What young tenants in Spain pay, and why their overburden fell

Abstract

Two sets of figures about young people and rent circulate in Spain.

• The CJE's. The Consejo de la Juventud de España (CJE) reports that a young wage earner would need 98.7% of their salary to rent a home alone, and 33.6% to rent a room in a shared flat.
• Eurostat's. The share of market-rent tenants who spend more than 40% of their disposable income on housing fell from 40.9% in 2021 to 26.8% in 2025. Meanwhile 68.7% of people aged 18–34 lived with their parents, the highest share Eurostat has published for Spain.

We used INE's Living Conditions Survey (ECV) microdata to measure what young tenants pay, to test whether the falling overburden is a selection effect, and to set the figures side by side.

What young tenants pay. In 2025, among people aged 18–34 who had left the parental home and rented at market price (n = 1,048 people in 783 households):

• the median rent paid was 630 € a month, 24.7% of household disposable income;
• the median housing cost burden was 30.8%;
• 31.5% were overburdened (95% interval 26.8–36.1). At ages 18–29 the figure was 34.0% (n = 438);
• 63.0% of those living alone were overburdened.

Our method reproduces Eurostat's overburden by tenure exactly in 76 of 84 tenure-year cells from 2008 to 2025, and within 0.3 points in all of them.

Selection on leaving home is real. It cannot explain the national fall; for young tenants it cannot be ruled out.

• Who leaves. In the panel for 2022–2025, young people in the top quarter of personal income left the parental home four times as often as those without income (23.3% against 5.8% a year).
• All tenants. Overburden fell by similar amounts at every age: by 13.3 points at 35–49 and 11.2 among young tenants. It also fell within households that rented in consecutive years, from 46.5% to 39.1% between 2022 and 2023.
• The bound. Fewer young people leaving home can account for at most about a tenth of the fall for all tenants (10.1%, or 23.1% if every extra stayer had rented). For young tenants the same bound exceeds their whole fall, so it rules nothing out.

Tenant households' incomes grew faster than their housing costs. Using the income growth observed within the same households, the overburden would have been 34.8% in 2025 had incomes grown only as fast as costs.

The CJE's figures answer a different question. They ask whether a young wage earner could rent at today's asking prices on one salary. On that question the microdata agree: only 0.6% of people aged 16–29 have an income at which the asking rent for a flat would take 40% or less. The measured burden asks what those who already rent pay.

• For tenants aged 16–29, the median rent paid is 596 €, about half the asking rent of 1,176 €. It is 24.4% of their household's median income.
• The difference has three parts. Measured against the ECV's median young earner, the parts are asking against paid rent (36.5%), the tenants' higher own income (20.8%), and sharing a household (42.7%). The split depends on the income benchmark.
• The CJE's own survey figure. Its ECV-based figure of 780 € of housing costs is consistent with ours (770–785 €). Its 48.9% overburden is not reproduced.

1. Introduction

In May 2026 the CJE's Observatorio de Emancipación reported:

• «Una persona joven asalariada tiene que destinar el 98,7 % de su salario para pagar un alquiler en solitario, incluso compartir piso (opción que ya no es una elección sino una necesidad) requiere el 33,6 % de nuestro salario superando el umbral de cualquier límite de asequibilidad que fijan los organismos internacionales (que por cierto fijan para el acceso a una vivienda, no a una habitación en un piso compartido).»;
• that «La tasa de emancipación juvenil en 2025 se situó en el 14,5%, la tasa más baja registrada».

The 98.7% figure was widely reported.

Eurostat's official indicator, built from the same survey that the CJE now uses, tells another story. The housing cost overburden rate of Spanish tenants paying market rent fell from 40.9% in 2021 to 26.8% in 2025. A reader may take one figure as wrong. Neither is: they measure different things.

This study does three things.

1. It measures the burden of young tenants from the microdata, by tenure, year, age band and income.
2. It tests the most natural reconciliation. If only those who can pay leave home, the tenants we observe are a selected group, and their burden can fall while access gets harder.
3. It shows, step by step, how the CJE's figures and the measured burden relate.

2. What each figure measures

The CJE's 98.7% and 33.6%. Each is a ratio of two inputs.

