Greg Smith (William Temple Foundation)
The relationship between religious identity and voting patterns is a matter of considerable interest in democratic societies across the world. This has been heightened in recent years by the rise of populist politics and nativist narratives, whereby traditional religious and national identities are deployed in electoral political competition. One only needs to think of the link between Christian nationalism and white evangelical voters in the USA, or of the Hindutva movement which influences the supporters of Modi in the world's largest democracy, India. Orban's Hungary, and Erdogan's Turkey provide further examples showing that religion and politics do mix. In the UK the associations between faith and politics are less obvious, more nuanced and poorly researched. This research note examines some of the academic literature, public data sources and research gaps on the topic of religious identity, belief and practice in Britain over the last decade and suggests further lines of enquiry. It concentrates almost entirely on England, for simplicity's sake, and because the other nations of the UK have significantly different religious and political cultures. The focus is mainly on the 2016 Brexit Referendum and the 2024 General Election.
The Changing Religious Makeup of England
Before examining voting patterns it is important to set the context by looking at religious trends since the start of the millennium. The big story is the rapid decline in the proportion of the population identifying as Christian. In the 2001 Census 72% of English people said they were Christian, by 2021 this had fallen to 46%. Representative social surveys confirm this trend and it is clear that fewer than half the population now consider themselves Christian, and growing numbers, especially among the younger generations opt for "no religion". But the rise of the "nones" is not the only story. In the same period minority religions have grown (Muslims from 3% in 2001 to just under 7% in 2021). Religious diversity has increased, especially in parts of metropolitan cities such as London, Manchester, Birmingham and Leicester, where in a context of ethnic superdiversity the white British (Christian / None) population has become a minority. This religious diversity comprises faith communities linked to major world faiths (Islam, Hinduism, Sikhism, Judaism) but also includes a thriving and growing representation of Christians worshipping in both diaspora and mainstream congregations.1 There is however some surprising recent evidence of a reversal of the decline of worshipping Christians between 2018 and 2024, concentrated among 18-24 year olds, and especially among young men. The findings come in The Quiet Revival, a survey commissioned by Bible Society and conducted by YouGov.
Voting Patterns and Religion.
a) Brexit
There are signs that religion remains an important predictor of political behaviour and contributed to how people voted both in the 2016 Referendum on EU membership and in the 2024 General Election. However, the research evidence remains rather sketchy and the interpretations of what is available can be contested. The most detailed analysis of the Brexit vote by religion is the 2018 paper by Smith and Woodhead2summarized in an LSE blog The main data set used was a yougov exit poll, subjected to rigorous statistical analysis, which was supplemented by a panel survey of Evangelical Christians. The headline finding that 66% of Anglicans voted "leave" was widely publicised. However a careful reading of the paper provides some nuance and some qualifications. The category "Anglican" might be better interpreted as those respondents who in answer to a tick box question told interviewers that they were "CofE", and implies nothing about strength of belief, church attendance or other religious practice. Indeed other data sets covered in the paper suggest that regular church attenders and committed evangelicals were much
1 Goodhew, D. and Cooper, A.P., 2018. The Desecularisation of the City: London's churches, 1980 to the present. In The Desecularisation of the City (pp. 3-38). Routledge.
2 Greg Smith & Linda Woodhead (2018) Religion and Brexit:
populism and the Church of England, Religion, State & Society, 46:3, 206-223, DOI: 10.1080/09637494.2018.1483861
more likely than average to vote Remain. The paper also suggests that it is possible that the direction of the association could be reversed; i.e. that leave voters (as people with a strong sense of English identity) were far more likely than remainers to state their religious affiliation as CofE. The other weakness of the data used in this paper is that the sample size of minority faith communities (other Christian denominations and the major World Faiths) were too small to allow reliable analysis.
One way of triangulating the results is to examine the relationship between the religious makeup of an area and the strength of support for Brexit. With the publication of the 2021 Census data we were able to look at this, albeit with a five year gap between the two sets of data. In order to make a preliminary assessment we examined the ONS 2021 Census data by constituencies seeking to correlate it with voting patterns in the Brexit referendum of 2016. There are some indications that the highest leave vote was in constituencies where 90%+ are Christian or no religion (correlated with very high proportions of "white British" on the ethnicity question. . Table 1 shows the top ten pro-Brexit constituencies, 8 of which fit this pattern. There are two exceptions, Walsall North and Stoke on Trent, urban areas which have significant populations of (ethno)religious minorities yet still voted overwhelmingly to leave,.
TABLE 1 EU REFERENDUM: CONSTITUENCY RESULTS Estimates and known results of the EU Referendum
In contrast the most strongly Remain constituencies in England are all urban and tend to be ethnically and religiously diverse. With three exceptions the top 15 are within Greater London, They include constituencies in Brighton and Bristol (districts which are in the top three for percentage of residents with "no religion"). However the third ranking district on this factor, Norwich is not strongly remain The Norwich South constituency voted 40.5% to leave, while Norwich North had a strong leave majority of 56.7%).
