The Household Impacts of Migration

Transcription

1 Paper to be presented at The 2001 British Household Panel Survey Research Conference, Colchester, July 5-7, 2001 The Household Impacts of Migration Tania Ford*, Glen Bramley* and Tony Champion** DRAFT VERSION ONLY- NOT FOR QUOTATION WITHOUT PERMISSION OF THE AUTHORS Contact Details: * School of Planning and Housing, Edinburgh College of Art/Heriot-Watt University, Edinburgh, UK, EH1 2HJ. T.Ford@eca.ac.uk /G.Bramley@eca.ac.uk **Department of Geography, University of Newcastle, Newcastle upon Tyne, UK, NE1 7RU. Tony.Champion@ncl.ac.uk.

2 2 Introduction It is generally recognised that any household change other than a birth or a death implies some residential mobility. The creation of a new household is often accompanied by geographical mobility, as when a child leaves the parental home to form a new household or when a couple parts and at least one partner leaves to establish a separate household. Mobility and household transitions are closely linked but difficult to investigate and many studies of household change have largely ignored their spatial dimension and geographical impact. One reason for this has been the lack of appropriate longitudinal data, which follows people over time. Analysis of people's living arrangements has had to rely primarily on associations derived from crosssectional 'snapshots' to make inferences about these processes. This paper begins to explore the relationship between migration, household formation or dissolution based on the analysis of the British Household Panel Survey (BHPS). The BHPS is unique in that it allows identification of the patterns of household formation and dissolution and the types of moves involved. Here we aim to identify the extent to which migrant households are distinctive as compared with continuing resident households and in doing so, estimate the household impacts of migration. The next section reviews the literature dealing with migration and household formation and this is followed by a brief description of the data. A typology of household transitions is then developed and adopted in an attempt to begin to assess the household impacts of migration. The paper concludes with a summary of the findings and issues for further research. Migration and Household Formation: The Literature Much of the current research treats migration as an outcome of household change. Often migration is seen as a possible by-product of a shift in the structure of the household. It is generally recognised that the creation of a new household is in many cases accompanied by geographical mobility and much of the current literature focuses on particular groups involved ie. young people leaving the parental home, divorced people, elderly people. There appears to be very little which looks at the relationship from a migration point of view. Many studies focus on particular groups involved in household change, hence addressing the distinct influences and behavioural patterns associated with each group. A large body of research deals with the issue of young people leaving the parental home. Buck and Scott (1993) use event history analysis to model the differences in the age-specific probabilities of leaving home in the US, while Murphy and Wang (1998) use the British Household Panel Survey (BHPS) to address children leaving the parental home from the viewpoint of the family rather than that of the child. Other studies which focus on young people's household formation include Ermisch (1999), Ermisch et al 1995; Berrington and Murphy 1994; Goldschieder and DaVanzo 1989). These studies address a variety of issues associated with the type of situation young people leave the parental home to move into, such as marriage or cohabitation, student accommodation or to live alone. A spatial dimension is introduced to some extent, in a West German study (Mayer and Schwartz 1989) which addresses the distance young people moved after leaving the parental home.

