Migration & mobilities strand abstracts

Strand organiser: Hill Kulu, University of Liverpool

Internal migration modelling - Monday 8 September 1.30pm

Spatial Analysis of internal migration in Luxembourg
Stamatis Kalogirou, Department of Geography, Harokopio University, Geoffrey Caruso, Institute of Geography and Spatial Planning, IPSE, University of Luxembourg

The aims of this paper are to look at the spatial patterns and to model internal migration in Luxembourg. Between 2010 and 2011, 17344 people or 3.7% of the total population moved from one municipality (commune) to another within Luxembourg, showing a very dynamic migratory system. Luxembourg City is a net migration looser in terms of internal migration: 3128 people moved out of Luxembourg and 1428 people moved in to the capital in the same period. This paper presents an innovative migration flow map while attempting to explain, for the first time in Luxembourg, the driving forces of migratory moves at a fine geographical scale, most of the literature being focussed on the strong international migration, thus hiding local residential processes. The data analysed refer to the migration flows between the 116 local authorities (communes) in Luxembourg and the 12-months migration question. We present models for out-migration, in-migration and migration flows. For the latter, we fit unconstrained gravity models using Poisson, and Negative Binomial (NB) regressions, and their Zero Inflated variants in order to account for overdispersion and the large number of zero flows. The empirical results are very interesting and show processes that differ from other countries. It appears that cultural and urban planning factors rather than labour force factors affect the decision to migrate within Luxembourg, which is plausible given the location, the size and the ethnic background of the population residing in Luxembourg.

Email: skalo@hua.gr

The association between student migration patterns and observed levels of geographic deprivation and social exclusion across the United Kingdom
Neil Bailey, Jakub Bijak, Sylke Schnepf, University of Southampton

Previous studies have commonly found evidence that links the levels of deprivation and inequality with differences between individuals’ socio-economic and ethnic background. While there is also evidence that these associations differ between the geographical areas of the UK. This paper aims to investigate the impact of student migration trends on the levels of deprivation and social inequality at the origin and destination areas by answer the following research questions: ‘How does student mobility impact on existing levels of geographical inequalities and deprivation at the local authority level’ and ‘How does geographical inequality and deprivation impact on the migration transition experienced by a student entering a HEI?’ With the use of the Student Record Dataset of the Higher Education Statistics Agency, which contains detailed information on every student recorded as attending an institute of higher education in the UK, this paper uses a series of statistical models to gain an in-depth understanding of student migration flows, levels of deprivation and inequality within the UK. The first approach uses multivariate regression techniques to model the migration outcomes of students in the UK against their individual (micro) characteristics (age, ethnicity, socio-economic status, parental education) and area (macro) level variables of deprivation and inequality. The second technique uses a spatial interaction model of the student flows between local authorities (LA) where distance and origin /destination deprivation variables are used to add informative information to our model of the transitions undertaken by students in the UK. The results of this paper will be ready in time for the conference in September.

Email: Neil.Bailey@soton.ac.uk

Mode of transport to work or study: What does Scotland’s Census 2011 data tells us?
Daniela Ene, National Records of Scotland

The purpose of this presentation is to outline the main trends in modes of transport for people who are economically active and for school children and students. The data used is the Census 2011 data in Scotland. The Scottish Census collects information on how does people usually travel to their main place of study including school. Firstly, we will outline the differences between the question asked in Scotland’s Census and the question asked in England and Wales’ Census; we will draw a comparison with the question asked in the 2001 Scotland’s Census. Then we will present the known data issues and their impact on the quality of the data. Finally, we will present the analysis which will comprise of: distribution of modes of transport at council area level; comparison with 2001 data; distribution by the Scottish Index of Multi Deprivation; distribution by the urban/rural indicator. This presentation will end with some guidance on how to access the data available for modes of transport from the 2011 Census website.

