When the Towers Go Dark: What Cleveland and China Can Learn from Each Other
Downtown Cleveland has an impressive skyline for a city its size. Key Tower, 200 Public Square, the old BP Building. It's all serious architecture, built to hold thousands of workers and still structurally sound. Many of those floors are empty and have been for years. Not temporarily empty, not between-tenants empty, but empty with no demand. The demand that initially propelled their construction is gone, and no real estate developer can figure out the math to justify future uses for them.
Way on the other side of the world is Yujiapu, Tianjin's would-be Wall Street on the Haihe River. During Cleveland's heyday nobody was even alive who would later think to built Yujiapu. By the time its towers were constructed Cleveland had already been in decline for many decades. And yet they share a skyline problem. In Yujiapu there are also towers of 40, 50, 60 floors of glass and steel. Theirs were built fast and built well. They are almost entirely empty, and in many cases not even finished works of architecture, with visible cores, unfinished cladding, and in some cases merely floor plates stacked and unfinished, by the dozens. Yujiapu is an infamous ghost city. Yujiapu's metro line arrived in 2022. The tenants have not.
These are two cities with two skylines full of towers nobody uses. The comparison writes itself. But the reason Cleveland's towers are empty and the reason Yujiapu's towers are empty are completely different, and the difference matters more than the visual similarity. The solutions do not transfer. The lessons, such as they are, run in only one direction.
How Cleveland's Towers Went Empty
Cleveland's vacancy problem is the residue of a century of growth followed by sixty years of contraction. The city built its downtown for a population and an economy that no longer exist at the scale they once did, with little hope of coming back. At its postwar peak Cleveland had nearly a million residents. Today it has under 370,000. The office towers were not overbuilt relative to the economy that justified them. That economy simply left.
The causes are the collapse of Great Lakes manufacturing, the automation of what remained, the federal highway and mortgage policies that subsidized suburban flight and systematically defunded urban cores. Cleveland did not fail exclusively through bad planning. It was subject to forces operating at a national scale that no city-level decision could have offset. The towers are empty because the employers moved out and the employers moved out because their industries died or were replaced by international competitors, and the workers either left the city altogether or followed cheaper housing because the federal government made that possible and advantageous. By the time Cleveland understood what was happening, it had already happened.
Cleveland's rebuild feels permanently partial. The Cleveland Clinic anchors a medical economy that didn't exist in the steel era. Ohio City and Tremont work as neighborhoods. The waterfront has been partially reclaimed. But the population has not come back, and there is no plausible scenario in which it does at the scale required to fill what was built. Some of those towers have been converted to residential. Some have been demolished. The rest sit in a kind of productive limbo, neither fully abandoned nor fully occupied, sustained by a regional office market that is under pressure from remote work patterns. Meanwhile, new development projects eat up any demand that does exist for Class A office or rental residential downtown.
How Yujiapu's Towers Went Empty
Yujiapu's vacancy is structurally different. Those towers were not built for an economy that then left. They were built for an economy that was never there in the first place. The pitch was straightforward and enormous: construct China's answer to Manhattan on the Haihe River, close enough to Beijing via high-speed rail that the capital's financial activity would naturally migrate south to be closer to the Tianjin port. Build the skyline, build the infrastructure, and the tenants will follow.
They didn't. Not at the pace or scale the projections required. The reasons are familiar to anyone who has watched China's property sector over the past decade: local governments funded by land sales had every incentive to build and almost none to wait, so development ran years and sometimes decades ahead of organic demand. Yujiapu was a bet on a financial geography that the market, left to its own devices, was not going to produce on the timeline the planners assumed.
The property crisis that erupted with Evergrande in 2021 and has ground through the sector ever since has closed off the most optimistic scenarios for places like Yujiapu. China's population began shrinking in 2022. The demographic expansion that the entire model was calibrated for is over. The towers are not going to fill up because the conditions that could have filled them are not coming.
What Do You Do With a Tower Nobody Wants?
Cleveland has a genuine head start, because it has been working on the problem for thirty years and has accumulated knowledge about what works and what doesn't.
The most important thing Cleveland learned is that office-to-residential conversion is harder than it looks and more necessary than almost anything else. The floor plates on older towers were designed for open-plan offices with perimeter windows and dark cores. Converting those to apartments requires carving out light wells, reconfiguring mechanical systems and navigating building codes that were not written with this use case in mind. Cleveland has done it, expensively and slowly, on a number of its older towers. The 668-unit conversion of the old Ameritrust complex added residential density to a downtown that needed it badly. Conversion is hard, but conversion is possible if the financing structure is right. Getting the financing structure right requires a mix of historic tax credits, low-income housing credits and municipal land bank tools that most cities have to assemble from scratch the first time.
Yujiapu's towers present a harder conversion problem. They are newer, for one thing, and they were built to Chinese commercial standards that make residential conversion even more structurally complicated than in the American context. But the underlying logic applies: if office demand isn't coming, the choice is between conversion, demolition and managed vacancy. Managed vacancy is not a strategy, especially with this many towers. Permanent "ghost city" status isn't really a solution. Demolition destroys sunk capital. Conversion, for all its difficulty, at least produces something that can generate ongoing economic activity.
The second thing Cleveland learned is that a downtown cannot survive on office workers alone, and that trying to rebuild it as primarily an office district is a losing strategy in almost every post-industrial American city. The downtowns that have stabilized are the ones that built a residential base dense enough to support ground-floor retail, food and entertainment independent of the nine-to-five office population. Cleveland's downtown residential population has grown significantly over the past two decades, and that growth has made the district more resilient to office vacancy than it would otherwise be. The towers that converted to residential pulled daytime-only districts into something that functions at night and on weekends.