3 3 A further set of literature examines household transitions in the middle years of life, such as cohabitation, divorce and separation. Several studies deal with effects of marriage breakdown and divorce (Grundy 1985; 1992; Sullivan 1986; Holmans et al 1987). Grundy (1992) argues that marital breakdown almost invariably involves a move for at least one partner, however migration itself increases the risk of marital breakdown and the formation of a new union following divorce may often involve another move. Warnes (1992) reports that long distance moves are most likely among married couples, those with dependent children and those who have been most mobile earlier in life. Flowerdew et al (1999) explore household changes accompanying migration of divorced people, and compares the circumstances of divorced and other migrants. Although these studies do examine household transitions, it is suggested (Clark and Onaka 1983; Murphy 1996) that the growth of single-person and nontraditional households has brought into question the applicability of a standard household history. The elderly are the focus of a further set of literature. Glaser and Grundy (1998) conclude that older people living alone or in complex households were more likely to move than those in other types of living arrangements. Hayes and Al-Hamad (1997) examine co-resident households which have been formed by the movement of people under the age of 65 yrs into the homes of the elderly and Al-Hamad et al (1997) explore the extent to which migration of the elderly is linked to changing family living situations. Distance of move is addressed and they conclude that most moves are short distance, but long-distance moves made by elderly people are mainly to join children, with shorter distance moves dominating other relationship groups. Various studies (Burch and Matthews 1987; Haurin et al 1993; Hendershott and Smith 1989; Odland and Ellis 1988) address the economic issues involved with household formation and dissolution. Mincer (1978) adopts an economic definition of family ties relevant to migration decisions and concludes that family ties deter migration and that migration rates diminish with distance. Ermisch and Overton (1985) develop a new method for the analysis of household formation, which suggests that the probabilities of sharing or of forming separate households relates to the structure and composition of individual household types and characteristics of their members. Although they do distinguish geographical migration as an important process, the analysis is cross-sectional only focusing on the socio-economic determinants of household formation and fission, using Minimal Household Units (MHUs) as units of analysis. Further, Ermisch (1988) argues that economic variables can influence household formation indirectly by influencing the composition of MHUs. Odland (1981) establishes a micro-economic approach to destination choice and argues that changes in circumstances of the household can cause destinations to be added or deleted from the set of chosen destinations. Much of the literature acknowledges that mobility and household transitions are closely linked but difficult to investigate and many studies of household change have largely ignored the spatial dimension and geographical impact. Notable exceptions include Ogden and Hall (2000) who see migration as playing an important mediating role in household change and Hall et al (1999) who argue that migration plays a key mediating role both in producing and reflecting household change and as a major contributor to their geography. In the Dutch context, Ter Heide and Scholten (1988) recognise a direct connection between migration and household formation and

4 4 dissolution and Grundy (1992) addresses the link between household formation and migration behaviour. From a methodological viewpoint, both Duncan and Hill (1985) and Laurie and Sullivan (1991) argue that over time the composition of households change significantly and often the most dramatic changes such as divorce or children leaving home produce equally dramatic changes in economic status and geographic location. Both argue that attempts to define 'longitudinal households' as households that do not change is problematic and will result in significant bias and loses sight of the dynamics of change which such a study is expressly designed to identify. Various studies have utilised panel data such as the BHPS to gain a longitudinal perspective (Boheim and Taylor 1999; Murphy and Wang 1998; Ermisch and Francesconi 1999; Ermisch et al 1995). Most of the work to date (Buck et al 1994, Buck 2000) treats the individual as the primary unit of analysis. Household characteristics are attached to the individual in an attempt to explore changes in household structure over time. As a longitudinal study, the BHPS provides a unique source for identifying and quantifying the frequency of the various different kinds of moves which contribute to migration and household formation, dissolution and transformation. The British Household Panel Survey Longitudinal data such as the BHPS provide a rich source of information for exploring transitions over time. Where cross-sectional data provide a 'snapshot' of the population, longitudinal data offer the possibility of a 'movie' (Berthoud 2000). The BHPS was set up with the aim of providing a nationally representative sample of the population of Great Britain living in private households in 1991 (Buck et al 1994). It consists of an original sample of around 5,000 households interviewed anually from 1991 to As a longitudinal survey, moves by all individual members of the sample of households are tracked. If individual members of the sample households move, they are followed into their new households and in the following year interview all members of the household, whether or not they were members of the original sample. This makes the BHPS a unique source for identifying and quantifying the frequency of the various different kinds of moves which contribute to migration and household formation, dissolution and transformation. Furthermore, the panel data facilitate the study of the sequence of household moves and provide important information on the events associated with each change. While this source does suffer from certain limitations of sample size, sample attrition and geographical disaggregation, it is capable of facilitating much useful analysis of migration flows. This paper uses the BHPS to examine the links between migration and household change in Britain in the 1990s. The analysis focuses on changes within two pairs of BHPS waves: waves 2-3 (1992 to 1993) and waves 6-7 (1996 to1997). Issues of missing information and sample attrition in the first and last waves led to this choice. Nevertheless, these two periods of 4-5 years apart represent conditions of relative economic recession and buoyancy respectively. Household transitions are defined on the basis of moves between addresses and household groups between successive waves. Migration is defined in this context as one-year moves between SAR district areas, the lowest spatial unit identifiable in the BHPS. This comprises a total of 278 zones in Great Britain.

