الصفحات

الأربعاء، 14 مارس 2018

The life course and residential mobility in British housing markets



The life course and residential mobility in British housing markets

William A V Clark

Department of Geography, University of California, Box 951524, Los Angeles, CA 90095-1524,USA; e-mail: wclark@geog.ucla.edu

Youqin Huang

Department of Geography and Planning, State University of New York, Albany, NY 12222,USA; e-mail: yhuang@albany.edu

Environment and Planning A 2003, volume 35, pages 323 - 339

Abstract. There is a substantial research literature on residential mobility in general, and the role of housing space in triggering moves in particular. The authors extend that research to mobility in British housing markets, using data from the British Household Panel Survey. They confirm the applicability of the general residential mobility model and also confirm the value both of pooled cross-sectional and of true longitudinal models of residential change. Age, tenure, and room stress (housing-space requirements) are found to be significant predictors of moving. In addition, the life course `triggers' of marital-status change and, in some situations, birth of a child play important roles in moving within housing markets in the United Kingdom. The same model, with somewhat lower levels of fit, is also significant for the London region. Variables that measure the desire to move and neighborhood satisfaction also play a role in predicting local moves: those who like their neighborhood are generally less likely to move. The results offer support for the view that residential mobility is a demographically driven process which also reflects the connection with neighborhood contexts.

Introduction 

   By now a large and well-established body of research has documented the process of residential mobility, and the place that the search for more housing space plays in that process. We know how the search for more housing space is intertwined with lifecourse `triggers', which in turn create the need for greater housing consumption. Specific triggers, such as a change in marital status or the birth of a child, are important life-course events, which are in turn translated into the acquisition of larger and often more expensive dwellings.

   Residential moves are in general short-distance moves which do not disrupt general patterns of living, and are, with few exceptions, within a specific labor and housing market. Unlike interregional moves which are often prompted by a change in job, most residential moves are prompted by housing considerations, though there is evidence that jobs may change along with changes in residential location, even within relatively restricted labor and housing markets.

  Previous studies of residential mobility have examined the process in the United States, Europe, and New Zealand and Australia. Because British housing markets have a higher proportion of public housing, are more similar to the Dutch context than that in the United States, and because the UK housing market underwent a major restructuring with the privatization of council housing, it is an interesting context within which to examine the applicability of the standard residential-mobility model. In addition, the data from the British Household Panel Survey make it possible to carry out a further test of the comparative value of pooled cross-sectional and longitudinal models of residential choice.

Background and context 

   An extensive literature has documented the relationship between household size and the space that the household `consumes'. As households increase in size they require more space, and residential mobility is the process whereby households adjust household size to the housing stock. Research in the United States and Holland (Clark and Dieleman, 1996; Clark et al, 2000), Germany (Clark and Drever, 2001; Frick, 1996) and, to a lesser extent, in Britain (Boheim and Taylor, 1999), has shown that there is a direct link between the need for more space and household-relocation decisions.

   Previous research can be broadly divided into studies which focus on the decision to move (Brown and Moore, 1970; Clark et al, 1984; 2000; Hanushek and Quigley, 1978; Murie, 1997), those which specifically examine tenure choice (Clark and Dieleman, 1996; Dynarski, 1985), and those which examine the interlinked nature of moving and tenure choice (Clark and Onaka, 1985; Henderson and Ioannides, 1987; 1989). The behavioral underpinnings of the models are the household's desire to come to a better matching between the household's space requirements and the space they occupy. The papers cited above and others along similar lines (Green et al, 1997) have established that younger households move more frequently than do older households, that the wealthier and better educated households are more mobile, and that the amount of housing consumed increases with age and household income. Specific analyses of the connection between moving and room stressöthe rooms needed by a household to be in equilibrium housing consumptionöshow that moving is more likely if the household is in disequilibrium, and that mobility is also `triggered' by marital-status change and the addition of children (Clark and Dieleman, 1996). 

