Why young people cannot buy a home: housing costs 3.6 to 14 years of income, land drove 80% of the rise since 1950, and by 2040 demography will make homes cheaper only outside big cities

01 · OverviewHow unaffordable housing is in 2026: from 3.6 (Edmonton) to 14 (Hong Kong) years of income, depending on the city

The simplest measure of affordability is how many annual incomes a home costs. The Demographia International Housing Affordability report, published by Chapman University (US), divides the median house price by the median pre-tax household income. The 2026 edition covers 96 large markets in eight countries, using data for Q3 2025 10 M. The scale: 3 and under is affordable, 3.1–4 moderately unaffordable, 4.1–5 seriously unaffordable, 5.1–8.9 severely unaffordable, and 9 and above impossibly unaffordable 34 M.

By this measure none of the three large English-speaking countries is affordable. The median is 4.5 in the US, 5.2 in Britain and 5.4 in Canada; Canada is one of two countries in the report where affordability did not improve over the year 10 M. The dearest markets are Hong Kong (14.1), Sydney (14.0), San Jose (11.3), Adelaide (11.2) and Vancouver (10.8) 11 M. But the spread within a single country is wider than between countries: in Canada Vancouver is at 10.8, Toronto at 7.6 and Edmonton at 3.6, the third most affordable of the 96 markets 35,36 M.

In dollars it looks like this. By Redfin’s calculation, to buy a typical home in the US and spend no more than 30% of income on housing, a buyer in June 2026 needed to earn $109,796 a year, slightly below the record of $110,382 a year earlier. For a “starter” home the figure is $70,693 37 M. The median US household income for 2025 is $87,460 38 H. So for a typical home the median household would need to earn roughly a quarter more, while a starter home is affordable by the 30% rule: it requires $70,693, or about 81% of median income. Redfin’s calculation assumes a 15% down payment.

The second measure compares a country with its own past. The OECD calculates a price-to-income index per head of population and shows how far it is above or below the same country’s long-run average. In 2025 the largest excesses were in Canada (150% of the average), Australia (148%), the Netherlands (141%) and Portugal (133%). The US is at 112%, Britain at 116% and the OECD as a whole at 108%. Below their own averages are Italy, Japan, Germany, Finland and South Korea (48%) 12 H.

Over a quarter of a century there have been two periods of growth. The first ended with the crisis of 2007–2008: the price-to-income index in the US fell from 131 (2006) to 90 (2012), in Ireland from 163 to 79 and in Spain from 153 to 100 12 H. The second began with the zero interest rates of the 2010s and the pandemic, and ended with the rate rises of 2022. Since then New Zealand has returned almost to its 2015 level (the index fell from 136 in 2021 to 103 in 2025), Canada has retreated from 146 in 2022 to 131 in 2025, and real prices in Germany have fallen by 17% 12 H. In Portugal, Spain and the Netherlands, however, prices relative to incomes are rising again after a pause in 2023–2024: in Portugal the index rose from 147 to 167 in 2025.

For young people these averages hide the main point. The overall homeownership rate in the US did not change over the year (65.0%), but among households with a head under 35 it fell from 36.4% to 35.2%, a drop larger than the sampling error (±0.7 percentage points) 4 H. The median age of a first-time buyer is 40, of a repeat buyer 62 3 H. Older owners keep their homes, while the young buy later or not at all.

02 · MethodologyHow to measure housing affordability: four indicators, and each misses something

Since 1991 the price-to-income ratio has been indicator H1 in the system of housing indicators of the World Bank and the UN; the name “median multiple” was given to it later by Demographia 39 M. It is convenient for comparisons but ignores the interest rate, the deposit and taxes. So analysts use three more measures. Each answers a different question.

The 30% rule has a bureaucratic origin. In 1969 the Brooke Amendment capped rents in US public housing at 25% of a tenant’s income; in 1981 the cap was raised to 30% for budgetary reasons 40 M. Eurostat counts households that spend more than 40% of disposable income on housing as facing housing cost overburden. In 2024 the share among 15–29-year-olds in the EU was 9.7%, and among the whole population 8.2% 41 M.

The measures can give opposite verdicts. The best example is the interest rate. At Freddie Mac rates, a $1,000 monthly payment serviced a 30-year loan of about $72,000 in 1981 (rate 16.6%), $238,000 in 2021 (2.96%) and $146,000 in October 2026 (7.32%) (our calculation) 13 L. That is, over 40 years the same payment came to service a loan more than three times as large, and over five years one almost 40% smaller (−39%).

As a result, the price-to-income ratio in the US rose moderately over 2000–2025, while the payment-to-income ratio swung widely. By our calculation, the mortgage payment on a median new home (20% down payment) took about 29% of median household income in 2000 (rate 8.05%), 20% in 2020 (3.11%), 33% in 2023 (6.81%) and 29% in 2025 38,13 L. The sense of crisis deepened because in 2020–2021 homes were bought at record-low rates. The International Monetary Fund (IMF) estimates that the affordability index in the US fell from about 150 in 2021 to about 85 in 2024, and in Britain from 105 to about 70 (100 is the affordability threshold) 42 H.

For the young, the deposit is often as big a barrier as the payment. In Australia, according to the country’s National Housing Supply and Affordability Council, saving a deposit takes 11.2 years, and a new loan takes 45.9% of median household income 43 H. In the US a median household that saves 10% of its income accumulates a 20% down payment on a $409,000 home (the median, by an estimate from Harvard’s Joint Center for Housing Studies (JCHS)) in about 9.4 years if prices do not rise (our calculation) 38 L; if prices and incomes rise by 3% a year, it takes 11 years (calculator in section 11). In England 31% of first-time buyers received help from family or friends, and another 6% used an inheritance 44 H. The calculators in section 11 let you enter your own figures.

03 · HistoryFrom 1870: 80 years of flat prices, 70 years of growth, and land drove almost all of it

The longest comparable series of house prices was compiled by the economists Katharina Knoll, Moritz Schularick and Thomas Steger. They gathered annual data for 14 advanced economies from 1870 to 2012 5 H; the series was later extended to 2020 in the Jordà–Schularick–Taylor macrohistory database 6 M. The result: from the late 19th century to the middle of the 20th, real house prices in most countries did not change overall. They returned to their 1913 pre-war level, on average, only in the 1960s. After the Second World War real prices rose by more than 2% a year, and between 1900 and 2012 they roughly tripled, almost all of it after 1950 5 H.

One of the longest known series for a single street covers the houses on the Herengracht canal in Amsterdam from 1628 to 2008. By this measure the real price of a prestigious house in the centre of a rich trading city merely doubled over 350 years 45 M. Against that background, real prices that tripled in the 60 years after 1950 are a historical exception, not a norm that can be mechanically extended into the future.

Where does the growth come from? The price of a house can be split into two parts: the cost of construction and the value of the land beneath it. According to the calculation by Knoll, Schularick and Steger, from 1950 to 2012 real house prices rose 3.4 times, construction costs 1.6 times, and the price of residential land 7.3 times. Land explains about 80% of the rise 5 H. Until the 1960s real land prices in most countries barely changed: railways, trams and then cars kept widening the area from which people could commute to work 5 H.

Compared with wages, the picture is more complicated. We divided the house price index by the average wage index in the Jordà–Schularick–Taylor database. From 1900 to 1970 housing became cheaper relative to wages in most countries: in the US the index fell from 299 to 90, in the Netherlands from 332 to 110, in Norway from 551 to 101 (1990 = 100). After 1970 the paths diverged. In Britain, the Netherlands, Norway and Australia housing became roughly twice as expensive relative to wages. In the US (+31%) and France (+13%) the rise was moderate. In Germany and Japan housing became cheaper (our calculation) 6 L.