5 5 The Typology of Transitions To facilitate a clear descriptive account of the relationship between migration and household formation, the aim of this section is to set up a basic scheme which can account for the main types of individual/household moves/changes. It is suggested that household transitions can be broken down into a number of categories and that these are best considered in terms of origins and destinations. Each origin category can be linked to several alternative destination categories and the net household change resulting from the transition will be the combination of the effects at the origin and destination. These may be in the same or different geographical zones. The scheme itself is relatively simple, recognising five origin and five destination categories. Nevertheless, the number of origin/destination combinations is still considerable, totalling 13 in this case. The five origin transition categories are as follows: Dissolution(DIS): a previously existing household ceases to exist (in the same form) because all of the former members have either exited the system (emigration or moved to institution) or moved separately (ie.excludes wholly moving households). Split Households (SPLIT): when an adult or group leaves an existing household and the remainder of that household continues to reside at the same address. Departure (DEP): when the household moves out of its previous address. Remains (REM): when the household remains in residence at the same address. Appearance (APP): when an individual adult or group enters the private household population through immigration, de-institutionalisation or attaining the age of 16 (the latter group are identified separately as APP16). The five destination categories are as follows: New Household (NEW): when one or more individual adults form a new household unit (all moving to do so). Joining a Household (JOIN): when an individual adult or group moves to join an existing household at its current address. Arrival (ARR): when a whole pre-existing household moves to a new address. Remains (REM): when a household remains in residence at the same address (this is the same as the corresponding origin category). Exit (Exit): when an individual adult leaves the private household population through death, institutionalisation or emigration. It should be noted that, whereas the origin and destination categories involve households, the combined transitions focus on individuals or groups of individuals who move together. This is consistent with the analysis of migration, which deals

6 6 primarily with individuals, but relates individual moves to the household circumstances at origin and destination. It is also consistent with the nature of panel data such as the BHPS, where individual moves are related to their household context pre- and post-move. A group is treated as the same household if it stays at the same address or if it moves as a whole. The BHPS collects limited information about the previous circumstances of individuals (temporary sample members-tsms) who join the original survey members (OSMs), or about TSMs if and when they subsequently leave the OSM's household. In addition, there is some attrition over time in the original sample, so that some members disappear without definite information as to where they have gone. Thus, the data we are really analysing is the BHPS sample as a system, rather than the whole private household population of England. This means that the categories of APP and EXIT are more widely defined than was originally conceived. In the case of APP, they now include people joining sample households from elsewhere in the private household population and in the case of EXIT people leaving the sample but possibly remaining in the private household population of England. The main transitions involving combinations of the five origin and five destination categories are listed in Table 1. The total sample population are allocated to each transition in the two interwave periods and and also the share within each transition group who were migrants (ie. moving between areas) in the second period. We are only interested in adults (aged 16+yrs), because individually or in groups, adults generally form the building blocks for households. Some adults will have children attached to them, but we are not interested here in tracing moves by children separately. In effect, the presence of children is a secondary attribute of individuals or family units, which may help to explain their behaviour. It is expected that some of these transitions will be more common than others across the sample population. The most common transition, accounting for around threequarters of adults in both periods (73.1 percent in both periods, although these are not necessarily the same people) is REM-REM, where no-one in the household moves. Wholly moving households are also quite common (DEP-ARR), accounting for around 6 percent in both periods. Many of the other transitions generate quite small numbers in any one year. The next most common transition is SPLIT-REM, where part of a splitting household remains in situ. The other parts of splitting households divide between NEW, JOIN and EXIT. In the earlier period, new entrants joining existing groups (JOIN) appear to be more common. However, in the second period, new entrants appear more likely to go into NEW households. Despite these differences, the incidence of the different transitions appear relatively stable between the two periods. It is important to note that the typology probably exaggerates the number of dissolutions and new households because of the treatment of TSMs in the BHPS. A common problem with migration data stems from the inability to identify multiple moves within any single year (or between waves) and hence must be treated as one move. Table 1 suggests that migrations only occur in four of the transitions (DEP- ARR, DEP-NEW, SPLIT-NEW and DIS-NEW), although in a small number of SPLIT-JOIN cases also migrated. However, the BHPS does not allow us to measure all migrations, where the origin area of APP cases is not known and in any cases of EXIT (some of whom might migrate to institutional accommodation and

