In their quest for the commonalities and differences between cities and companies, Geoffrey West and his team came across the crucial interplay between those two social structures. And from those findings, we can see that the role of cities in innovation is a lot stronger than we usually realise.
A grand idea
The heart of the matter is superlinear scaling, a unique characteristic of cities that has no analog in biology or in other social systems: a city with double the number of citizens will generate more than twice as much wealth, will be more than twice as productive, will deliver more than twice as much innovation. This phenomenon occurs regardless of city location, and it has massive effects.
If you think about the development of a specific city that grows over time, we can predict that by the time it reaches twice its current size, its wealth generation, productivity and innovation will have more than doubled. That’s how size matters for a city. Or think about an individual who moves from his current location to a city twice as big: in that new location, the individual will be about 15% more wealthy, productive, innovative. That’s why location matters for the individual.
Sounds like a simple success story, but unfortunately the story doesn’t end here. Superlinear scaling in cities applies to desirable criteria such as wealth creation, productivity and innovation. But it applies to negative criteria like disease, crime and pollution as well. And there’s no way to pick and chose: with growing city size, you will always have a superlinear increase of both, the good and the bad effects.
These are counter-acting forces at work –the positive forces make cities attractive places to go to, while the negatives repel people. In Geoffrey West’s view, we actually observe a race between the good and the bad effects of city growth. And innovation is the key to keep cities alive, to avoid their collapse and to keep them functioning as social structures. What is more, West points out that steady innovation at today’s rate will not suffice: rather we need to continuously accelerate innovation. To keep the innovation engine running, we must rev it up.
This outcome of Geoffrey West’s investigations has far-reaching implications, including for long-term sustainability. Or could his conclusion simply be far-fetched? Let’s see whether we can find evidence to support his view. I’d suggest to tackle that question from two angles, one conceptual and one data-driven.
Testing the idea
At the conceptual level, we can consider the Paradox of Development suggested by Ian Morris. In his investigations he observed that social development creates the conditions that undermine further social development: the solutions to existing problems create yet other problems, and their solutions will generate further problems. So Morris expresses a similar concept of two counter-acting forces that are fatefully connected. What is interesting here is the striking similarity across scientific disciplines: the findings of the historian and archaeologist Morris on social development converge with the observations of theoretical physicist West on cities as complex adaptive systems.
While such conceptual similarities clearly support West’s findings, additional underpinning data would make his case even stronger. And here again, we should consider two questions: How did urbanisation develop? And what does a history of innovation look like?
West emphasised that urbanisation has now reached a level where the majority of the humankind lives in cities. In fact we’ve “achieved” the transition from a rural to an urban population majority only a short moment ago, in 2006. Yet that is only one single data point. When looking for a long-data analysis of urbanisation, we can call on Ian Morris once again. He had thoroughly investigated the social development of humankind since the last ice age (if you are interested, you’ll find his full data and analysis here, it’s an 8.9 MB, 233 pages file). His concept of social development uses “organisation” as one out of four key traits, and his yardstick for organisation actually is the size of the biggest city. In the following chart I used his original data, just omitting the beginnings from 14,000 BCE to zoom in on the last 2000 years of human history.
The exact definitions Morris used to compare the evolution of social development in Eastern and Western societies are not relevant for our discussion. Neither is the numerical value he derived to compare city sizes across space and time. What is essential is that Morris delivered a useful index of urbanisation, based on his archaeological and historical investigations. And the result is compelling. This long-range analysis reminds us that urbanisation as we experience it today is a historically very young phenomenon. For the longest time in human history, city sizes remained rather stable. Only very recently, in the 18th century, something seems to have happened that was bound to change our cityscapes dramatically.
In the next step, we can then try to figure out the historic development of innovation. I had to scratch my head a little, but then I remembered an article published in The Economist back in February 1999, which tackled the idea of accelerating innovation. The following chart, essentially taken from that article, illustrates how society has experienced phases of significant innovation intensity. These are the long waves of creative destruction that Joseph Schumpeter introduced to economics to account for the innovation driven aspects of economic growth.