   Research from the US Panel Study of Income Dynamics and the German SocioEconomic Panel has shown that, overall, housing quality has been increasing over time and that most households are enjoying more space and better quality housing than was true a decade ago (Clark and Drever, 2001; Clark et al, 2000). Interestingly, even though the gains are greater for the white population, there is evidence of significant gains for minority households in the USA and for foreign-born households in Germany as well. As expected, the studies of homeowners versus renters show that the gains for homeowners are significantly greater than the gains for renters in the United States, and that mobile households do much better than stayers.

   The most recent research on residential mobility has been set within the context of the life course. The life-course paradigm emphasizes that changes in one dimension of the household-aging process are necessarily linked to changes in other dimensions (Clark and Dieleman, 1996). Thus, changes in household composition are closely linked to changes in occupational careers, and these in turn are translated into changes in housing tenure and housing consumption, so forming the `housing career' (Champion and Fielding, 1992; Fielding, 1992). An important methodology for examining changes in the life course focuses on the events themselves, and measures the intervals between events. Event-history analysis uses event-history data (data on the change from one state to anotheröbeing an owner or being a renter, moving or staying), from which it is possible to construct or examine life courses (Allison, 1984). 

   Within research on the life course there has been a debate about the relative value of cross-sectional and longitudinal models to estimate the coefficients for the effect of room stress on mobility. Davies and Pickles (1985), using simulated data, argued that crosssectional estimation procedures yielded biased coefficients for the size and direction of room stress. However, an empirical analysis which compared the coefficients from pooled cross-sectional and longitudinal models confirmed the usefulness of both modeling strategies (Clark, 1992). In this paper we provide another test of the usefulness of both estimation procedures.

British housing markets 

   The British housing market has a much larger proportion of its housing stock in social, or what is called `public' housing in the United States, and in this sense is somewhere between the Scandinavian countries, where very large sections of the housing market are under the control of housing authorities, and the United States where there is very little public housing. Even though social housing is subsidized housing, it does not necessarily have the same connotation as public housing does in the United States.

   Decisions by the Conservative British governments in the 1980s had important impacts on the housing market and its functioning. The proportion of the housing stock in private rental increased from 9.5% in 1989 to 10.3% in 1996 but, even more important, a substantial proportion of the council housing stock was made available for purchase by sitting tenants (Murie, 1997; Wilcox, 1997). The result was a reduction in the size of the social housing sector. The combination of the increasing tendency of better-off tenants to move out of council housing and the sale to sitting tenants of the council's more desirable properties in the Right to Buy program has created an increasing stigmatization of the council sector (Burrows, 1999; Forrest and Murie, 1988; Pawson and Bramley, 2000). Specific studies of residential moves in British housing markets which have paid particular attention to social (council) housing have suggested that selective mobility is turning council housing into a residualized section of the housing market (Pawson and Bramley, 2000). Jones and Murie (1999) suggested that there has been a hollowing out of council housing, with only the very young and older less affluent households left in the social housing stock.

   The recent changes in the British housing market must be set within the implicit preference for homeownership documented by Clapham (1996) and Saunders (1990). The sale of council housing was an explicit policy favoring ownership, and recent governments have not moved to change the policy of encouraging ownership although Britain, unlike the United States, no longer has a policy of mortgage deduction from ordinary income in income tax assessment.(1)

  Household formation and composition have also been changing in Britain in the decades leading up to the end of the 21st century. Green et al (1997) have documented the increase in the number of single-person households, and the fact that cohabitation and marriage are taking place at later ages than was true formerly.(2) Overall there is an increase in single-person households. These changes are in turn linked with changes in housing tenure, including a general shift away from renting property to owning, although for younger cohorts private renting is still the main tenure choice.