7 7 others of whom might migrate at the time of disappearing from the sample through non-response or non-tracing). Migration is most common for the SPLIT-NEW transitions, where it applies to 50 percent of cases in both periods. In migrants made up 53 percent of adults in DEP-NEW transitions and 49 percent in DIS-NEW transitions, but these proportions decline to 33 percent and 38 percent respectively in Clearly, the share of migration in transitions varies over time. In terms of the geographical impact of these household transitions, in the later period, SPLIT-NEW transitions may be characterised by more urbanising movements such as young people leaving the parental home to set up new separate households. On the other hand, counterurbanising moves are more common among DEP-ARR cases (wholly moving households) and DIS-NEW transitions (household break-up and reformation). However, these patterns are not entirely consistent across the two periods. Inter-regional migrations are more common for SPLIT-NEW transitions and slightly more common for DEP-ARR cases, while intra-regional moves are more common for the other migrant transitions. Table 1: Main Household Transitions: and Description % all 16+yrs % all 16+yrs % of group migrating Transition group REM > REM Continuing unchanged household REM > JOIN Continuing household joined by others DEP > ARR Whole Household moves DEP > NEW Whole Household moves to form new household with other mover(s) SPLIT > REM Part of splitting household remaining in situ SPLIT > NEW Part of splitting household moving to form new household with other mover(s) SPLIT > JOIN Part of splitting household moving to join other non-mover(s) SPLIT > EXIT Part of splitting household leaving the system DIS > NEW Part of dissolving household moving to form new household with other mover(s) DIS > EXIT Part or all of dissolving household leaving the system APP > NEW New entrant to BHPS forming new household with moving BHPS OSM(s) APP > JOIN New entrant to BHPS joining non-moving BHPS OSM(s) APP16 > REM Attaining age of 16 within BHPS sample and remaining in situ Total (n) 10,108 10, Source: calculated from the BHPS Note: Only the main transition groups are listed. The final column refers to the proportion of adults in the transition group that moved between, as opposed to within an area. A key advantage of the BHPS is that it allows analysis of the socio-demographic characteristics of the transition groups, whilst also distinguishing migrants within selected transition groups. Table 2 shows the socio-demographic profile of the main

8 8 transition groups based on the characteristics in wave 7 (1997) post-transition. Perhaps most relevant for the purposes of this paper are the differences between migrants and non-migrants within transition categories. Table 2: Main Household Transitions-General Characteristics and Differences for Migrants, Transition group REM > REM REM > JOIN DEP > ARR DEP > NEW Socio-demographic characteristics (after transition) All Household types; older; owner occupiers Couple, lone parents, unrelated & other adults; higher income; higher SEG; working; owner occupiers Singles; couple families; owners & private renters Childless couples, unrelated adults; higher SEG; working; HE qualifications SPLIT > REM Couple, older children; lone parent; older; higher SEG; working SPLIT > NEW Single, childless; unrelated adults; younger; smaller households; lower income; private renters SPLIT > JOIN Lone parents, unrelated & other adults; larger households; lower SEG SPLIT > EXIT NA DIS > NEW Single; childless, unrelated adults; younger smaller household; lower income; private renters DIS > EXIT NA APP > NEW Childless couples, unrelated adults; ethnic; private renters APP > JOIN Childless couples, unrelated adults; ethnic; private renters; HE qualifications APP16 > REM Couple with children, lone parents; very young in large higher income households; lower SEG (individual); not working parents; ethnic; owner occupiers Source: calculated from the BHPS Differences for Migrants Older; higher incomes &SEG; less working; more HE qualifications; more private renters Younger; smaller households; lower SEG; less working; less HE qualifications; more private renters Lower income; less working; more private renters Older, smaller households; higher incomes & SEG; more working; more HE qualifications Among wholly moving households (DEP-ARR), migrants tend to be older and better off in terms of income, occupation and qualifications, although they are less likely to be working and more likely to be renting privately immediately after the move. People moving into new households (DEP-NEW) are likely to be younger, living in smaller households with lower occupations and qualifications. They are also less likely to be working and more likely to rent privately. Migrants within the SPLIT- NEW group are likely to be young people leaving the parental home, and tend have lower incomes, are less likely to be working and more likely to rent privately. For the DIS-NEW group, migrants are older, living in smaller households, more often working and with higher occupations and qualifications. From this brief description it

