Nobody's City
Who's in charge here?
I have spent the better part of thirty years working on cities, and one question keeps arriving uninvited. Some cities have a single person who is responsible for them — not nominally responsible, actually responsible, in the sense that the buck stops at one desk, the voters can remove the person sitting at it, and consequential decisions get made without four agencies first agreeing. Other cities do not. The gap between these two types of city is not a gap in resources or ambition or the quality of the people working inside them. It is a gap in institutional design. Where it came from, and why it has proved so hard to close, is what this essay is about.
The potholes on Moi Avenue in Nairobi CBD are legendary for breaking axles and swallowing motorbikes. I have counted at least four agencies that could plausibly fix it: the county government, two national roads authorities whose mandates overlap, and an environment agency with jurisdiction over the drainage culvert that keeps collapsing.
All four exist. None of them has fixed Moi Avenue.
Ask the county and they will tell you, politely, that the budget allocation went to the CBD. Ask the roads authorities and each will question whether this particular stretch falls under their mandate. Ask the environment agency and they will send an inspector. The inspector will write a report. The report will be filed. The road will still be broken when the long rains arrive in April, and the drainage culvert will collapse again, and matatus will thread around the gap with the practiced patience of people who have learned to expect this.
This is what the question looks like on the ground. The padlock is working fine. It just was not designed to open from the inside.
To see the other possibility, go to New York in 2009. Michael Bloomberg decided to pedestrianize Times Square and told the Department of Transportation to do it. They put up orange cones within weeks. The cones became permanent planters within months. One person decided, one agency executed, and eleven million square feet of Manhattan street space changed its function permanently. Bloomberg could do this because New York’s mayor controls the city’s capital budget, its agencies, and enough of the political ground beneath him that consequential unilateral decisions are possible and survivable.
Nairobi’s governor holds a title roughly equivalent to mayor. Under Kenya’s 2010 Constitution, 47 county governments received devolved authority over health, agriculture, and local services. They did not receive the national agencies, which kept their staff, their mandates, and their reporting lines to Cabinet Secretaries. So when it rains hard enough that the Ngong River rises and Westlands floods back toward Chiromo Road, the county handles one part of the response, the Water Resources Authority handles another, the Kenya Red Cross shows up independently, and the question of who is responsible for the embankment that failed remains genuinely unresolved.
This is the puzzle. Across urban sub-Saharan Africa and South Asia, this fractured, nobody-in-charge condition is normal. Across the United States, a single empowered urban executive is normal. The divergence is not recent and it is not accidental. To understand it, you have to go back to 1899 and a railway campsite on the edge of the Kikuyu escarpment.
Nairobi was surveyed that year as Mile 327 of the Uganda Railway, chosen by engineers because it was the last flat ground before the steep climb toward Uganda. The city that grew around it was organized around the railway depot and the colonial administrative offices that sat in what is now the CBD, near the site of the old Provincial Commissioner’s compound off Harambee Avenue.
The Provincial Commissioner was the institutional answer to a specific question: how do you govern a large territory with a tiny administrative class, extract resources from it efficiently, and prevent the local population from organizing political resistance? The answer was to concentrate authority upward and outward, toward London, never downward toward the city itself. The PC was appointed by the colonial government, answerable to the colonial government, and specifically designed to stand above local politics. He administered. He did not represent. A mayor represents a constituency and can be removed by it. The PC represented the Crown and could be removed only by the Crown.
The scholar Mahmood Mamdani calls what resulted the “bifurcated state”: European settlers in Nairobi had access to elected municipal institutions while the African population was governed through a chain of appointed native authorities. The two systems shared a city but operated on different legal ground. A strong, locally elected urban executive — someone who actually controlled the budget and the agencies and answered to everyone living in the city — would have threatened this arrangement at its foundation. So the colonial administration produced, instead, a Municipal Board of Nairobi, appointed rather than elected until 1938, and even then elected on a franchise that excluded the majority of the city’s residents until independence.
New Delhi, built almost simultaneously, encodes the same logic in its street grid. The British planners sited Government House at the apex of a great triangular boulevard that ran outward toward the secretariat buildings. The geometry pointed toward the administration. American cities of the same era pointed toward City Hall, which sat at the head of a public plaza. Both street plans are still legible today. Both still organize how power flows.
