Defining and Measuring Displacement: Is Relocation from Restructured Neighbourhoods Always Unwelcome and Disruptive?

Abstract

Current regeneration policy has been described as ‘state-led gentrification’, with comparisons made with the ‘social disruption’ caused by slum clearance of the 1950s and 1960s. This article takes issue with this approach in relation to the study of the restructuring of social housing areas. The terms ‘forced relocation’ and ‘displacement’ are often too crude to describe what actually happens within processes of restructuring and the effects upon residents. Displacement in particular has important dimensions other than the physical one of moving. Evidence from a recent study of people who have moved out of restructured areas shows that although there is some evidence of physical displacement, there is little evidence of social or psychosocial displacement after relocation. Prior attitudes to moving and aspects of the process of relocation—the degree of choice and distance involved—are important moderators of the outcomes. Issues of time and context are insufficiently taken into consideration in studies and accounts of restructuring, relocation and displacement.

Keywords:

• Restructuring
• relocation
• displacement
• state-led gentrification

1 Introduction: The Social Consequences of State-Led Housing Restructuring

Observers and commentators on state housing policy in the UK have consistently remarked on the negative social effects of restructuring programmes. Thomas noted that ‘The criticisms levelled at slum clearance during the 1880s and 1890s were very similar to those voiced nearly a hundred years later—those of planning blight, social displacement and high costs’ (1986, p. 62). In considering how state policy moved from ‘comprehensive redevelopment’ in the 1950s and 1960s to housing improvement in the 1970s, Thomas highlights the role played by ‘evidence on social dislocation created by clearance programmes … and in particular the disruption of community’, citing in particular the classic studies by Young & Willmott (1957) and Jennings (1962), which have equally had a strong influence on academic approaches to considering state restructuring.

Paris & Blackaby (1979) also note that ‘comprehensive redevelopment has frequently been accused of the “destruction of communities” and established neighbourhoods’ and that ‘For many, slum clearance came to mean a forcible displacement to an unfamiliar high rise flat – a “prison in the sky”– without friends and relations nearby…’ (p. 18). Ravetz (2001) further explains that the break-up of communities occurred because slum clearance residents often had fewer and less secure housing rights than other applicants for council housing (being single people, furnished tenants and immigrants), and were seen as ‘undesirable’ so that the authorities opted to ‘scatter them piecemeal in the hope that they would not drag down their neighbours, but rather be elevated by them’ (p. 132).

Similar accounts have been given of the effects of what was termed ‘urban renewal’ in the USA, starting in the 1930s and 1940s but proceeding rapidly through the 1950s, 1960s and 1970s under programmes funded through 1950s legislation on either housing (for slum clearance reasons) or highways (for expressway and suburban development reasons). Large parts of the inner city of many major US cities were rebuilt in this way, with the negative effects already being termed ‘negro removal’ by the 1960s. A recent account of the redevelopment of the Lower Hill District in Pittsburgh in the 1950s, for example, describes how an ethnically mixed area was demolished, leading to greater racial segregation and problems for relocated individuals who did not receive much information, practical help or financial assistance once told they had to move (Fitzpatrick, 2000). The destructive effects of urban renewal were the subject of classic sociological studies, with two of the most well known and influential appearing in the early 1960s, namely Jacobs' (1961) The Death and Life of Great American Cities in the case of New York, and Gans' (1962) The Urban Villagers in the case of Boston. Gans tells how the ethnically diverse West End district of 20 000 people was cleared in less than 2 years for reasons of both public health and anticipated higher post-redevelopment tax returns, with residents reporting how ‘It isn't right to scatter the community to all four winds’ (Gans, 1999, p. 11). Later reflections also confirm the failure of US urban renewal on the grounds of the crudeness of the slum clearance process and the impracticalities of complete redevelopment (e.g. Lang & Sohmer, 2000).1

This ‘destruction of communities’ narrative from the 1950s, 1960s and 1970s, supported as it is by “classic” sociological studies in both the UK and USA, has remained very dominant and influential in the field of housing and community studies. So much so in fact that contemporary research on these issues often runs the risk of finding what it seeks. The aim of this paper is to show how, if one expounds and broadens the meaning of relocation and displacement in theory and practice, considers the influence of timing and context and adopts more rigorous research designs and analytical approaches, then a more balanced assessment may be made of both the process and outcomes of restructuring and relocation.