Two exceptional countries are instructive for the Ukrainian reader. In Germany real prices rose by about 4% a year from 1961 to the early 1980s, and then fell by about 0.8% a year until 2012 5 H. Fewer than half of households there own their home; the rest rent 46 M. In Japan, prices based on urban land values rose by almost 10% a year from 1949 to the late 1980s, and then fell by about 3% a year for two decades 5 H; by our calculation on JST data, the real index fell by 40% between 1991 and 2005 6 L. Both examples show that a long rise in prices is not a law of nature.

Mass homeownership appeared not because houses became cheaper in real terms, but because credit terms changed and incomes grew. In the early 1920s a buyer in the US could be asked for a deposit of up to 40% of the price and a loan of no more than 12 years; terms depended on the lender and the state 47 H. In 1933–1936 the state-owned HOLC refinanced about 1m mortgages, and from 1934 the Federal Housing Administration insured loans of 20–30 years with a deposit from 10% 47 H. From 1940 to 1960 the US homeownership rate rose from 43.6% to 61.9%, even though the real median value of an owner-occupied single-family home rose by 91% (our calculation from two census tables) 48 M. In England the homeownership rate rose from 25% in 1918 to 51% in 1971; in the 1950s and 1960s wages grew faster than house prices 49 H.

The timeline below is a short prehistory up to 2000: from the transport that made land cheaper to credit reforms and the first big price falls.

04 · Milestones2000–2026: from cheap credit to high rates, 63 events

The past 26 years fall into four periods. In 2000–2007 credit became cheaper and banks readily issued mortgages: US household mortgage debt rose to 73% of GDP (our calculation from Federal Reserve and BEA data) 50 L, and the homeownership rate among households headed by someone under 35 peaked at 43.1% (2004) 4 H. In 2008–2015 the crisis sharply cut prices in the US, Ireland and Spain, US construction fell to 1.9 new homes per 1,000 residents (our calculation from Census and BEA data) 14 L, and in England almost half of households headed by someone aged 25–34 (48.2%) ended up in private renting 44 H.

In 2016–2021 cheap credit (US mortgage rates of 2.96% to 4.54%) and the pandemic pushed prices to records in most rich countries. Since 2022 rates have risen: the average US mortgage rate went from 2.96% to 6.81% in two years 13 H. Prices fell in some countries (Germany, New Zealand, Canada, China), but affordability improved in few: in the US the loan payment on a median new home rose from 24% of median income in 2021 to 33% in 2023 (our calculation) 38,13 L. In Ukraine this period coincided with war, loss of population and the launch of the state mortgage scheme yeOselia.

05 · CausesWhy housing became more expensive: land near jobs, low rates, more credit and less building

Economists agree on the list of causes but disagree about their weight. Below are the factors in order of how strong the evidence for them is.

1. Constrained supply: cities do not let people build where the jobs are

Edward Glaeser and Joseph Gyourko compare the price of housing with the cost of building it and call the difference a “regulatory tax”. Their conclusion: in many US cities the gap is large, and the most likely explanation is restrictions on building, although the evidence is not conclusive 51 H. Chang-Tai Hsieh and Enrico Moretti estimated that restrictions on building in the most productive cities lowered US economic growth in 1964–2009 by 36% 52 H. But in 2026 another economist repeated their calculation, found errors in the code and obtained a GDP gain roughly 100 times smaller than the 3.7% that reforms were supposed to deliver 53 H. So the scale of the damage to the economy is in doubt; the mechanism itself is not.

US construction per head has halved since the peak of 1972: 9.5 new homes per 1,000 residents in 1972, 6.5 in 2005, 1.9 in 2011 and 4.8 in 2024 (our calculation from Census and BEA data) 14 L. Estimates of the US housing shortage range from 3.7m (Freddie Mac) to 5.5m homes (National Association of Realtors, 2020); the Congressional Research Service notes that there is no single agreed figure for the overall deficit 54 H. The number of small houses has fallen sharply: in 1976 more than 400,000 single-family homes of up to 130 m² were built (over 35% of new homes), in 2020 about 65,000 (7%), and in 2024 101,000 (10%) 55 H. Labour productivity in US construction is estimated to have fallen by about 1% a year since 1970 56 M.

2. Low rates: 40 years of decline

The real (inflation-adjusted) yield on 10-year US Treasury bonds fell from 6.9% in 1982 to about zero in 2012 and to −0.1% in 2020 57 M. When rates fall, the same payment services a larger loan, and buyers bid prices up. David Miles and Victoria Monro of the Bank of England calculated that the fall in the real risk-free rate in 1985–2018 more than fully explains the rise in house prices relative to incomes in Britain, and that a lasting rise in the rate of 1 percentage point lowers real prices by about 20% in the long run 58 H.

Rates and supply are not mutually exclusive. Rates explain why prices rose in almost all rich countries at once. Restrictions on building explain why prices in some cities rise far faster than in others: in San Francisco by 458% over 1950–2000, against 132% on average for the 50 largest metropolitan areas (section 08). The rise in rates in 2022 created a new obstacle: homeowners with cheap mortgages do not sell. In Q2 2026, 49% of US mortgages carried a rate below 4% 59 M.

3. More credit

Mortgage lending averaged, across advanced economies (10–14 countries, as the sample varies), 15% of GDP in 1950, 39% in 1990 and 76% in 2020 (our calculation from Jordà–Schularick–Taylor data) 6 L. In Australia it rose from 6% in 1970 to 91% in 2020. In the US household mortgage debt rose from 16% of GDP in 1952 to 73% in 2007 and, after the crisis, fell to 45% in 2025 (our calculation from Federal Reserve and BEA data) 50 L.

4. Smaller households and migration

Demand for housing is determined by the number of households, not of people. In the US there were 3.3 people per household in 1968 and 2.5 in 2025. Over that time the number of households rose by 122% and the population by 70% (our calculation) 60 L. Had households stayed the size they were in 1968, the US would need about 31m fewer homes. In Britain net migration has accounted for 99% of population growth since 2020 61 M; in Australia in 2025 it accounted for about 73% 62 M.

5. Investors, foreigners and short-term lets are smaller factors than they seem

After the foreign-buyer tax in Vancouver (August 2016), sales fell by 19% in a month but returned to their usual level in 2017; in 2021 foreigners accounted for 1.1% of housing transactions in the province 63 L. In areas with more foreign buyers, prices fell by about 6% more than elsewhere 64 M. According to Barron, Kung and Proserpio, a 1% increase in Airbnb listings raises rents by 0.018% and prices by 0.026% 65 H. The effect is noticeable in tourist areas but does not explain the rise in housing costs across whole countries.

Demand subsidies can raise prices

If the state helps buyers and more cannot be built, the help is capitalised into the price. Britain’s Help to Buy scheme, according to Carozzi, Hilber and Yu, raised prices in London, where building is difficult, and did not increase construction; near the border with Wales, where building is easier, it increased construction with no noticeable effect on prices 66 H. In Poland the Safe Credit 2% scheme ran from July 2023 to the start of January 2024; real prices rose by 9.4% from Q2 to Q4 2023 after falling by 10.5% over the previous five quarters 67,68 M. This is a coincidence in time, not a proven cause, but it fits the mechanism. For Ukraine’s yeOselia the risk is the same (section 10).

