ImmigrationDBVerified immigration data

Clean air barely shows up in the rent

By · Published · Updated

Question: in places with cleaner air, do people pay more rent?

Answer: slightly, and far too weakly to plan a move around. Across 392 German housing-market districts, Spearman's correlation between the average advertised rent and the annual PM2.5 mean is −0.12. Lower PM2.5 means cleaner air, so the negative sign points towards marginally higher rents where the air is better. The coefficient is small enough that air quality explains almost nothing about what a place costs.

The more useful result is the other one: 31 districts, containing 143 cities with a published air reading, sit below the German median on both measures — cheaper than typical and cleaner than typical at the same time.

What we measured

Each row is a place that publishes both an average rent per square metre per month and an annual PM2.5 mean in micrograms per cubic metre. Lower PM2.5 means cleaner air. The packet contains 1,836 paired cities across eight countries, of which 1,798 are German.

The rent measure needs naming precisely, because it is not the number a German tenant sees on an existing lease. It is the BBSR Wiedervermietungsmieten inserierter Wohnungen — the average advertised rent for re-letting an existing flat, compiled from property portals and newspaper listings, excluding new-build. It is what someone searching today is quoted. It excludes utilities and operating charges, so it is not a total housing bill either.

That series is published at district level, and this is the fact that shapes the whole analysis. The 1,797 German city rows that carry it hold only 399 distinct values: a median of four cities share one district figure, and one district covers 27. Treating those as 1,797 independent observations would count the same rent up to 27 times — the error statisticians call pseudo-replication, and it inflates apparent precision without adding a single new observation. So the correlation is computed once per district — 392 districts with both measures — using each district's advertised rent against the median PM2.5 of its member cities.

Germany is also the only country with enough coverage to report. The remaining 38 cities come from Finland, Spain, Greece, Hungary, Croatia, France and Austria, through seven different national publication systems. They appear as context and nothing more.

The observation periods are not uniform. German rents are dated 2024; German PM2.5 readings run from 2018 to 2024, most of them 2024. The comparison is cross-sectional: it asks how published values line up today, not how either has changed.

What the data shows

At district grain the relationship is weak and negative: Spearman −0.12, Pearson −0.11, across 392 districts. The median district advertises €9/m² per month and records 8 µg/m³ of PM2.5.

Computed the other way — one row per city, ignoring that the rent repeats — the coefficient is −0.14 across 1,797 cities. The two numbers are close, which is reassuring: the finding does not depend on the mistake. But only the district figure is honest about how much independent evidence stands behind it.

The scatter is the real answer to the question. Districts advertising the same rent sit several micrograms apart on annual PM2.5, and districts with near-identical air readings differ by several euros per square metre. There is a faint average tendency and no price schedule: nothing here lets a mover convert a micro­gram of PM2.5 into an expected rent.

The practical output is the bargain quadrant — districts strictly below the German median on both measures. Thirty-one of 392 qualify, and they contain 143 cities that publish their own air reading. The cheapest and cleanest of them advertise €6/m² per month at 6–7 µg/m³: the areas around Eschwege, Sangerhausen, Holzminden and Salzwedel. The Siegen–Wittgenstein belt — Lennestadt, Bad Berleburg, Freudenberg, Hilchenbach, Siegen and Kreuztal — sits in the same quadrant at €8/m².

These are not the cheapest places in Germany, nor the cleanest. They are places where both published values are favourable at once, which is a different and more useful screen.

Why a city page may show a different rent

Anyone who follows a link from this article to a city page will meet a second number, and the difference is not an error:

The Mietspiegel describes rents across the standing stock, including tenancies signed years ago under Germany's strong sitting-tenant protections. The advertised series describes what is on the market this month. In Berlin the gap between the two is more than a factor of two — which is itself the single most important housing fact in this dataset, and a reminder that "average rent in Berlin" is not one number.

Only 31 of the 1,836 paired cities have an official reference rent at all, which is why this analysis uses the advertised series: it is the only measure with the coverage to support a comparison.

Where it breaks down

Outside Germany the evidence runs out. Finland contributes 11 cities, Spain 10, Greece and Hungary five each, Croatia four, France two and Austria one. None comes close to a reportable sample, and each arrives through a different national system — official municipal reference tables in Spain, an approved rent observatory in France, a statistical office series in Finland, commercial market reports in Greece, Hungary and Croatia. Their rows can illustrate a single place; they cannot support a claim about a country.

Density also blocks any reading of this coefficient as the price of clean air. Across 388 districts with density data, advertised rent correlates with density at 0.48, while PM2.5 correlates with density at 0.11. Denser places are markedly more expensive and, weakly, more polluted — two associations that pull in the opposite direction to the headline result. Something other than air quality is doing the work, and this packet does not identify it.