However, four decades after some of the classic urban renewal studies, we find similar language being used in commentaries on urban regeneration policy of the 1990s and 2000s. Slater (2006) bemoaned the lack of attention given in recent years to the study of displacement within gentrification research, attributing this to the methodological difficulty of measuring it. He argued that this was a failing, or blind spot, on the part of researchers because gentrification and its core component, displacement, were happening writ large, only they were ‘disguised as “social mix”’ (p. 751). He identified a new phenomenon, or new incarnation of gentrification:

The current era of neoliberal urban policy, together with a drive towards homeownership, privatization and the break-up of ‘concentrated poverty’ … has seen the global, state-led process of gentrification via the promotion of social or tenure ‘mixing’…in formerly disinvested neighbourhoods populated by working-class and/or low-income tenants. (pp. 749–750)

Just as Warde (1991) specified ‘displacement of one group of residents with another of higher social status’ as the first of four ‘criteria for identifying gentrification’ (p. 225), so Slater states that ‘Displacement is and always will be vital to an understanding of gentrification, in terms of retaining definitional coherence and of retaining a critical perspective on the process’ (2006, p. 748). He goes on to call for more ‘critical takes on policy-led gentrification in Europe’ involving the search for displacement and gentrification in policy discourse and policy impacts.

This call was answered and supported by Lees & Ley (2008) who confirmed the view that ‘public-policy-led gentrification’ was the third model of gentrification (Cameron & Coaffee, 2005), again ‘disguised’ as ‘urban renaissance’ whereby ‘mixed income strategies are part of the neo-liberal restructuring of cities’ (Lees & Ley, 2008, p. 2382, citing Lipman, 2008); this was in contrast to a ‘benign’ view of the all-round benefits of such an approach, which would rather see it as ‘third-way gentrification’ (Elorza, 2007). Echoing Slater, Lees and Ley saw the neo-liberal approach as happening ‘in cities world-wide’ with ‘State-led gentrification … being promoted in the name of community regeneration (in the face of supposed social/community breakdown/degeneration) through policies of mixed communities’ (p. 2381).

Demonstrating our earlier remarks about the long-term influence of post-war urban sociology, we note that in arguing their case that we need to focus on the effects of state-led gentrification, Lees & Ley (2008, p. 2381) make a firm connection to the earlier era:

There are also parallels with the large urban renewal programmes that led to the destruction of inner-city communities enacted in many countries in the 1950s and 1960s … It seems that planners and policy-makers have undergone some form of amnesia with respect to the massive criticism that these post-war slum clearances engendered concerning the destruction of community networks (for example, Young & Willmot, 1957).

We will return later to the question of whether it is appropriate or helpful to draw such strong comparisons with much earlier social and policy eras in seeking to understand current developments. Before doing so, we turn in the next section to consider how relocation and displacement have been studied in the recent period of restructuring of social housing areas. This leads us to reflect on the meaning of the two key elements of ‘forced relocation’ and ‘displacement’ in theory and practice. We then describe the objectives and methods of our own study of relocation through restructuring in Glasgow, before presenting the findings. The paper proceeds with a discussion of our findings on the outcomes of relocation, and ends with some conclusions about the implications of our work for the way restructuring and relocation are critically discussed and researched in future.

2 Recent Studies of Restructuring and Relocation

Over the past decade or so there has been renewed interest in studying the effects of area regeneration programmes on residents who move as part of the restructuring process. Much of the recent research stems from studies in countries with significant area-based regeneration programmes, namely HOPE VI in the USA (e.g. Popkin et al., 2004), The Big Cities Policy in The Netherlands (e.g. Bolt & Van Kempen, 2010) and The New Deal for Communities Programme in the UK (e.g. CLG, 2007), but there has also been work on the redevelopment of Asian towns and cities. Two key concerns in this field of inquiry are that ‘forced relocation’ occurs, and that this, either in and of itself constitutes ‘displacement’, or, less automatically, results in ‘displacement’ for those who relocate. As we shall see, both these issues are less than straightforward to define.

Goetz (2002) drew attention to the distinction between ‘voluntary mobility’ and ‘involuntary mobility’, or what he also termed ‘forced relocation’ in US renewal programmes. With references to the HOPE VI programme, Goetz defined ‘the involuntary approach as one in which families are forcibly moved out of their previous homes’ (p. 108). He also refers to ‘those involuntarily displaced by the demolition of public housing’ and calls the entire group ‘displaced families’ (p. 109). Goetz went on to show, using data from Minneapolis/St Paul, that on some issues (but by no means all), outcomes differed between voluntary and involuntary mover groups, with involuntary movers more satisfied with some aspects of neighbourhood amenities and services, but less likely than voluntary movers to report neighbouring behaviours among adults or good social relations among children. The unusual thing to remember here, though, is that in the US case it is the involuntary movers who had more choice of destination than the voluntary movers, due to restrictions placed upon where participants in the voluntary, Moving to Opportunity (MTO) Programme could go (i.e. to low-poverty neighbourhoods).

In a later review of research across HOPE VI Programme sites, Goetz (2010) reported that overall, the studies showed only a ‘modest level of benefits’, especially at the individual rather than the community level. Two key reasons for this were put forward, namely that the involuntary movement programme did not necessarily relocate people to much better neighbourhoods and many of the individual-level benefits required ‘active’ engagement on the part of movers, rather than being ‘passive’ benefits that could be accrued irrespective of the movers' own inclinations to engage (e.g. feelings of greater safety).

