Ce que nous pouvons montrer, et ce que nous ne pouvons pas
Une base de preuves en construction pour l'Église européenne, assemblée à partir de sources nommées et publiée avec ses propres lacunes. Chaque chiffre ci-dessous indique de quel type d'énoncé il s'agit.
Ces pages sont pour l'instant une traduction assistée par machine du texte anglais, en attente d'une relecture théologique par un francophone natif. Les chiffres cités renvoient à la bibliographie, qui reste en anglais.
This page collects the figures we are willing to publish about the church in northern and western Europe, together with the year each figure describes, the population it covers and the source it came from. It is not a summary of the field. It is the subset of the field that survived checking.
The distinction matters because the most widely circulated European church statistic of the past two years was withdrawn in March 2026 and is still in circulation. Leaders are making decisions about property, staffing and succession on evidence that in some cases no longer exists. So each statement below is labelled: a measured figure, an analyst's reading of one, a modelled forecast, or something we could not confirm. Where a source's own wording is narrower than the way it is usually quoted, the narrower wording is used and the difference is stated.
The gaps are set out at the same length as the findings. A research programme is defined as much by what it cannot yet answer as by what it can, and several of the questions that matter most for European churches between now and 2040 have no published European data behind them at all.
Comment lire un chiffre sur cette page
A figure published by a named source, retrieved at that source. The year and the population it describes are stated with it.
A reading of data rather than a measurement. Traceable to the evidence cited, but arguable, and other readings of the same figures exist.
A modelled forecast, not an observation. Published projections of European church decline have so far run behind what was subsequently measured, so these are the least reliable statements on the page.
A claim in circulation that we could not verify at a primary source. Recorded here because its absence matters, and because it is repeated elsewhere as though settled. Unverified is not the same as disproven.
Membership and affiliation
Register-based membership is the best-measured thing in European church life, because in several countries the state keeps the count. It is also the measure that says least about practice.
Church of Sweden membership fell from 7,754,784 people, or 95.2 per cent of the Swedish population, in 1972 to 5,426,053, or 51.4 per cent, in 2024.
The Church of Sweden's own long-range projection puts its membership at 4,384,637 people, or 39.8 per cent of the Swedish population, in 2040, and at 3,868,857 in 2051.
The Evangelical Church in Germany recorded 17.4 million members at 31 December 2025, a fall of 3.2 per cent on the previous year, against roughly 350,000 resignations and about 105,000 baptisms in that year.
At the end of 2025 the two large German churches together accounted for 43.8 per cent of the German population. Annual losses ran at roughly 1 per cent a year before 2016 and have run at roughly 3 per cent a year since 2022.
Pew Research Center measured 505 million Christians in Europe in 2020, 67 per cent of the region's population, down from 75 per cent in 2010, while the religiously unaffiliated rose to 190 million, or 25 per cent.
The Center for the Study of Global Christianity estimates 553,294,000 Christians in Europe in 2026, against 562,875,000 in 2000, and models 398,231,000 by 2075. The series counts affiliation rather than practice, includes Russia and eastern Europe, and is therefore not comparable with attendance figures.
Attendance, practice and belonging
Attendance and membership move at different speeds and in some countries in different directions. Where a source counts something other than membership, the definition is given.
Average weekly attendance across the Church of England fell from 864,000 in 2019 to 702,000 in 2024, a decline of 19 per cent, and average Sunday attendance from 716,000 to 581,000 over the same period.
Church of England baptisms and thanksgivings fell from 89,600 in 2019 to 61,600 in 2024, weddings and services of prayer and dedication from 31,700 to 21,600, and funerals from 114,700 to 87,400.
John Curtice's analysis for the National Centre for Social Research, published in May 2026, concludes that church attendance in Britain remains below its pre-pandemic level.
In the Netherlands 44 per cent of people aged 15 and over identified with a religious group in 2024, against 55 per cent in 2010, and 14 per cent attended a service at least monthly. Statistics Netherlands described the decline in religious belonging as having stagnated.
The Swiss Federal Statistical Office's 2024 survey on language, religion and culture records 38 per cent of the Swiss population believing in one God, against 46 per cent ten years earlier. The 36 per cent with no religious affiliation reported alongside it comes from the 2023 structural survey, not from the 2024 survey.
