A neighbourhood can gain better drains, safer buildings and a new bus stop while becoming harder for the people who live there to afford. That possibility was central to the 2026 World Cities Day discussion in Fuzhou, China. In its November 2 account of the global observance, UN-Habitat described a debate over how to improve existing neighbourhoods and ensure their residents share the benefits. It is a useful test for a plan anywhere, from the repair of one housing block to the redevelopment of a whole district.
Start by comparing two things over time. First, follow the buildings and services. Are homes safer? Is the bus stop easier to reach? Then follow the households that were there when the project began. Can they afford to stay? If they moved during work, did they return, settle somewhere they wanted to live, or lose contact with the project? The answers require different records. A completed building can be counted from a construction log. A household's path often cannot.
What the Fuzhou discussion established
World Cities Day is observed on October 31. The global programme in Fuzhou spanned October 31 to November 2, and UN-Habitat's account published November 2 describes discussions and addresses from the observance. The dates refer to different things: the annual observance, the programme and the publication of the account. The release records what participants discussed. It does not show that any particular project has delivered the outcomes the speakers called for.
The theme paired urban regeneration with adequate housing. That pairing matters. Improving a district might include restoring buildings, replacing unsafe structures, creating homes on vacant land, upgrading drainage, or connecting people to services. Adequate housing asks what households can actually use and keep. A clean streetscape and an occupied apartment measure different outcomes.
UN-Habitat's World Cities Report 2026 puts affordability, displacement, informal housing, climate risk and the conditions of everyday life in the same discussion. Its estimate of up to 3.4 billion people experiencing some form of housing inadequacy is an upper estimate with a broad definition. It should not be read as a count of people sleeping outside, nor as a count of households that must move. The relevant question for a local project is which specific problems it can identify and change.
Establish who was there before the first move
A credible comparison begins before work or relocation. Establish a date, draw the project boundary and state who counts as a resident. Owners, leaseholders, subtenants and people with informal arrangements can all experience the work differently. A list built from property titles will miss households that rent, and a list built only from named leases can miss people whose housing arrangements are undocumented. The project needs a way to recognize those gaps without publishing private household information.
Record basic conditions with enough detail to compare them later. A count of homes is useful, but so are the number occupied, the range of rents or housing charges, travel time to regular destinations, accessibility barriers, and the interruptions residents already experience. If a street floods every wet season, record where and when. If a bus runs infrequently, note the times relevant to actual school or work trips. If the existing information is incomplete, state what could be measured and what remains uncertain.
The choice of baseline date matters because people can move before construction begins. Announcements, lease changes, property sales and early service interruptions can all precede demolition or repairs. If the project starts counting only on the day the contractor arrives, it may miss households that have already left. An earlier reliable count and a later check can make that change visible.
Measure the cost of living in the improved neighbourhood
A rent figure by itself tells only part of the story. Compare the costs that affect what a household can keep paying. A different flat may have lower utility bills after repair, but a higher deposit or new recurring building charges. A move can change transport spending and the time needed for school, work, care or shopping. These amounts need not all move in the same direction.
Consider an illustrative household paying 500 units in rent, 70 for utilities and 50 for its usual travel, or 620 altogether. It is offered a repaired flat at 550 with utilities of 40 and travel of 50, a total of 640. Another offer has rent of 490, utilities of 60 and travel of 110, totalling 660. The lower advertised rent belongs to the more expensive of the two offers on these assumptions. One-time moving costs and deposits would need a separate comparison. Neither total tells us whether the household can absorb the increase without knowing its resources and other obligations.
This is a way to compare options, not a formula for every city. A project can show the distribution of changes by income or tenure and the proportion of households facing significant increases under a locally explained measure. An average alone can hide that some residents saved money while others could no longer afford the district.
Ask what happens when a resident has to leave during work
Some improvements can take place while people remain at home. Others require a temporary move or rebuilding. A relocation plan should say who qualifies for which support, where residents can live during work, what happens if the project runs late, and what terms apply to a return. A statement that residents can come back leaves important questions open. Which residents? To what kind of home? At what cost? When can they accept the offer?
Follow the difference between an offer and an outcome. A project may record that a household received a letter. That does not establish that the household could accept the terms, that it wanted to return, or that it did return. Keep separate counts for eligible households, offers made, acceptances, moves back, moves elsewhere by choice and outcomes still unknown. Use the same baseline cohort or explain each change in the denominator.