The American city developed along a different track for reasons that had little to do with democratic virtue and a great deal to do with debt. In the nineteenth century, if Boston or Chicago wanted a sewer or a water main, it had to borrow. To borrow, it needed a credible executive who could promise that the debt would be repaid, someone lenders could hold accountable. This produced the strong mayor, part CEO, part ward boss, who consolidated executive authority because capital markets demanded it. The political machines that ran American cities in the Gilded Age were corrupt in ways that would make a modern procurement officer flinch. They were also effective at laying pipes, building schools, and absorbing waves of immigrants into the urban economy. The accountability flowed both ways: the mayor could be voted out. The machine needed to deliver.
The colonial city inverted this logic entirely. Infrastructure was funded by the central colonial treasury. There was no local borrowing, no municipal bond market, no constituency of creditors demanding a strong local executive. The city was a node in a global supply chain, designed to move goods to the coast and maintain order cheaply. The institutional structure followed from that purpose.
At independence, the District Commissioner traded his pith helmet for a suit.
In 1963, Jomo Kenyatta inherited the Provincial Commissioner’s office, the District Officer corps, the Public Works Department, and the entire administrative geography that the British had built for extraction and control. Decentralizing this to city mayors would have created dozens of potential rivals to the national government. It would also have required building new institutions from scratch while simultaneously running a country. Keeping the prefectural structure looked like the rational choice. It was, in the short term.
The Nairobi City Council that survived into independence was progressively hollowed out over the following decades. The central government dissolved it in 1983 and ran the city directly through a Commission for eight years. It was reconstituted, reorganized, and finally replaced by the Nairobi City County Government under the 2010 Constitution. The county government has real formal powers. It collected KSh 10.4 billion in own-source revenue in the 2021-22 fiscal year. It also shares the city with multiple national agencies — revenue, power, housing, water — each reporting to a Cabinet Secretary rather than to an elected city executive. No single person controls all of these.
The same pattern repeated elsewhere. India’s 74th Constitutional Amendment, passed in 1992, mandated that states transfer 18 categories of urban functions to city governments, including land-use planning and urban poverty alleviation. A 2013 assessment by the Administrative Staff College of India found that most states had transferred fewer than half the listed functions, kept control of the budgets that matter, and continued appointing municipal commissioners who answer to the state government rather than to elected city officials. Thirty years after the amendment, the gap between the law’s intent and its implementation remains the defining feature of Indian urban governance.
The puzzle I started with turns out to have a clean answer, even if acting on it is hard. The cities that developed strong empowered local government did so because their institutional histories created a constituency for it: lenders who needed a credible borrower, residents who could vote the executive out, a political culture that had never been organized around extraction. The cities that did not develop it were deliberately designed to prevent it. The prevention was effective.
This matters now because the cities of sub-Saharan Africa and South Asia are in the middle of the largest urban expansion in human history. By 2050, urban sub-Saharan Africa will hold roughly 1.4 billion people, up from about 700 million today, according to UN-Habitat’s 2022 World Cities Report. The street patterns, drainage gradients, and land-use decisions being made right now will be essentially permanent within a generation. The governance structure inside which those decisions get made shapes everything else. A city whose road maintenance requires four-agency coordination will also require four-agency coordination for its bus network, its affordable housing program, and its flood defenses.
The exceptions are instructive. Curitiba’s bus rapid transit system works because Jaime Lerner, as mayor in the 1970s, had the authority to close streets over the objections of business owners, redirect capital budgets, and instruct a single agency to execute. Medellín built cable cars up hillsides that no private developer would touch because Sergio Fajardo, as mayor from 2004 to 2007, controlled the capital budget and the urban planning agency and the public utility that anchored the surrounding development. Both cities are held up as models of what postcolonial cities can do. The governance condition that made them possible gets less attention than the cable cars and the buses.
The April rains come to Nairobi every year, turning the beautiful city lush green. The culvert under Peponi Road fills, backs up, and spills across the carriageway somewhere between Niks Donuts and Westgate Mall. The water finds the lowest point, which is the road itself, because the original drainage design assumed a city a fraction of its current size.
Four agencies share responsibility for this. The question I have spent thirty years circling turns out to be the question underneath the flooded road: who, exactly, is in charge here? The answer was settled in London, in the 1890s, by people designing an administrative system for a railway depot that did not yet exist as a city. They were not designing for Peponi Road. They were designing against the possibility that anyone living on it would ever have the authority to fix it themselves.
The institutional descendants of that design, with different titles and different business cards, still run the drainage.