Friends and neighbours – if once, political opinions were shaped by discussions over garden fences, are they now hardened by the discussions between friends on social media sites?
Many commentators presumed that the advent of digital communication would result in the neighbourhood withering away as a source of influence; and that in future, people would vote on the basis of individual demographics such as age, education and social class, not on the shared memories of the physical communities to which they belong. Parties’ share of the vote would continue to vary geographically, but only to the degree that the demographic composition of local areas differed from the national average.
These expectations are confounded by the experience of Britain’s main political parties. If such a shift has occurred, it is more limited than political commentators had predicted. Most Britons continue to vote in the same way as their immediate neighbours, not their personal circumstances. That is the message of my book, The Predictive Postcode – A Geodemographic Classification of British Society.
This finding explains why Mosaic continues to play a vital role in the way parties target individual communications during their General Election campaigns. Mosaic is the system most extensively used to classify people according to the type of neighbourhood in which they live. If people living in the Mosaic type ‘Bungalow Retirement’ support Brexit to a far greater extent than can be explained by the age and class of ‘Bungalow Retirement’ postcodes, and if in the 2017 General Election Labour achieved higher than average swings in every constituency rich in the Mosaic type “Liberal Opinion”, then it can be understood why Mosaic is a useful basis for classifying electors.
However, despite Mosaic having been embedded in the algorithms political parties use to drive their communications for a number of General Election campaigns, neighbourhood classification is still surprisingly little understood in academia, certainly by comparison with commercial organisations, and even less by civil servants. Whereas elsewhere the mining for behavioural patterns within big data sets is increasingly taking over from the questionnaire as a source of fresh insights; academics, civil servants and the market research industry continue to seek inspiration from the results of survey responses using classifiers based on respondents’ occupation, i.e. their social class, or the ranking of their neighbourhood on an index of deprivation.
The Predictive Postcode aims to introduce geo-demographics to many of the professions that themselves cluster together in neighbourhoods of ‘Liberal Opinion’ – such as journalists, academics, opinion-formers, and market researchers – by illustrating how it can be used to analyse data relevant to a series of contemporary social issues. For example, the transformation of rural Britain into a playground for the middle classes; the increased social polarisation between successful and declining seaside resorts; the divergent paths of different types of minority-ethnic neighbourhood; the low horizons that characterise young people living on the large, low-rise council estates on the peripheries of provincial cities; and the impact of the emerging Liberal Metropolitan Elite on the breakdown of the relationship between social class and party affiliation.
It is evident that social class is no longer the principal engine of political division (see here) when one compares the gap between the Conservative and Labour shares of the vote at the 2017 General Election for various social classes, and that for parliamentary constituencies. For social classes, the variation in the Conservative lead over Labour is a paltry 33 points (+19 to -14). The variations at constituency level is a massive 132 points (+56 to -77).
However, the pattern is not just one where differences in party support are exaggerated by prevailing local influences; that the doctor in Liverpool Walton is influenced by the life-experiences of his local patients whilst the postman in Hampshire North East adopts the political viewpoint of those he delivers letters to. The difference in the parties’ performances in York Central and in Harlow reveal how electors are much more favourably disposed to Labour in a seat with an over-representation of ‘Urban Intelligence’ than ones in seats of equivalent income, class or status, characterised by what Mosaic describes as ‘Blue Collar Enterprise’.
If it is the character of that neighbourhood as well as its social composition which contributes additional power to the prediction of voting behaviour, over and beyond personal characteristics on their own, it is instructive to identify in this table those constituencies whose Social Rank (where a low number indicates high social class) aligns least well with their Conservatives Rank (where a low number indicates Conservatives doing well compared with Labour). Conventional wisdom would suggest the two indices should be closely aligned.
While Labour does far better than one would expect on the basis of social class composition in some university seats (Manchester Withington, Sheffield Hallam), the biggest misalignment between class and voting occurs in two blocks of seats, one in North London the other south of the Thames, both characterised by the Mosaic grouping ‘Urban Intelligence’.
Conversely, most seats with particularly low social class, but where Conservative candidates have built big majorities, are located mostly around the Wash and in other coastal communities in the East of England. All are associated with, but not caused by, a high UKIP vote prior to 2017.
One of the practical advantages of a classification that can be tied to postcodes is the ease with which it is possible to analyse consumer behaviour, values and attitudes. Survey respondents can be coded by the type of area in which they live and any consumer behaviour that is digitally captured can be cross-analysed by Mosaic assuming that the postcode of the consumer is known, which it increasingly is. This makes it possible to draw up a rich qualitative understanding of any segment.
In relation to ‘Urban Intelligence’, therefore, we can use evidence to show that this group, in their daily existence, is significantly more exposed to people from different backgrounds, whether in terms of race, religion, housing tenure or deprivation, than its equivalents in suburban or rural locations. And, though it is not especially dependent on state benefits, its wellbeing is particularly dependent on the collective provision of services, public transport, clean air and access to open space.
Market research data shows it is a group which, on account of its high income, lack of time and often cramped housing, is among the pioneers of personal outsourcing – eating in restaurants, being entertained in the cinema or theatre, having clothes dry-cleaned. Its ethos contrasts with the self-reliance practiced in the countryside with its preference for physical and practical activities such as DIY, cooking, gardening and the exercising of pets, where, on the basis of information on charitable giving, people appear to have greater difficulty imaging the plight of people different from themselves.
In terms of its employment, the ‘Urban Intelligence’ group is disproportionately involved in highly specialised professional activities often of a normative or affective nature, in the media and entertainment for example, in politics, charities or lobbying, or in the arts. Career advancement within each of these occupational groupings often depends on an ability to identify with the emotional demands of a range of different demographic groupings, not just an ability to accurately apply rule-based professional standards.
Labour has already lost its Westminster representation of the Mosaic type ‘New Town Materialism’. Viewed in this light, it becomes particularly clear how difficult it has become to reconcile the values of ‘Urban Intelligence’ with those of ‘Rustbelt Resilience’ and ‘Ex-Industrial Legacy’, groups recognised by Nick Timothy as potential targets for a new One Nation Conservatism and significantly present in those seats which the Conservatives did capture from Labour in 2017.
The Predictive Postcode is available to purchase via Sage Publications, with a special 20% discount offer until 30 April 2018, with the code UK16AF15.