06 · RentingGeneration rent: young people buy later, rent for longer and live with their parents for longer

In the US, the homeownership rate among households headed by someone under 35 was 41.2% in 1982, peaked at 43.1% in 2004, fell to 34.5% in 2016 and stood at 37.0% in 2025 4 H. Households headed by 30–34-year-olds were hit harder: the rate fell from 57.1% in 1982 to 46.6% in 2025. Among 35–44-year-olds it fell from 69.3% in 2005 to 58.5% in 2015 and has recovered only in part. Meanwhile, among people aged 65 and over it rose from 74.4% to 78.6%, so the gap between the oldest and the youngest widened from 33 to 42 percentage points 4 H.

In England, “generation rent” peaked in 2013–2014: 48.2% of households headed by someone aged 25–34 rented privately (21.4% in 2003–2004), and 35.8% were homeowners. By 2024–2025 the homeownership rate had recovered to 47.7%, but the average age of a first-time buyer keeps rising: 31.4 years in 2003–2004 and 33.9 in 2024–2025 44 H. Private renters spend on average 40% of income on rent (including housing benefit), homeowners with a mortgage 20%. 58% of private renters say they expect to buy a home at some point 44 H. According to the Resolution Foundation (2016), about 63% of Britons born in 1946–1965 owned their own home by the age of 30 75 M.

Rent is rising faster than incomes. In the US, the consumer price index for rent rose by 128% over 2000–2024, 25% more than inflation, while real median household income rose by 16% (our calculation; the series are deflated with different price indices) 76,38 L. The number of flats renting for under $1,000 a month fell from 8.3m to 5.8m over 2014–2024 (the threshold is in nominal dollars). In 2024 a record 22.7m renter households (49%) spent more than 30% of income on housing, and more than 12m spent more than half 16 M. In 2025 the number of new households in the US fell by 16%; Harvard researchers link this to young people being unable to afford to rent on their own 16 M.

In Europe the gap between prices and rents is even wider. Over 2015–2025, house prices in the EU rose by 62%, while rents in consumer prices rose by 21% 15 H. In Hungary house prices rose 3.7-fold (provisional data), and in Portugal, Lithuania, Bulgaria and Czechia 2.5-fold. Finland is the only country in our sample where prices fell over the decade while rents rose, by 18%. Buying became far more expensive than renting, so buying a home grew steadily harder even for those who paid their rent without difficulty.

The consequence is that young people live with their parents for longer. In 2025 the average age at which young people left the parental home in the EU was 26.3 years: from 21.3 in Finland and 21.8 in Denmark to 30.2 in Spain and Italy and 31.5 in Croatia. Since 2008 it has risen by 1.8 years in Spain and by 2.4 years in Greece, while in Germany and France it has barely changed 77 H.

The share of 25–34-year-olds who live with their parents or are financially tied to the parental household was 30.1% in the EU in 2025. In Croatia it was 62.3%, in Greece 56.3%, in Poland 51.0%, in Italy 50.7% and in Spain 50.2%. In Germany it was 13.5%, in France 14.8%, in Finland 4.0% and in Denmark 3.9% 17 M. In Spain the share has risen from 33.2% in 2004, in Poland from 36.4% in 2005.

Is renting bad? Not necessarily. In Germany fewer than half of households own their home 46 M, and young people leave the parental home at 24 77 H. The problem is not renting itself. In the US, rent takes a growing share of income, and wealth is increasingly concentrated among older generations. US households headed by someone under 40 held 6.6% of all wealth in 2026, against 12.0% in 1989 (according to one secondary source) 78 L. Where renters hand the landlord a large deposit, young people carry the risk: in South Korea, 41,559 victims of jeonse deposit fraud had been recognised by September 2026, 75.9% of them under 40 79 M.

07 · WorldDeveloping countries and Central Europe: in fast-growing capitals, prices are already at Western levels, but incomes are not

The problem of affordable housing exists not only in rich countries, but it looks different there. According to UN-Habitat, 2.8bn people, about 40% of humanity, lack adequate housing, secure land rights or basic water and sanitation; more than 1.12bn live in slums or informal settlements, and about 300m are homeless 23 H. In many poor countries of Asia and Africa, mortgage debt is below 1% of GDP, against 83% in the Netherlands (World Bank data for 2013) 80 M. There, people build homes themselves, gradually and without credit, and a high homeownership rate does not mean that housing is affordable.

China: prices fell by a quarter, but from a very high level

According to the Bank for International Settlements (BIS) index, real house prices in China fell by 24.7% from the peak in Q3 2021 to Q1 2026, and nominal prices by 21.9% 18 H. In September 2026, new-home prices in 70 cities were on average about 3% lower than a year earlier, the 39th consecutive month of annual declines 81 M. Sales of new homes fell from more than 17trn yuan in 2021 to 8.39trn in 2025, the lowest since 2015 82 M.

But the fall began from a level that young people could not reach. In the first half of 2019, an apartment in Shenzhen cost 36.1 years of income, in Beijing 24.9 and in Shanghai 24.6 (an estimate by the research institute E-House) 19 M. A fall of a quarter does not make such housing affordable. The share of property in Chinese households’ assets, according to Goldman Sachs, fell from 67% in 2021 to 52% in 2026 83 M. According to a 2012 PIIE estimate, about 89% of urban Chinese households own their home, and a large part of this housing is a legacy of the privatisation of state flats: according to a 2026 IMF working paper, market-bought housing accounts for only about 40% of occupied urban dwellings 84 M. New buyers pay market prices, while older owners live in cheap privatised housing.

Turkey: nominal price growth is mostly inflation

From 2010 to Q1 2026, Turkey’s nominal house price index rose 46-fold, while the real index rose 2.2-fold. The real peak came in Q2 2023; since then, real prices have fallen by 17.9% 21 H. In 2025, nominal prices rose by 29%, while real prices fell by 1.4% 85 M. Prices in lira rise so fast that saving for a deposit is hard; our research has no data on the wages and borrowing of young Turks.

Central and Eastern Europe: some of the fastest price growth in the EU

In Q1 2026, real house prices compared with 2010 were 2.1 times higher in Hungary, 1.6 times in Czechia, 1.5 times in Bulgaria, 1.5 times in Croatia and 1.2 times in Poland 86 H. In Bulgaria, real prices rose by 37.6% over three years (from Q1 2023), and in Q2 2026 nominal prices rose by 15.5% year on year, second in the EU after Portugal. On 1 January 2026 the country adopted the euro 86,87 M. The average mortgage rate is 2.77% in Bulgaria, 9.08% in Hungary and 6.91% in Poland 22 H.

Each year Deloitte calculates how many average gross annual salaries a new 70 m² flat costs. In the 2026 edition, the most expensive city is Vlorë (Albania) at more than 21 salaries, followed by Prague at almost 17; the cheapest is Turin at 4.7 22 H. In the 2025 edition (2024 prices), Prague (15.0) was the third most expensive after Amsterdam (15.4) and Athens (15.3); Bratislava stood at 12.3, Budapest at 11.4, Warsaw at 9.7 and Krakow at 9.3 88 M.

India, Brazil, Mexico, South Africa, Vietnam

In Mumbai in 2025, the mortgage payment took 47% of household income, in Ahmedabad 18% 89 M. Real prices in India are 8% below their 2019 peak, in Brazil 23% below the 2013 peak, and in South Africa 27% below the 2007 peak 90 M. In these countries the constraint is not price but income and credit: by November 2024 Brazil’s Minha Casa Minha Vida programme had signed contracts for more than 1.1m homes, and the housing shortfall is estimated at 6–7m 91 L. In Vietnam, prices of new flats in Hanoi rose by 26.5% year on year in Q1 2026 92 M; according to CBRE, a 60 m² flat in Hanoi costs about 25 years of income (our calculation) 93 L. That is close to the 2019 levels of Beijing and Shanghai (24.6–24.9) and below Shenzhen (36.1); the methods differ, so the comparison is approximate.

