Drive an hour outside some of the world's fastest-growing metros and you'll find them: gleaming apartment towers with no tenants, six-lane boulevards with no traffic, shopping malls where footsteps echo. China gets the most attention, but ghost cities exist in Spain, Angola, Egypt, and the American Sun Belt. They aren't accidents.

Ghost cities are what happens when demographic projections meet political ambition. Someone drew a line on a chart showing millions of future residents, and that line justified concrete. Sometimes the residents arrive decades later. Sometimes they never do. Understanding why we keep building for populations that may not exist tells us something important about how societies gamble on their own futures.

Population projections look scientific, but they rest on assumptions that quietly favor growth. Planners typically extend recent trends forward, assuming that a decade of 3% annual growth will continue for another twenty years. Demographers call this the momentum fallacy: treating a temporary surge as a permanent trajectory.

The math compounds quickly. A city growing 5% annually doubles in fourteen years. Planners see the doubling and design infrastructure for it. But growth rates almost never hold. Migration slows as opportunities elsewhere improve. Birth rates fall as incomes rise. The demographic transition, which every developing region eventually experiences, is remarkably consistent, yet planners often build as if their region will be the exception.

There's also a data problem. Projections made during boom years get baked into 20-year master plans. By the time better data arrives, the concrete is already poured. Ordos Kangbashi in Inner Mongolia was designed for a million residents based on coal-boom projections. When coal prices collapsed, the projections didn't update fast enough to stop construction.

Takeaway

Straight lines on population charts almost always bend. The safest projections assume that today's growth rate is temporary, not permanent.

Even when officials know projections are optimistic, they often build anyway. The reason is straightforward: construction generates jobs, tax revenue, and visible achievement now, while the consequences of overbuilding arrive years later, usually on someone else's watch.

In China, local officials were long promoted based on GDP growth, and construction offered the fastest path there. A mayor who built a new district for 500,000 people got credited for the growth immediately, whether or not the residents materialized. In Spain's pre-2008 boom, mayors approved developments that expanded their tax base overnight. In Egypt and Saudi Arabia, new capitals become legacy projects for leaders who won't be around to see them filled.

The incentive structure rewards ambition, not accuracy. A conservative planner who correctly predicts modest growth gets no monument. An optimistic one who builds a ghost city at least gets a ribbon-cutting. Combine this with cheap credit and land available for reclassification, and the machine builds itself. The residents are almost a secondary concern.

Takeaway

When the rewards for building arrive today and the costs of overbuilding arrive tomorrow, expect more building than tomorrow can absorb.

Not every ghost city stays empty. Some fill slowly as broader demographic forces catch up to the concrete. Pudong in Shanghai was mocked as a ghost district in the 1990s; today it houses millions and hosts one of the world's busiest financial centers. Kangbashi, once the poster child for Chinese overbuilding, is now roughly half occupied and functioning as a real city.

The pattern that predicts recovery is connection. Ghost cities integrated into functioning regional economies, served by transit, and located near existing job centers eventually attract residents. Isolated developments in the desert or steppe rarely do. Population follows opportunity, and opportunity clusters around existing networks. You can build the buildings, but you can't manufacture the reasons people move.

The recovery timeline is often generational. Cities designed to fill in ten years may fill in thirty, or never. The financial losses along the way, borne by banks, governments, and often ordinary savers, are real. But the physical infrastructure sometimes finds its purpose eventually, becoming the skeleton for something the original planners never quite imagined.

Takeaway

Buildings don't create demand; they wait for it. The cities that recover are the ones connected to something people actually want to reach.

Ghost cities are monuments to a specific kind of confidence: the belief that we can plan a future we haven't yet earned. They reveal how demographic projections become political tools, and how the costs of being wrong get spread across generations.

The next time you see a projection promising doubled populations or endless growth, ask who benefits from believing it now. Demographic change is real, but it's slower, messier, and more conditional than master plans suggest. Concrete is patient, but populations have their own timelines.