How Cities Actually Grow

The conventional story of urban growth goes like this: cities attract people because they offer jobs, and they offer jobs because employers want to locate near other employers and near a skilled workforce. Agglomeration economies, in the academic jargon. This story is not wrong, but it is incomplete in ways that matter enormously when you try to understand why particular cities grew in particular periods rather than why cities in general tend to grow in general.

The actual history of urban expansion is more contingent, more physical, and more political than the clean economic model implies. Cities don’t grow because markets decide they should. They grow because of specific decisions about water, railroads, port rights, and administrative boundaries made by specific people at specific moments — and because geographic features that were irrelevant in one technological era became decisive in another.

Consider Chicago. By the standard agglomeration story, Chicago grew because it was well located for trade and attracted a workforce that attracted more employers. True enough. But the specific timing and scale of Chicago’s growth in the 1850s and 1860s was determined overwhelmingly by the decisions of railroad executives who chose to route their lines through Chicago rather than Galena or Milwaukee, decisions that were themselves shaped by land speculation by early Chicago settlers who had bought up the relevant parcels. The railroads followed the land speculation at least as much as the land speculation followed the railroads. The “natural” advantages of Chicago’s location existed for centuries before the city grew; what changed was the political and financial infrastructure around specific real estate interests.

The same dynamic repeats across urban history. Los Angeles in 1900 was a mid-size city of about 100,000 people, smaller than San Francisco by a factor of four. Its growth into the second-largest American metropolitan area was not economically inevitable. It required the Owens Valley water project, which was itself the product of a specific political decision by the Los Angeles Department of Water and Power to acquire water rights in a distant valley through a process that contemporary observers called, not entirely inaccurately, a fraud. William Mulholland’s aqueduct, completed in 1913, solved a resource constraint that would have capped Los Angeles’s growth at perhaps 300,000 people. Everything that followed was built on that foundation.

This matters for understanding contemporary urban growth because the lesson is not that economics is irrelevant but that economic logic operates within physical and political constraints that determine which cities can take advantage of favorable economic conditions. In the twenty-first century, the binding constraints have shifted. Water remains a constraint in the American Southwest and is becoming one in South and Southeast Asia. But the new binding constraint in most wealthy-country cities is land use regulation — specifically, the accumulation of zoning rules, environmental review requirements, and local approval processes that make it extremely difficult to add housing supply in response to demand.

The economic consequences of this are now well documented. Chang-Tai Hsieh and Enrico Moretti’s research, published initially in 2019 and substantially updated since, estimated that housing supply restrictions in high-productivity cities like San Francisco, New York, and San Jose cost the American economy roughly $2 trillion per year in foregone GDP — workers who couldn’t move to where their labor would be most productive because housing costs made it impossible. That number is almost certainly an underestimate of the global phenomenon, because the same dynamics operate in London, Sydney, Amsterdam, and Tokyo (though Tokyo’s more permissive zoning has made it a partial exception to the pattern).

What makes contemporary land use restrictions different from earlier physical constraints is that they are self-imposed by existing residents acting through democratic processes. The water limit on Los Angeles’s growth was a physical fact that required an engineering solution. The housing limit on San Francisco’s growth is a political fact that requires a political solution, and political solutions are much harder because the people creating the constraint are also the voters who would have to approve removing it. Homeowners in high-cost cities have a direct financial interest in restricting supply, since their homes appreciate faster in scarcity conditions. This creates what economists call a “homeowner veto” — a structural feature of urban governance that operates independently of party politics and that no standard economic incentive can easily dislodge.

The geographic dimension of contemporary urban growth adds another layer of complexity. The technology sector’s partial shift to remote work, which accelerated sharply after 2020 and has since stabilized at a level meaningfully higher than pre-2020 norms, is producing a slow redistribution of high-income workers from superstar cities to second-tier metros. Boise, Austin, Raleigh, Tallinn, Porto, and Lisbon have all experienced this dynamic to varying degrees. The pattern is not random: it favors cities with good climate (or at least tolerable climate), decent university presence, and enough existing amenity infrastructure to attract workers who expect certain lifestyle standards. This is a genuine shift in the logic of urban growth, but it is probably less transformative than its boosters claim. The industries that drive the highest-value economic activity still concentrate physically, even if some of the workers involved in those industries can live farther from the center. Finance remains in New York and London not because all financial workers need to be there but because the deal-making, relationship-building, and information-gathering that generates the highest returns requires physical proximity.

The city-building story of the next two decades will be told primarily in the Global South, where urbanization is happening at a pace and scale that dwarfs anything in the historical record of wealthy countries. Sub-Saharan Africa added roughly 500 million urban residents between 2000 and 2025, and is projected to add another 900 million by 2050. These are not cities growing around existing infrastructure; they are cities being assembled simultaneously with the infrastructure. Lagos, Kinshasa, Dar es Salaam, and Kigali are undergoing in decades transitions that took European and American cities centuries.

What the historical record suggests about this process is simultaneously reassuring and alarming. The reassuring part: cities in very poor countries have consistently raised living standards for migrants from rural areas even when urban conditions are difficult by wealthy-country standards. The productivity premium of urban labor over rural labor holds across vastly different income levels; simply being in a city, with access to denser markets and more complex economic networks, raises incomes. The alarming part: the specific form that African and South Asian urbanization is taking — informal settlements with weak property rights, inadequate water and sanitation infrastructure, high exposure to flooding and heat stress — creates path dependencies that are very difficult to escape. Jakarta spent decades trying to retrofit adequate drainage infrastructure into a city that was built without it. The attempt largely failed, which is part of why Indonesia built an entirely new capital.

The political economy of urban growth has one more dimension that rarely gets the attention it deserves: municipal boundaries. The jurisdiction that governs a city determines who counts as a taxpayer and who counts as a beneficiary of public services, which shapes both the fiscal capacity of the city and the political incentives of its government. American cities are unusually constrained by fixed municipal boundaries that were often drawn in the nineteenth century, which means that the tax base of the central city and the tax base of the metropolitan area diverged as suburbanization proceeded. Detroit in 2013 went bankrupt partly because it was legally a city of 700,000 people in a metropolitan area of 4.3 million, and the 3.6 million people in the suburbs owed nothing to the city’s debts.

European cities generally avoided this through more fluid municipal governance: the boundaries of Paris, Vienna, and Stockholm expanded in the twentieth century to capture suburban growth, and regional governance structures allow metropolitan-scale coordination. This is not simply a better administrative arrangement in the abstract — it produces meaningfully different fiscal outcomes and meaningfully different capacity to invest in the infrastructure that drives agglomeration.

The lesson from all of this is not that urban growth is impossible to influence through policy. Cities can be shaped, directed, and accelerated. But the levers that work are usually not the ones planners focus on. Comprehensive plans, zoning codes, and design standards matter relatively little compared to infrastructure investment decisions, boundary and annexation rules, and the property rights framework that determines whether and how land markets can respond to demand.

The cities that thrived historically did so because they had physical advantages that other cities lacked, or because specific political decisions removed constraints at critical moments, or both. The cities that will thrive in the next fifty years will be the ones that figure out how to remove their own self-imposed constraints before the window of growth opportunity closes.

Some cities will manage this. Most won’t, which is also how history has always worked. The geography of prosperity is never permanent, which is what makes urban history so instructive and so frequently ignored.