A corollary of this important distinction comes when HOPE VI is also described as ‘forced displacement’ which ‘interrupts social support networks that are important to very low-income families, and actually impedes their ability to experience benefits of relocation’ (p. 152). Evidence is cited to show that involuntary movers miss their old social relationships and have difficulty creating replacement ones in their new locations (e.g. Clampet-Lundquist, 2004; Greenbaum et al., 2008). A key part of the argument is also that HOPE VI, by its very nature, constitutes ‘displacement’, because ‘existing residents must be relocated from the site prior to demolition’ so that ‘HOPE VI begins with the displacement of the residents’ (Goetz, 2010, pp. 144–145).

Dutch research approaches the issue of restructuring and relocation from a similar perspective, whereby ‘all [these] movers can be considered as displacees’ (Kleinhans, 2003, p. 481), applying a definition of displacement from gentrification research whereby ‘displacement occurs when any household is forced to move from its residence by conditions which affect the dwelling or its immediate surroundings’ and this is beyond the household's control, including by physical means such as demolition and upgrading (p. 480, based on Grier & Grier, 1978; LeGates & Hartman, 1981; Marcuse, 1986).

However, Kleinhans implicitly (rather than explicitly) questioned the notion of displacement by identifying a number of factors that would influence whether a ‘displacee’ would perceive positive outcomes from relocation, not simply objective differences in housing circumstances. The key factors he expounds, based on a range of literature, can be seen as relating to four things. First, housing intentions, and in particular whether someone was planning to move in any case (Popp, 1976). Second, a resident's understanding of renewal and their appreciation of the necessity for both renewal and forced relocation (van Kempen & Priemus, 1999). Third, personality, or an individual's ‘preparedness for change’, wherein ‘dispositional optimists’ and confident personalities will act to achieve better outcomes for themselves post-move, whilst others have difficulty adapting to change (Ekstrom, 1994; Fried, 1967; Lazarus, 1991). Lastly, the process of movement and change itself and the degree of choice and control residents can exercise within it (Allen, 2000).

In subsequent survey research on ‘forced relocation’ from neighbourhoods in The Hague, some of these hypotheses were capable of being tested. Thus, both a degree of understanding and approval of renewal and relocation, and a prior intention to move, were positively associated with reported, post-relocation dwelling progress (Kleinhans & van der Laan Bouma-Doff, 2008). Similarly, an understanding of the need for dwelling demolition and not having a strong preference to remain in the same or adjacent neighbourhood were both positively associated with post-relocation perceived neighbourhood improvement (Doff & Kleinhans, 2011). In earlier qualitative research on ‘forced relocation’ in The Hague and Utrecht, Kleinhans (2003) similarly found that prior attitudes influenced residents' views of the relocation process and of their new housing circumstances. Any negative effects of relocation on social ties were limited, because ‘many respondents considered local social ties relatively unimportant’ (Kleinhans, 2003, p. 495).

Research on relocation processes and outcomes in Asian cities has had a slightly different focus, relating restructuring efforts to processes of economic development and institutional reforms. In contrast to many European and US studies, predominantly positive accounts are given of relocation in Singapore and Hong Kong. In the case of Singapore, the demolition of older flats and rehousing to new, taller buildings since the mid-1990s is reported to have resulted in high rates of satisfaction (Yuen et al., 2006). This is attributed to a number of factors including the popularity of high-rise, improved building and neighbourhood facilities, and the fact that most people acquired a larger dwelling as a result. Interestingly, the element of choice provided in the renewal programme is also important, whereby residents can continue to live in the same neighbourhood or choose another location (Lau, 1998). In addition to similar factors about attitudes towards and the quality of high-rise developments, in the case of Hong Kong, the success of resettlement programmes from the 1950s to the 1980s may also have been due to the poor conditions in the squatter settlements they replaced; the fact that rehousing rights were for a long time attached to occupancy of squatter huts caused squatter settlements to continue to expand for some time after resettlement began (la Grange & Pretorius, 2005; Yeung & Wong, 2003).

The Chinese case is reported in less positive terms. Here, economic development and an emerging property market which has spread from coastal to inland cities has led to housing clearance in central city areas to make way for commercial, industrial and residential redevelopments for both profit and city image reasons (Han & Wang, 2003). Two concerns raised about the rehousing of residents to replace affordable housing have been that the ‘affordable’ housing is too expensive for many eligible groups and is therefore occupied by people in less need, and that affordable housing developments in peripheral locations may lead to new patterns of social segregation (Stephens, 2010; Wang & Murie, 2000), in echoes of the criticisms of US urban renewal for similar reasons a few decades previously.

