Partisan geographic sorting in the UK: do political views influence to where we choose to move?

Georgios Efthyvoulou - Associate Fellow, SPERI, and Senior Lecturer in Economics, University of Sheffield

Vincenzo Bove - Professor of Political Science, University of Warwick

Harry Pickard - Lecturer in Economics, Newcastle University

Partisan geographic sorting has important implications for the type of information that individuals receive, the attitudes they form and the social interactions they experience. Despite a considerable amount of research on the economic factors that influence migration decisions, we still lack a comprehensive understanding of the role that community ties, information, beliefs and values play in shaping internal migration patterns. 

In a recently published study in the Journal of Economic Geography, we examine whether changes in political similarity between two geographic areas in the UK affect their migration flows; that is, how many people move from one district to another in a given year. To do that, we leverage variations in political preferences across pairs of local authority districts over time, and combine this information with annual data on their migration flows from the Office for National Statistics. The resulting dataset is the most comprehensive ever used to investigate the self-grouping of British citizens into politically similar communities – it covers all possible origin and destination districts in England and Wales between 2002 and 2015. 

We consider two alternative measures of political similarity. First, we identify whether a district-pair’s local councils are controlled by the same party in a given year. Second, we construct a ‘continuous’ measure of partisan spread between the two districts, calculated by the difference in the share of local council seats held by the Conservative party (or the Labour party) in a given year, with lower values reflecting higher political proximity. 

We account for additional factors that influence migration flows between two districts, such as geographic proximity, relative wages and unemployment rates, and pair-specific differences in ethnic mix, age structure, education levels, industrial composition, religious composition and genetic background. We also perform various tests to address ‘endogeneity concerns’; that is, the possibility that political similarity between two districts is influenced by migration patterns, or the possibility of unobserved factors affecting our results.

Our analysis reveals that migration flows between two Labour or two Conservative districts are 5% higher than those between other pairs of districts. Similarly, we find that an increase in political proximity between two districts in a given year (as captured by our ‘continuous’ measure) leads to significantly more moves between the two districts in that year.

Figure 1 shows that the most important variable in predicting migration flows between two districts is migration flows the year before (‘Past migration flows’), which acts as measure of migrant networks. This is to be expected as the cost of adapting to a new society is largely mitigated by the presence of family members and friends who are familiar with both the origin and the destination areas. Yet, political similarity (‘Proximity in party shares’) ranks sixth overall and exerts about the same influence as relative wages and proximity in ethnic mix.

The desire for ‘homophily’

To understand more about why we tend to move to districts that share our political views, we look at individual-level data for the same period, 2002–2015, from the British Household Panel Study (BHPS) and its successor, Understanding Society (UKHLS).

We investigate whether individuals’ perceptions and attitudes towards the location where they live are affected by the extent of their political alignment with their district of residence. In particular, we compare a measure of respondents’ political alignment to their district with their answers to questions on whether they would like to move and how satisfied they are with their neighbourhood.

We find that politically-aligned individuals are about 3 percentage points less likely to report a preference to move. They are also 2-4 percentage points more likely to provide positive responses to questions about whether they plan to reside in the area for a long time, feel a sense of belonging in their neighbourhood, and feel that they are similar to their neighbours. This offers evidence that one of the key channels underpinning our district-level results is the desire for ‘homophily’: living in areas with ideological views similar to our own contributes to a sense of ‘fitting in’ and ‘feeling at home’. 

However, the desire to fit in is unlikely to be the driving factor for relocation. Put differently, political alignment is an insufficient ‘push’ factor to cause the decision to migrate. Using our individual-level dataset, we show that this is indeed the case: the desire for political affinity affects where we choose to move to, but not our decision to move in the first place.

Conclusions

Our study shows that politically like-minded individuals in the UK tend to cluster within the same communities. This has negative social consequences. For one, by limiting the development of (and exposure to) diverse viewpoints, it perpetuates a politically homogeneous environment that can foster a hostile political culture. It can lead to a less politically educated electorate and intensify political polarisation. The rising levels of ‘affective polarisation’ in Western democracies, including the UK, is a cause for concern, especially in the wake of the Brexit referendum which has further exacerbated this trend.

Partisan geographic sorting also damages democratic performance because it reduces the number of politically competitive settings. This poses a threat to a government’s accountability function, as competitive elections are the means by which citizens reward or punish the performance of their representatives. In a first-past-the-post system, it also leads parties to embrace ‘minimalist’ electoral coalition strategies where policy platforms are tailored towards the geographic areas where they have a realistic chance of winning.

As the timer for the next UK election ticks down, our findings suggest that political parties could have a difficult job on their hands: self-selected regional political polarisation will make it more difficult to sway districts from set opinions and ideals.