9 9 becomes apparent that across all transitions, there is no a clear association between socio-demographic attributes and migration. Household Impacts of Transitions It is possible to broadly quantify the impacts of these household transitions. Transitions originating with a DEP or a DIS result in a reduction of households at the origin, whilst transitions ending with an ARR or a NEW at the destination result in an increase in households at the destination. The number of households disappearing or appearing in each case can be calculated by dividing by the number of adults present in the relevant household. Table 3 shows the results of this analysis by transition category, distinguishing migrant from non-migrant transitions. Table 3: Household Changes Based on Transitions and Migration Estimated for England, Transition Group & Migrant Status Origin Change 000s Destination Change 000s Net Change 000s Net Change Nonmigrants % Net Change Migrant % Net Change Total MISC Non-migrant Migrant REM > REM REM > JOIN DEP > ARR Non-migrant Migrant DEP > NEW Non-migrant Migrant SPLIT > REM SPLIT > NEW Non-migrant Migrant SPLIT > JOIN SPLIT >EXIT DIS > NEW Non-migrant Migrant DIS > EXIT APP > NEW APP > JOIN APP16> REM TOTAL Non-migrants TOTAL Migrants TOTAL ALL Source: calculated from the BHPS Note: Final three columns show the estimated household changes as a proportion of all existing households in England (set at 20 million). Migrant refers to between-area movers, non-migrant to within-area movers. Numbers are roughly grossed up to correspond to an English total of c.20 million households, to give a feel for magnitudes, with net changes also expressed as

10 10 percentages of existing households. As previously mentioned, a precisely accurate account may not be given for some transitions (eg. EXIT). Table 3 includes for completeness a category of MISC referring to minor transitions and transitions involving origins and destinations not allocated in the typology of main transitions used in Tables 1 and 2. It is clear from Table 3 that although wholly moving households (DEP-ARR) generate large numbers of household disappearances at their origin, there are an equal number of appearances at their destination, giving no total net change. For this transition group, non-migrant households outnumber migrant households by around three to two. DIS-EXIT removes 354,000 households from the system, while APP-NEW generates a similar number of 370,000 new households. Both of these figures may be overestimates. Departure to join other movers in new households (DEP-NEW) generates a small net loss in household numbers, while dissolution followed by moves into new households with other movers (DIS-NEW) generates a positive net impact on household numbers, the majority of whom are non-migrants. Leaving an existing household to form a new household (SPLIT-NEW) generates a large number of extra households (431,000), but with only a small majority migrants. Overall, the net impact of migrant transitions is slightly greater than the net impact of non-migrant transitions. Nevertheless this analysis is entirely exploratory and the total net household impact derived from this analysis is considerably greater than the typical average estimated growth in household numbers in England. The impact of household transitions also varies geographically and hence it is important to address the spatial impact. Some insight can be gained from an analysis of net household changes generated by transitions across the simple six level urban hierarchy shown in Table 4. Table 4: Net Household Change Based on Transitions by Urban Hierarchy Level, and Urban Hierarchy Level Net % Change Net % Change Inner London/Principal Metropolitan Cities Outer London/Other Metropolitan Districts Large Non-metropolitan Cities Other Urban Mixed Urban-rural Rural/Coastal Source: Calculated from the BHPS Note: % refers to the proportion of existing households in BHPS. The six urban hierarchical levels are aggregates based on those used by OPCS/ONS (see OPCS 1991): 'Other urban' comprises Small Non-metropolitan Cities, Industrial Districts and Districts with New Towns. 'Rural/Coastal' comprises Remoter Mainly Rural Districts and Port Resort & Retirement Districts. This shows a generally familiar picture of net changes favouring areas further down the urban hierarchy. However, there are some notable changes between the two interwave periods. Inner London and the Principal Metropolitan Cities showed a considerable increase in their net loss of households between the two periods. Although Outer London and Other Metropolitan Cities were losing households in the