• The numerator is an asking rent. For the 98.7%: «Precio de oferta de las viviendas libres en alquiler que recogen los informes que publica el portal Idealista.com, aplicando una superficie media de 80 metros cuadrados construidos»; in 2025 that was 1,176 € a month. For the 33.6%: the median asking rent for a room in a shared flat, 400 €.
• The denominator is the median net salary of a young wage earner aged 16–29: 14,292.22 € a year, or 1,191.02 € a month. The CJE builds it from INE's quarterly labour cost survey, the wage structure by age and sex, and «el modelo de conversión de salarios brutos a netos de la ECV».
• What it measures. The CJE's earlier methodological note describes the indicator as the relation between a young person's purchasing power «y el importe mensual de un alquiler de mercado». It measures the entry cost of renting, a flat or a room, at current asking prices on one median salary. The CJE publishes the same ratio for a young household's income: 45.3%.

Eurostat's overburden rate. It is the share of people living in households whose housing costs exceed 40% of disposable income, both net of housing allowances. Housing costs include rent or mortgage interest, utilities and community fees.

• Rent paid, not asked. Costs are those currently paid, under whatever contract the household has, often signed years earlier.
• A household's income, not a salary. Income is the household's disposable income in the previous calendar year, from all members and all sources, excluding imputed rent.
• People already housed. The population is people in private households. Someone who cannot afford to leave the parental home appears only as a member of the parents' household.

The CJE also publishes the measured version. For renter households with a young member aged 16–29, it reports «780 € mensuales … que representaba el 30 % de los ingresos netos del hogar joven». It also reports that 48.9% spent more than 40% of their income on housing.

3. Prior work

What we searched. On 4 October 2026 we ran 21 searches on DuckDuckGo, Crossref and OpenAlex. A general web-search tool was unavailable: its session quota was used up. Every search is in data/prior_work_search.csv, and the documents read are in data/prior_work.csv.

The selection argument and the reasons for the fall have been stated before.

• The CJE. Manuel Mejías Leiva, a CJE researcher, told Newtral on 30 September 2026 that «los datos de la Encuesta de Condiciones de Vida no miden cuánto cuesta acceder hoy a una vivienda, sino cuánto pesa el gasto en vivienda sobre los ingresos de los hogares que ya están viviendo en ella». That is the distinction behind our second and third questions, stated by the CJE itself.
• elDiario.es (Ainhoa Díez, Yuly Jara and Raúl Sánchez, 29 September 2026), from ECV microdata.
  • It found that the share of tenants paying more than 30% of income in rent fell from 36.4% to 26.2% in ten years.
  • It explained this mainly by income growth at the bottom: the minimum wage, falling unemployment, the labour reform and the minimum income scheme.
  • It added the «ahogados invisibles»: young people who cannot leave home and tenants who return.
• DatosRTVE (2 October 2026), with the same microdata.
  • It built a market rent from Idealista's asking price per m² and a mean rented-dwelling area of 78.6 m² from SERPAVI. That rent was 1.7 times the rent households declare.
  • It used household income for the emancipated and personal income for those at home.
  • It noted that «solo compartiendo gastos con la pareja o compañeros se puede costear un alquiler», and warned that effort indicators «solo describen a quien ya vive por su cuenta».

This is the closest prior work. It does not decompose the CJE's figure or bound selection.

• FEDEA (Conde-Ruiz and Pinto, September 2026).
  • It computes the overburden of market tenants from the ECV (26.8% in 2025), poverty after housing costs, and their concentration in young and foreign-born households.
  • It already offers a composition explanation for the fall: «Parte de esta mejora coincide con el fuerte crecimiento en número de hogares y personas de este régimen, que introduce mayor heterogeneidad interna».

Burden from the ECV and other surveys is not new either.

• FEDEA (Pinto, February 2026) covers burden by tenure and household type, 2019–2024.
• Banco de España Occasional Paper 2432 (2024) documents the rise of renting among young households and Spain's high rent effort in the EU. It links that effort to delayed emancipation.
• Funcas measures young renters' effort from the household budget survey (EPF).

Context. Eurofound (2025) concludes that the housing crisis impedes household formation. Echaves-García and Martínez del Olmo (2021) relate falling emancipation to rental effort. Colom Andrés and Molés Machí (2026) decompose changes in young Spaniards' living arrangements by education and employment.

