Windward wealth divide

Why is the poorer half of so many old industrial cities the eastern half?

London's East End and its West End. Paris's wealthy sixteenth arrondissement in the west and its working districts in the east. The same lopsidedness turns up in Manchester, Glasgow, Newcastle — and in far smaller towns nobody has ever written a novel about.

It is easy to say "the poor areas are in the east" as though it were a fact about compass directions. But nothing about east is intrinsically cheap. Something in these cities must have made one side systematically less pleasant to live on, and gone on doing it for long enough to sort people by what they could afford to avoid.

What could differ between two sides of the same town? Not the soil, usually. Not the distance to work. Ask instead what a nineteenth-century industrial city produced in enormous quantity that nobody wanted: smoke. Coal smoke from furnaces, foundries, gasworks, brickfields, and from every domestic grate besides — soot that blackened washing and buildings, sulphurous air, and, layered on top, the smell of tanneries, breweries, slaughterhouses and sewage.

Now: does smoke stay where it is made? No. It goes where the wind takes it. And in the mid-latitudes of north-west Europe the prevailing wind is from the west and south-west, most days of most years. So the ground east of a works received the plume; the ground west of it received relatively clean air, because the air had not yet passed over the works.

That is a real, persistent difference in the quality of a place. What follows from it? Consider who has choices. A mill hand takes housing near the mill and near cheap rents. A merchant or manufacturer can live wherever a carriage will reach and will pay for air his children can breathe. Both act sensibly, and the joint result is a sorting: bids for upwind land rise, bids for downwind land fall, and the two sides diverge in rent, in the quality of what gets built, and in who lives there.

Is this just an appealing story? It has been tested more carefully than most such claims. Economic historians studying English and Welsh towns have reconstructed where the coal-burning works stood, modelled how their plumes would have dispersed given each town's own wind and terrain, and asked whether modelled historical pollution predicts where the poor neighbourhoods ended up. It does — and importantly, the prediction is downwind, not east. Where a town's winds ran differently, the deprived quarter sits accordingly. That is what turns a curiosity into a mechanism: the pattern follows the causal variable rather than the compass.

Then the part that makes it matter now. British smoke control arrived with the Clean Air Act of 1956, and heavy industry has since left most of these cities entirely. The disamenity is gone. The divide is not. Why would an effect outlive its cause?

Because sorting builds things. Once one side is the poor side, it acquires the housing stock of a poor side — small, dense, cheaply built — and the schools, shops, transport and reputation that follow the residents. Each of those outlasts the smoke by generations and keeps recruiting the same population. The pollution wrote the initial condition, and the city has been copying it forward ever since. That is path dependence in a very literal sense.

This is the part where the neat version needs cutting down, because the pattern is statistical and its causes are tangled.

Wind is not the only thing that varies across a city, and several of the others point the same way, which makes them hard to separate. Rivers run one way, so wharves, docks and the trades that need water tend to sit downstream — often, though not always, the same side as downwind. Terrain matters: higher ground has cleaner, cooler air and better drainage, and the wealthy quarter of many cities is simply the uphill one. Railways and canals cut lines that industry clusters along. And landownership matters more than people expect — London's westward growth was already under way before the age of coal, pulled by the court at Westminster and shaped by large estates that could develop their land for a genteel market.

So the honest claim is narrower than the slogan. Prevailing winds are one well-evidenced contributor to which side of an industrial city became poor, with an effect large enough to detect across many towns and persistent enough to still show in today's deprivation statistics. They are not the only cause, and in any single city another factor may dominate. Counter-examples exist and do not overturn the general finding.

One more caution, since the mechanism is easy to run backwards. Sorting was not only about air. Downwind land was cheap, which is why industry and workers' housing went there together, so the causation runs in a loop rather than one way. And the modern residue is not merely historical: air quality today still tends to be worse in more deprived areas of the same cities, now from traffic rather than furnaces, which layers a live inequality on top of an inherited one.

• Whether the same gradient appears in cities whose prevailing winds run north-south, and how strongly.
• Why cleaning the air did not reverse the sorting, when the amenity difference that created it disappeared.

Every term the collection defines is gathered in the glossary.