3 ‘Forced Relocation’ and ‘Displacement’

We have begun to see that two of the key components in restructuring processes and studies thereof, namely ‘forced relocation’ and ‘displacement’, are not as clear-cut as often depicted. In the first place, it is somewhat too simple to label one type of move as ‘forced’ or ‘involuntary’ and others ‘voluntary’, with the distinction often made on the basis of the programme characteristics (e.g. HOPE VI vs. MTO). Most obviously, some of those people required to move under a renewal programme may in fact want to move and see relocation as an opportunity for ‘betterment’ both in housing and in other personal terms. This point was made some time ago by Gibson & Langstaff (1982) in discussing the post-war clearance programme, citing the work of English, Madigan & Norman (1976), namely that ‘… substantial support for clearance can be found from those who view rehousing as providing the fulfilment of their aspirations for better housing’ (p. 45).

In addition to this potential desire for ‘betterment’, residents' perceptions of the ‘forced’ nature of displacement may also depend on two other factors. First, the strength of their place attachment within their current location, which may in turn be influenced by their means of arrival, i.e. how they got there in the first place, and whether this was a willing or desired move. It is also affected by their experiences whilst living there, good or bad, and of particular importance to place attachment within deprived areas is the capacity of individuals to overcome any challenging circumstances they might have faced during their time in the area (Livingston et al., 2010).

Second, residents' views of the ‘forced’ nature of their move will be influenced by the process involved in moving, and in this regard we can expect differences today compared with the process described earlier in relation to mid-twentieth century slum clearance. This is because of the development of equality within western societies and the advent of legislation to outlaw discrimination and to promote equal treatment between groups in the provision of goods and services. In addition, there has been a significant growth in consumerism, consumer choice and rights, within both market and state sectors and a recognition that consumers of public services need information, advice and support to help them make the right choices. Consumerist approaches have been adopted in the social housing sector, and decisional support mechanisms to assist consumers in relocation situations are being developed (Baker, 2008). These broad social trends of equality and consumerism would tend to suggest that relocation processes today might not be as brutal or insensitive as in past decades.

Furthermore, the association between programme type and level of choice is not straightforward. Thus, rather unexpectedly, in the US case it is the ‘involuntary movers’ who have more freedom of destination location than the ‘voluntary movers’, although, more curiously still, this results in a narrower range of destinations for the ‘forced movers’ (Goetz, 2002). In Europe, a different set of circumstances exists. In the Dutch case ‘forced relocators’ are free to choose their destinations under the Delft Model of housing allocation (van Daalen & van der Land, 2008), albeit with ‘urgency status’. In the UK case, those people required to move under renewal programmes are often found new homes and locations by their landlords who are undertaking the demolition (rather than being free to do so themselves anywhere within the rented housing market). Thus, social housing systems, and processes therein, vary between countries so that the extent of choice and control within relocation processes differs in ways which make a crude distinction between ‘forced’ and ‘voluntary relocation’ unhelpful to the explanation of outcomes across contexts.

Equally, displacement is not a term to be used lightly. We can deduce this from the fact that the United Nations (2004) not long ago established its Guiding Principles on Internal Displacement, pointing out that many of the causes of such displacement were similar to those which create refugees, i.e. people forced or obliged to flee their homes due to conflict, violence, human rights violations and natural disasters. The UN placed a prohibition on ‘arbitrary displacement’, including ethnic cleansing, punishment and large-scale development projects not justified by a compelling public interest. It is theoretically possible that the latter could include the restructuring of social housing areas, which would mean that UN principles for treatment, such as adequate information, consent, compensation and review, could and should be applied to regeneration. In general, however, it would seem that the scale of movement and severity of impacts involved for internally displaced persons, or IDPs as the UN refers to them, are probably of a different order of magnitude to what might be expected in or through most regeneration programmes in advanced nations; thus, the UN's aim is to avoid ‘prolonged displacement’ and achieve ‘sustainable returns’ for IDPs, particularly where a culture or way of life is threatened as a result (UNHCR, 2010).

Hence, our starting point is to be cautious about labelling all regeneration activity as involving displacement. Automatically using the language of displacement is not neutral in its approach (more so when one considers the IDP concerns of the UN); displacement is a loaded term with highly negative connotations. It immediately raises the stakes involved in the consideration of a public policy programme and risks conflating regeneration and estate restructuring with much more serious forms of residential disruption, such as those resulting from Hurricane Katrina in the USA, the Three Gorges Dam project in China or ethnic conflicts in unstable nations. It is also far from self-evident that the regeneration of social housing areas comprises displacement: this is more of a question than a truism.

Despite the concern in the 1960s and 1970s with the ‘social displacement’ caused by the comprehensive redevelopment of inner city areas, the majority of research on urban restructuring in the last two decades has focused on what have been termed ‘physical’ and ‘exclusionary’ displacement, pertaining to a gentrification research agenda, i.e. people having to move by virtue of the demolition or upgrading of properties and being unable to access or re-access local properties due to the changes of value and ownership which have ensued (Marcuse, 1986). The recent assessment of displacement has typically involved measuring the number and social class groupings of outmovers from restructured areas (e.g. Freeman & Braconi, 2004; Newman & Wyly, 2006; see Atkinson & Wulff, 2009, for a review of such displacement studies); this approach is informed by the notion that gentrification entails the replacement of the working class by higher social groups within an ‘upgraded’ area (Warde, 1991). However, just as with ‘forced relocation’, this focus on movement, or physical displacement, reflects an under-developed notion of what ‘displacement’ may comprise.