Table 2 – Constituency Leave vote
Nonetheless a high level of religious diversity did not guarantee a high Remain vote. Leicester has three constituencies, two of which reflected the UK national vote of just over 50% leave, and the other was below average at 40% leave. The two Harrow constituencies had only a slight minority of leave voters (45% and 46%). The two Brent constituencies voted 43% leave, and constituencies in Newham and
Redbridge had similar rates. At the other end of the diversity scale constituencies in the most Christian areas of south Lancashire / Merseyside had slightly above average majorities to leave.
The tentative conclusion is that religious make up of an area made some difference in the Brexit vote, but the patterns were not simple. The more important factors favouring Leave seem to be regional setting, lower levels of urbanisation, lower levels of education, older age profile, and high proportions of white British residents.
General Election July 2024
For the general election of July 2024 as we began this paper we were waiting to see a detailed statistical analysis of exit polls, but first impressions show religious identity is significant, and still sometimes divisive. There have been a couple of blogs based on exit polls
https://www.electionanalysis.uk/uk-election-analysis-2024/section-2 -voters-polls-and-results/religion-and-voting-behaviour-in-the-2024-g eneral-election/
https://news.exeter.ac.uk/faculty-of-humanities-arts-and-social-sciences/religion-remained-an-important-predictor-of-voting-behaviour-in -the-2024-general-election-analysis-shows/
This analysis (as in the 2016 exit poll) only looks at major Christian denominations and "nones" because of inadequate samples in minority faiths. And conclude that Anglicans (CofE tick box?) tended more than average to vote Conservative (but not Reform).
A better range of Data from the British election study (wave 29) is now available. It is an online panel survey with a large sample (over 25,00) and a detailed breakdown of Christian denominations with some sophisticated weighting to ensure it is representative of the electorate. There are enough followers of Islam, Hinduism, Judaism and Pentecostalism to see meaningful patterns of voting in these groups. But there is no indication of active religious practice or firmness of religious belief which may be important especially among Anglicans and Roman Catholics.
TABLE 3 (From British Election Study Wave 29)
And our own blog
https://williamtemplefoundation.org.uk/faith-and-voting-the-uk-gen eral-election-2024/
gives some impressions of how constituency results reflected local demography.
The constituency results (see this BBC news page for maps of the results and party vote shares by constituency) clearly have some correlation with the 2021 census (voluntary) question on religion
(which is a very basic one indicating nothing more than a household identifying with a particular religious category or "no religion"). Looking at area based data for religion (which can be done on this ONS census map page) needs to be done with caution, especially when constituency boundaries do not match the local authority areas which are reported in census data. However, having looked at the Census data on religion in England and Wales only, a few things stand out, and generate some hypotheses for further investigation.
Areas with the highest proportion of "No Religion" produced some interesting results. Two of the Green MPs are in cities which are top of the "no religion" league, Brighton (55% No religion) and Bristol (51%). No religion in these cities correlates with a population with high proportions of young graduates and professional, and significantly large LGBTQI communities. The other two Green MPs were elected in more rural areas in Suffolk and Herefordshire, where "no religion" rates were around or slightly above average, but with little religious diversity. It is possible local issues and/or tactical voting may have played a part here.
On the other hand, Reform MPs were elected in Ashfield which is also in the top five districts for "no religion" (around 50%) while Clacton, Boston, Basildon and Great Yarmouth are also above average no religion but also well above average Christian. All these constituencies, and many others where Reform came second with a high proportion of votes, (in places such as Hull and the North east of England) are very un-diverse religiously, with a white majority, older electorate, and a more working class history. They also strongly voted "leave" in the Brexit referendum.
On the other hand the areas recording the highest proportion of Christians are in South Lancashire and Merseyside. They have low levels of religious diversity but had massive proportions of Labour votes, and relatively poor performances by Reform. This is perhaps associated with strong residual Roman Catholic loyalties.
In constituencies with high proportions of Muslim voters the Labour vote share was generally lower than previously, with the war in Gaza being a salient issue and favoured independent ("Muslim Vote") candidates. In many of these constituencies the campaigning was reported as being brutal and frightening. Independents ousted Labour from Blackburn, Dewsbury and Batley, Birmingham Perry Barr and Leicester South. However, in most multicultural constituencies Labour held on to their seats, for example in East Ham veteran evangelical MP Sir Stephen Timms held on with a reduced majority, but still commanded 51.6% of the votes cast. Leicester East, one of the most religiously diverse constituencies in England was an outlier. It became the only seat the Conservatives took from Labour, with a strong showing from Liberal Democrats and Independents that reduced the Labour vote by nearly 30%. Interestingly Harrow East which is one of the areas with the highest proportion of Hindus in England, is one of a handful of London constituencies that remain in Conservative hands, with only a 1% fall of the Tory vote.