   The increased emphasis on a strong private rental sector is a recognition of the importance of market forces in creating housing opportunities. Kemp and Keoghan (2001) point out that government housing policy now sees private rental housing as a source of flexible, mainly short-term, accommodation. In this sense, the private rental housing market is a transitional form of housingöa stepping stone to owner-occupation and social housing. In this sense too, it is more like the US housing market. Newly formed households will spend some time in the private rental sector while they accumulate capital and down-payment money to move up to the ownership sector, or become eligible for social housing. This function of the private rental sector fits with our notions of a housing career, or housing ladder, in which there is a `hierarchy of tenures', a hierarchy through which households move as they transit through the life course (Kemp and Keoghan, 2001). The notion of the housing career, in which new households move into the private rental sector before they access the owner housing market, and then in due course move up to larger and more expensive owner-occupation, is well accepted, though it is certainly not a simple or linear progression.
_______

(1) The mortgage income deduction was removed in 1990 (Hamnett, 1999). (2) The study by Green et al (1997) also includes detailed tables on residential mobility, tenure choice, and graphs of household formation and dissolution.

   Indeed, Hamnett (1999) has emphasized the complexity of the housing market and that the sequences of change are more complicated than the simple notion of a ladder of success. Households which experience composition changes, especially divorce, often move back down the ladder, and many households with lower incomes cannot ever make the move from renting to owning. Nevertheless, the housing career is a useful working concept and, in association with the life course, provides a way of examining the changes in the housing market. Approaching moves in the housing market as consumption driven is quite different from the perspective that changes are created by an investment strategy of buying and selling to create housing wealth. This last perspective, suggested in the United States by Sternlieb and Hughes (1980) and in Britain by Saunders (1990) argues that many housing moves are created by the investment climate rather than by demographic changes. Hamnett (1999) provides some evidence from survey data which argues in favor of the demographic explanation for residential change. To the extent that a model of mobility incorporating demographic variables is a good fit to the British mobility data it is an argument in favor of the classic age and space-consumption explanations of mobility. At the same time, we can acknowledge that investment decisions may be playing a role within the basic demographic processes.(3)

   Recent research on residential mobility in the private rental sector has also shown that the processes of mobility and tenure change are much more fluid than is suggested by the simple notion of a housing-career ladder. Many low-income households find themselves permanently `stuck' in the private rental sector, often in low-quality housing. There is also a significant reverse flow of households who move from owner-occupation to the private rental sector, and a flow from social housing back to the private rental sector (Kemp and Keoghan, 2001). From the perspective of the present study, the research suggests that the British market seems to be in a transitional phaseöa phase which will likely continue the process of making the housing market, and moves within it, more like that in the United States, Canada, and Australia. It is in this context of a more complex housing stock with social housing, private rental housing, and private ownership housing, increasingly subject to market forces, that it is important to know the extent to which models of mobility which emphasize household decisionmaking are relevant.

   The British research literature has paid much more attention to the role of neighborhoods in general, including the interconnection with residential choice and household decisionmaking. In a review of the significance of neighborhood, Kearns and Parkinson (2001) argue that the type of neighborhood can be a source both of opportunity and of constraint (page 2106). It can foster belonging and attachment and, of course, by extension play a role in potential mobility. Neighborhoods can also be `traps' which make upward mobility very difficult. The research by Brower (1996), Butler and Robson (2001), and Forrest and Kearns (2001) was designed to investigate the way in which neighborhoods shape life chances. Thus British research has directly confronted the context within which mobility occurs, and the research in this paper attempts to address the context as well as the individual decisionmaking processes.(4)
_______

(3) It is not possible to construct a formal test of the housing consumption versus investment-driven decisionmaking. 
(4) The Economic and Social Research Council has funded a Center for Neighborhood research


Questions and data 

  The focus in this research is on the level of space consumption and we use a standard model developed for the Panel Study of Income Dynamics (PSID) in the United States. For each household it is possible to calculate the required number of roomsöbased on family size and compositionöand compare that number with the number of rooms actually occupied. The difference can be identified as a measure of `room stress' or space deficit. This measure can be traced over time, and by income, to reveal the extent to which the housing system is in balance with housing consumption. Where are the imbalances and which households are most affected?.