08 · CitiesCapitals versus small towns: the gap does not widen every year, but it is the periphery that empties

Why do homes in some cities get more expensive twice as fast as in others? In their study of “superstar cities”, Joseph Gyourko, Christopher Mayer and Todd Sinai showed that real prices in San Francisco rose by 3.5% a year in 1950–2000. That was 2 percentage points faster than the average for the 50 largest metro areas. Over 50 years prices rose by 458%, against 132% on average for the 50 largest metro areas 94 H. Las Vegas grew to the size of San Francisco over the same period, but prices there rose more slowly than average, because the city allowed building. High prices arise where demand is large and building is not allowed.

In Britain the average home price in London was twice that of the North East of England in 1995, 4.2 times in 2016 and 3.4 times in 2025 95 H. Between 1995 and 2025 prices in London rose by 583% and in the North East by 303%. But since December 2016 London has become only 9% more expensive, and the North East 35%. Real prices in London are about 20% below the 2016 peak, according to a 2025 estimate by UBS 26 M. The gap in prices is larger than the gap in affordability: a home in London costs 10.6 years of earnings, in the North East 5.0, because wages in London are higher 2 H.

In the US the average price index of five expensive metro areas (New York, Los Angeles, San Francisco, Miami, Seattle) rose 5.3-fold from the end of 1995 to the end of 2025. For three Rust Belt metros (Detroit, Pittsburgh, Buffalo) the rise was 3.1-fold, and for the US as a whole 3.7-fold (our calculation: a simple unweighted average of metro-area indices) 96 L. But since 2019 cheaper metro areas have grown faster (+60% against +48%), and San Francisco has become only 15.8% more expensive, against 59.5% nationally 96 H. A metro area that is losing population can lose value for decades: in Detroit prices fell by 41% from the 2005 peak to 2011, and returned to their nominal peak only in 2020 96 H.

Remote work has changed the distribution of demand, but not in the way that was expected. Demand moved from the centres of large cities to their suburbs, and hardly at all to small towns and villages 97 H. The share of working days that Americans work from home was 27.1% in 2025 (about 7% in December 2019, according to a different survey) and about 26% in January–September 2026 98 H.

Japan: the clearest example

Japan has been ageing and shrinking since the 2010s. In October 2023 the country had 9m vacant homes — a record 13.8% of the housing stock, twice as many as in 1993. Of these, 3.9m are abandoned: not for sale, not for rent and not second homes 8 H. In the prefectures of Wakayama and Tokushima 21.2% of homes are vacant 27 M. Meanwhile official land prices in Tokyo prefecture rose by 6.5% in 2026 — the most among the 47 prefectures — and for the first time in 18 years Tokyo topped that list 99 M. The average price of a new flat in the 23 wards of Tokyo rose by 21.8% in 2025, to ¥136m 9 M.

South Korea shows the same thing even more sharply. The Seoul metropolitan area (Seoul, Incheon, Gyeonggi province) was home to 50.7% of the country’s population in 2023 100 M. The birth rate in Seoul is 0.63 children per woman, the lowest in South Korea, while official appraised flat prices in Seoul rose by 18.7% in 2026 101,102 M. In Busan in June 2025 there were 2,663 completed but unsold homes — the most since 2010 103 L.

In China, according to Oxford Economics, prices in smaller tier-three cities have fallen by no more than 30% from the 2021 peak, and in Beijing and Shanghai by less than 10% 20 M. But other indicators give a different order: in March 2026 secondary-market prices fell year on year by 7.4% in the four largest cities and by 6.4% in 35 smaller ones 104 M. In Germany in Q2 2026 flat prices in the seven largest cities fell year on year by 0.4%, in other large cities rose by 2.0%, and in densely populated rural districts fell by 1.8% 105 H. There is currently no stable capital-city premium in the German data.

The conclusion of this section is cautious. Capitals are not immune to falls: London, Paris (about 12% below its record at the end of 2020) and San Francisco have become cheaper or are flat 106 M. But where the population is shrinking, the periphery reacts first: first the number of vacant homes grows, then prices fall. Capitals go through cycles; the periphery of ageing countries goes through a prolonged decline.

Which cities win: the size threshold in the west and east of the EU and in Ukraine

There is no single size above which cities grow; it depends on the region. We calculated population change in EU cities in 2014–2024 by size class from Eurostat data (Urban Audit). In the western EU every class of city with 50,000 or more residents gained population on average. In the east only cities of 500,000 or more did: 94% of cities of 50,000–100,000 and 84% of cities of 100,000–250,000 lost residents (our calculation) 107 L.

Population change of cities within their administrative boundaries from 2014 to 2024, average by size class; in brackets, the share of cities that lost residents. Our calculation from Eurostat Urban Audit (urb_cpop1, 327 cities after cleaning series breaks).

In the west the threshold is lower still. In Germany no class of municipality shrank on average, not even villages of under 1,000 people (+0.7%); in Sweden municipalities grow from about 5,000 residents, in Portugal from about 10,000 and in Finland from about 20,000 (our calculation) 108,109 L. Italy is the exception: every class of municipality shrank. In the east a different pattern holds: cities of 20,000–250,000 lose residents within their boundaries while the villages and towns around them grow. In Poland 61 of 66 cities with county rights lost population in 2024; only Warsaw, Kraków and Rzeszów grew, while the counties around Poznań, Wrocław and Gdańsk gained 110 M. It is therefore better to measure a city together with its commuting zone: for such zones the growth threshold in the EU was about 250,000, and about 500,000 in the east (our calculation) 111 L.

Prices follow size in level but not in pace. In eight EU countries a flat in a city of under 100,000 costs on average 0.91 times the price in the country’s median city, 1.23 times in a city of 250,000–500,000 and 1.74 times in a city of over 1m (our calculation) 112 L. Since 2019–2022, however, mid-sized cities and suburbs have been getting dearer fastest: in Germany at the end of 2025 flats in large cities outside the seven biggest rose 5.5% year on year, against 2.6% in the seven biggest 113 M.

Where people move. Young people go to capitals: in Poland the number of 20–34-year-olds in the capital region fell by 9.9% in ten years, against 29.3% in the rest of the country (our calculation) 114 L. Foreign immigrants go to the biggest cities: the foreign-born make up 20.1% of residents of EU cities and 7.8% of rural residents 115 H. People moving from elsewhere in their own country, including from villages, more often settle in small and mid-sized cities and suburbs. In the eastern EU villages lose people both through departures (−2.0 per 1,000 residents a year) and natural decline (−3.6), whereas in the west net migration to rural areas is positive thanks to foreigners (our calculation) 116 L.

Ukraine resembles the eastern EU, with the war on top. Safety sets prices first and size second: a one-room flat costs $78,000 in Uzhhorod (115,000 people), $77,000 in Lviv and $75,000 in Kyiv, but $25,000 in Kharkiv (1.4m) 117 M. Regional centres of 100,000–400,000 in the west and centre are getting dearer fastest: over a year +20% in Khmelnytskyi, +17% in Vinnytsia, +13% in Ternopil and Uzhhorod, +4% in Kyiv 117 M. Jobs are even more concentrated than people: Kyiv and its oblast account for 40.6% of robota.ua vacancies 118 M. 46% of IDPs live in Dnipropetrovsk and Kharkiv oblasts, Kyiv city and Kyiv oblast, and 13% in the seven western oblasts; of those who returned home, 46% live in large cities and suburbs 119 H.