In a very useful, yet atypical, perspective, Davidson (2008) argued against a focus solely on physical, or what he termed ‘direct displacement’, because the absence of such direct effects may hide other more subtle processes of displacement. Like Marcuse, Davidson talked about ‘indirect economic displacement’, such as where house price rises occur as a result of adjacent housing and cultural developments whereby the affordability of housing in an area falls. But he also identified two other kinds of effects upon lower income residents remaining in restructured areas, namely: ‘neighbourhood resource displacement’, as local shops, services and meeting places are upgraded or changed in ways which lead incumbent residents to no longer associate with them; and ‘community displacement’, as the governance and identity of a place are changed so that incumbent residents feel a ‘loss of sense of place’ (p. 2399). As Atkinson earlier remarked, changes that ‘are often perceived as improving an area’ can also constitute a ‘geography of privilege’ from which original residents feel separated (2000, pp. 321–322).

For outmovers from restructured areas, as opposed to those remaining in situ, we can conceive of displacement as having four potential dimensions; this is in addition to assessing to what extent their relocation is ‘forced’ and/or ‘voluntary’ or desired.

	Physical displacement describes the relocation of residents to a place which is separate from their original location, so that there is perceived distance between the two and they are no longer part of the same neighbourhood as before.
	Functional displacement refers to the extent to which the local services and amenities a household typically uses are less available, or harder to access, in or from the new location; part of this functional displacement is the destination-area equivalent of Davidson's (2008) origin-area ‘neighbourhood resource displacement’. The daily routines of those relocated are also disrupted by the change.
	Social displacement takes us back to the concerns of the 1950s and 1960s, namely that outmovers suffer the loss of their pre-existing social networks, and social support mechanisms, i.e. they are disconnected from their social relations and fail to replace them adequately with new ones in their new location.
	Psychological displacement refers both to the sense of loss felt by outmovers for their old home (Fried, 1963), community and neighbourhood, and to an absent, or reduced sense of identity in, and with, their new place of residence. Psychosocial outcomes such as sense of community, trust, progress and status would be diminished in their new location compared with their previous one due to lower familiarity, reduced ‘fit’ between person and place, and less well developed place attachment.

We now turn to how we intend to explore these issues of ‘forced relocation’ and ‘displacement’ in our own study of regeneration and its impacts.

4 The Study and Methods

In approaching our own study of restructuring, we have sought to maintain an open mind about what is going on and how it is received and experienced by those affected. Our main research questions are therefore as follows:

1.	To what extent can relocation be considered ‘forced’, unwelcome and an imposition on residents?
2.	Does relocation deliver better residential outcomes for residents?
3.	To what extent does relocation result in displacement2 for Outmovers, in terms of: • physical displacement; • social disruption; and • psychosocial disruption.
4.	Are perceived outcomes for Outmovers dependent on issues of residents' prior attitude (i.e. desire to move) and process (degree of choice and distance involved)?

4.1 Regeneration and Clearance in Glasgow

The current phase of area regeneration in the city of Glasgow developed following the transfer of the city's social housing stock from the City Council to the Glasgow Housing Association (GHA) in 2003 (see Gibb, 2003). By 2005/2006, GHA had determined how it was going to deliver on the promises made to tenants at transfer and meet its business plan objectives. This entailed improving its housing stock and producing a ‘sustainable housing system’ in the city by integrating that stock into ‘successful neighbourhoods’ (GHA, 2005a). Through examining the structural condition of its properties and analysing management information on void rates, demand and turnover, GHA identified ‘a limited number of locations in the city each containing a large group of homes that are unsustainable’ (p. 25). In collaboration with the city council, it was decided that eight transformational regeneration areas (TRAs) would be declared across the city, each involving special planning exercises in anticipation of widespread demolition and redevelopment of the housing and neighbourhoods, together with seven local regeneration areas (LRAs), which were smaller in scale and where demolition would be less extensive (GCC, 2007; GHA, 2005b; GHA, 2006a). Together, these 15 areas, which have a large presence of high-rise flats, contained a population of 35 000 people, equivalent to 6 per cent of the city's total. GHA planned to demolish 19 100 dwellings by March 2015; of these, 9900 were approved for demolition in the period 2003–2008, 40 per cent of them being high-rise (GHA, 2006b), mostly within the declared regeneration areas. A review of the programme in April 2009 reported that 9886 units had been demolished in the 6 years since stock transfer (GHA, 2009).