What is new here. Narrowly:

• a bound on one selection channel, an age comparison, a panel of continuing tenants, and the income gradient of leaving home across four panels;
• a step-by-step comparison with the CJE's 98.7% under two income benchmarks;
• a check of the CJE's own ECV figures;
• a compilation, from INE's releases, of the ECV's fieldwork changes and their size.

4. Data and method

The full method is in METHOD.md.

Data.

• ECV cross-sectional microdata 2004–2025, with income from 2008 (base 2013).
• ECV longitudinal files covering 2013–2025.
• EPA microdata for the second halves of 2023–2025.
• Eurostat's published tables and INE's rental price index (IPVA).
• The CJE's 2025 report and its 2023 and 2024 second-semester reports.

Definitions.

• Population: all members of private households, including those temporarily absent, as in Eurostat's indicators.
• Burden: (12 × housing cost − allowances) / (disposable income − allowances). Overburden means above 40%. A household with no positive income counts as overburdened.
• Young: by age at interview.
• Emancipated: neither father nor mother in the household.
• "Young tenants": emancipated people aged 18–34 (or 18–29, or 16–29 to match the CJE) in dwellings rented at market price.
• Weights and intervals: INE's cross-sectional person weights. Intervals are normal intervals from a household bootstrap within regions (200 replicates).
• Small cells: groups of fewer than 10 households publish no medians or rates.

Replication. With this population our overburden by tenure (including the total) matches Eurostat's exactly in 76 of 84 tenure-year cells for 2008–2025, and within 0.3 points in all. For market-rent tenants it is 40.9% in 2021 and 26.8% in 2025, as Eurostat. By age, with age at the end of the income year, 52 of 54 cells match exactly. The share of 18–34s living with a parent matches ilc_lvps08 in 2021 (64.5%) and 2025 (68.7%). The scout figures quoted in the introduction (40.9%, 26.8% and 68.7%) were confirmed against Eurostat's API.

5. Results

5.1 What young tenants pay

n: emancipated people aged 18–34 renting at market price. Income: household disposable income of the previous year.

2025.

• Main figures. Young tenants aged 18–34 paid a median 630 € a month, 24.7% of household income. Their median housing cost burden was 30.8%. 31.5% were overburdened (95% interval 26.8–36.1); at 18–29, 34.0% (n = 438).
• Other definitions. With age measured at the end of the income year, the overburdened share is 31.4%. Counting households whose responsible person is 18–34 gives 41.4% (574 households): young people who head their own household are more exposed.
• By household type. One in four young tenants lives alone, and 63.0% of them are overburdened (n = 227). Couples without children: 20.2%.
• By national income quintile. 65.5% of young tenants in the poorest fifth are overburdened, and 0.4% in the richest.

Change since 2021. Overburden among young tenants fell from 42.7% (n = 852) to 31.5% between the 2021 and 2025 surveys. The interval of the change is wide: −20.9 to −1.5 points.

5.2 Why the overburden fell: five tests of selection

The hypothesis. The share of 18–34s living with a parent rose from 64.5% (2021) to 68.7% (2025). If those who stay are those who could not have paid, the tenants who remain are better off, and their measured burden falls even as access gets harder.

1. Leaving home selects on income.

Quarters of positive personal net income among all people aged 18–34 in each year.

• Who stopped leaving. Between 2021 and 2025, emancipation fell most among young people with low or middling incomes: from 19.5% to 9.6% in the lowest quarter, and from 51.3% to 43.4% in the third. In the top quarter it fell from 60.0% to 56.6%.
• Young tenants moved up. The two upper quarters supplied 53.3% of young tenants in 2021 and 59.5% in 2025.
• The panel. The longitudinal files observe young people at home in one year and outside it, or recorded as moved, the next.
  • In 2022–2025 the top quarter of personal income left at 23.3% a year, against 5.8% for those with no income (1,150 departures among 10,937 person-years at risk): about four times as often.
  • In 2013–2016 the ratio was 2.7 (17.2% against 6.3%), and in 2016–2019 3.2 (17.7% against 5.6%).
  • At ages 25–34 the gradient steepened, from 17.2% against 11.8% (2013–2016) to 25.3% against 8.2% (2022–2025). At 18–24 it did not (17.3% against 4.6%, then 14.6% against 5.2%).
  • The 2019–2022 panel, which spans the pandemic and codes movers differently, shows a flat gradient (7.3% against 5.2%).
  • Levels are not comparable across panels; ratios are.

