When ABBA sang "The winner takes it all/The loser standing small", they were referring to a divorce. But their lyrics equally well describe the structure of the global economy.
The word "superstar", traditionally used to describe musicians and actors, has been increasingly applied to corporate executives, companies and industries. It is used to denote individuals or organisations that account for an outsized share of economic output, while their peers see market shares slip.
Consider this example: how many people use the second-best internet search engine? Even if the difference between the best and second best in terms of performance is only marginal, and there is no cost to switch, why wouldn't you use the one that is slightly better?
Once that initial advantage is established – by Google in this instance – then that superstar company can continue to invest in its technology to solidify its place as a market leader. This means there is even less incentive for a customer to switch to one of its competitors.
Superstar cities
Over the past 2 decades, cities have been exhibiting similar superstar characteristics, in that certain locations are producing more than we might expect from their share of the national population. In a 2018 report, the McKinsey Global Institute (MGI) found that 50 cities, which make up just 8% of the world’s population, account for more than 20% of the world’s output.
Perhaps unsurprisingly, these cities also appear to be among the most integrated and most innovative, and are home to the headquarters of nearly half of the world's largest firms. As Harvard economist Edward Glaeser wrote in The Triumph of the City: "the central paradox of the modern metropolis [is that] proximity has become ever more valuable as the cost of connecting across long distances has fallen." That suggests even COVID-19 is unlikely to alter this trend.
But how pervasive are superstar effects in cities, and what are the factors behind them? If location didn't matter then economic output would be spread evenly across a country; the MGI study suggests that this is certainly not the case, however.
"50 cities, which make up just 8% of the world’s population, account for more than 20% of the world’s output"
Commonalities between superstars
Cities themselves may also benefit from institutions such as the legal system or arrangements such as trade deals, which are often determined at the national level. This can make it difficult to assess the prevalence of superstar effects across borders, because success may be attributed to national factors, such as security or the law, or to local ones, such as human capital and infrastructure, in varying degrees.
For example, would Luanda have developed differently from Los Angeles if each city had had the same legal system? And how would Pune change if it could access the same trade agreements as Paris?
To attempt to answer this, I developed a framework of analysis that assesses the economic concentration in cities in any country, controlling for factors that are not determined at the local level to enable international comparison. The result was the superstar city statistic (SCS).
This is still only 1 metric to evaluate the success of cities, though, and there remains significant scope for further research into the factors behind economic concentration.
Applying the formula
The results clearly indicate that superstar effects are globally pervasive. For instance, residents of London, Munich, New York, Paris and San Francisco produce on average 30–80% more economic value annually than residents of other cities in their respective countries. At the most extreme, economic concentration in San José – which hosts the headquarters of global tech companies such as Google and Apple – is such that the average resident produces almost 3 times as much as someone working in the average US city.
These superstar effects are not only visible in developed countries. A sample of 124 African cities shows that the SCS for Nairobi is similar to that of Paris and Munich, with residents producing 65% more economic output annually than Kenyans living in other urban areas.
It is also important to highlight the dynamics at the bottom of the distribution, as well as the top. In nearly half of US metropolitan areas, a resident produces 25% or less than the country’s average city-dweller. Data also shows that US economic concentration has clearly increased over the past 2 decades, and there is good reason to believe that this trend would also be observed globally.
"A sample of 124 African cities shows that the SCS for Nairobi is similar to that of Paris and Munich"
Further consolidation ahead
All of this suggests that agglomeration economies are increasingly present in certain cities, potentially at the expense of others. To paraphrase ABBA, these cities continue to take a greater share of production away from smaller ones.
The results have a clear implication for the private sector when deciding where to invest. However, the implications for policy are less clear.
A natural question for policymakers would be whether this inequality between cities – this winner-takes-all dynamic – is to be encouraged or restrained. Should a public official operating under budget constraints attempt to pick winners, and make place-based investments where they are likely to yield the highest returns? Or should policymakers attempt to level the playing field between cities?
Further research into the factors behind economic concentration would likely clarify things. However, social preference rather than economic efficiency may be the ultimate determinant.