Swedish state subsidy statistics count betjänade, defined as members of a religious community or people who take part regularly in its activities and self-reported by each denomination. The gap from membership is large: Equmeniakyrkan reported 96,460 betjänade for 2024, a year in which its own audited accounts recorded 57,118 members.
Young adults, and a claim that was withdrawn
This is the most contaminated topic in the field. One widely cited study has been retracted, several others rest on opt-in online panels, and the strongest evidence comes from probability-based series that were never translated into English.
The Bible Society and YouGov report The Quiet Revival, published in April 2025 and widely cited for the claim that monthly church attendance among British 18–24-year-olds rose from 4 per cent in 2018 to 16 per cent in 2024, was retracted in March 2026 after YouGov confirmed that quality controls on the 2024 sample had not been functioning. YouGov accepted responsibility for the result.
Pew Research Center's January 2026 analysis of British random-probability surveys found that the share of Christians aged 18–34 attending church at least monthly fell from 8 per cent in 2018 to 6 per cent in 2024, and that Christian identification in the same age group fell from 37 per cent in early 2018 to 28 per cent in summer 2025.
John Curtice's December 2025 analysis of the British Social Attitudes series concludes that the survey evidence does not support the claim of a religious revival in Britain.
Magnus Hagevi's 2025 analysis of the SOM Institute's Swedish survey series finds that belief in God and regular prayer among Swedes aged 16–24 have risen from a trough in the early 2010s to sit at or above the levels recorded for older age groups in 2024. The series is published as graphs rather than tables, so no precise values can be quoted from it.
Hagevi's own summary is that the secularisation of the Swedish people appears to have stopped, or at least to have taken a pause. He does not claim that the pause is permanent, and youth service attendance in 2024 remains well below its late-1980s level.
The Swedish youth and civil society agency's 2025 survey Så tycker unga records 21.2 per cent of Swedes aged 16–29 describing themselves as believers and 49.4 per cent as not believers, with 21.7 per cent saying religion is very or quite important to them.
Conversion, entry and return
The two European growth signals that survive scrutiny come from an institutional census and a national statistics office rather than from surveys. Both are clear about direction and small in absolute scale.
Adult baptisms at Easter recorded by the French bishops' conference catechumenate survey rose from 4,124 in 2016 to 13,234 in 2026, alongside more than 8,100 adolescent baptisms recorded across 89 dioceses in 2026.
In the 2026 French catechumenate survey the largest single age band among adult catechumens is 18–25, and women outnumber men in every adult age band except those aged over 65.
Set against a French population of roughly 68 million, the 13,234 adult baptisms of 2026 amount to about 0.02 per cent of the population in that year, which makes the series a strong signal about who is converting and a weak one about how many.
Active joins to the Church of Sweden rose from just under 6,000 a year in 2010 to about 15,000 in 2024, and 17 per cent of confirmands were baptised at confirmation in 2024, against 10 per cent in 2021.
The Evangelical Lutheran Church of Finland recorded 25,000 people joining in 2025, up from 23,000 the previous year, against 50,400 leaving, down from 51,300. The largest joining age group was 30–39.
The Church of Norway recorded 4.6 million service participants in 2025, an increase of 1.5 per cent achieved across roughly 1,300 fewer services, so that average attendance per service rose from 80 to 84.
Congregations: closure, planting and retention
The peer-reviewed European evidence on church planting is more sceptical than the sector that cites it. New congregations convert at many times the denominational rate and lose members at a rate that cancels most of it.
The Swedish free church survey Frikyrkoundersökningen records the number of free church congregations falling from 2,224 to 2,024 over its most recent five-year period, with 273 congregations closing and 77 newly established.
The same survey finds 89 Swedish municipalities in which free church membership is now below 1 per cent of the population, and records decline in the metropolitan regions for the first time.
Equmeniakyrkan's audited accounts record 641 congregations and 59,573 members in 2021 falling to 605 congregations and 57,118 members in 2024, with an operating result of minus 24.3 million Swedish kronor in 2024.
Stefan Paas's study of the Bund Freier evangelischer Gemeinden, covering nearly 500 German churches and about 40,000 members between 2011 and 2015, records overall annual growth of 0.76 per cent, one convert per 119 members a year, and growth of 42.3 per cent among churches planted between 2006 and 2015.
The same study records annual losses of 8 to 10 per cent from departures in those younger congregations, a pattern Paas calls a fast metabolism: churches that attract newcomers at many times the denominational rate without retaining them, so that conversion growth does not become net growth.