2. Tenants of every age saw similar falls, which is a comparison, not a placebo.

• What the comparison shows. Tenants aged 35–64 are not subject to selection into leaving home, and their overburden fell as much. So that selection cannot be what drove the national figure. Of the 14.1-point fall for all market tenants, a shift in the age mix accounts for 0.3 points, and young emancipated tenants' own fall for 1.7.
• What it does not show. It is not a placebo for the young tenants' fall.
  • Their housing costs rose 27.1% from 2021 to 2025, against 14.7% for tenants aged 35–64 (rents: 27.9% against 20.0%). Their overburden fell about as much only because their household incomes rose slightly more (47.1% against 46.3%), which is what selection would produce.
  • Selection into renting also operates at older ages.
  • The comparison has little power: the young-minus-35–49 difference in the change is 2.1 points, with a 95% interval of −8.8 to 13.0.
• By income quintile. Splitting the all-tenant fall by national income quintile, tenants moving up the national distribution account for 2.0 points (14.1%). Falls within quintiles account for 12.1 points: from 74.3% to 55.1% in the poorest fifth, and from 32.9% to 20.1% in the second.

3. The same tenant households became less overburdened. In the panel, households that paid market rent in two consecutive years:

• Within the same households. Selection into leaving home cannot lower the burden of a household that stays in the sample with the same members and the same tenancy.
• Larger than the cross-section's fall. Summed over the four transitions, the fall within continuing households is 16.9 points. Among all market-rent households in the cross-section it is 11.3 points (46.6% to 35.3%, household-weighted). So households entering and leaving the panel pushed the burden up, not down.
• Attrition. Households that move untraced or stop renting drop out. If they were the ones whose burden worsened, the within-household fall is overstated. Earlier transitions show rises as well as falls (2013→2014 and 2020→2021).

4. Shift-share on young tenants. Holding the overburden of each relative-income group fixed and letting only the mix change:

• 18–34, 2021 → 2025. Composition accounts for 2.8 of the 11.2-point fall (24.8%). Adding age subgroups to the cells changes this little.
• 2019 → 2025: 1.9 of 9.2 points.
• 18–29: 1.8 of 9.5 points.

The composition term is a floor on selection, not an estimate of it. The largest falls within cells are where leaving home collapsed: tenants' overburden fell from 75.7% to 47.3% in the lowest income quarter and from 70.6% to 43.1% among those without income, while it rose from 10.7% to 13.8% in the top quarter. That is the pattern that selection on unobserved resources, such as a partner's or a family's, would produce.

5. A bound on one channel. The bound caps one channel only: fewer young people leaving home, which changes who is a tenant.

• The extra stayers. The rise in the share living with a parent between 2021 and 2025 is the equivalent of 386,160 people aged 18–34 at 2025 population.
• The worst case. Suppose they had left home, rented at the rate other emancipated young people do (165,288 tenants) and all been overburdened.
• All tenants. The 2025 overburden would be 28.2% instead of 26.8%: 10.1% of the fall since
  1. If all had rented, 23.1%. This channel can explain at most about a tenth of the national fall, partly because young emancipated tenants are only about 15% of market tenants.
• Young tenants. The same bound gives 71.3% of their fall, or 144.1% if all had rented, and 124.0% at ages 18–29 even at the tenancy rate. It cannot exclude that this channel explains all of young tenants' fall.
• Two variants. If the extra tenants were overburdened only as often as young tenants in the two lowest income groups, the share for young tenants is 13.5% using those groups' 2025 rates and 43.2% using their 2021 rates. Neither is a central estimate.

What the tests say.

• Leaving home selects on income, more steeply in 2022–2025 than in the earlier panels with comparable coding.
• For the national fall, selection into leaving home is a minor channel. The fall is as large among older tenants, larger within continuing households, and the one-channel bound caps it at about a tenth.
• For young tenants' own fall, the evidence does not separate selection from other causes. Their costs rose faster than older tenants', the within-cell falls are where leaving collapsed, and the bound exceeds their whole fall.

5.3 Incomes against costs already paid

The two growth rates.

Incomes rose faster than costs.