Email: daniela.ene@gro-scotland.gsi.gov.uk

Causes and consequences of international migration - Tuesday 9 September 11.00am

Quantifying the effects of visa restrictions on global migration flows
Adam Dennett, Centre for Advanced Spatial Analysis, University College London, Pablo Mateos, CIESAS, Mexico

Sovereign states have always exercised the right to control entry to their territory, even if for many hundreds of years this meant little as people tended to be born, live and die within a very small geographical area. But in an ever more populous and mobile world, passports and visa restrictions are paying an increasingly important role in affecting demographic evolution through controlling the volume and characteristics of international migration flows. Indeed the inequalities of who can move where are characterised by some as a form of ‘global apartheid’, and so understanding the geography and effects of visa and passport restrictions is as important as it is timely. This paper makes use new data on global visa restrictions, and explores the effects of full and partial restrictions on bilateral migration movements, measured through widely available migrant stock data and new migration flow estimates. These effects are examined through the calibration of both global and more localised spatial interaction models, with ‘local’ groupings of countries identified through the clustering of zones of restriction and corridors of free movement. Hitherto, global migration models (Cohen et al., 2008; Lewer and van den Berg, 2008; Bijak, 2008) have largely ignored the effects of political restrictions embodied in the complex network of global visa restrictions and access, and so this work offers new perspectives in this field.

Email: a.dennett@ucl.ac.uk

Does it matter why immigrants came here? Original motives, the labour market, and national identity
Stuart Campbell, Institute of Education, University of London

The importance of the original motives for migration has been asserted repeatedly in the economics of migration literature, but direct measures of such motives have seldom been included in empirical models of immigrant outcomes. I show that original motives for migration are important predictors of labour market participation, wages, and feelings of national identity among settled immigrants in the UK. Using new measures from a large survey sample, I find that those who originally came as work or student immigrants are the most likely to participate in the labour market, and that the employed among them also earn the highest wages. Those who originally came as family immigrants or refugees are less likely to participate, and, if employed, are paid less. However, refugees and family immigrants are the most likely to adopt the native national identity. These differences may partly be driven by a selected outflow of immigrants, but remain after accounting for region of origin, time spent in the country, and other relevant demographic and human capital characteristics. My results suggest that economic integration among immigrants does not always coincide with a new feeling of national affiliation: indeed, it is possible for the best economically adjusted immigrants to feel the most foreign in their adopted country, and for the most economically marginalised to feel the most at home.

Email: scampbell@ioe.ac.uk

The Long-Term Economic Impacts of Reducing Migration: the Case of the UK Migration Policy

Katerina Lisenkova, National Institute of Economic and Social Research Centre for Macroeconomics, Marcel Mérette, University of Ottawa, Miguel Sanchez-Martinez, National Institute of Economic and Social Research

This paper uses an OLG-CGE model for the UK to illustrate the long-term effect of migration on the economy. We use the current Conservative Party migration target to reduce net migration “from hundreds of thousands to tens of thousands” as an illustration. Achieving this target would require reducing recent net migration numbers by a factor of about 2. In presented simulations, we compare a baseline scenario, which incorporates the principal 2010-based ONS population projections, with a lower migration scenario, which assumes that net migration is reduced by around 50%. The results show that such a significant reduction in net migration has strong negative effects on the economy. The level of both GDP and GDP per person fall during the simulation period by 11.0% and 2.7% respectively. Moreover, this policy has a significant impact on public finances. To keep the government budget balanced, the labour income tax rate has to be increased by 2.2 percentage points in the lower migration scenario.

Email: k.lisenkova@niesr.ac.uk

Welfare Participation: A comparison between immigrants and natives within the United Kingdom
Nele van der Wielen, University of Southampton

International migration is a hotly debated topic. Romania and Bulgaria joined the European Union (EU) in 2007; however they were subject to transitory restrictions until 2014. Since 2014 the citizens of Bulgaria and Romania have permission to work in the United Kingdom (UK) and any other EU member states without restrictions. Using the Labour Force Survey from the fourth quarter of 2012 this research analyses benefit claims among Eastern European immigrants, immigrants from the old EU member states and natives in the UK. Particular focus is on different social economic factors that influence welfare participation. Binary logistic regression has been used to analyse individual-level correlates of any kind of benefit claim among UK born and immigrants, while multinomial logistic regression has been utilised to examine the impacts on different kinds of benefits claimed. Results of logistic regression show that, compared to natives, the social benefit claims are higher among immigrants from Eastern European countries that became member states of the EU in 2004. However, those immigrants have a smaller likelihood than natives to claim unemployment related benefit or income support indicating that the decision to migrate is not related to potential benefit support.

