Big cities are economic powerhouses. New York City by itself generates nearly 5 percent of America’s GDP with only 2.5 percent of the country’s population.
But how much of New York’s success is because of its size? Do large cities make humans more productive by bringing them together—thus fostering cooperation, innovation, and skill growth? Or do they merely attract the type of ambitious people who would succeed wherever they lived, while keeping out lower earners through high housing prices?
The answer probably involves a little of both, but economists have spent a lot of time trying to pin down the details. One controversial 2019 paper, for example, claimed that large cities’ refusal to allow more housing “lowered aggregate US growth by 36 percent from 1964 to 2009.” A new working paper, from Brown economists Matthew Turner and David N. Weil and published by the National Bureau of Economic Research, reaches a different conclusion.
The researchers found that if every metro area’s size had been capped at 1 million in 1900, the nation’s economic output in 2010 would have been 8 percent lower—a significant blow, yes, but a modest effect given such an enormous restriction. More than half of the U.S. population lives in places whose populations would have been capped in this alternate reality.
The paper’s estimates are based, in part, on existing research. Previous studies have estimated cities’ effects on productivity by measuring how much more workers get paid in bigger metros. Others have approximated cities’ effects on innovation and growth by analyzing whether inventors are more productive when located in a bigger cluster of peers. Turner and Weil collected these estimates and picked a sensible middle ground for each important number. Then they plugged the parameters into an economic model, along with data on metro areas’ historical size and economic performance.
The authors conclude that the relevant relationships are rather weak. According to their model, per-person economic output grows only about 4 percent as quickly, and per-person patenting about 6 percent as quickly, as city size. Their model totals these effects across all the nation’s metro areas and compounds them over time, leading to the headline result: if cities’ populations were capped at 1 million people from 1900 onward, national output in 2010 would have been only 8 percent lower.
Turner and Weil’s results, however, depend heavily on their parameters, such as the 4 percent and 6 percent figures. If you adjust those inputs slightly, as the authors do for curiosity’s sake, you get different results. Increasing them to 8 percent and 20 percent, respectively, leads to a 17 percent reduction in output from a 1 million-resident cap; more aggressive changes to the model generate an output reduction of a third.
What are the takeaways here? Even if we assume that Turner and Weil’s preferred small estimates are correct, it’s hardly a fatal blow to supporters of big cities or advocates of greater urban housing supply. While their research indicates only small-to-moderate innovation and output benefits associated with city size, those are still benefits. And economic justifications are not the only reasons to support thriving big cities with plentiful and affordable housing.
The study is also a reminder, though, that economic models can generate an extremely wide range of estimates based on what’s fed into them. We shouldn’t take them as gospel—whether they’re being used to claim surprisingly large or surprisingly small benefits for big cities.
Photo by CHARLY TRIBALLEAU/AFP via Getty Images