• The cross-section. Between the 2021 and 2025 surveys, market-rent households' median income rose 43.5% and their median housing cost 17.5%. The windows differ: incomes are for 2020 and 2024, costs at interview, about September 2021 and April 2025.
• Within households. Over the same transitions, the chained median income growth of continuing tenant households was 35.3%.
• The counterfactual. Had tenants' incomes grown only as fast as their housing costs, using the within-household growth, 34.8% of tenants (40.3% of young tenants) would have been overburdened in 2025, instead of 26.8%.
• Who rents. With the cross-section's income growth the figure is 37.7%. The 2.9-point difference reflects who rents, not how incomes grew.

The starting year. The 2021 survey records 2020 incomes, the pandemic year, and its overburden (40.9%) is a local peak. From 2019 (37.4%) the fall is 10.6 points.

Sitting tenants pay less than the market.

• The law. Annual updates of existing contracts were limited by RDL 6/2022, art. 46: to the change in the Índice de Garantía de Competitividad until 31 December 2023, then «no podrá ser superior al tres por ciento» in 2024. Both applied in the absence of a new agreement, and as a cap for large landlords.
  • Since 2025 an INE index is the «límite de referencia» for updates (Ley de Arrendamientos Urbanos, additional provision 11).
  • Real Decreto-ley 26/2026, of 29 September, would have limited updates to 2% until the end of
  • The Congress repealed it on 2 October 2026, together with Real Decreto-ley 27/2026 of the same date, so neither is in force.
• The index. INE's IPVA, built from rents declared to the tax agency, rose 7.5% for existing contracts and 19.3% for new contracts between 2021 and 2024.
• Fewer new contracts. The weight of new contracts in the index fell from 17.6% to 12.3%.
• The survey. The ECV therefore measures the rents of a stock of tenancies that is turning over more slowly.

The survey moved. INE moved the ECV's fieldwork:

• 2017–2021: «tercer cuatrimestre»;
• 2022: «Segundo cuatrimestre de 2022»;
• 2023–2025: «de febrero a mayo».

The lag shrank. Costs were measured, on average, 14.1 months after the middle of the income year in 2021, but 9.4 months after it in 2025.

• Size. Growing 2025 costs over the extra 4.7 months at 2.8% or 5% a year raises the 2025 overburden from 26.8% to 27.6% or 28.1%.
• Timing. The largest single drop, from 39.4% to 30.6% between 2022 and 2023, coincides with the move to spring interviewing and with 2022 incomes. The panel shows the same households' overburden falling from 46.5% to 39.1% across that very change.

5.4 The CJE's figures and the measured burden

We swap one CJE input at a time for what young tenants aged 16–29 actually have (ECV 2025; 438 people in 345 households with a recorded rent and income. The 18–29 group in §5.1 also numbers 438 by coincidence: the 16–29 group adds two 16–17-year-olds and drops two people without a recorded rent or income):

Where the figures differ. The log of the ratio of asking rent to a young earner's income, minus the log of rent paid over household income, splits into three parts. Their sizes depend on which young earner is the benchmark.

• Asking rent against rent paid. The rent young tenants actually pay is about half the asking rent.
  • Part of it is the contract stock: rents fixed years ago and updated under legal limits.
  • Part is what young tenants rent, which is not necessarily an average 80 m² flat.
  • Size. We cannot measure floor area in the ECV. INE's IPVA weights put the middle of rent expenditure in the 75–90 m² band, and DatosRTVE reports a SERPAVI mean of 78.6 m² for rented dwellings. 80 m² is therefore close to a typical rented dwelling.
  • New tenancies. The 2025 module identifies those who moved in the last 12 months, but the samples are small (62 at 16–29, 129 at 18–34) and point in opposite directions.
• The benchmark against the tenants' own income.
  • The two benchmarks. The CJE's salary (1,191 €) is built from a labour-cost survey. The ECV's median for young earners (749 €) counts annual income, including part-year work.
  • Tenants have more own income. Within the ECV, young tenants' own net income (1,102 € a month, 2024 incomes, all sources) is well above the young earners' median, consistent with the selection in §5.2. Against the CJE's salary the difference disappears.
• One income against a household's.
  • Who they live with. Among tenants aged 16–29, 24.4% live alone, 24.6% as a couple, 15.9% as a couple with children, and 27.7% with flatmates or other relatives.
  • Sharing, as the CJE frames it. The CJE prices sharing at 400 € for a room, 33.6% of its salary. Young tenants who share pay a median 180 € each (household rent divided by members aged 16 and over), 19.3% of their own income (n = 102 people in 76 households). That figure counts sharing with relatives as well as flatmates.
  • Renting alone. The CJE scenario for the 98.7% is renting alone. The young people who do so pay a median 500 €, spend 35.4% of their income on rent, and 62.8% of them are overburdened (n = 100).