Email: n.van-der-wielen@soton.ac.uk

Migration and Mobilities: Spatial demography - Tuesday 9 September 1.30pm

Exploring change in population profiles across England and Wales, 2001-2011
Chris Lloyd, University of Liverpool

The paper focuses on the population profiles of Output Areas (OAs) within England and Wales. Differences between areas can be examined using multivariate statistical approaches such as principal components analysis (PCA). This paper analyses a set of demographic and socioeconomic variables (including age, housing tenure, National Statistics Socio-economic Classification (NS-SeC), limiting long term illness (LLTI), and ethnicity) for 2001 and 2011. Using standard PCA, in 2001 the strongest trends related to economic status. The first PC (explaining 39% of the variance) differentiates areas with high rates of car ownership, qualifications, owner occupied households and employment from areas with low rates of these. The second PC (22% of the variance) differentiates areas with younger age profiles from those with a large proportion of people identifying with a White ethnic group and higher rates of LLTIs. These trends were similar in 2011, although ethnicity (in terms of White/non White groups) was relatively less important in terms of how far it differentiated areas. These findings can be enriched by using geographical weighting schemes to modify PCA. In this way, it is possible to assess the relative importance of (combinations of) variables with respect to how far they differentiate areas. Analysis using geographically weighted PCA (GWPCA) shows that there are considerable differences in the features which distinguish neighbourhoods across England and Wales. The paper contributes to an understanding of some of the complex ways in which the population geography of England and Wales have changed between the two most recent Censuses.

Email: c.d.lloyd@liverpool.ac.uk

A multilevel model of origin to destination residential movement in England and Wales
Michael Thomas, John Stillwell, Myles Gould, School of Geography, University of Leeds

Using individual record data from a large national research opinion survey, this paper approaches the analysis of internal migration in a novel way, seeking to quantify the factors influencing the distance travelled by residential movers in England and Wales at different levels of analysis – the individual, neighbourhood and city region – using multilevel regression models. The results demonstrate that whilst micro-level variables such as ethnicity, educational attainment, occupation type and household income are significant influences, differences in origin and destination context, particularly at the level of the macrogeographic city region level, play an important additional role in enacting variations in the distance moved. When cast as an additive cross-classified origin and destination model a typical migrant is pulled over significantly longer distances towards rural/coastal (amenity-rich) city region destinations and, at the same time, is pushed significantly longer distances if the origin city region happens to be a metropolitan core. Therefore, building on previous aggregate based analyses of residential mobility in the UK, it is suggested that the findings reflect differences in place-based attractively, differences that appear to be a key motivating factor behind what is a continuing pattern of urban-rural shift. In conclusion it is suggested that such findings support the contention that residential mobility is increasingly a means through which a considerable number of people seek to satisfy their leisure, lifestyle and consumption desires; a situation which has driven, and continues to drive, a quite substantial redistribution of the population towards the amenity-rich environments of England and Wales’ coast and countryside.

Email: gymjt@leeds.ac.uk

Spatial variation in human rights violations and population distribution at the chiefdom level: A case study of Sierra Leone, 1985-2004
Amie Kamanda, University of Southampton

This paper examines the reasons behind the differences in percentage change in population size of chiefdoms from 1985 to 2004 in Sierra Leone by looking at human rights violations at this administrative level during the 1991-2002 civil war. Results from the country’s post-war 2004 population census showed large population gains in the south and west compared to substantial population loss in the east and north. Geospatial analysis and k-means cluster analysis of the Sierra Leone Truth and Reconciliation Commission dataset which has 40,242 counts violations is undertaken to address 'Why were there variations between the patterns of population distribution observed at the chiefdom level in Sierra Leone, 1985-2004?' Analysis of the data shows that forced displacement was among the most reported violation by war victims. Moreover, the violations occurred in three distinct phases. During Phase I, the conflict was confined in the south and east. Towards the end of the conflict, violations were concentrated in the north and east, and remained very low in the south. The k-means cluster analysis found a similar classification of the grouping of the different types of violations. These findings suggest that variations between patterns of population distribution are attributable to the geographical differences in the intensity of the conflict. High counts of violations in the east and north of the country in the final stages of the war prevented many displaced persons from resettling in their chiefdom of origin whereas low counts in the east and west, where the state had established control, enabled displaced persons to