    In previous models of space consumption it has been found that age and family size are negatively related to mobility and gains in housing quality, whereas income and tenure change are positively related to increases in space. These will be the basic independent variables in the model for residential mobility in the British housing market. Tenure change is a critical part of the process of making housing gains, and has been the subject of substantial research by economists and demographers. In the context of the present study, tenure change will be used as an independent variable to assess the link between shifting from tenant to ownership status and the consequent gains in housing consumption.

   Longitudinal research has made significant use of the data from the PSID. This rich data set has year-to-year data on a wide variety of housing and household measures, including data on income and family-change variables. The PSID has collected data since 1968 and there is now a thirty-year panel data set. The British Household Panel Sample (BHPS) was initiated in 1991 and is similar in structure and content to the PSID. There are substantial data on households, housing, and a variety of variables that measure aspects of consumer expenditures. The BHPS now comprises ten waves and it is possible to replicate the housing-consumption studies that have been prepared with the PSID. In addition, the BHPS has important geographic detail on local housing markets. In our analysis of residential mobility we use data on moves within 57 defined labor markets. To create these labor markets we aggregated the 279 local authority districts into 57, larger, labor-and-housing markets. Thus the London region is made up of 32 smaller local districts. The actual aggregation is listed in the appendix.

  In our analysis we use measures of age, tenure, housing consumption, and variables which capture the composition of the household and changes in household structure and compositionöchanges in marital status, and births (table 1, over). Age is an essential predictor in models of residential mobility: younger households move more frequently than do older households, and as aging occurs the probability of moving declines. Tenure is a second critical differentiator in models of residential mobility: owners, with more locational capital, move less frequently than do renters. The cost of moving for homeowners is an additional constraint on moving.

   The mismatch between a family's housing needs and the actual space they have available is also a mobility predictor. Part of the life course is household-composition change, and as additional family members are added the amount of space required increases. The room-stress variable, a measure of the space needed by the household, was created in a manner similar to that used in the PSID and used for the German Socio-Economic Panel (GSOEP) (Clark and Drever, 2001). The room-stress variable measures the difference between actual and required rooms. The required number of room is a measure of the minimum number of rooms a family should have in order to


avoid space stress.(5) The variable ranges from negative (underconsumption) to positive (overconsumption of space). The variable is created by dividing the actual number of rooms by the required number of rooms, and 1 is subtracted from the total. Following previous work, we also include measures of age squared and room-stress squared, as the interaction of age and mobility and housing consumption and mobility are hypothesized to be curvilinear. Thus, either too little or too much space can generate residential moves.

   An extended model of mobility includes measures of marital status, as married households have lower probabilities of moving. Two `trigger' variables, the birth of a child and marital-status change, are also included. Because higher income is likely to facilitate moving, family income is included; and ethnicity is included as a control variable. To investigate the significance of neighborhood, two variablesö``like the neighborhood'' and ``prefer to move''öare included. The first of these was expected to be negatively related to mobility whereas a preference to move was expected to be positively related to mobility. This last variable is not truly a measure of neighborhood context alone as a variety of factors influence the preference for a move; yet it does reflect levels of satisfaction with current location and housing.

   In our research we asked four questions: (1) are gains in housing space generated by residential shifts; (2) does the basic residential model produce similar results when estimated for moves in British housing markets; (3) are neighborhood measures significant predictors in the mobility process; and (4) can we conclude that the residential-mobility model process is driven by housing-consumption decisions rather than by investment strategy?

   The model The discrete time logit model is a standard approach to measuring the connection between moves and the values for age, room stress, and measures of income, marital
_______

(5) Two rooms are allocated for each head of household with or without a spouse. Then, one room is added for each additional married couple or single person aged 18 or over; one room is added for every two boys under 18, and one room for every two girls under 18. If there is an odd number of children then the numbers are rounded up. If there is an odd number of girls and an odd number of boys, then those under 10 years of age are paired regardless of sex (see Clark, 1992, page 1297). We recognize that there are cultural differences in the way in which households apportion space/bedrooms to children but the general structure used in the PSID (and in GSOEP) appears to reflect actual practice, and we believe is appropriate for British households.