Paas and Vos's 2016 analysis of three small Reformed denominations in the Netherlands found that younger congregations gained roughly four times as many converts and five times as many returnees as older ones, which the authors attribute to location, outreach effort and entrepreneurial leadership.
Volunteers, buildings and money
Volunteer capacity moves independently of attendance and is measured far less often. Buildings are the largest long-term liability in European Christianity and the hardest to alter.
The number of volunteers in the Church of Sweden fell from over 70,000 in 2019 to nearly 35,000 in 2023, roughly halving in four years, according to research by Sara Fransson reported in November 2024.
Church of Sweden research into voluntary work in children's and youth ministry found the proportion of volunteer leaders in confirmation work nearly doubling between 1998 and 2019, over a period in which confirmation among 15-year-olds fell from 47 to 22 per cent, and found that many parishes have no clear strategy for volunteer work.
The National Churches Trust's 2025 survey of 3,628 UK churches found 83 per cent saying successful building management depends on active volunteers, 45 per cent reporting that a lack of volunteer time limits the community activity they can provide, against 33 per cent in 2010, and an average congregation relying on 265 hours of unpaid service a month.
The same survey counts about 38,500 churches, chapels and meeting houses in the UK in 2025, roughly half of them listed, with one in ten needing urgent repair within twelve months and more than a third of roofs at risk or needing urgent repair.
One in twenty churches in that survey said they felt they would definitely or probably not be used as a place of worship in five years' time. That is a self-reported expectation about future worship use, not an assessment of closure risk, and the report does not express it as a percentage.
The Freiburg projection published by the Evangelical Church in Germany in 2019, with a 2017 baseline, models Protestant membership falling from 21.5 million to 10.5 million by 2060 and projects church tax revenue alongside it. The figure covers Protestant membership only; the study publishes no combined Protestant and Catholic total.
Leadership supply and wellbeing
The only quantitative clergy wellbeing data we could verify for Europe is British. A widely repeated Swedish finding is recorded here as unconfirmed rather than dropped, because it is being cited elsewhere as settled.
The Church of England's Living Ministry panel survey, whose fifth and final wave was published in December 2025 with around 500 clergy responding, found 40 per cent of clergy feeling isolated in their ministry context.
The same survey found 16 per cent of clergy reporting symptoms consistent with probable depression and a further 13 per cent with possible or mild depression. Press coverage summing these to around 29 per cent quotes a figure the report itself does not print.
The same survey found 62 per cent of clergy rating their overall health as good or excellent and 72 per cent feeling they are fulfilling their sense of vocation, alongside 6 per cent meeting its burnout threshold and 41 per cent reporting a sense of personal ineffectiveness.
The claim that Swedish clergy top the occupational ranking for stress-related illness is widely repeated but cannot at present be verified. It appears to originate in a 2014 register study by the Swedish Social Insurance Agency, and the specific risk ratios attributed to that study's 2020 successor do not appear in the retrievable text of the later report.
The Church of Norway recorded 6,700 participants aged 13–17 in its leadership courses in 2025, 10 per cent more than the previous year and 48 per cent above the 2019 level.
The digital and AI question
The livestream question closed in 2024. The AI question has not opened in Europe at all: the only survey evidence available is from the United States.
The Church of England's Statistics for Mission 2024 records Church at Home attendance in 2024 at approximately 30 per cent of its 2021 level, a fall of about 70 per cent from the pandemic peak.
The same return records roughly 5 per cent of Church of England worshipping communities in 2024 as consisting entirely of online worshippers.
A Barna survey of 442 US Protestant pastors conducted in December 2025 and published by Gloo in June 2026 found 87 per cent using AI in ministry, most often for brainstorming (50 per cent), graphic design (37 per cent) and biblical or theological research (36 per cent).
The same survey found 71 per cent of those pastors cautious about AI, 79 per cent worried that it could act as a replacement for God and 63 per cent worried that it could replace pastoral roles.
Ce à quoi cette base de preuves ne peut pas encore répondre
Ce sont les questions qu'un programme européen de recherche sur la santé des Églises existerait pour fermer. Elles sont listées ici plutôt que dans un fichier interne parce qu'une page de recherche qui ne publie que ce qu'elle a trouvé est de la publicité.