Email: ak904@soton.ac.uk

Women who out-earn their partners: Regional differences across Europe
Agnese Vitali, University of Southampton, Bruno Arpino, Universitat Pompeu Fabra

Women who out-earn their partners represent a relatively new phenomenon in Europe. In this paper we study the contextual correlates of this phenomenon. In particular, we focus on the role of male unemployment rate and gender attitudes as possible explanatory factors of the variation in the prevalence of female breadwinner couple across European regions and countries. We employ multilevel models on data from the fifth round (2010) of the European Social Survey, complemented with data from the Eurostat database. Preliminary descriptive results show a high level of spatial heterogeneity across countries and across regions of the same country.

Email: A.Vitali@soton.ac.uk

Where is it? When is it? What is it like? What can we do with it?
Paul Norman, School of Geography, University of Leeds

You're the Cheshire cat, aren't you?” “I was the Cheshire cat,” he replied with a slightly aggrieved air, “but they moved the county boundaries, so technically speaking I'm now the ‘Unitary Authority of Warrington Cat’, but it doesn't have the same ring to it.” (Fforde 2002: 164) This work takes small area data from the 1971, 1981, 1991, 2001 and 2011 Censuses across GB and shows how boundary definitions have changed over time and demonstrates how the geographies can be harmonised to contemporary areas (Super Output Areas and Datazones) and how area measures (e.g. deprivation) can be calculated in a comparable way and therefore how areas change their characteristics over time (e.g.become less deprived than other areas). These changing area characteristics can be used to show how relationships with other variables changes over time (e.g. educational achievement, health) and can be linked to individual data to show how people’s trajectory across area types relates to their life course experiences. Fforde, Jasper (2002) Lost in a good book. Hodder & Stoughton: London

Email: p.d.norman@leeds.ac.uk

Migration and mobility: Internal migration and residential mobility - Tuesday 9 September 4.45pm

The changing propensity to move home in England and Wales, 1971-2011: A micro-level analysis
Tony Champion, Newcastle University, Tom Cooke, Connecticut, Kevin Lynch, ONS, Ian Shuttleworth, Queens Belfast

Studying how migration rates have altered over the longer term, and why, is pressing for policy reasons as well as for understanding social change. This paper uses the ONS Longitudinal Study of England and Wales to examine the propensity of people to change address between censuses and compares these 10-year address change rates across the four decades from 1971-1981 to 2001-2011. The descriptive results are presented both for the aggregate population and for a selection of population subgroups that can be identified on a consistent basis across the censuses involved. They are also broken down by the distance between start and end of decade addresses. The headline result is that the overall rate of 10-year address changing has dropped by almost one-fifth between the 1970s and the 2000s – from 55% to 45% of all the ‘decade survivors’. This decline in rate is almost entirely due to the reduced rate of moving 10km or less, whereas as indicated by the NHSCR-derived data the rates of longer-distance address changing have held up, unlike in the USA. Even so, this latter finding is significant because a rise in longer-distance migration might have been expected from the population shift into composition towards higher-skill (and traditionally more migratory) occupational groups. Almost all population groups have shared in the decline in rates, the main exceptions being those starting the decade as private renters and those living in communal establishments. Regression analysis and Oaxaca-Blinder modelling are then used to gauge the separate contribution of each personal attribute.