11 11 first period, they were gaining households in the second period. While Other Urban areas and Rural/Coastal increased their net gains in the second period, Mixed Urbanrural areas (the largest group) saw a decline in the later period. Conclusion Not only is migration selective as to the type of household, but there is also a direct relationship between migration and household formation and dissolution. Except in the case of a single person, household formation and dissolution always involves one or more persons moving house. If such a move takes place across area boundaries, migration takes place. The relationships between individuals and groups are the essence of household formation and insights into the processes and causes of household changes are crucial. The analysis presented here has attempted to shed some light on these issues by providing a unique way of looking at household change. The analysis presented in this paper has been entirely exploratory, but nonetheless informative. Whilst only beginning to explore the household impacts of migration, a key conclusion of this analysis is that migration does increase the possibility of new household formation. Furthermore, not only do the characteristics of migrants vary across all household transitions, but the impact of household transitions also varies geographically. These conclusions highlight the need for more descriptive analysis and modelling to increase the explanatory value and to account for causal factors. Consideration of both the causes and the consequences of differences in household formation between migrants and the remaining population, requires a distinction to be made between selection effects and behavioural changes. It is well documented that migration is a highly selective process, influenced by such characteristics as age, sex, and stage in the life cycle of the individual. It is argued that the differences between migrants and non-migrants influence other aspects of behaviour such as household formation propensities. The argument suggests that areas that experience significant in-migration will accumulate people with these traits, while areas exporting substantial migrants will experience a deficit of such people. Hence further analysis dealing with these issues is crucial if we are to gain a clearer understanding of the household impacts of migration.

12 12 References Al-Hamad. A., Flowerdew, R. and Hayes, L, 1997, Migration of Elderly People to Join Existing Households: Some Evidence from the 1991 Household Sample of Anonymised Records, Environment and Planning A, 29, pp Berthoud, R, 2000, 'Introduction: The Dynamics of Social Change' Chapter 1 in Berthoud, R and Gershuny, J (eds), 2000, Seven Years in the Lives of British families: Evidence on the Dynamics of Social Change from the British Household Panel Survey, The Policy Press, Bristol. Berrington, A. and Murphy, M, 1994, Changes in the Living Arrangements of Young Adults in Britain During the 1980s, European Sociological Review, Vol. 10 (3), pp Boheim, R. and Taylor, M, 1999, Residential Mobility, Housing Tenure and the Labour Market in Britain, Report of the Institute for Labour Research programme on 'Labour Market Dynamics in a Changing Environment' funded by the Leverhulme Trust. Buck, N. and Scott, J, 1993, She's Leaving Home: But Why? An Analysis of Young People Leaving the Parental Home, Journal of Marriage and the Family, 55, pp Buck, N. Gershuny, J, Rose, D. and Scott, J,(eds) 1994, Changing Households: The BHPS 1990 to 1992, University of Essex, Colchester. Buck, N, 2000, 'Housing, Location and Residential Mobility' Chapter 6 in Berthoud, R and Gershuny, J (eds), 2000, Seven Years in the Lives of British families: Evidence on the Dynamics of Social Change from the British Household Panel Survey, The Policy Press, Bristol. Burch, T.K. and Matthews, B.J, 1987, Household Formation in Developed Societies, Population and Development Review, 13(3), pp Clark, W.A.V. and Onaka, J.L, 1983, Life Cycle and Housing Adjustment as Explanations of Residential Mobility, Urban Studies, Vol. 20, pp DaVanzo, J. and Goldsheider, F.K, 1990, Coming Home Again: Returns to the Parental Home of Young Adults, Population Studies, Vol. 44, pp Duncan, G.J. and Hill, M.S, 1985, Conceptions of Longitudinal Households: Fertile or Futile?, Journal of Economic and Social Measurement, Vol. 13, pp Ermisch, J.F. and Overton, E, 1985, Minimal Household Units: A New Approach to the Analysis of Household Formation, Population Studies, 39, pp Ermisch, J, 1988, An Economic Perspective on Household Modelling, Chapter 3 in Keilman, N., Kuijsten, A. and Vossen, A. (eds), Modelling Household Formation and Dissolution, Clarendon Press, Oxford.

