many patents (i.e. super-linear returns), because you have twice as many people at twice the density.
Sorry, you are quite right, linear per capita, but over a population if y = total patents, and x = total population;
y = (1 / k)*x ^ 1.3
Where k is a constant value.
Sorry, should be y = (1.3^Log[x/k] x)/(1 + Log[1.3])+k
Where k is the frquency at which a single person living alone in that area would generate a patent, and log is a base 2 log
I’ve often wondered whether the mainstream economics literature on agglomeration economies has underplayed the social dimension to cities. It seems clear to me that one of the key benefits of cities is being able to meet a shedload of relatively intelligent people relatively easily, who in turn can form new ideas/businesses.
I’ve also often wondered to what degree social agglomeration economies depend on the level of openness and trust, i.e. people’s willingness to meet and trust new people. This is possibly one reason why cities in emerging countries, such as Jakarta, don’t seem to benefit so much from agglomeration economies.
Leo Hollis in his new book Cities Are Good For You has a whole chapter called: Trust in the city. I think you would find this interesting Stu.
One important component of the network density that make cities so effective is the role of “weak ties”. This is a different form of social capital, one made not of connections based on time, intimacy etc, but one based on a bridging capacity to connect networks with other networks. JJ talks about this very critical urban phenomena using the colorful characters of her neighborhood as the conduits that provide the access to information, activities and opportunities.
I think individually, and perhaps collectively, it is important for people to learn to look beyond their strong ties of family, class, fields of expertise, etc in order to leverage the value of human capital. I think this generally relates to the strong correlation with highly innovative cities also being very tolerant and diverse.
And like Patrick I am fascinated with the role of internet technology, in particular social media in this equation. For cities on the verge of urbanisation, like Auckland, I think social media introduces the “third place” benefits and gives people a taste of the wider networks available to them. So instead of replacing face-to-face connections, social media and technology is actually facilitating and promoting it.
.
yes! Completely agree.
It is not only fascinating but exciting.
Jakarta does not have a dense pedestrian-dominated core. Its main office employment is along wide vehicle-dominated roads that were great when Jakarta was much smaller in the Dutch era. The poor (and tourists) might walk, but those in senior positions need to show their cars to prove their status. The middle class take motorcycles, and motorcycle sales outnumber car sales by about 10 to one in Indonesia. Motorcycles are also a quicker way of getting through the traffic even for those wealthy enough to afford cars, so once everyone knows they have a car they can use a motorcycle most days. The motorcycle in Indonesian cities is like the car in Los Angeles. Everyone has them, even schoolkids going to school. To go from one shop to another that are 100m apart, a motorcycle is used. There are few pedestrians. Motorcycles require much less fuel than cars, can be stored in lounge-rooms, but lead to a lot more bad accidents.
Innovation in Jakarta is unlikely to be captured by the number patents, in a country where software piracy is the norm. Unfortunately the form of innovation that provides the greatest rewards is that from rent-seeking behaviour. The standard kick-back on government contracts is 40%, which the contractors need to forward into Singapore bank accounts of the public service staff arranging the contracts. Jakarta has strong class boundaries between public servants of various ranks,the super-wealthy Chinese, the super-wealthy of Indonesian ethnic groups, the middle class of various ethnic groups, and the rest. Each has to show their status, and riding a bus or train isn’t the way to do it.
So while Jakarta is a large city it has lots of internal barriers of class and ethnicity. Add to that a burdensome bureaucracy, business practices that in the West would be called corruption, little public financial support for public transport, chronic congestion, and strong private transport based on motorcycles. The current Governor of Jakarta is wanting to improve its public transport, but faces the same difficulties as Los Angeles of a lack of past PT investment and only a weak pedestrian-dominated core.
Stu, “one of the key benefits of cities is being able to meet a shedload of relatively intelligent people relatively easily, who in turn can form new ideas/businesses”. That is also true for meeting relatively corrupt or amoral/immoral or relatively stupid people hence the comment in the quoted article “As they grow in population, all kinds of positive outcomes like patents and GDP and innovation (and negative ones like STDs and crime) grow at an exponential factor of 1.1 to 1.3.”
This perhaps explains why illicit drugs use has always been associated with big cities such as New York, Chicago, Philadelphia, Boston and Detroit most noticably with alcohol during prohibition and hard drugs in the 1960s and 1970s.