The CJE's question, answered with the same microdata. Of all people aged 16–29, 0.6% have a personal net income at which 1,176 € would take 40% or less of it, and 0.2% at which it would take 30% or less. Among young wage earners the share is 0.8% (0.7% on wages alone). At 18–34 the population shares are 2.4% and 0.9%.

The CJE's own ECV figures.

• Consistent. For renter households with an emancipated member aged 16–29, the mean monthly housing cost (HH070: rent plus utilities and fees) is 770–785 € under the four combinations of age definition and tenure we tried. The CJE reports a «gasto medio en alquiler» of 780 €.
• Arithmetic. The CJE's 30% follows arithmetically from its 780 € and its own income figure.
• Not reproduced. Its 48.9% overburden: we get 35.4–40.1% with that household definition, and 47.0–51.0% only with households whose responsible person is 16–29, whose mean cost (722–748 €) then does not match.

5.5 The emancipation chart

The splice is documented. The CJE's chart «Tasa de Emancipación Joven en España (2006-2025)» carries the note: «Se muestra la tasa de emancipación del segundo semestre de cada edición del Observatorio de Emancipación hasta 2024. En 2025, se muestra a la tasa de emancipación calculada con la nueva metodología.»

• The CJE's methodological note says the ECV-based figures «no son directamente comparables con las publicadas en ediciones anteriores del Observatorio, que se basaban en la Encuesta de Población Activa».
• The year-on-year fall it reports for 2025 (−1.13 points) is computed within the ECV.
• The chart joins two series, and says so.

Does the record survive the change of survey?

• ECV. The share of 16–29s with no parent in the household in 2025 is the lowest since 2004 under both age definitions: 13.2% by age at interview, 14.6% by age at the end of the income year (the CJE's 14.5%). Eurostat's own series for 18–34s living with parents is also at its maximum.
• EPA. We could not build an EPA series that reproduces the CJE's EPA rates. Our definition gives 17.26% and 17.25% in the second halves of 2024 and 2025, but it also misses the CJE's 2023–2024 change (−0.36 points against −1.8). So it cannot say whether the 2025 fall would show in the EPA.

6. Discussion

The figures answer different questions.

• The CJE's 98.7% and 33.6% ask whether a young person on a median salary could rent a whole flat, or a room, at today's asking prices. For a flat the answer is no, and the microdata agree: fewer than 1 in 100 people aged 16–29 earn enough to rent one within 40% of their income.
• The ECV-based burden asks what people who rent pay out of what their household earns. For young tenants the answer is about a quarter in rent and three tenths in all housing costs. About one in three is overburdened, and nearly two in three of those who live alone.
• Both are true. They describe the same market from two sides: entry and incumbency.

Selection is real. Its weight depends on which fall.

• Leaving home selects more on income. Young tenants are increasingly drawn from the upper half of the young income distribution.
• For the national fall in tenants' overburden it is a minor channel. The fall is as large among older tenants and larger within continuing households, and the one-channel bound caps it at about a tenth.
• The main mechanism there. Tenants' incomes, recovering from 2020 and lifted at the bottom, outgrew the costs of tenancies whose rents were held back by legal limits on updates and by slower turnover. Part of it is also who rents, as FEDEA noted.
• For young tenants' own fall, selection cannot be ruled out.

Incumbency and entry. A fall in overburden among incumbents can coexist with a rising entry cost. The record share of young adults living with their parents and the falling overburden are not one phenomenon. Both are consistent with the widening gap between the rents of new and of existing contracts that INE's index shows.

For users of the indicators.

• Eurostat's overburden rate is not a measure of access, and starting from 2021 overstates its fall.
• The ECV's fieldwork moved twice between 2021 and 2023.
• The CJE's indicator is a measure of access, not of what young tenants pay. The CJE publishes both kinds of figure, and its researcher stated the distinction in public.