13 13 Ermisch, J, 1999, Prices, Parents and Young People's Household Formation, Journal of Urban Economics, 45, pp Ermisch, J. and Francesconi, M, 1999, Cohabitation in Great Britain: Not for Long, but Here to Stay, Institute for Social and Economic Research, University of Essex, Colchester. Ermisch, J., Di Salvo, P. and Joshi, H, 1995, Household Formation and Housing Tenure Decisions of Young People, Occasional Paper 95-1ESRC Research Centre on Micro-Social Change, Essex. Flowerdew, R., Al-Hamad, A. and Hayes, L, 1999, The Residential Mobility of Divorced People, in McRae, S, (ed), Changing Britain: Families and Households in the 1990s, Oxford University Press, Oxford. Glaser, K. and Grundy, E, 1998, Migration and Household Change in the population Aged 65 and Over, , International Journal of Population Geography, 4, pp Goldshieder, F.K. and DaVanzo, J, 1989, Pathways to Independent Living in Early Adulthood: Marriage, Semiautonomy and Premarital Residential Independence, Demography, 26 (4), pp Grundy, E, 1985, Divorce, Widowhood, Remarriage and Geographic Mobility Among Women, Journal of Biosocial Science, 17, pp Grundy, E, 1992, The Household Dimension in Migration Research, in Champion, T. and Fielding, T (eds), Migration Processes and Patterns: Vol. 1. Research Progress and Prospects, Belhaven Press, London. Hall, R., Ogden, P.E. and Hill, C, 1999, Living Alone: Evidence from England and Wales and France for the Last Two Decades, in McRae, S, (ed), Changing Britain: Families and Households in the 1990s, Oxford University Press, Oxford. Haurin, D.R, Hendershott, P.H. and Kim, D, 1993, The Impact of Real Rents and Wages on Household Formation, The Review of Economics and Statistics, 75 (2), pp Hayes. L. and Al-Hamad, A, 1997, Residential Movement into Elderly Person Households: Evidence from the 1991 Household sample of Anonymised Records, Environment and Planning A, pp Hendershott, P.H. and Smith, M.T, 1989, Transfer Programs and Aggregate Household Formations, Population Research and Policy Review, Vol. 8, pp Holmans, A.E., Nandy, S. and Brown, A.C, 1987, Household Formation and Dissolution and Housing tenure: A Longitudinal Perspective, Social Trends, 17, pp

14 14 Laurie, H. and Sullivan, O, 1991, Combining Qualitative and Quantitative Data in the Longitudinal Study of Household Allocations, The Sociological Review, Vol. 39(1), pp Mayer, K.L. and Schwarz, K, 1989, The Process of Leaving the Parental Home: Some German Data, in Grebenik, E, Hohn, C. and Mackensen, R (eds), Later Phases of the Family Cycle: Demographic Aspects, Clarendon Press, Oxford. Mincer, J, 1978, Family Migration Decisions, Journal of Political Economy, 86 (5), pp Murphy, M, 1996, The Dynamic Household as a Logical Concept and its Use in Demography, European Journal of Population, Vol. 12, pp Murphy, M. and Wang, D, 1998, Family and Socio-Demographic Influences on Patterns of Leaving Home in Postwar Britain, Demography, Vol. 35 (3), pp Odland, J, 1981, A Household Production Approach to Destination Choice, Economic Georgraphy, Vol. 57(3), pp Odland, J. and Ellis, M, 1988, Household Organization and the Interregional Variation of Out-migration Rates, Demography, Vol25 (4), pp Ogden, P.E. and Hall, R, 2000, Households, Reurbanisation and the Rise of Living Alone in the Principal French Cities, , Urban Studies, 37(2), pp OPCS, 1991, Key Population and Vital Statistics: Local Authority Areas. Population and Vital Statistics by Area of Usual Residence in England and Wales, Office of Population Censuses and Surveys Series VS, No. 17, PP1, No. 13. HSMO, London. Sullivan, O, 1986, Housing Movements of the Divorced and Separated, Housing Studies, Vol. 1, pp Ter Heide, H. and Scholten, H, 1988, Applications of Household Models in Regional Planning, Chapter 14 in Keilman, N., Kuijsten, A. and Vossen, A. (eds), Modelling Household Formation and Dissolution, Clarendon Press, Oxford. Warnes, T, 1992, Migration and the Life Course, in Champion, T. and Fielding, T (eds), Migration Processes and Patterns: Vol. 1. Research Progress and Prospects, Belhaven Press, London.