Heat killed an estimated 2,700 people in England and Wales this summer. The housing stock is the variable.

West London recorded 35.1°C on the hottest day of the May heat wave, the kind of number that used to belong in a continental forecast. By the second week of July, researchers had assembled the human bill: roughly 2,700 deaths across England and Wales linked to heat in May and June, around 550 in the first wave and roughly 2,200 in the second, according to a Monday analysis cited by The Independent and Phys.org.

The point of the count is not the heat itself. Britain will see more of it, and the meteorological explanations are no longer in serious dispute. The point is what the heat does once it meets a housing stock built for a different climate, a rental market that treats cooling as a luxury amenity, and a public-health apparatus that does not classify heatwaves as a notifiable event. Read the mortality map and you are reading a property map.

Two waves, two shapes

The two heat events of late spring and early summer behaved differently, and the difference shows in the bodies. The May spike was short, sharp, and concentrated: a single day of 35.1°C in west London and a quick climb in the death counter, then a return to seasonal temperatures. The June wave was longer, slower, and deadlier. Researchers estimate roughly 2,200 heat-linked deaths in that second event, four times the May toll, according to Phys.org's summary of the study released Monday.

The asymmetry is familiar to anyone who tracks excess mortality in temperate countries. A one-day burst overwhelms the people most at the margin: elderly residents in top-floor flats, outdoor workers, people whose medication interacts badly with sudden heat. A multi-week siege reaches further, into households that survived the first wave by sleeping on cool floors and now face a second one with no reserves of money, sleep, or neighbours willing to check on them. The June curve, in other words, is a stress test of social infrastructure, not just of thermoregulation.

The building is the body

Britain does not have an air-conditioning problem in the sense that Riyadh or Phoenix has one. It has a retrofit problem. The country's housing stock is dominated by brick and stone construction from the late nineteenth and early twentieth centuries, heavy on thermal mass and light on the through-ventilation that older timber buildings provide. Double-glazing has spread since the 1990s, but double-glazing seals a building the wrong way for a heatwave: it cuts draughts in winter and traps still air in summer. Internal shutters are rare. External shading is rarer.

A second structural fact overlays the first. Cooling, in the British housing market, is overwhelmingly a private good. The country's split between owner-occupied and privately rented stock means that the household most likely to be elderly, on a fixed income, and living in a poorly insulated flat is also the household least likely to have a working air-conditioning unit installed by a landlord. Portable units, where they exist, draw current through wiring that was never specified for sustained summer load. Window units are often prohibited in lease terms. The cooling infrastructure is, in a literal sense, someone else's capital expense.

What a public-health frame changes

For two decades the British policy conversation about heat has run through the climate-adaptation lane: warmer summers, more extremes, prepare accordingly. That frame is correct and incomplete. The same excess mortality appears year after year, in heat domes that are not formally declared, in populations that overlap almost perfectly with the population in poorly rated Energy Performance Certificate (EPC) band D and below housing. The Independent's reporting on the latest study makes the point bluntly: heat is a public-health emergency that arrives without the protocol of one.

The lever that would move the curve fastest is not a forecast model. It is a building one. Insulating lofts, fitting external shading on south-facing façades, replacing single-pane glazing with ventilation-aware alternatives, and reframing cooling as a fit-and-thermal-comfort obligation on landlords rather than a tenant add-on: these are interventions that look like climate policy and behave like health policy. The cost is upfront and concentrated; the dividend is diffuse and decades long, which is exactly why the interventions tend to stall in five-year budget cycles.

A counter-reading, and the case for restraint

The sceptical case deserves airtime. The 2,700 figure is an estimate, not a count; it relies on the standard method of comparing observed deaths against a baseline drawn from non-heat weeks, and that method captures not only direct heatstroke but also the cardiovascular and renal deaths that accelerate in warm weather. Some fraction of those deaths would have happened within days or weeks regardless. Critics also note that the British climate, even in 2026, produces fewer cumulative heat-days than Mediterranean or North African peers at the same latitude, and that adaptation policy should be proportional to that gap rather than imported from hotter jurisdictions.

The defence of the alarm is straightforward: even after generous discounting, a non-trivial excess mortality that arrives every summer, on a known schedule, in a known housing stock, is not a natural-disaster story. It is a policy choice, made every time a retrofit programme is delayed, every time a cooling unit is treated as a tenant luxury, every time a heat-health alert system runs on voluntary action rather than statutory obligation. The sceptical reading sharpens the question; it does not dissolve it.

What to watch next

The next test is procedural. The UK Health Security Agency publishes its heat-period mortality reports with a lag of weeks, and the Office for National Statistics reconciles the figures later still. Watch for whether the post-mortem on these two waves produces a notifiable-event classification for heatwaves, the equivalent of what exists for cold snaps, and whether the autumn budget carries funding for the retrofit rate that the EPC trajectory implies. The deaths in May and June are already in the register. The next ones are not yet booked in.

This publication framed the 2,700-death estimate as a housing-infrastructure story rather than a climate-anomaly story, on the grounds that the meteorology is settled and the policy lever sits in the building stock.