The legal requirement under the Housing (Scotland) Act 2001 is that tenants required to move due to clearance, and regeneration must be offered ‘suitable alternative accommodation’, which GHA defines in terms of location, size, character, affordability, safety and accessibility (GHA, 2003). In addition, GHA undertakes only to move people to properties which meet, or have been improved to, the ‘Warm Homes Standard’ (above the national minimum standard for social housing).3 In terms of prioritisation within the clearance process, the key factors evident in the relevant policy statements are length of residency, for which extra points enable residents to access higher quality alternative accommodation; overcrowding at the current address and any health and safety risks resulting from remaining one of few residents in a clearance block. For each regeneration area, a clearance strategy is produced which sets out residents' preferences for rehousing and identifies the local housing alternatives available, ranging from GHA improved properties, through, to a lesser extent, GHA new-build homes, to homes provided by other registered social landlords who are co-operating with GHA to facilitate the city's regeneration programme (GHA, 2005c). GHA proceeds by holding rehousing interviews with the affected residents and undertaking to make them three offers of suitable alternative accommodation ‘within a reasonable timescale’ of the block in question being declared as ‘under clearance’ (GHA, 2003). After this, court action may be considered to remove someone who refuses all offers made to them, although this is uncommon.

Residents required to move under clearance procedures receive home loss and disturbance payments, as required by law, from GHA of £2750 each (GHA, 2008). For those who remain living in the regeneration areas for longer, some improvement works are carried out to their homes to make them more liveable in the meantime; this was true for a third of social sector tenants in the TRAs in 2008, and half of those in the LRAs (GoWell, 2010, Table 5.3). The most common works to properties in regeneration areas were new doors and locks and, to a lesser extent, improvements to heating systems.

4.2 Surveys and Samples Used

As part of ongoing research into the health and well-being impacts of housing improvement and regeneration in Glasgow,4 we have conducted surveys on 2 occasions in 6 of the city's 15 regeneration areas (3 TRAs and 3 LRAs). We are also following residents who move out of these areas as regeneration progresses. For this analysis, we compare data from two sets of interviews obtained in the period 2008–2009, as follows:

Remainers: 678 interviews were conducted with householders in the 6 regeneration areas in the period June–September 2008 (54.6 per cent response rate). These interviews were undertaken within households where we had previously conducted an interview with the householder or their partner in mid-2006, and who were still living at the same address. Thus, we could identify them as ‘Remainers’ who were living through the regeneration process in situ.

Outmovers: 224 interviews were conducted with householders in the period March–May 2009 (39.9 per cent response rate). These interviews were undertaken with households whom we knew had been living in any of the six regeneration areas 3 years earlier (March–April 2006), but who no longer lived in their original location.

Both samples were weighted by age and gender according to known characteristics of the populations from which they were drawn (the 2008 Remainer population and the 2006 Outmover group as a whole, respectively). Further details of the survey and data handling procedures are given in GoWell (2011).

4.3 Measures

4.3.1 Relocation

Outmovers were asked several questions about the relocation process itself, covering issues of desire, choice, distance and problems. First, they were asked what they considered to be the main reason for moving home, and those who said they moved due to the demolition and clearance of their old home were further asked whether they had wanted to move home beforehand in any case. Outmovers were further asked how much choice they had (a lot, some, none, d/k) about several things: the area they moved to, the home they moved into, and the fixtures and fittings inside the home (such as style of kitchen and bathroom). With regard to location and distance, Outmovers stated whether their new neighbourhood was, in their view, part of the same neighbourhood they lived in before, an adjoining or nearby neighbourhood, or ‘a long way away’ from their previous neighbourhood. We also independently measured the distance moved via GIS using centroids of the postcodes of the origin and destination dwellings. Finally, Outmovers were asked whether any of the following were problems when they moved: the upheaval and disturbance, the costs involved and being kept informed about where and when they might move.

4.3.2 Residential outcomes

Both groups of respondents were asked about residential satisfaction and quality. Remainers and Outmovers were asked to state how satisfied they were with their current home and neighbourhood (on a five-point Likert scale from ‘very satisfied’ to ‘very dissatisfied’). For housing, both groups were asked to rate the quality of their homes (on a five-point Likert scale from ‘very good’ to ‘very poor’) across 11 items (e.g. space, heating, external appearance, security). For neighbourhood quality, they were asked to rate three items relating to the environment (attractiveness of buildings, attractiveness of the environment and whether it was quiet and peaceful), nine local services and amenities (again, from ‘very good’ to ‘very poor’) and 11 items of anti-social behaviour (‘serious problem’, ‘slight problem’ or ‘not a problem’). In addition, Outmovers were asked to compare the quality of their new and previous home and neighbourhood, and to state whether they would be happy to remain in their new neighbourhood or wished to move back to their previous area or to move on to another area.