status, the changes in marital status, and the birth of children. The logit, or log odds, is the ratio of two probabilities for any two mutually exclusive states. For a given probability of an event, P, the odds are defined as P=(1 ÿ P ). The logit is derived by using the natural base of the logs: thus ln [P=(1 ÿ P)]. In the discrete time logit model the assumption is that, for any person in the population the odds of the event occurring (the hazard) at each discrete time ti(i ˆ 1, 2, .::) is proportional to the odds of the event occurring for some specific individuals who represent a set of baseline states of covariates, such that:


 where l(ti ; X) is the conditional probability of having an event at time ti for a given covariate vector X ˆ (X1 , .::, Xk ), and the bk (k ˆ 1, .::, K ) are parameters. The baseline hazard function l0 (ti) is characterized by conditional probabilities for cases in which the covariant vector X ˆ= 0. The implication of this is that the odds of an event occurring at each discrete point in time are higher by the exponential power of 
k bk Xk for the subjects which are characterized by covariates X in comparison with subjects in the baseline group. With increasingly fine measurements of time, the ratio of the two odds approaches the ratio of the two rates [l(ti ; X )=l0 (ti)] and the result is a continuous-time proportional-hazards model. In the situation where the conditional probabilities are sufficiently small, the logit model provides an approximation to the continuous-time hazards model. As a logistic regression, the relationship in equation (1) becomes


  where ai is the log odds for the baseline group, and the parameters can be estimated with a logistic regression.

Mobility rates and housing consumption in British housing markets 

  Mobility rates in Britain, as in Europe in general, are significantly lower than those in the United States. Intraurban migration rates are currently about 10% a year, an increase from about 7% or 8% per year a decade ago (figure 1). Interregional rates have not changed very much, and are about 2.7% a year for the past decade.


Figure 1. Mobility and interregional migration rates 1991 ^ 92 to 1998 ^ 99.


Figure 2. Distribution of overconsumption and underconsumption of housing space (actual ÿ required rooms) in British housing markets in 1991 and 1999.

   Most households in Britain are in equilibrium or are consuming more housing than they require according to the measure of room stress (figure 2). However, less than 20% of all households consume more than one additional room more than they need.(6) The distribution of consumption has not changed over the past decade.

   The changes in levels of consumption are consistent with our theoretical expectationsöincreases in space with mobility. On average, the rooms per person measure increases with mobility from 1.7 to 1.9. If we think of three age groups of householdsö head of household aged 19 ^ 29, household forming; head aged 30 ^ 49, household expanding; and head age 50‡, household contractingöwe would expect space gains with mobility in the first two, and decreases in the last group, which is what we do in fact find (figure 3). The average number of rooms per person increases for the first two age groups for all housing markets and also for London specifically [figure 3(a)]. The older group, who may have made space gains from a household member leaving, decrease their space when they move.

   Greater details on the effects of mobility are given in figure 3(b). Here we see that single-person households tend to lose space when they move, but of course they are consuming much more space per person than do larger households, and the small losses are probably not consequential. Two-person and three-person households show space gains across the age groups up to age 50. Three-person households over aged 50 also gain space, though here the average gains are negligible.

   Further detail on the nature of housing change and mobility is reflected in an analysis of housing-space changes with tenure change (table 2). We can conceptualize the process of the housing career as first involving rent-to-rent changes, at a second stage rent-to-own shifts, followed by further changes in ownership (own-to-own), and finally own-to-rent, or down-market shifts. Initial changes within rental do not yield gains in housing space, although moves from rental to ownership and moves within ownership do increase the room space. Interestingly, the own-to-rent shifts do not yield space losses per person. Although it is not obvious why the rooms per person should increase with this transition, at least some of these moves are likely to involve household-composition change.

  Overall, the results are consistent with our expectations that local moves by households are driven by gains in consumption in the housing market. The categorization by age provides additional information on the nature of consumption and the extent of the gains.
___________

(6) Overconsumption is about the same as in Germany, but is much less than the levels of overconsumption by households in the United States (Clark et al, 2000).