Email: tony.champion@ncl.ac.uk

Residential mobility through the lifecourse: an exploration using the British Cohort Study 1970
Nissa Finney, University of Manchester, Naomi Tyrrell, Plymouth University

Children today are growing up in a mobile society. Yet relatively little is known about the characteristics of child migrants in the UK (compared with those who do not migrate), or about the impacts of child migration on later life. This paper first lays out a theoretical framework for understanding migration from a connected lifecourse perspective. It develops the concept of migration capital and proposes that this is accumulated in childhood through migration experiences and can provide advantages in later life. The paper then goes on to test this notion by addressing 3 research questions: 1. What are the characteristics of residentially mobile and immobile children (or, who develops migration capital in childhood, and who does not?); 2. Is residential mobility during childhood associated with residential mobility in later life (or, does the acquisition of migration capital in childhood lead to a greater propensity to migrate in adulthood)?; 3. Can differences be observed in later life wellbeing between those with residentially mobile life courses and those with residentially immobile life courses (or, is migration capital associated with positive life outcomes)? The paper uses the British Cohort Study 1970, examining the residential mobility of the sample in their childhood in the 1970s and their residential mobility in adulthood to age 42; and their wellbeing in adulthood. In providing a fresh theoretical lens and original longitudinal analysis, thereby responding to calls for progressive approaches to family migration and the incorporation of life course theories in migration scholarship, this paper aims to stimulate debate about the direction of the research agenda for studies concerned with migration and lifecourse.

Email: Nissa.Finney@manchester.ac.uk

Ethnic variations in the association between residential mobility and cardiovascular disease hospitalisations in Auckland, New Zealand
Daniel J Exeter, Arier Chi Lun Lee, Tania Riddell, Corina Grey, Sue Wells, School of Population Health, The University of Auckland, Matire Harwood, Te Kupenga Hauora Māori, The University of Auckland

Numerous studies from New Zealand and elsewhere have demonstrated that residents living in the most disadvantaged communities experience a disproportionate burden of poor health conditions. Many of the studies that draw these conclusions use cross-sectional data and fail to account for the role that selective migration has on health. Every New Zealander has a unique national health identifier (NHI) that is used to record their interactions with the state-funded health system. In this study, we used encrypted the NHIs and link data recorded in routine health datasets (e.g. Primary Health Organisation (PHO) registrations, pharmaceutical dispensing, hospitalisations and mortality) to construct a cohort of approximately 670,000 patients aged ≥30 years old living in Auckland, between 1/1/2006 and 31/12/2012, to investigate ethnic disparities in cardiovascular disease (CVD) hospitalisations. Residential mobility was measured using the census Meshblock (MB – similar in size to OAs in the UK) of usual residence from the PHO database for every calendar quarter between 1/1/2006 and 31/12/2012. The ethnicity of patients was obtained from the PHO database and prioritised into 5 groups: Maori, Pacific, Asian, Indian, and New Zealand & Other Ethnicities combined (NZEO). We describe the geography of residential mobility in the Auckland Region, outlining the demographic differences between those individuals who remained living in the same neighbourhood (MB) and those who moved during our study period. Using a measure of area-level deprivation (NZDep2006), we investigate the interaction effects that ethnicity and deprivation mobility have on the odds of movers being hospitalised for CVD compared to stayers.

Email: d.exeter@auckland.ac.nz

Densification, Displacement and/or Development: short and long term impacts of new migrants from rich/ poor countries on the population growth in London and across the Greater South East
Ian Gordon, London School of Economics

International and internal migration flows have tended to be studied quite separately, though evidence has accumulated since the 1990s of strong interdependence between them. Specifically studies of gateway regions have pointed to strong net outflows from these to other regions in periods of large scale immigration. A common expectation is that these might reflect displacement effects within spatial labour markets – or else some version of 'white flight'. In this paper, which focuses on the experience of London and its extended region (the Greater South East) in recent decades, the emphasis is, however, on (chains of) displacement within overlapping spatial housing markets. In each of these markets the impacts are partitioned between: induced increases in the accommodation stock (development); more intensive occupation of the existing stock (densification), and altered patterns of household movement (displacement). Within the core city-region, the densification and displacement are found to be the dominant responses, though their relative importance is radically different in relation to inflows of migrants from rich or from poor countries, and to shift dynamically over time. Densification is found to be the prime cause of recent population growth within the core (i.e. Greater London), though its specific association with recent inflows from poor countries raise questions about its longer term sustainability as a mode of accommodating growth there. Results are presented from analyses both of LSOA level changes between the 2001 and 2011 Censuses and of regional level time series from 1981-2011.

Email: I.R.Gordon@lse.ac.uk