Schumpeter developed the idea of “the gale of creative destruction” in the early 1940s to explain how innovation transformed the economy from within, destroying old structures and establishing new ones. When building this concept, he could only consider the first three waves. But from today’s venture point, almost seven decades later, we look back at nearly five complete waves; and we can see the validity of Schumpeter’s concept.
While the ups and downs of innovation intensity remained relatively unchanged, Schumpeterian Waves actually got significantly shorter, or rather: faster. The 5th wave is predicted to last only half as long as the 1st did. Which demonstrates a significant acceleration of innovation that occurred since the Industrial Revolution.
If you now take Morris’ data on urbanisation for the period 1700 to 2000 and overlay the Schumpeterian Waves, the result reveals a striking correlation: the acceleration of innovation coincided with the massive urbanisation since the Industrial Revolution. While I hesitate to postulate a definitive causation between urbanisation and accelerated innovation, I do take this result as a confirmation of Geoffrey West’s thesis that city growth and the speed of innovation are intimately linked.
Furthermore, accelerated innovation is not simply a novel theoretical notion; on the contrary, it has been a reality for more than two centuries. It just stayed below the radar for many decades, because the resulting changes initially were too small to be observed.
Where does all of that lead us?
I must admit that I have more questions than answers.
One set of questions relates to our potential success: How can we sustain this acceleration of innovation? How fast could this innovation engine actually go? Are there speed limits? And if there aren’t, what is the effect of innovation acceleration on society? What would it mean if individuals experienced several technological revolutions in just one lifetime?
The other range of questions addresses the potential failure: How sustainable is this innovation acceleration? What happens if we cannot sustain it anymore? What would failure look like? And what would the effect on our cities be? Would cities just stagnate? Or partially collapse? And how do my earlier considerations on energy constraints and complexity limits fit with these new ideas?
There’s obviously a lot more to think about …
understanding innovation 
A remarkable piece of work. Truly instructive!
As always, thank you very much for the encouragement
Pleasure is mine. May I reblog it some time soon?
That would be my pleasure. As I said, I have arrived at more questions than answers, hence I’d be delighted to share those questions widely and stimulate others to think about them. Thanks again.
Reblogged this on ashokbhatia and commented:
In emerging economies, rapid urbanization poses unprecedented challenges. A positive spin-off of the same appears to be an acceleration in the rate of innovation.
Perhaps, urban spaces enable the coming together of creative minds, thereby fuelling innovation?
Here is a thought-provoking post on the subject.
Allow me to debate on some of the subtopics that are touched in this article. The grand idea of super-linear scaling of the combination of city size and productivity/innovation is a fact from the perspective that we put down some measurable data of two conditions that we pre-assume are directly related. This does not mean that the one is the cause of the other or vice versa. The urbanization has its own root causes throughout the course of history and the productivity and innovation as well. And speaking of a large city, what defines a large city? How big is a big city and comparing to what type of small city? Even in the middle ages there were proportionally large cities with people trying to move to these cities (for totally different reasons) and the difference in size between the then big city and the small town was relatively bigger than the same difference today. It seems that today we have a great number of let’s say middle size cities and the urbanization has more economic rather than safety and security motives or incentives. During the Renaissance period a modern financial and banking system was developed in Europe starting from the famous Italian Medici bank that financed quite some innovative activities of the time. Loans, sponsors, private and state financing, as well as the circulation of paper money later on, gave the means for human interaction on a much faster pace bringing products and services in the center of social life. Merchants started exchanging goods in urban places because more people were attracted by the investments of the wealthy. Capital driven social economy was already in place; the basic concept that continues today. But the investments could not become innovation automatically. The scientific knowledge was missing, and this created the need for educational facilities such as universities and schools of theoretical and practical science that promoted innovative ideas. And this again could not have been developed without the existence of economic prosperity and wealthy sponsors acting as investors. The knowledge that was generated was used then for creative innovations and the populated areas like big cities were the ideal audience in a market driven society. So, maybe economy has played equally (if not more) important role in the development and innovation, and if one would proceed with the same research analysis comparing this time the history of financing and banking systems (instead of large cities) with the development and innovation, the results (or at least the pattern) would be very similar. The graph lines might coexist in the fluctuation of the Schumpeterian waves.. but I would suggest not to approach this correlation in a “cause and effect” pattern.