Figure 3. Housing consumption (rooms per person) by age of household head before and after a move (a) for the whole of the United Kingdom, and for Greater London, and (b) for different household sizes.

Table 2. Change in housing consumption (mean rooms per person) by change in tenure.


  Analysis and outcomes Following the arguments outlined in the introduction and in the section on modeling residential mobility, the model was fitted both to pooled cross-sectional and to longitudinal data. The creation of both pooled cross-sectional and longitudinal estimates establishes the generalizability of the results and provides additional nuances on residential mobility. We also, following earlier work, analyzed households with the same head as well as all households in total.

   The estimates from the pooled cross-sectional (table 3, over) and longitudinal (table 4, over) models are consistent with previous models of mobility and are consistent with our hypotheses of decreases in mobility with age and ownership, and with positive room stress. The variables are significant and the model overall has a respectable rescaled R2 . The results are presented for all households and for same-head households, but our discussion will focus primarily on all households as the results are not fundamentally different. The correctly predicted case numbers are 77.9% for

Table 3. Estimates from a pooled cross-sectional model of residential mobility for all households and for households with the same head.


Table 4. Estimates from a longitudinal discrete time logit model for all households and for households with the same head.


the basic pooled model and 82.5% for the full model including family-composition change, ethnicity, and neighborhood measures. In the longitudinal model the results are nearly the same, though room stress is a more powerful predictor in the longitudinal model and tenure is less strong here. In addition, although it has the correct sign, the neighborhood variable is not significant for all households.

  It is important that the pooled cross-sectional and longitudinal basic models are not structurally very different and this provides additional support for the work that has shown the consistency between the results from cross-sectional and longitudinal analyses. Age and room stress are significant, as are the quadratic terms. Room stress and room-stress squared have larger coefficients in the time-dependent models, suggesting that their role is more accurately captured in the time-dependent structures. However, the overall fit of the model is lower for the longitudinal formation. That `birth' is significant in the longitudinal models and not in the pooled cross-sectional models hints at the greater complexity captured by the longitudinal models, and at the fact that trigger effects may not always be captured in a pooled cross-sectional model.

    The full models with additional variables are also significant. In each case the overall predictions from the model are increased by about 5%. Although the increase in explanation is small, it is important to note that the trigger variablesöchange in marital status, and, in the longitudinal model, birth of a childöare significant. Income is significant both in the pooled cross-sectional and in the longitudinal models. It is in the measures of change that the longitudinal model may be more able to capture the nuances of time-dependent changes and their impacts on residential relocation. Ethnicity is significant: white households are more likely to move. The neighborhood variable is of the right sign and significant in the pooled cross-sectional model.

  Additional information on the role of neighborhood is contained in a table showing reasons for preferring to move (table 5). Although housing is important, as expected, the proportion of the respondents who cite `neighborhood' as a possible reason for moving is a powerful reminder of the context of mobility. Clearly age and space needs are deciding reasons, but at the margins households are acutely aware of their local contexts.

Table 5. Possible reasons why households prefer to move: all households and households with the same head.


  The same models were estimated for moves within the London region. The London housing market is known for its significantly higher prices and large numbers of single and foreign-born residents. The fit of the model and the relative contribution of the independent variables are of interest for their ability to predict change in this specialized market (tables 6 and 7, over). 

  There are some notable differences in the London results, although the basic pooled cross-sectional model for all households is similar to that for the United Kingdom as a whole. Although room stress is significant for all households, room-stress squared is not: households with excess space do not adjust to equilibrium. For same-head households in London, the room-stress measures are not significant, but the tenure effect is much more powerful. One plausible interpretation is that ownership is a surrogate for space, and space-consumption effects are captured by the shift to ownership.

  The longitudinal model for London is very similar to the pooled cross-sectional model, again confirming the validity of both methods. Room stress is significant for

Table 6. Estimates from a pooled cross-sectional model of residential mobility, for all households and for households with the same head, in London.