7. Limitations

• Cross-sections and short panels. The cross-sectional ECV cannot follow anyone. The longitudinal files follow households for at most four years.
  • Most young people who leave home are recorded as moved and not traced, so we see their income before they leave, not after. Their destination is unknown.
• Panel levels. Leaving rates are not comparable in level across the four panels; we compare ratios between income groups. Continuing tenants are a subset of tenants, and attrition may bias their change (§5.2).
• Selection on observables. Personal income is measured in the previous calendar year. Our tests use relative income groups, age and tenure. Selection on unobserved resources is not captured, and the one-channel bound does not cover it.
• Small samples. Young tenants number about 540–1,100 a year.
  • Intervals are wide: the bootstrap uses 200 replicates and normal intervals.
  • The 2021–2025 files carry no primary sampling units. For 2019, which does, clustering on them changed the SE little.
  • The age comparison has little power. The year-to-year series for 18–29 is noisy.
• Timing. Income refers to the year before the survey, rent and costs to the interview. Both the lag and the collection mode changed between 2017 and 2023.
• What we could not measure.
  • There is no floor area in the ECV.
  • SERPAVI and Eurofound's report were not accessible to our crawler.
  • The rents of new tenancies are measured only in the small 2025 module.
• What we could not reproduce. The CJE's EPA emancipation levels and its 48.9%. The definitions behind them are not published in enough detail.

8. Data, code and licences

• Code (scripts/, tests/): Apache-2.0. Our data and text: CC BY 4.0.
• INE microdata are used under INE's terms: «toda persona que utilice ficheros de microdatos se compromete a citar … al INE como fuente del dato primario (fuente: INE, www.ine.es)».
  • The estimates derived here are our responsibility.
  • The terms do not say that the files may be redistributed, so scripts/run.sh fetch downloads them from ine.es.
  • Vintage: as downloaded on 4 October 2026; hashes in scripts/locked_inputs.json.
• data/ holds aggregates only. scripts/run.sh checks the 54 locked inputs and rebuilds data/ from them.
• Eurostat data are reused under its notice: «Reuse of statistical data … is authorised provided the source is acknowledged». CJE figures and quotations are cited for comment.

9. Automation and review

AI agents (Claude models) did the work on 4 and 5 October 2026:

• read the sources and quoted them;
• downloaded the INE microdata, Eurostat tables, INE releases and the CJE and FEDEA reports;
• wrote the code and ran the analysis;
• searched the prior work;
• drafted the text.

The quotations in data/quotes.csv are checked by script against the locked source files, and every headline number in the README and in this paper is recomputed from data/ by scripts/check_headlines.py.

Independent review. An independent AI agent reviewed the study adversarially. It recomputed every headline number from INE's raw files with its own code, rebuilt data/ byte-identically, and re-checked the quotations and the prior work. Its findings, all applied:

• temporarily absent household members had been dropped, which was the whole gap to Eurostat;
• the bound had been described as capping selection in general;
• the age comparison had been treated as a placebo;
• the within-cell term had been read as "not selection";
• the income counterfactual mixed in the changing pool of tenants;
• the CJE's sentence had been quoted without its shared-flat figure;
• the decomposition depended on one benchmark;
• "reproduced exactly" overclaimed;
• missing rents had been counted as zero in one table;
• small cells published single-household values;
• fetch.py could silently accept a new vintage.

The changes are reflected throughout this version, which has not yet been re-reviewed.

10. Competing interests

None known. EasyByte Hub S. Coop. Mad., which runs EasyxLab, has no commercial relationship with INE, Eurostat, the CJE, Idealista, the Ministry of Housing, or any organisation that campaigns on housing or represents landlords or tenants.

11. Ethics

• Microdata. We used INE's anonymised public-use microdata under its terms. We publish aggregates only, each with its sample size. Groups of fewer than 10 households publish counts only. No attempt was made to identify anyone.
• Polite access. We fetched public pages and files with a declared User-Agent, one request at a time, at least 1.2 seconds apart per host, after reading robots.txt. Hosts that refused it were not used.
• No contact. We contacted no one.
• Tone. The CJE is a public body representing young people. Its figures are documented and answer a legitimate question, and on that question the microdata agree with it. This study does not say it is wrong. It says what question it answers, and measures a different one.