This is a lesson Yujiapu has only experimented with by adding new residential construction proximate to the Yujiapu district, but not through conversions. A pure financial district that fails to attract finance has no fallback. A mixed-use district that happens to have very large floor plates has options. The question is whether China's planning system, which is very good at designating zones and building infrastructure, can pivot to permitting and financing the kind of mixed-use density through conversion of vacant floor plates that didn't appear in the original plan. There are examples elsewhere in China of exactly this kind of rezoning happening at speed when political will is present. The tools exist.
Redefining What a Downtown Is
The bigger question both cities are working through, mostly without naming it directly, is what a downtown is actually for when it is no longer primarily a place where large employers concentrate workers in tall buildings.
Cleveland's answer has been to lean into institutions: the Cleveland Clinic, the universities, the museums, the sports venues. These are uses that require physical presence, generate foot traffic across multiple hours of the day and anchor the surrounding district in ways that office tenants no longer reliably do. The medical corridor on the eastern edge of downtown is not glamorous urban design, but it is economically durable in a way that speculative office development is not. People need hospitals regardless of what happens to remote work.
The institutional anchor model is one that Yujiapu could adapt more deliberately than it has. Right now the district's institutional anchors are thin: the factoring industry association, some government offices, the transit infrastructure itself. That is not enough to generate the kind of consistent activity that makes a place feel inhabited. But Tianjin has universities, cultural institutions and government agencies that could theoretically be relocated into Yujiapu in ways that would change the district's daily rhythm. The challenge is political as much as logistical: relocating institutions is disruptive and expensive, and Chinese local governments have proven more willing to build new districts than to reorganize existing ones.
The other thing both cities are learning is that the ground floor matters more than any other floor. A tower with empty upper floors and an active ground floor is a functioning piece of urban fabric. A tower with empty upper floors and a dead ground floor is a problem that radiates outward and drags down everything around it. Cleveland has had success in some corridors with ground-floor activation: covered markets, food halls, cultural venues, pop-up retail programs that use vacant ground-floor space to generate activity while longer-term tenants are recruited. These are not solutions at the scale of the vacancy above them. But they hold the urban fabric together while the larger problem is worked on, and they are far cheaper than letting the ground floor sit dark.
Yujiapu's ground floor situation is worse than Cleveland's because Yujiapu is much emptier than Cleveland. The district was designed as a financial center, which means the buildings' bases were planned for the kind of formal, low-traffic uses that a bank lobby or corporate entrance requires. Today, Yujiapu contains almost no restaurants. Activating that ground floor for retail or food requires physical reconfiguration and a retail leasing strategy that the original development plan did not contemplate. Cleveland's experience suggests this is solvable, and that the investment in getting it right pays back in the district's overall functionality faster than almost anything else.
The Financing Problem
Cleveland's toolkit for financial remediation, assembled over decades, relies on a patchwork of federal tax credits, state programs, philanthropic capital and private developer risk-taking. It is slow, requires significant local expertise to navigate and breaks down regularly when federal programs change or private capital decides Cleveland is not competitive with other markets. The city has made it work on individual projects, sometimes spectacularly. It has not made it work at the scale the overall vacancy problem requires.
Yujiapu's financing problem is in some ways more tractable and in some ways worse. The state can direct capital at scale and speed that no American city can match. But the property crisis has compromised the local government finances that would normally fund this kind of intervention, and the political math around admitting that a flagship development project needs fundamental restructuring is complicated in a system where that project was championed at senior levels. The capital is potentially available. The institutional willingness to deploy it in ways that amount to an acknowledgment of the original plan's failure is less certain.
Both cities are, in different ways, waiting for a financial structure adequate to the problem. Cleveland is waiting for federal policy that treats post-industrial vacancy as the infrastructure problem it actually is. Yujiapu is waiting for a political moment in which the state decides that restructuring the district is a priority worth the cost and the implicit admission it requires.
What Each City Can Actually Learn From the Other
Cleveland's most exportable lesson is the value of patience combined with a clear theory of what the district is becoming. The near west side neighborhoods recovered because the city held land, maintained basic services, kept the ground-floor activity going and waited for enough private risk-takers to accumulate critical mass. That is a model that requires municipal patience and financial durability that not every city has. Cleveland developed both out of necessity. The lesson for Yujiapu is not to expect a single intervention to solve the problem, and not to declare victory based on a few high-profile conversions while the broader district remains under-activated.
Yujiapu's most exportable lesson is the value of infrastructure investment made ahead of recovery rather than in response to it. Cleveland's downtown revival has been persistently hampered by transit gaps, infrastructure deferred during the lean years and public realm quality that lags behind what private developers need to take risks in a district. China's willingness to build the metro line, improve the public realm and invest in connectivity before demand fully materializes is a model that American cities, constrained by capital budgets and political cycles, rarely manage to replicate. The infrastructure creates optionality. Cleveland has learned this slowly and expensively.
Neither city has solved the problem of towers that went dark for reasons their builders couldn't have anticipated. But both are far enough into the attempt that the comparison is worth more than the visual similarity of their skylines. The physical problem is hard. The toolkit for addressing it, assembled from thirty years of Cleveland's trial and error and China's capacity to intervene at scale, is more developed than it was. That is not a satisfying conclusion.