Table 7. Estimates from a longitudinal discrete time logit model, for all households and for households with the same head, in London.


all households but not for same-head households, whereas room-stress squared is not significant in either model. Surprisingly, birth of a child does not serve as a trigger for a move as it does in the national longitudinal model, and marital-status change is significant only for the all-households model, but not for same-head households. In an expensive and tight housing market such as that in London, the desire to move, as indicated by room stress and changes in household composition, may be difficult to fulfill. In contrast, socioeconomic indicators such as family income and age of household head, indicative of the ability to move, are significant in all models.

Summary and observations 

  There is now a large body of literature on mobility in general and on the need for space and the way in which changes in life cycle trigger moves in particular. In this paper we have extended the research to the United Kingdom, utilizing data from the British Household Panel Survey (1991 ^ 99), and we find results similar to those for the United States and Germany. Whereas the average housing consumption has not changed much during the 1990s, households do in general gain space after a residential move, except for older households (with heads aged 550 years) and single-person households. To evaluate the relative roles of different factors on mobility and the implications of different modeling strategies, we estimated both a pooled cross-sectional and a longitudinal discrete time logistic regression for all households and for same-head households, and for the nation as a whole and for the London area.

   For the nation as a whole, room stress is a significant predictor of moving. Households with underconsumption of housing are more likely to move, and those with excess housing are also more likely to moveöprobably to reduce housing consumption. As expected, homeownership discourages mobility. Second, life cycle and changes in household composition are important determinants of mobility. Older, better-off, and married people, and households with a birth or martial-status change are more likely to move. Interestingly, the triggering effect of birth is captured only in the longitudinal model but not in the pooled cross-sectional model. Although both methods provide similar results, the longitudinal method thus seems better able to capture complicated housing decisions.

  The models for London share similarities and also reveal differences between the London housing market and the rest of the country. On the one hand, age, family income, and housing tenure have similar effects on mobility as in the national models. On the other hand, room stress is significant only for all households, not for samehead households; and room-stress squared is not significant in either model. Given the high prices and tight housing market in London, fewer households are likely to consume excess housing and, if they do, they are probably unlikely to move to reduce their housing consumption. In addition, a birth does not trigger a move. Although the dynamics underlying preference to move are more or less the same across housing markets, local contextual effects are thus important in determining the observed mobility.

    The results of this research lend credibility to the general argument that demographic changes underlie much of the logic of residential mobility. Although it is not a direct test of the investment hypothesis, the powerful role of tenure change and space needs provide a basis for continuing to believe that households are sensitive to the basic housing process of adjusting their housing consumption to remove the disequilibrium that arises from changing family needs. These forces are working in the more complex British housing markets as they do in other housing markets.

Acknowledgements. We would like to thank two anonymous referees for comments which were helpful in revising the paper. We are also grateful to Anne Green who provided suggestions and materials on mobility and neighborhoods. The research in the paper was supported in part by funds from the UCLA Academic Senate Council on Research. 

References

Allison, P D, 1984 Event History Analysis: Regression for Longitudinal Event Data (Sage, Beverly Hills, CA). 

Boheim, R, Taylor, M, 1999, “Residential mobility, housing tenure and the labor market in Britain”, WP 99–16, Economic and Social Research Council, Polaris House, North Star Avenue, Swindon SN2 1US. 

Brower, S, 1996 Good Neighbourhoods (Praeger, Westport, CT). 

Brown, L, Moore, E, 1970, “The intra-urban migration process: A Perspective” Geografisker Annaler 521–13. 

Burrows, R, 1998, “Residential mobility and the owner-occupied market in England” Housing Finance 3715–23. 

Burrows, R, 1999, “Residential mobility and residualization in social housing in England” Journal of Social Policy 28 27–52. 

Butler, T, Robson, G, 2001, “Social capital, gentrification and neighborhood change in London: A Comparison of three South London neighborhoods” Urban Studies 38 2145–2162. 