AI adoption in European churches
No survey of AI use among European churches, clergy or congregations could be located. The only quantitative evidence available anywhere is from the United States. Since US adoption reached 87 per cent of Protestant pastors within roughly three years, European use is likely to be substantial and is entirely unmeasured.
Clergy wellbeing outside Britain
All clergy wellbeing data we could verify for Europe is British. For Germany, France, the Netherlands, Belgium and Switzerland we found no comparable quantitative study. The Swedish occupational-risk figures in circulation rest on a table appendix that has not been separately published.
Migrant-led congregations in continental Europe
There is no systematic count of migrant-led congregations for Germany, France, the Netherlands, Belgium or Switzerland. The academic literature is qualitative and case-based. Aggregate denominational statistics conceal both the growth and the high closure rate of these congregations.
Pentecostal and charismatic totals for Europe
Global and country-level estimates are published, but no Europe aggregate appears on open sources, and Pew's regional Pentecostal breakdown omits Europe. Any European total in circulation is derived rather than measured.
Volunteer trends outside Sweden and the UK
The Swedish collapse between 2019 and 2023 and the British constraint figures are the only volunteer series we could verify. There is no equivalent time series for Norway, Denmark or Finland, and no study isolates the cause of the Swedish fall.
Church planting counts for the other Nordic countries
Sweden publishes plant and closure counts through its five-year free church survey. Norway, Denmark and Finland do not publish a comparable series, so the Nordic planting picture rests on one country.
Digital engagement outside England
The Church of England publishes detailed digital engagement figures. No comparable statistics were found for German, French, Dutch, Belgian or Swiss churches, so the finding that livestreamed worship collapsed while discovery and devotional use grew rests on a single national case.
Belgian attendance after 2022
The most recent Belgian institutional attendance figures we could verify are for 2022. Nothing verifiable was found for 2023 to 2026.
A pan-European closure dataset
No such dataset exists. Each national body defines closure, deconsecration, repurposing, demolition and transfer differently, so any Europe-wide total in circulation is an aggregation of incompatible national figures.
Anything with 2040 as its terminal year
No major research body publishes a projection ending in 2040. Available projections end at 2050, 2060 or 2075, and two independent checks show that published projections have run behind subsequently measured decline. We therefore do not interpolate a 2040 figure from them.
The replacement British survey
A repeat of the retracted 2024 survey was commissioned in spring 2026. Until its results are published and its methodology can be examined, the British young-adult picture rests on random-probability series alone.
Le programme que cela exigerait
Un rapport annuel sur l'état de l'Église européenne est un objectif raisonnable à long terme et n'est pas quelque chose que ce cabinet peut produire aujourd'hui. L'infrastructure vient d'abord, et elle ressemble à ceci.
National registers and state statistics first
The strongest European church data is not produced by churches. Population registers, national statistical offices and state subsidy returns give counts that do not depend on anyone volunteering an answer. Where a register figure exists it is preferred to a survey figure, and where the two disagree both are reported.
Institutional censuses second
Counts made by an institution of its own activity — baptisms, confirmations, joins and leavings, service attendance returns — are complete rather than sampled, and can be followed year on year. Their weakness is definitional drift, so the definition in force in each year has to be recorded alongside the number.
Probability-based academic series third
Long-running academic surveys with probability samples are the only usable source for belief, prayer and attitude. Opt-in online panels are recorded as opt-in and are never presented as measurement, because the failure that produced the 2026 retraction was specific to that method and specific to young respondents.
Commissioned primary work for the named gaps
Several of the gaps above cannot be closed by better searching, only by fieldwork: European AI adoption, clergy wellbeing outside Britain, migrant-led congregation counts, and volunteer trends in Norway, Denmark and Finland. Each would need a defined population, a sampling frame and published methodology before any figure from it could be quoted.
What an annual report would require
An annual report on the European church would need three things that do not yet exist together: a stable set of definitions that national bodies would accept, at least one comparable series in each country covered, and a published verification procedure applied before rather than after publication. Until those are in place, the honest form is a standing register of verified findings and named gaps, revised as sources are retrieved or withdrawn.
The rule that governs revision
A figure may be published only when it has been retrieved at its own source, and it is removed when that source is withdrawn or contradicted. Corrections are recorded in place rather than absorbed silently, so that the difference between what was believed and what was checked stays visible.
Les numéros entre crochets renvoient à la bibliographie de la page Sources, où chaque entrée porte son statut de vérification. Ouvrir la bibliographie