For example, suppose 100 households lived in a building before work. The project reaches 80 afterward. Of those, 60 return and 20 live elsewhere. Reporting that 75 percent of contacted households returned is mathematically accurate, but says nothing about the remaining 20. Reporting 60 known returns out of the original 100, with 20 outcomes unknown, gives the reader the information needed to judge the result. Neither example is a claim about Fuzhou or any real programme.
Check whether the new services work at the hours people need them
A rebuilt footpath is a physical improvement. Whether people can use it is a separate question. Walk the whole route between homes and a destination, including the crossing where the improved pavement ends. Look for a usable route for a wheelchair or a person with a pushchair. Consider shade, lighting and drainage at the places where people must wait. If a route is being evaluated for a school or shift job, check it at the relevant time. A route that works at noon may serve a different set of people from one that works before dawn.
Maintenance belongs in the assessment. A drain inspected once on opening day cannot tell you what happens after a heavy rain or after a year without cleaning. Record who maintains it, how problems are reported and whether they were fixed. The same applies to lifts, entrances and shared lighting. For a lift user, the number of days it is out of service may matter more than the fact that it appears on a list of completed installations.
Projects should also explain the limits of their evidence. One route observation does not represent every resident. A short monitoring period may miss seasonal failures. Reports can distinguish between completed work, checks actually carried out and outcomes awaiting a later visit.
Give participation a decision to change
A meeting is easy to count. A change to a decision is more informative. Ask what was open to influence and when residents heard about it. Were they shown a choice between repairing in phases and moving everyone at once? Could they comment on who qualifies for return or where a temporary bus stop should go? Did the project publish an answer to concerns that it did not adopt?
A practical record can list the issue raised, the options still open, the decision, who made it and why. Suppose residents point out that the proposed temporary crossing falls beyond the stop used by schoolchildren. The response might move the crossing, change the bus stop or explain why another route was chosen. Any of these responses can be examined. A count of attendance cannot show whether the crossing question changed anything.
The way a project collects input affects who appears. A single weekday meeting may reach fewer shift workers or carers. A meeting upstairs without a usable lift leaves some people out. A confidential way to report problems can matter to someone unsure whether their housing arrangement will be recognized. Participation evidence is strongest when the project can explain whose input it reached, what it did with it and whose experience remains missing.
Read a progress report in two columns
Make one column for work delivered and another for residents' outcomes. In the first, put homes repaired, new homes completed, lifts installed, crossings rebuilt or services restored. In the second, put the number of baseline households still in the area, their changes in costs, known relocation outcomes, disruption during works and access to important destinations. Record the dates and the source beside each figure. Keep the unknowns visible.
If a project reports 200 new homes and the original site had 120 occupied households, those numbers cannot tell you how many residents benefited. The new homes may include different sizes, different charges and different eligibility rules. Compare how many were offered to the original households, who could take an offer and what happened to everyone else. At the same time, count the genuinely new households served. An assessment can recognize an increase in supply and still examine who could reach it.
Look for repeated checks. Some changes appear on opening day. Others take several rent cycles or a rainy season to become visible. Compare the same outcomes at a clear set of intervals and explain changes in data collection. A later check may answer questions that the completion report can only leave open.
A further distinction is between a change observed and a change caused by the project. Suppose travel times fell after a housing upgrade, but the transit authority also introduced a more frequent service across the city. The improvement still matters to residents. It would take more evidence to say how much credit belongs to the housing project. Show the dates of both changes and the trips measured. If a suitable comparison is available, explain what makes it comparable and what differs. Where the evidence only supports a before-and-after description, give readers that description and leave the causal question open.
Use the test before the plan is final
At the start of a proposal, ask for the date and boundary of the baseline, the options for people who have to move, and the evidence the project will collect after opening. Ask which office or organisation will hold the records and how residents can correct a mistake in their own entry. A useful answer might identify that some residents are missing from current records and explain how the project will reach them. It need not pretend to have a perfect count.
When reading announcements, watch the verbs. A government can promise, fund, approve, start, finish or verify a change. Those actions are distinct. An approved budget does not tell you whether repairs happened. A photograph of repairs does not establish that residents can afford to stay. A later household survey does not automatically explain why an outcome changed. Matching the claim to the evidence makes a progress story easier to assess.
The practical question raised in Fuzhou is what it would look like to improve a place in a way its residents can use. The answer might include new homes, safer buildings and more reliable services. To judge whether those benefits reached people who were already there, follow their routes, costs and housing outcomes as carefully as the construction itself.
For another example of separating an announced figure from what it measures, read humanitarian aid funding figures in context.
Sources checked November 3, 2026. The event account is UN-Habitat's own published report. Global housing context comes from its World Cities Report 2026. The cost and relocation scenarios are illustrative, with no city data implied.