Champion, T, Fielding, A, 1992 Migration Processes and Patterns: Volume 1. Research Progress and Prospects (Belhaven Press, London). 
Clapham, D, 1996, “Housing and the economy: Broadening comparative housing research” Urban Studies 33 631–647. 

Clark, W A V, 1992, “Comparing cross-sectional and longitudinal analyses of residential mobility and migration” Environment and Planning A 24 1291–1302. 

Clark, W A V, Dieleman, F M, 1996, “Households and housing: choice and outcomes in the housing market”, Rutgers University, Center for Urban Policy Research, New Brunswick, NJ. 

Clark, W A V, Drever, A I, 2001, “Do immigrants improve their housing quality when they move? Evidence from the German Socio-Economic Panel” Vierteljahrshefte zur Wirtschaftsforschung 70 87–94. 

Clark, W A V, Onaka, J, 1985, “An empirical test of a joint model of residential mobility and housing choice”Environment and Planning A 17 915–930. 

Clark, W A V, Deurloo, M C, Dieleman, F M, 1984, “Housing consumption and residential mobility” Annals of the Association of American Geographers 74 29–43. 

Clark, W A V, Deurloo, M G, Dieleman, F M, 2000, “Housing consumption and residential crowding in US housing markets” Journal of Urban Affairs 22 49–63. 

Davies, R B, Pickles, A R, 1985, “Longitudinal versus cross-sectional methods for behavioural research: A First-round knockout” Environment and Planning A 17 1315–1329. 

Dynarski, M, 1985, “Housing demand and disequilibrium” Journal of Urban Economics 17 42–57. 

Fielding, A J, 1992, “Migration and social mobility: South east England as an escalator region” Regional Studies 26 11–15. 

Forrest, F, Murie, A, 1988 Selling the Welfare State (Routledge, London). 

Forrest, R, Kearns, A, 2001, “Social cohesion, social capital and the neighborhood” Urban Studies 382125–2143. 

Frick, J, 1996 Lebenslagen im Wandel: Determinanten kleinräumlicher Mobilität in Westdeutschland [Changing living situations: determinants of residential mobility in West Germany] (Campus, Frankfurt). 

Green, A E, Elias, P, Hogarth, T, Holmans, A, McKnight, A, Owen, D, 1997, “Housing, family and working lives”, Institute for Employment Research, University of Warwick, Coventry. 

Hamnett, C, 1999 Winners and Losers: Modern Homeownership in Modern Britain (UCL Press, London). 

Hanushek, E, Quigley, J, 1978, “Housing disequilibrium and residential mobility”, in Population Mobility and Residential Change Eds Clark, W A V, Moore, E G, Studies in Geography 25 (Northwestern University,Evanston, IL) pp 51–98. 

Henderson, J, Ioannides, Y, 1987, “Owner occupancy, investment versus consumption demand” Journal of Urban Economics 21 228–241. 

Henderson, J, Ioannides, Y, 1989, “Dynamic aspects of consumer decisions in housing markets” Journal of Urban Economics 26 212–230. 

Jones, C, Murie, A, 1999 Reviewing the Right to Buy (University of Birmingham Press, Birmingham). 

Kearns, A, Parkinson, M, 2001, “The significance of neighborhood” Urban Studies 38 2103–2110. 

Kemp, P, Keoghan, M, 2001, “Movement into and out of the private rental sector in England” Housing Studies 16 21–37. 

Murie, A, 1997, “The social rented sector, housing and welfare state in the UK” Housing Studies 12437–461. 

Pawson, H, Bramley, G, 2000, “Understanding recent trends in residential mobility in council housing in England” Urban Studies 37 1231–1259. 

Saunders, P, 1990 A Nation of Homeowners (Unwin Hyman, London). 

Sternlieb, G, Hughes, J, 1980 America's housing (Center for Urban Policy Research, New Brunswick, NJ). 

Wilcox, S (Ed.), 1997 Housing Review 1997–1988 Joseph Rowntree Foundation, 40 Water End, York YO3 6LP. 


Full Text


Or 



ليست هناك تعليقات:

إرسال تعليق