Innovation in music, cinema and broadcasting are major exceptions to the agglomeration rule. The promoters live in big cities but the few of the creators do, There is plenty of evidence that the same is true for inventors such as Bell and Ford who grew up on farms and moved to cities to manufacture their inventions. Which begs the question, is the patent correlation due really proof of greater inventiveness in dense cites or simply because specialist patent attorneys are concentrated in the densest, wealthiest cites.
the geography of music. http://www.theatlanticcities.com/arts-and-lifestyle/2011/12/geography-top-music-hits/750/
The economy of the City of Detroit allowed the mass production of the Model T. It couldn’t have been done on the farm. Inventors are very rarely lone genuiuses.
Kent, ” So yes, we’re starting with a subjective assessment, one that is heavily skewed toward English language acts in the indie, alternative, and hip-hop genres.” Hence Memphis is on the list but Nashville isn’t. Its interesting that the list is almost reversed for hits per capita.
I did actually say that Ford had to move to Detroit to manufacture his invention, the important thing is if he had grown up in Detroit he would never have seen the need to develop a car that would also function as a tractor and a stationary engine for mobile threshers and timber mills, which gave the model T it’s critcal market volume. Necessity is the mother of invention.
The fact that Packard and Ford both set up mass production of cars in detroit in 1903 is not altogether surprising. Detroit had already become the US centre for heavy steel products such as coal ranges due to its convenient location between the Great Lakes and the Union Pacific railroad, ie bulk ore and coal shipped in and completed product railroaded out. The traditional importance of transport efficient in industry agglomeration.
Kevyn, Ford didn’t move to Detroit to manufacture his invention. It’s what created it. He copied the idea. Charles King invented the first car in Detroit. Ford famously would follow him around on his bike to figure out how it worked. There was also the Dodge Brothers, Fishers, Dunbar etc already working in an established industry.
I’d be keen to hear more about the cultural innovation that occurs outside cities.
Another reason that Detroit was well placed for the auto-boom was that it was a centre for the boat engine industry on the Great Lakes.
I may also add that these features of Jakarta probably apply to many other large cities in developing countries. Each needs to be well understood before adding it to a scatterplot of innovation vs size.
yes I agree. These developing ,mega-cities do tend to have a number of characteristic/unique historical/cultural features that make it very difficult to generalise from.
And another one.
“The model suggests that the right response to urban sprawl is to focus spending on a hot inner core, which ought to do most to boost output.”
http://www.economist.com/blogs/babbage/2013/06/urbanisation
The really interesting thing about the internet and spatial form is that is pretty clear that every thinker on this issue was exactly 180 degrees wrong about the likely outcome. From the 60/70s on everyone was predicting the technology that allows more and more powerful distant communication would lead to everyone on their own ranch miles apart working away at screens or holograms and never bothering again with the inconvenience of the city. [funny vids on u-chube show these predictions]
Well we are in this future now and the exact reverse is happening, 2011 was the first year that more Americans moved to urban areas than suburban ones, and 2012 continued that trend. We live in an age of unparalleled urbanisation.
It seems the ease of virtual contact has led to a desire for more actual contact. With more people more often.
NZ seems a little behind the US in this but it seems the internet shopping is hurting bigbox retail the most…. we’re still building them; pain to come methinks.
Once you’ve bought the idea of the big box, then internet retail is the ultimate big box retailer. Small specialist boutiques are the retail experience the internet can’t replicate, and those only work in dense urban centres.
Yes exactly. And there is no reason to believe that we’re still not just following the US on these trends as we were in the car age too…
A Japanese professor told me last week that the young in Japan want to go to Tokyo, which leaves the other cities and regions with lots of old people. Could this be another feature of cities that the economists haven’t yet captured, that the larger cities have the deepest market for the most talented young graduates ? An employer in this market can recruit the A level students and retain them in the organisation, whereas a regional employer may need to settle for a B grade graduate ? (They would prefer an A grader who is willing to relocate there because of family connections in the region, but not many apply.) After 20 years the business in the larger city would staffed mainly by A-graders, and the regions by B-graders.
It was surprising to hear this from Japan, where I expect the regional cities would have good PT, high density, and relatively deep employment markets. It’s just that Tokyo is denser and deeper.
That is certainly the case in NZ in the legal profession – A graders generally go to firms in Auckland and maybe Wellington (for some Public Law work). Christchurch and Dunedin may get some A graders with the Unis there but the provinces generally end up with the rest.
I know also that provincial lawyers have huge problems finding young people to take on partnerships – they will generally use the small provincial firms to get experience and then take off ASAP to the city or overseas. Provincial firms may sell out later to a larger firm in the area or to a lawyer in their 40s who wants a small rural practise.
This is despite the fact that rural practises can actually be quite lucrative as there is often a captive market.