The world is increasingly spiky. Inequality is growing and it is increasingly geographic in nature. We know that people tend to make more money in urban areas compared to rural areas – even when they possess the exact same level of education. The returns to being smart and educated are simply greater in cities.

But they also depend on the size of the city. Mark Muro and Jacob Whiton of Brookings recently published data looking at labor market performance – by metro size – from 2009-2015 (right after the financial crisis). What they found is that larger metropolitan areas simply performed better than smaller ones.

In summary:

City size matters because it’s a major influence on city prosperity and adaptability as well as local worker fortunes. Bigger cities are more productive. They are more innovative. They draw better-educated workers by offering higher wages.

The situation is even more pronounced across the pond. According to the New York Times (quote from Richard Florida), a third of Britain’s gross domestic product comes from London alone.

What is far less clear is what should be done to address the decline of some of the smaller cities in America – cities that are stagnating and feeling left behind. But perhaps the first step is acknowledging what has happened and what remains feasible in today’s global economy.

Here is another quote from the above NY Times article:

Mr. Trump’s promise to relieve the pain by reviving the coal and steel industries, by keeping immigrants out of the country and by raising barriers against manufactured imports is only a rhetorical balm to satisfy an angry base seeking to reclaim a prosperous past that is no longer available.

That rhetorical balm.

Josh Lehner of the Oregon Office of Economic Analysis published a study earlier this summer where he looked at employment growth according to city size across the US.

What he found since the Great Recession of 2008-2010 is that larger metros – with populations greater than 1 million people – have rebounded the fastest. They are shown in the light blue line below:

According to Josh:

This is at least partially due to the fact that all those good economic things — agglomeration effects, knowledge spillovers, clustering, etc — happen in certain locations, which are usually bigger cities.

However, if you go back to the 1980s, you find that this trend isn’t consistent. Large metros outperformed in the late 90s. But they were more or less on par with smaller metros during the housing boom of the early 2000s and actually under performed in the early 90s recession.

One possible explanation for this – which Josh proposes – is that the recent housing boom acted as a sort of equalizer for smaller metros. It created stronger population growth outside of the bigger cities.

I buy that.

But then does that mean that going forward big cities will continue to outperform? Is this going to be more or less the new norm? Intuitively, I would think yes.

You can find Josh’s blog post, here. Richard Florida also wrote one for CityLab, here.

The MIT Senseable City Lab recently teamed up with a few other research groups to investigate the relationship between human interactions and city size. If you happen to be a member of the Journal of the Royal Society Interface, you can download the full report here. But in true ATC fashion, I’m going to give you the Coles Notes version here.

What the study did was look at billions of anonymized mobile phone data in both Portugal and the UK in order to determine how our real life social networks change with city size. And what they found is a pretty consistent relationship:

[T]his study reveals a fundamental pattern: our social connections scale with city size. The larger the town you live in, the more people you call and the more calls you make. The scaling of this relation is “super-linear,” which means that on average, if you double the size of a town, the sum of phone contacts in the city will more than double – in a mathematically predictable way.

What’s interesting about this finding is that it starts to explain how cities–and the clustering of people–can act as fertile ground for the exchange of ideas and knowledge. The bigger the city the more people you probably know.

But what I’m curious about (I don’t have the report) is if there’s some kind of upper limit. Presumably this “super-linear” relationship tapers off after a certain city size, because there has got to be limits to the number of people we can maintain productive relationships with.

According to British anthropologist Robin Dunbar, that number was 150 people.