Our estimate for 2040 L: in the western EU almost all cities grow, except remote villages. In the eastern EU and Ukraine growth comes in centres of about 500,000 or more together with their suburbs, and in Ukraine also in safe regional centres of 100,000–400,000. Cities of 50,000–250,000 from which one cannot commute daily to a big city will lose residents, and prices there will stall or fall. What matters is not the number of residents but the commuting time to a large labour market: a town of 40,000 forty minutes from Kyiv grows; a district centre of 60,000 two hours from its regional capital does not.

09 · DemographyWhen and where homes will start to get cheaper because of low birth rates: on the periphery, already; in the capitals of countries with immigration, not before 2040

The idea that fewer children mean cheaper housing is not new. In 1988 the economists Gregory Mankiw and David Weil estimated that, if the historical relationship between prices and cohort size held, real house prices in the US would fall substantially over the next 20 years, because the generation of buyers born after the postwar baby boom is smaller 120 H. The forecast did not come true: prices rose, in our estimate because of lower interest rates, wider access to credit and smaller households. This is the main caution for all demographic forecasts of prices.

The effect exists, but it is small. Using data on 22 advanced economies for 1970–2009, the Bank for International Settlements (BIS) estimated that population ageing would slow real house-price growth in the US by about 0.8 percentage points a year over 40 years, and by more in Europe and Japan 25 H. For Japan, researchers put the demographic component of the change in land prices in 2012–2040 at −2.4% a year 121 H. For comparison: interest rates and supply move prices by 10–20% in a few years, as in 2022–2025 in Germany, New Zealand and Canada.

Who shrinks by 2040

According to the medium variant of the UN’s 2024 projection, the population aged 25–64 — the main buyers of homes — will shrink between 2025 and 2040 by 18.2% in South Korea, 17.7% in Italy, 15.2% in Japan, 14.5% in Germany, 13.2% in Spain, 10.9% in China, 9.8% in Poland and 9.0% in Ukraine. In countries with immigration it will grow: by 3.6% in Britain, 4.9% in the US, 5.7% in Canada and 8.3% in Australia 7 M.

Households will shrink later than people

Demand for housing is determined by the number of households, and these are getting smaller: there are more people living alone and more couples without children. So the number of households starts to fall 10–20 years later than the number of people. In Japan the population has been shrinking since the 2010s, but the number of households, according to the projection of the National Institute of Population and Social Security Research, will peak only in 2030 (57.7m) and fall to 52.6m in 2050; the share of one-person households will rise from 38% to 44% 24 M. In Tokyo the peak is expected in 2035 122 M. At the same time Japan keeps building: in 2023 there were 1.16 homes per household 8 H.

In countries with immigration, households keep growing. In the US, according to the projection of Harvard’s Joint Center for Housing Studies (JCHS), 8.6m households will be added in 2025–2035 and 5.1m in 2035–2045. With low immigration the figure is 6.9m in 2025–2035 (we could not confirm the estimate of 3.2m for 2035–2045) 123 M. In England, according to the projection of the Office for National Statistics (ONS), the number of households will rise from 23.5m in 2022 to 25.9m in 2032 and 28.9m in 2047; in 2022–2032 that is about 242,000 a year 124 H.

What this means by 2040

We see three groups of countries (our estimate based on the data above) L:

• Countries with immigration (US, Britain, Canada, Australia). The number of households, according to projections for the US and England, keeps rising through 2040 and beyond; in Canada and Australia the population aged 25–64 is growing. There will be no demographic fall in prices here before the 2040s; affordability will be determined by interest rates and building.
• Countries where the population is already shrinking (Japan, South Korea, Italy, Germany, Spain, China, Poland). There will be 10–18% fewer buyers aged 25–64 by 2040, and the number of households will start to fall from about 2030–2035. Pressure on prices will increase in the 2030s, but above all on the periphery: in Japan vacant homes are concentrated far from big cities, while prices in Tokyo are rising.
• Countries losing population through war and emigration (Ukraine). Here demography is already shifting demand: towards western cities and the capital, not away from them (section 10).

Even in Japan the data do not confirm scenarios of a sharp fall. The share of vacant homes rose from just 13.5% to 13.8% in 2013–2023 8 H; forecasts that one home in three will be vacant by the end of the 2030s are known only from press reports and contradict one another. In China the population has been shrinking for the fourth year in a row: in 2025, 7.92m children were born and 11.31m people died 125 M. The IMF estimates that baseline demand for new housing in China in 2024–2033 will be 35–55% lower than in the previous decade 126 H. This is an example of a large economy where demography is already reducing demand for new housing — but alongside the developers’ crisis, not instead of it.

10 · UkraineUkraine: over 90% of households own their home, but a one-room flat in Kyiv costs 5–7.5 years of salary, and only 0.4% of households have a mortgage

Ukraine’s situation is unlike the American or the western European one. In June 1992 state housing was transferred to its residents free of charge. As a result, over 90% of households own their home. According to a 2019 Cedos survey in cities of 100,000 or more residents, only 25% of homeowners bought their home; 32% inherited it, 26% received it from the state through the housing queue, and of those who bought, only 4% used a loan 28 M. By Housing Europe’s estimate, private renting makes up 4.8% of the housing stock and social housing 0.1%; there are about 1,100 social homes in the country 127 M. A young person without a parental flat enters a market with almost no affordable rent and no credit.

Prices and salaries

As of 1 July 2026 the median price of a one-room flat on Kyiv’s resale market is $69,900 (+8% over a year), or about $1,690 per m² 29 L. In August the median for a two-room flat was $107,000 and for a three-room flat $155,000 128 M. By LUN’s calculation from salaries in Work.ua job postings, a one-room flat in Kyiv costs about 7.5 average annual salaries; at the end of March 2026 the figure was 7.6, against 8 a year earlier 29,129 M. On the average Kyiv salary from the State Statistics Service (Derzhstat) (UAH 47,115 a month in May 2026), it comes to about 5 years (our calculation) 130 L. Lviv’s resale market ($1,700 per m²) has caught up with Kyiv’s ($1,690), and new-builds in Lviv ($1,500) are dearer than Kyiv’s ($1,330) 29 M.

In dollars, housing in Kyiv is cheaper than at the 2008 peak: then various agencies quoted $2,500–3,400 per m² 131 M. The methods differ, so an exact comparison is impossible. Salaries in dollars, meanwhile, have risen: the average Kyiv salary from $434 a month in 2010 to $954 in 2025, and in Ukraine as a whole from $283 to $623 (our calculation from Derzhstat data and the National Bank of Ukraine exchange rate) 132,133 L. By our calculation, a 50 m² flat cost 16.7 average Kyiv salaries in 2014 and 24 in 2015, and now costs about 6 L. So on the price-to-salary measure, affordability in Kyiv is now better than in 2014–2015 (we have no salary data for 2008). But young people can barely take advantage of this, because mortgages finance only about 3% of housing transactions.

Rent for a one-room flat in Kyiv in Q2 2026 is about UAH 18,000 a month, almost unchanged on a year earlier; in Lviv it is UAH 21,000 and in Uzhhorod UAH 22,500 29 M. This is roughly 38% of the average Kyiv salary (UAH 47,115 in May 2026) and 58% of the Ukrainian average (UAH 30,961) (our calculation) 130 L. In frontline Kharkiv and Zaporizhzhia rents rose over the year by 75% and 60%, from a very low base. By RDNA5’s estimate, half of internally displaced people (IDPs) who rent spend more than a third of their income on it, and 18% spend more than half 31 M.

yeOselia: the state mortgage scheme that has become nearly all mortgage lending

The programme launched in October 2022. A 3% rate applies to soldiers, security-service staff, medics, teachers and scientists (since December 2025, also to mobilised people); 7% applies to IDPs, veterans and, since August 2023, anyone without a home. The term is up to 20 years, and the deposit is from 20% (from 10% for people under 25) 134 M.