How to cite

EasyxLab (2026). What young tenants in Spain pay, and why their overburden fell. Study S24. EasyByte Hub S. Coop. Mad. https://github.com/easybytehub/easyxlab (INE microdata as of 4 October 2026).

References

Official statistics and documents

• INE. Encuesta de Condiciones de Vida.
  • Cross-sectional microdata, base 2004 (2004–2007) and base 2013 (2008–2025).
  • Longitudinal microdata 2013–2016, 2016–2019, 2019–2022, 2022–2025. https://www.ine.es (operation 30453).
  • Press releases ECV 2017, 2019, 2021, 2022, 2023, 2024, 2025.
  • Módulo sobre dificultades de acceso a la vivienda, 2025 (27 April 2026).
• INE. Encuesta de Población Activa: microdata, 3rd and 4th quarters 2023–2025, and the base-2021 weights annex for 2023.
• INE. Índice de Precios de Vivienda en Alquiler (IPVA), tables 59004, 59008, 59063.
• Eurostat. ilc_lvho07c, ilc_lvho07a, ilc_lvps08 (updated 17 September 2026); yth_demo_030 (updated 16 April 2026).
• Consejo de la Juventud de España. Observatorio de Emancipación.
  • 2025, balance general (PDF created 31 May 2026).
  • 2º semestre 2024: informe estatal and nota metodológica.
  • 2º semestre 2023: informe estatal.
  • https://www.cje.org/observatorio-de-emancipacion/
• BOE.
  • Ley 29/1994, de Arrendamientos Urbanos (consolidated, last updated 2 October 2026), disposición adicional undécima.
  • Real Decreto-ley 6/2022, art. 46 (consolidated).
  • Real Decreto-ley 26/2026 and Real Decreto-ley 27/2026, of 29 September, and the Congress resolutions of 2 October 2026 ordering publication of their «derogación» (BOE-A-2026-20526, BOE-A-2026-20527).

Prior work

• Newtral (30 September 2026). Cuánto presupuesto se dedica a la vivienda: así es la brecha entre…
• elDiario.es, A. Díez, Y. Jara and R. Sánchez (29 September 2026). Quiénes son los casi tres millones de personas que viven ahogadas por el alquiler en España.
• DatosRTVE (2 October 2026). Los jóvenes frente al muro del alquiler en España.
• Conde-Ruiz, J. I. and Pinto, Á. (2026). El Coste de la Vivienda y la Pobreza Oculta en España: 2014–2025. FEDEA, Estudios sobre la Economía Española 2026/28.
• Pinto, F. (2026). Esfuerzo residencial y desigualdad económica en España: evidencia 2019–2024. FEDEA, eee2026-08.
• Khametshin, D., López Rodríguez, D. and Pérez García, L. (2024). El mercado del alquiler de vivienda residencial en España. Banco de España, Documentos Ocasionales 2432. https://doi.org/10.53479/37872
• Romero-Jordán, D. (2024). ¿Cuál es el esfuerzo por vivir de alquiler en España? Cuadernos de Información Económica 299, Funcas.
• Funcas (18 March 2026). Radiografía del elevado sobreesfuerzo de los jóvenes para alquilar vivienda.
• Hyland, M. et al. (2025). Foundational challenges: The housing struggles of Europe's youth. Eurofound. https://doi.org/10.2806/6354911
• Echaves-García, A. and Martínez del Olmo, A. (2021). Emancipación residencial y acceso de los jóvenes al alquiler en España. CyTET. https://doi.org/10.37230/CyTET.2021.M21.02
• Colom Andrés, M. C. and Molés Machí, M. C. (2026). Changes in Living Arrangements of Young Spanish Adults Over Time. Comparative Population Studies. https://doi.org/10.12765/cpos-2026-06

Cite this study

EasyxLab (2026). What young tenants in Spain pay, and why their overburden fell. Study S24. EasyByte Hub S. Coop. Mad. https://github.com/easybytehub/easyxlab/tree/main/studies/s24-youth-rent-burden

@techreport{easyxlab_s24,
  title       = {What young tenants in Spain pay, and why their overburden fell},
  author      = {{EasyxLab}},
  institution = {EasyByte Hub S. Coop. Mad.},
  number      = {S24},
  year        = {2026},
  url         = {https://github.com/easybytehub/easyxlab/tree/main/studies/s24-youth-rent-burden}
}