4.3.3 Social outcomes

For social outcomes we examined both neighbourly behaviours and available social support. Both study groups were asked five questions about neighbourliness: how many people in the neighbourhood they knew (from ‘most people’ to ‘no-one’); how often they spoke to their neighbours (from ‘most days’ to ‘never’) and to what extent they visited their neighbours in their home, exchanged favours with their neighbours, and stopped and talked to people in the neighbourhood (‘a great deal’, ‘a fair amount’, ‘not very much’ or ‘not at all’). In addition, Outmovers were asked whether their ‘closest’ neighbours from where they lived previously still lived ‘very nearby’ to them, and how they felt about that (‘happy’, ‘unhappy’ or ‘don't mind either way’).

With regard to social support, both groups were asked how many people outside their own home they could ask for different kinds of help: practical support (‘to go to the shops if unwell’), financial support (‘to lend money for a few days’) and emotional support (‘to give advice and support in a crisis’). Respondents could reply ‘none’, ‘one or two’ or ‘more than two’ or that they ‘would not ask’ for help.

4.3.4 Psychological and psychosocial outcomes

Psychosocial outcomes were measured in relation to the home and neighbourhood by asking both groups of respondents to state their level of agreement (from ‘strongly agree’ to ‘strongly disagree’ on a five-point Likert scale) with eight statements about benefits derived from the home (including privacy, control, sense of progress, safety, retreat, freedom, desirability, expression of personality and values) and one statement about the neighbourhood (sense of personal progress).

Other psychological outcomes included sense of community and feelings of trust. Sense of community was measured by asking respondents three questions: to what extent they enjoyed living where they were, felt they belonged to the neighbourhood and felt part of the community (‘a great deal’, ‘a fair amount’, ‘not very much’ or ‘not at all’). In addition, Outmovers were asked how the ‘feeling of community’ in their new location compared with where they had been before.

Trust was measured through questions about safety, reliance on informal social control and perceived honesty of local people. Specifically, the three questions were how safe they would feel walking in the neighbourhood after dark (from ‘very safe’ to ‘very unsafe’), whether they thought it was likely that someone would intervene to stop youths harassing someone locally and whether it was likely that a wallet or purse lost in the local area would be returned intact (‘strongly agree’ to ‘strongly disagree’).

4.4 Analyses

Four types of analyses were undertaken. The patterns of responses given by the Outmovers were examined in order to elicit how they felt about the relocation process itself, and in particular to consider, across the questions, whether relocation can be considered a ‘forced’ event. Second, the responses of the Remainers and Outmovers were compared in respect of the residential, social and psychosocial outcomes to see whether any significant differences existed in their current circumstances, i.e. which group appeared better off and in which respects. Third, some of the residential outcomes (home and neighbourhood satisfaction) and psychosocial outcomes (sense of community and trust responses) were examined for the Outmover group in relation to the attitude and process issues outlined earlier, namely desire to move beforehand, degree of choice and perceived distance moved. Significant differences between Outmovers and Remainers in the frequencies of classes of categorical variables or between the means of continuous variables were sought using χ² contingency tests and independent samples t-tests, respectively.

Lastly, we further examined the social and psychosocial outcomes relevant to the issue of displacement. To do this, we combined the responses to questions relating to particular topics into the following scales: neighbourliness, social support, sense of community, trust, psychosocial benefits of autonomy and psychosocial benefits of status. The constituent items for each scale were drawn from the outcome variables described above and are summarised in Table 1, which also gives the values of Cronbach's α for each scale. To calculate indices, in each case the values from the three- to five-class ordinal variables were summed and the total divided by the number of items comprising the index. Indices were standardised so that they ran from 0 (poorest rating) to 100 (best rating).

Table 1. Social and psychosocial outcome scales.

Scale	Constituent items	Cronbach's α
Neighbourliness	Know people, speak to neighbours, visit neighbours, borrow things, stop and talk to people	0.745
Social support	Someone to: go to shops, lend money, give advice	0.905
Sense of community	Enjoy living here, belong to neighbourhood, feel part of the community	0.856
Trust	Rely on someone to intervene, likely to have lost wallet returned	0.698
Autonomy	Privacy, control, safety, retreat, freedom	0.808
Status	Progress from home, home desired by others, home expresses personality, progress from neighbourhood	0.844

Multivariate analyses of variance were conducted with each of the six indices as dependent variables, with sample status (Outmover vs. Remainer) considered as an independent variable alongside gender, household structure (adult, single parent family, two parent family and older person), built form (multi-storey flat, other flat and house) and length of residence in the area ( < 1 year, 1–2 years, 3–5 years, 6–9 years and 10+ years). Regression models were developed by backward elimination, starting with a model containing all two-way interactions and removing variables until only significant terms (p < 0.05) remained. In this way, we could see whether being an Outmover (rather than a Remainer) might be a significant influence upon the outcome in question, over and above the effects of personal characteristics (in case there was any selection bias in who was in the Outmover group), type of dwelling (because we knew the Remainers were more likely to be living in high-rise flats where outcomes might be worse) and length of residence (which might affect outcomes like sense of community, but would be shorter for Outmovers than the majority of Remainers). The multivariate analysis was conducted using only British citizen cases because the vast majority, 93 per cent, of the Outmover sample were British citizens compared with 70 per cent of the Remainer sample.