The discrepancy in the number of loans that the press often cites is explained by dates. The figure of 18,122 loans worth UAH 30.1bn is as of 14 July 2025; 20,250 as of 30 September; 21,860 worth UAH 37.4bn as of 11 December; and 22,600 worth UAH 38.9bn at the end of 2025 135 M. We did not find the figure of 17,451 loans worth UAH 28.8bn in primary sources; judging by its size, it corresponds to a date before mid-July 2025 (our assumption) L. As of August 2026 about 28,000 loans worth ≈UAH 49bn had been issued 30 M. As of 14 July 2025, 45% of borrowers were aged 26–35 and 33% were aged 36–45 135 M.

yeOselia accounts for 93–97% of all new mortgages in the country. The banks’ mortgage portfolio on 1 June 2026 was UAH 50bn (+35% over a year), but only about 42,000 households, or 0.4%, have a mortgage, and credit finances about 3% of housing transactions 30 M. Twelve of 59 banks issue new mortgages 136 M. The state funds the programme with bonds at 11–16% a year (the December 2025 issues at 11–14%) and lends at 3–7% 137 M; the 2026 budget is UAH 17.1bn 138 M. Since 9 February 2026 price caps have applied: in Kyiv, up to UAH 66,178 per m² 139 M. On 15 October 2026 the government approved a youth model: 7% for the first three years, 10% from the fourth to the tenth year, 12% thereafter, minus 2 percentage points for each child born or adopted 140 H.

Section 05 showed that where building is hard, subsidies to buyers pass into prices. The share of new-builds among yeOselia loans rose from 33% in 2024 to 60% in 2025 141 M, so the programme partly finances construction, not just the resale of existing flats. eVidnovlennia certificates add demand for the same limited stock of housing: by June 2026, 206,000 families had received compensation totalling UAH 103.9bn, and about 33,000 families had already bought homes with certificates 142 M.

War, construction and demography

By RDNA5’s estimate, by the end of 2025 about 14% of the housing stock had been damaged or destroyed, affecting 3m households; rebuilding housing over ten years requires about $90bn 31 M. In 2025 the country completed 9.55m m² of housing (−2.1% over a year), more than half of it detached houses; in Kyiv 1.19m m² (−15.6%), and in Kyiv region 1.99m m² 143 M.

Ukraine’s population, by World Bank estimates, fell from 52.2m in 1991 to 44.3m in 2021 and 37.7m in 2023 32 M. The Ptoukha Institute for Demography estimated the population in July 2024 at 35.8m, of which 31.1m in territory under government control 144 M. The fertility rate is about one child per woman 32 M. At the end of 2025, 5.2m refugees remained abroad and 3.7m IDPs inside the country 145 H. The World Bank’s forecast, based on UN data, is 35.3m in 2040 and 32.0m in 2050; the government’s demographic strategy projects 28.9m in 2041 (known from press reports) 32,33 L. The rural population, by World Bank estimates, fell from 17.0m in 1990 to 13.7m in 2020 (−19%); the values for 2022–2025 (11.5–12.6m) are modelled 32 M.

What this means for housing to 2040 (our estimate) L. Demand concentrates in Kyiv, Lviv and the western regional centres, where IDPs have moved and which are safer. Prices there will probably hold, and affordability will be determined by salaries and credit, not by demography. In villages and small towns, especially in the east and south, there will be more housing than people, and it will lose value. The main unknown is how many refugees return after the war: the gap of about 6m people between the forecasts (35.3m and 28.9m) depends on it.

11 · ForecastBy 2040: no official price-to-income forecasts exist, and neither the Fed, Fannie Mae nor the IMF expects a return to the low rates of the 2010s

Official bodies do not publish long-term house-price forecasts to 2040. There are forecasts of interest rates, household numbers, construction and prices for one to five years. A picture can be assembled from them.

Rates: the low rates of the 2010s will not return

In September 2026 members of the Fed’s Open Market Committee estimated the long-run rate at 3.2% (central range 3.0–3.6%), and the Fed raised its rate for the first time in three years, to 3.75–4.00% 146 H. In October 2026 Fannie Mae forecast the 30-year mortgage rate at 7.5% throughout 2027 and house-price growth of only 0.6% 147 M. The IMF notes that most forecasts do not envisage long-term rates returning to pre-pandemic levels 42 H. So rates of about 3%, on which the prices of the 2010s rested, are absent from the forecasts.

Construction: targets are being missed

The Canada Mortgage and Housing Corporation (CMHC) estimates that to restore 2019 affordability by 2035 the country must build 430,000–480,000 homes a year; on current conditions it will build about 245,000. In the baseline scenario, housing costs in 2035 will take 52.7% of household income against 40.3% in 2019; even if construction doubles, they will take 41.1% 148 M. In England, about 438,000 homes were added in the government’s first 26 months in office — about 29% of the target of 1.5m, although 43% of the time has elapsed 149 M. Australia, according to Australia’s National Housing Supply and Affordability Council, will reach its target of 1.2m homes one to one and a half years later than planned 43 H.

Prices: modest growth in nominal terms, and in real terms from −1% (London) to +6% (Britain) over five years

In June 2026 Savills forecast a 2% fall in prices in Britain in 2026 and growth of 18.5% over the five years to 2030; for London 10.6%, and for the North West 25% 150 M. Taking account of the inflation assumption used by Savills, this is about +6% in real terms for Britain and about −1% for London (our calculation) L. In September 2026 UBS saw a high bubble risk only in Zurich and Tokyo; real prices in 23 big cities rose by an average of 0.5% over the year, in Seoul by more than 10%, while in Vancouver and Toronto they fell by about 10% 151 M. In China the IMF estimates underlying demand for new housing in 2024–2033 at 35–55% below that of the previous decade 126 H. Climate-driven rises in insurance costs, by First Street’s estimate, could cut the value of US housing by $1.47trn by 2055; insurance will become dearer most in Miami, Jacksonville, Tampa, New Orleans and Sacramento 152 M.

Our arithmetic: how price-to-income could fall by 2040

Affordability can return without a price crash, through inflation and rising incomes. Take Britain’s median according to Demographia — 5.2 years of income — and calculate what it would be in 2040 under different assumptions (our calculation) L:

To get back to a level of 4 by 2040, nominal prices must grow 2 percentage points more slowly than incomes for 14 years. After the 1989 crash, real prices in Britain fell by 30% in seven years 6 M; after 2007, Ireland’s price-to-income index halved in five years 12 H. So such a scenario is possible, but historically it has come through a crisis, not smoothly.

Check it with your own numbers

Our assessment: what is likely by 2040

12 · LimitsHonesty and limitations

13 · QuestionsFrequently asked questions

14 · MoreRelated reading

15 · SourcesSources and methodology

Each number in the text leads to its source. Confidence labels: H — primary source (state statistics, central bank, academic paper), M — secondary source or one checked against several, L — estimate, one unconfirmed source or our calculation. Methodology: nine research strands, each checked by a separate independent reviewer (202 items: 148 confirmed, 36 corrected, 18 could not be verified; none refuted). Then two reviewers checked the finished text. The series for the charts were taken from the research files automatically, without manual retyping.