The high-rent districts make the point concrete. Munich advertises €21/m² at 7.9 µg/m³; Berlin advertises €17/m² at 11.3 µg/m³. Both are expensive, and their air readings differ by more than three micrograms in the wrong direction for a clean-air premium. Whatever Berlin's rent is buying, it is not cleaner air than Munich's.

What this means if you are moving

Use the bargain quadrant as a filter and ignore the coefficient. Start from districts below the median on both rent and PM2.5, then check the things this article does not test: whether the jobs are there, how long the commute is, what the schools and clinics are like, and whether a specific flat exists at the advertised average.

Treat the advertised rent as the number you will be quoted, not the number you will end up paying across a tenancy, and check the city's Mietspiegel where one exists. Treat the annual PM2.5 mean as a background level, not a description of a particular street or a winter week.

For expensive places, ask what the premium buys. When high rent comes with high PM2.5, air quality plainly is not the explanation. When it comes with low PM2.5, the two values still do not show that one caused the other.

Limits

This is a correlation between two published averages, and it cannot establish causation in either direction.

The grain is a housing-market district, not a street or a building. Every city inside a district shares one advertised rent, and the PM2.5 figure used for the district is the median of its member cities' readings. Neighbourhood variation disappears at this resolution, in both variables — and a correlation measured over districts need not hold over the streets inside them. Nothing here describes what any individual household pays or breathes.

The rent measure is advertised re-letting rent, excluding utilities, operating charges and new-build. It is not total housing expenditure and not the rent of an existing tenancy. The published values are rounded to whole euros, so many districts tie — the ordering within a tie carries no information.

The years do not always match: rents are dated 2024, air readings run from 2018 to 2024. Germany dominates the evidence, and the seven other countries are too thin to report.

Method

The analysis starts from the paired packet of 1,836 cities and keeps every row that publishes both measures. Because the German rent series is compiled to district level, German rows are collapsed to their 392 distinct districts before any coefficient is computed: each district contributes its published advertised rent and the median PM2.5 of its member cities, and each district is weighted equally. Population is not used as a weight.

Spearman's rank correlation is the primary coefficient, since it is robust to the skewed rent distribution and to the tied whole-euro values. Pearson's coefficient is reported as a secondary check, and every coefficient carries its sample size. A country is reported separately only with at least 25 paired places, which leaves Germany alone.

The bargain quadrant uses strict inequalities against the German district medians of €9/m² and 8 µg/m³. Thirty-one districts qualify, containing 143 cities that publish an air reading. Density correlations use the 388 districts that also publish density, and are descriptive only.

Sources

Frequently asked questions

Does cleaner air cost more in rent?

Only slightly. Across 392 German housing-market districts, Spearman's correlation between the average advertised rent and annual PM2.5 is −0.12. Because lower PM2.5 means cleaner air, the direction points to a small premium, but the relationship is far too weak to predict what any individual place will cost.

Which German places combine cheaper rent with cleaner air?

Thirty-one of 392 districts sit below the German median on both measures, and those districts contain 143 cities with a published air reading. They include the areas around Eschwege, Sangerhausen, Holzminden, Salzwedel and Sondershausen, where advertised rents are €6/m² per month against a national district median of €9.

Why does this article show a different rent for Berlin than the Berlin city page?

They measure different things. This article uses BBSR advertised re-letting rents, which give Berlin €17/m² per month for 2024. The Berlin city page shows the official Mietspiegel reference rent of €7.71/m² per month, which covers existing tenancies rather than new listings. Advertised rents are what a mover is quoted; the Mietspiegel describes the standing stock.

Is this a European comparison?

No. Germany contributes 1,798 of the 1,836 paired cities, and it is the only country with enough coverage to report a coefficient. The 38 cities from seven other countries provide context and come from different national systems, so no cross-country rent comparison is made.

Why is the analysis reported for districts rather than cities?

Because that is the grain the rent data is published at. The BBSR advertised-rent series is compiled to district level, so the 1,797 German city rows carry only 399 distinct values — a median of four cities share each one. Correlating at city level counts the same rent up to 27 times and overstates the sample.

More research

More foreign residents, more crime? The answer changes at every border

Does a larger foreign population mean more recorded crime? It depends on the country: 0.06 in Germany and the US, 0.50 in Spain and Belgium, 0.59 in the Netherlands.

The cities gaining migrants are not the cities that are growing

Where are foreign populations growing fastest? Across 423 cities, growth is almost unrelated to whether the city is growing at all: 0.11 pooled, near zero by country.