So one would argue that it was not the urbanization and the size of the cities that increased the innovation over time, but rather the financing and the availability of wealth that was invested (regardless the size of the cities).
Finally, the way that the Schumpeterian Waves Accelerate graph is presented might be somehow deceiving.
I find it hard to believe that for example between 1900 and 1950 there is one and only pick point that actually defines the wave shape and everything else is either descending or ascending. The evolution of technology implies also some kind of recursiveness (to use a term from Brian Arthur) and the development and production of innovative or disruptive technology is a step process that evolves not in a specific linear or curvilinear motion. For the same reason I would not think that someone should interpret Schumpeter’s graph in a way implying that there was almost no creative innovation activity between 1985 and 1995 in areas of electronics, aviation, software and digital networks. But this is also the beauty of this human activity. We try to exchange ideas in our own unique ways giving a different interpretation on what we find or what we visualize as authentic thought or idea in our mind. The logic and nature of representation of our thoughts is thoroughly examined by Wittgenstein. In his “Tractatus Logico-Philosophicus” he states that the logical picture of the facts is the thought, and the latter is the significant proposition with sense.
Anyway, while you have successfully explained through Schumpeter, the innovation driven aspects of economic growth, my intent was to rather point out also the opposite: The economic driven aspects of innovation growth.
My apologies for daring challenging the findings of well distinguished and respectable scientists and researchers such as Morris, West and Schumpeter, and I thank you at the same time for allowing me this brain-storm type of comment. Keep up the great work.
Thank you for the brainstorming. There’s a lot to think about.
First, what is city size? I’d agree that size of a city has a lot of relative context. Size has different “meanings”, in historic times as well as in different location. Take Berlin as the biggest city in Germany with around four million people: in China that is not even worth putting it on a map. But Geoffrey West did look for an absolute yardstick, and that is number of citizens. And his findings suggest that this simple number is sufficient to explain the benefits that a relatively larger city would offer compared a relatively smaller city.
Next, economics. clearly a topic that would warrant an entire blog of its own. I appreciate your reminder of historic evolution of the banking system, and I would suggest that that actually was one of the most important innovations of all times. It created “money” as an abstract exchange unit with all the far-reaching effects unintended effects and benefit that we can only see today. And one of those effects was the facilitation of investments. That cannot be overestimated.
But while banking and investment is of such paramount importance still today, it was an innovation in its on right only in renaissance Italy. Its role and relevance for innovation today is that of an enabler of innovation, enabling research, enabling start-ups, enabling the pull-through of smart ideas to marketable goods and services. Nothing less, nothing more. For me, the economy serves a human purpose: facilitating the exchange of goods and services. It’s a stretch too far to see economics as THE human purpose, or even worse, to see society at the service of the economic.
Education and universities, that’s another interesting topic. Clearly, their development is tied to the availability of funding, i.e., rich donors. And it is tied to the spirit of the time: that knowledge can be gained through intellectual effort, that it can be stored and should be spread. But I’m not entirely convinced that this success story was driven solely by a need (which imply something like a purpose). I’m inclined to see an opportunity behind this development (not alone, but in concert with the need), the opportunity created by the political system of that time and that location: the relative wealth of the region, the competition between the cities and their leaders. Together these circumstances created the conditions for new kinds of intellectual endeavours that far exceed the traditional teachings and learnings of the Script and the seven Artes Liberales.
Lastly, Schumpeterian Waves. Of course such a chart draws a lot of attention. It condenses a lot of information, it dares interpretation, it focuses on some aspects and ignores others. For me personally, it depicts an abstract concept, a theoretical pattern. And that way the chart triggers discussion like ours now. I would agree that it is not perfect, I would e.g. want to think a little more about wave heights: are they all the same? or would shorter waves need to be higher?
But I would never suggest that any such depiction could ever deliver a complete view of all the different factors that are at play. Given the limitations of two-dimensional information representation, there is no way to get “everything” into one chart. We just have to keep that in mind whenever we look at graphs like that: they can only be partial representations of reality, and they necessarily incomplete.
Which leaves with some homework: think about the missing pieces and sketch those out. And then go back to earlier charts to find out whether all those concepts actually resonate. That’s still some journey ahead. And I’m happy that you are on board.