1. ONS — Housing affordability in England and Wales: 1997 to 2016. ons.gov.uk H
2. ONS — Housing affordability in England and Wales: 2025. ons.gov.uk H
3. National Association of Realtors — Profile of Home Buyers and Sellers 2025. nar.realtor H
4. Census Bureau — Housing Vacancy Survey, homeownership by age (Table 17 and Q2 2026 release). census.gov H
5. Knoll, Schularick, Steger — No Price Like Home: Global House Prices, 1870–2012, American Economic Review, 2017. dallasfed.org H
6. Jordà–Schularick–Taylor Macrohistory Database, release R6 (CC BY-NC-SA 4.0). macrohistory.net M
7. UN — World Population Prospects 2024, medium variant (via Our World in Data). ourworldindata.org M
8. Statistics Bureau of Japan — 2023 Housing and Land Survey. stat.go.jp H
9. Real Estate Economic Institute — new condo prices in Tokyo, 2025. nippon.com M
10. Demographia International Housing Affordability 2026 (Chapman University) — national medians, via Mortgage Professional. mpamag.com M
11. Demographia 2026 — ten least affordable markets, via Aussie. aussie.com.au M
12. OECD Analytical House Price Indicators. sdmx.oecd.org H
13. Freddie Mac Primary Mortgage Market Survey (via FRED). fred.stlouisfed.org H
14. Census Bureau / HUD — housing completions (via FRED). fred.stlouisfed.org H
15. Eurostat — house price index (prc_hpi_a) and HICP actual rentals (prc_hicp_aind). ec.europa.eu H
16. Harvard Joint Center for Housing Studies — State of the Nation’s Housing 2026 (via NLIHC). nlihc.org M
17. Eurostat — 25–34-year-olds living with their parents (ilc_lvps08). ec.europa.eu M
18. Bank for International Settlements — China residential property prices (via FRED). fred.stlouisfed.org H
19. E-House China Research Institute — price-to-income in Chinese cities, H1 2019. yicaiglobal.com M
20. Oxford Economics — Chinese house prices by city tier, September 2025. oxfordeconomics.com M
21. Bank for International Settlements — Turkey residential property prices (via FRED). fred.stlouisfed.org H
22. Deloitte Property Index 2026. deloitte.com H
23. UN-Habitat — 2025 Assembly (via UN News). news.un.org H
24. National Institute of Population and Social Security Research (IPSS) — household projections, 2024. ippjapan.org M
25. Takáts — Ageing and asset prices, BIS Working Paper 318, 2010. bis.org H
26. UBS Global Real Estate Bubble Index 2025. ubs.com M
27. 2023 Housing and Land Survey — vacancy by prefecture (via a secondary source). nippon.com M
28. Cedos — housing conditions survey, 2019. cedos.org.ua M
29. LUN Statistics — Q2 2026 report. lun.ua M
30. Association of Ukrainian Banks — mortgage market, 2026. aub.org.ua M
31. World Bank, EU, UN and Government of Ukraine — Fifth Rapid Damage and Needs Assessment (RDNA5), February 2026. cedos.org.ua M
32. World Bank — World Development Indicators and population projections. api.worldbank.org M
33. Ukraine’s Demographic Development Strategy to 2040 (via Ukrainska Pravda). life.pravda.com.ua L
34. Demographia — affordability ratings by median multiple. dailyhive.com M
35. Demographia 2026 — Toronto, Montreal, Calgary. dailyhive.com M
36. Demographia 2026 — Edmonton among the most affordable markets, via Daily Hive. dailyhive.com M
37. Redfin — income needed to afford a typical home, August 2026. redfin.com M
38. Census Bureau — median household income (via FRED). fred.stlouisfed.org H
39. Price-to-income ratio (World Bank / UNCHS Housing Indicators Programme) and the Demographia median multiple. en.wikipedia.org M
40. The Brooke Amendment (1969) and the 30% rule (1981). en.wikipedia.org M
41. Eurostat — housing cost overburden rate by age. ec.europa.eu M
42. IMF, Finance & Development — housing affordability, December 2024. imf.org H
43. National Housing Supply and Affordability Council (Australia) — 2026 report. nhsac.gov.au H
44. English Housing Survey 2024–25 — headline report and tables. assets.publishing.service.gov.uk H
45. Eichholtz — Herengracht house price index, Amsterdam (via IMF Finance & Development). imf.org M
46. Federal Agency for Civic Education (bpb) — homeownership in Germany (Mikrozensus 2022). bpb.de M
47. Federal Reserve Bank of Richmond — history of the US mortgage (HOLC, FHA). richmondfed.org H
48. Census Bureau — Historical Census of Housing tables (ownership and home values). www2.census.gov H
49. English Housing Survey 2013–14 — trends in tenure. assets.publishing.service.gov.uk H
50. Federal Reserve Financial Accounts (Z.1) — household mortgage debt. fred.stlouisfed.org H
51. Glaeser, Gyourko — The Economic Implications of Housing Supply, Journal of Economic Perspectives, 2018. aeaweb.org H
52. Hsieh, Moretti — Housing Constraints and Spatial Misallocation, AEJ: Macroeconomics, 2019. nber.org H
53. Greaney — comment on Hsieh and Moretti, AEJ: Macroeconomics, 2026. aeaweb.org H
54. Congressional Research Service IN12628 — US housing shortage estimates. congress.gov H
55. NAHB / Census — construction of small single-family homes. congress.gov H
56. Goolsbee, Syverson — US construction productivity (NBER, 2023; via Chicago Booth Review). nber.org M
57. Federal Reserve Bank of Cleveland — 10-year real interest rate. fred.stlouisfed.org M
58. Miles, Monro — UK house prices and three decades of decline in the risk-free real interest rate, Bank of England SWP 837. bankofengland.co.uk H
59. Wolf Street — share of mortgages below 4%, Q2 2026. wolfstreet.com M
60. FRED — US population and households (our calculation). fred.stlouisfed.org L
61. ONS — contribution of migration to UK population growth. migrationobservatory.ox.ac.uk M
62. Australian Bureau of Statistics — net overseas migration and population growth, 2025. macrobusiness.com.au M
63. BCREA — impact of the Vancouver foreign buyers tax. web.bcrea.bc.ca L
64. UBC Sauder — prices in foreign-buyer neighbourhoods after the 2016 tax. org-www.sauder.ubc.ca M
65. Barron, Kung, Proserpio — The Effect of Home-Sharing on House Prices and Rents, Marketing Science, 2021. pubsonline.informs.org H
66. Carozzi, Hilber, Yu — Help to Buy and house prices, Journal of Urban Economics, 2024. researchonline.lse.ac.uk H
67. Bank for International Settlements — Poland real residential property prices (via FRED). fred.stlouisfed.org H
68. Bank.pl — dates of the Safe Credit 2% programme. bank.pl M
69. Greenaway-McGrevy, Phillips, So — Auckland upzoning (working papers 2023–2024). motu-www.motu.org.nz M
70. Redfin and Zillow — rents and permits in Austin. texastribune.org M
71. TU Wien (2025) citing Statistik Wien — municipal housing in Vienna. repositum.tuwien.at M
72. SingStat — Population Trends 2026 (households in HDB dwellings). singstat.gov.sg M
73. Federal Constitutional Court of Germany — ruling on the Berlin rent cap, 2021. goerg.de M
74. Jofre-Monseny, Martínez-Mazza, Segú — Catalonia’s 2020 rent regulation. diposit.ub.edu M
75. Resolution Foundation — homeownership by generation, 2016. cityam.com M
76. Bureau of Labor Statistics — CPI rent of primary residence (via FRED). fred.stlouisfed.org H