Interpretation
There is not yet enough research evidence to enable us to be sure how faith affiliations and religious identities and practice map onto electoral politics in England with any degree of precision. In any case religious identities intersect with other factors, most noticeably, ethnic identity, age group, gender, locality and social class. But there are enough indications to suggest some hypotheses which could be tested empirically.
● Muslims in the past predominantly supported the Labour party (with a few educated / business class Muslims opting for the Tories. However increasing numbers have lost faith in Labour and are now more likely to support radical / identity politics candidates who are perceived as more likely to support Muslim interests on the global stage. In 2024 the BES suggested 34% voted Labour, while 21% voted Green and 16% supported an independent candidate.
● Hindus are more likely to vote Conservative especially if they have prospered through business and education, and moved out of the inner cities. In 2024 there may have been a Rishi Sunak effect. TOne might expect some pull towards populism, possibly linked to the politics of Hindutva in India, but actual voting suggests hesitancy about Reform, perhaps as it may be perceived as white / English / racist in ethos. In fact in 2024 according to the BES data 36% voted Labour, 34% Conservative (many more than the national average) 10% Lib Dem and only 7% for Reform.
● One would guess that the majority of Jewish people will vote for the Tories, as many have rejected Labour because of perceived anti-semitism and left wing opposition to Israeli war strategies. However, the BES data suggests that in 2024 Labour held on to a third of the Jewish vote, slightly more voted Conservative (more than the national average) 10% for Reform and 11% for the Lib Dems.
● Sikhs… possibly still tend to support Labour. (In the BES small sample of Sikhs, 59% did so.)
● Nones (No religion)..The area based constituency results suggest they are likely to break two ways. ○ Educated progressives… will tend to support Greens, Liberal Democrats or Left Leaning Labour candidates. ○ Working class (white English) "nones" will be attracted to populism with English nationalist connotations.. I.e. Reform (in 2019 Boris Tories). In the BES data those of no religion had larger than average proportions of left leaning voters (40% for labour, 15% Lib Dem, and 10% Green.) More detailed analysis shows that less educated, lower class "nones" do indeed lean to the populist right, while graduates and professionals lean to the left, a pattern similar to Anglicans / CofE (see Tables 4 & 5 in appendix.)
● Christians is too wide a category with diverse politics among them, and in most surveys little distinction between actively practicing and nominal affiliates.: One might expect.
○ Nominal Anglicans towards populism (Reform) ○ Practicing Anglicans .. Centrist Conservatives
○ Non Conformists Lib Dems - left Labour - or Greens ○ Roman Catholics weakly towards Labour
○ Evangelicals / Charismatics… slightly left of Centre but diverse voting patterns
○ Pentecostals trending towards populism.. (socially conservative, some Trumpite influence from USA)
The BES data for 2024 seem to support the bias of Anglicans towards the right, and of Roman Catholics towards the left of Centre parties. But the free churches (Methodist, Baptist and URC) lean more to the right than expected - perhaps because of their members tending to be elderly, and Pentecostals supporting Labour somewhat more than the national average.
Obviously political positions change over time and in response to the salience of particular hot button issues, so there is no expectation that these suggestions will amount to predictions of future election results.
However as politics continues to change away from economic and class issues towards contours of identity and social justice knowledge of the relationships between religion and politics will become even more significant.
Towards more sophisticated research.
In order to explore contemporary relationships between faith and politics in England with more precision and sophistication more research is needed, and standard political research needs to be modified.
1. Opinion Polls and surveys of political issues need to use more sophisticated categories around religion.
a. The tick box categories for religious affiliation need to have more and better categories (at least as many as in the BES) especially for the sub categories of Christian
b. Samples of minority faith communities need to be boosted in order to gather meaningful comparisons between groups.
c. Additional questions are needed to assess respondents level of religiosity
i. Strength of belief in the teachings of the religion
ii. Frequency of public practice… (noting that church attendance does not mean the same as praying in a mosque, or attending a Hindu festival)
iii. Private prayer / spirituality
D. Analysis should be multivariate to include the effects of variables such as gender,age group and social class.
2. Analysis of Census data for religion and ethnicity should be taken to a deeper level..looking at correlations with voting patterns in constituencies and Council wards.
3. Qualitative research using in depth interviewing and focus groups should be carried out to explore how particular local faith communities decide to vote.
4. Ethnographic work.. (Perhaps among political parties and activists) should explore if and how politicians pitch their policies and platforms to voters of different faith backgrounds.