77. Eurostat — estimated average age of young people leaving the parental household (yth_demo_030). ec.europa.eu H
78. Federal Reserve Distributional Financial Accounts — wealth by age (via a secondary source). eco3min.fr L
79. Korean Ministry of Land — recognised jeonse fraud victims (via a secondary source). koreajoongangdaily.com M
80. World Bank — mortgage debt to GDP by country income (blog, 2013). blogs.worldbank.org M
81. National Bureau of Statistics of China — 70-city new home prices, September 2026 (via Trading Economics). tradingeconomics.com M
82. National Bureau of Statistics of China via Caixin — 2025 new home sales. caixinglobal.com M
83. Goldman Sachs — property share of Chinese household assets (via a secondary source). scmp.com M
84. IMF — working paper on China’s urban housing segments, 2026. imf.org M
85. Global Property Guide — Turkish house prices in 2025. globalpropertyguide.com M
86. Bank for International Settlements — Central European real residential property prices (via FRED). fred.stlouisfed.org H
87. Eurostat via BTA — Bulgarian house prices, Q2 2026. bta.bg M
88. Deloitte Property Index 2025 (via Notes from Poland and Expats.cz). notesfrompoland.com M
89. Knight Frank — Affordability Index (India), 2025. business-standard.com M
90. Bank for International Settlements — real residential property prices for India, Brazil, Mexico and South Africa (via FRED). fred.stlouisfed.org M
91. Brazilian Ministry of Cities — Minha Casa Minha Vida programme (via Mercado & Consumo). mercadoeconsumo.com.br L
92. Savills Vietnam — new apartment prices, Q1 2026. vietcetera.com M
93. CBRE — housing affordability in Hanoi and Ho Chi Minh City, 2025 (via Dtinews). dtinews.dantri.com.vn M
94. Gyourko, Mayer, Sinai — Superstar Cities, AEJ: Economic Policy, 2013. nber.org H
95. HM Land Registry — UK House Price Index. landregistry.data.gov.uk H
96. FHFA All-Transactions House Price Index by metro (via FRED). fred.stlouisfed.org M
97. Ramani, Bloom — The Donut Effect of COVID-19 on Cities, NBER WP 28876. nber.org H
98. WFH Research (Barrero, Bloom, Davis) — Survey of Working Arrangements and Attitudes. fred.stlouisfed.org H
99. Japan’s Ministry of Land — official land prices 2026 (via Jiji / Adnkronos). english.adnkronos.com M
100. Statistics Korea — Seoul capital region population share (via Asia News Network). asianews.network M
101. Statistics Korea — 2025 fertility (via Reuters). freemalaysiatoday.com M
102. Korean Ministry of Land — official apartment prices 2026 (via a secondary source). koreajoongangdaily.com M
103. Unsold housing in Busan, June 2025 (via a secondary source). breezeinflow.com L
104. National Bureau of Statistics of China — prices by city tier, March 2026 (via a secondary source). aa.com.tr M
105. Destatis — house price index by district type, Q2 2026. destatis.de H
106. Notaires du Grand Paris — Paris flat prices (via secondary sources). notairesdugrandparis.fr M
107. Eurostat Urban Audit — city population (urb_cpop1), our calculation by size class. ec.europa.eu M
108. Eurostat GISCO — population of German municipalities in 2014 and 2024, our calculation. gisco-services.ec.europa.eu M
109. Eurostat GISCO — population of municipalities in Portugal, Sweden and Finland, our calculation. gisco-services.ec.europa.eu M
110. GUS (Poland) — city populations on 1 January 2025 (via Polsat News). polsatnews.pl M
111. Eurostat Urban Audit — functional urban area population (urb_lpop1), our calculation. ec.europa.eu M
112. Eurostat Urban Audit — flat prices by city (urb_clivcon), our calculation. ec.europa.eu M
113. IW Wohnindex, Q4 2025. bca.de M
114. Eurostat — population by age in NUTS2 regions (demo_r_pjangroup), our calculation. ec.europa.eu M
115. Eurostat — foreign-born population by degree of urbanisation (lfst_r_pgaumd). ec.europa.eu H
116. Eurostat — migration and natural change in NUTS3 regions (demo_r_gind3), our calculation. ec.europa.eu M
117. OLX Real Estate — median one-room flat prices, August 2026 (via Omega Lviv). omega.lviv.ua M
118. robota.ua — vacancies by region, Q2 2026 (via Interfax-Ukraine). interfax.com.ua M
119. IOM DTM — General Population Survey, rounds 24–25 (2026). dtm.iom.int H
120. Mankiw, Weil — The Baby Boom, the Baby Bust, and the Housing Market, NBER WP 2794, 1988. nber.org H
121. Saita, Shimizu, Watanabe — Aging and real estate prices, 2016. emerald.com H
122. Tokyo Metropolitan Government — household projection. toukei.metro.tokyo.lg.jp M
123. Harvard JCHS — household projections 2025–2045. jchs.harvard.edu M
124. ONS — household projections for England (2022-based). ons.gov.uk H
125. National Bureau of Statistics of China — 2025 births and population (via Caixin). caixinglobal.com M
126. IMF — estimate of fundamental housing demand in China, 2024. imf.org H
127. Housing Europe — Ukraine’s housing system, 2025. housingeurope.eu M
128. Glavcom — Kyiv flat prices, August 2026. glavcom.ua M
129. The Page — Kyiv one-room flat price in salaries, March 2026 (LUN data). thepage.ua L
130. State Statistics Service of Ukraine — average wage, May 2026. interfax.com.ua H
131. Ekonomichna Pravda — Kyiv housing prices in 2008. epravda.com.ua M
132. State Statistics Service of Ukraine — average wages in Ukraine and Kyiv (via Minfin). index.minfin.com.ua M
133. National Bank of Ukraine — annual average official exchange rate (via the World Bank). api.worldbank.org H
134. yeOselia programme terms. interfax.com.ua M
135. Ukrfinzhytlo and the Ministry of Economy via Interfax-Ukraine — yeOselia loan counts. interfax.com.ua M
136. Banks issuing mortgages, 2025 (via a secondary source). aub.org.ua M
137. Ukrfinzhytlo — bond funding of yeOselia. interfax.com.ua M
138. State budget 2026 — yeOselia allocation. interfax.com.ua M
139. Cabinet of Ministers — yeOselia price caps from 9 February 2026. epravda.com.ua M
140. Cabinet of Ministers — yeOselia youth model, 15 October 2026. kmu.gov.ua H
141. Share of new builds in yeOselia loans. interfax.com.ua M
142. Ministry for Development — eVidnovlennia programme, June 2026. borgexpert.com M
143. State Statistics Service of Ukraine — housing commissioned in 2025 (via Interfax-Ukraine). minfin.com.ua M
144. Ptoukha Institute for Demography — population estimate, July 2024 (via Fakty). fakty.com.ua M
145. UNHCR — refugees and internally displaced people from Ukraine. api.unhcr.org H
146. Federal Reserve — Summary of Economic Projections, September 2026. federalreserve.gov H
147. Fannie Mae Economic and Strategic Research — forecast, October 2026 (via MPA). mpamag.com M
148. CMHC — Housing Supply Gaps, June 2025 (via BNN Bloomberg). investmentexecutive.com M
149. Full Fact — 1.5 million homes tracker. fullfact.org M
150. Savills — Revised Mainstream House Price Forecasts, June 2026. pdf.savills.com M
151. UBS Global Real Estate Bubble Index 2026. ubs.com M
152. First Street — climate insurance impact on home values, 2025. impact.wharton.upenn.edu M

Read this briefing in Ukrainian →