All statistical analyses were carried out using SPSS v.15 (SPSS, Chicago, IL, USA).

4.5 Findings

4.5.1 Relocation

The majority of Outmovers (74.7 per cent) reported that the main reason they had moved was because their previous home was being demolished. The remainder gave housing (13.3 per cent), neighbourhood (7.1 per cent) or personal (4.9 per cent) reasons for moving. Those Outmovers who considered that they moved due to demolition were evenly divided between those who said that before they moved, they ‘had been wanting to move home or area in any case’ (49.4 per cent), and those who said they did not (42.3 per cent) or could not recall (8.3 per cent).

Outmovers did not generally move very far: the mean distance moved was 1.7 km, with 40.6 per cent moving up to 1 km and 21.5 per cent moving more than 2 km. Outmovers' perceptions of neighbourhood change and distance were, however, somewhat different from reality. A third (35 per cent) of Outmovers said they were still ‘part of the same neighbourhood as before’ even though they had mostly moved out of an estate with clearly identifiable boundaries. A quarter of Outmovers said they lived in an ‘adjoining or nearby neighbourhood’ and two-in-five (39 per cent) said they lived ‘a long way from their previous neighbourhood’, even though only a fifth of them had actually moved more than 2 km.

Outmovers reported having more choice about the area they moved to than about the home itself (type, size, etc.) or its internal features. A majority (55.6 per cent) said they had ‘some’ or ‘a lot’ of choice about the area, just under half (46.6 per cent) said they had choice about the home itself and around a third (36.4 per cent) said they had choice about the fixtures and fittings inside the home.

Sizeable minorities of the Outmover group reported experiencing problems with the process of moving. The most common problem, affecting more than two-in-five Outmovers, was the costs involved, even despite the home loss and disturbance payments made. As Table 2 shows, the reporting of problems varied between household types. Adult households (those without dependent children) reported the fewest problems and families with dependent children reported the most. Around half the families experienced problems of ‘being kept informed about where and when they might move’, and two-in-five single parent families reported problems of upheaval and disturbance, perhaps reflecting less assistance being available to them (see findings on social support, below).

Table 2. Movement problems by household type, Outmovers.

	Upheaval and disturbance	Costs involved	Being kept informed	N (minimum)
Adult household	21.9	38.3	19.3	114
Single-parent family	42.9	55.1	49.0	49
Two-parent family	27.5	55.3	50.0	38
Older person(s)	22.2	50.0	23.5	17
All	27.6	45.1	31.5	216
p	0.012	< 0.0001	< 0.0001
Note: The values in this table represent the percentage of respondents in each row citing the item as a ‘slight’ or ‘serious’ problem (row percentage).

4.5.2 Residential outcomes

On average, Outmovers rated the quality of the homes they moved into higher than Remainers rated the homes they remained living in. This was true across all 11 items respondents were asked to assess. The biggest differences between the two groups were in respect of heating, insulation, security and external appearance of the home; in each case, at least a fifth more Outmovers than Remainers rated the item as ‘very’ or ‘fairly good’. This is reflected in dwelling satisfaction, with 78 per cent of Outmovers being satisfied, compared with 60 per cent of Remainers. One of the main residential outcomes for Outmovers was that only a fifth of them lived in high-rise flats, compared with four-fifths of Remainers. The largest group of Outmovers (50 per cent) lived in low-rise flats in much smaller buildings, and a further 20 per cent lived in houses; thus, the move out of mass housing estates and to lower storeys was the main transition experienced by Outmovers (see GoWell, 2011). However, satisfaction levels were also significantly higher in the Outmover group when people in similar types of dwellings were compared.

Likewise, Outmovers were more satisfied with their neighbourhoods than Remainers: 85 per cent versus 66 per cent, respectively. In terms of quality, the most consistently reported differences were in terms of physical environmental quality, where for all three items (attractiveness of buildings, attractiveness of environment, and quiet and peacefulness) around three-fifths of Outmovers rated the item as ‘very’ or ‘fairly good’ compared with half this number doing so in the Remainer group. The ratings for local amenities and services were more variable. Of the nine items inquired about, only four showed significant differences in response between the two study groups, with significantly more Outmovers than Remainers rating schools, community venues, policing and street cleaning as ‘very’ or ‘fairly good’. Outmovers also made substantial gains in terms of the social environment. For 9 out of 11 items, significantly fewer Outmovers than Remainers identified anti-social behaviour problems in their local neighbourhood (the exceptions being problems families and house break-ins).

Overall, nearly three-quarters (73 per cent) of Outmovers rated their current home as ‘better’ or ‘much better’ than their previous one, whereas three-in-five (62 per cent) rated their current neighbourhood as better. Outmovers were reasonably settled, with nearly two-thirds (64 per cent) saying they ‘would be happy to stay in this area for the foreseeable future’, and a fifth not sure what they wished to do. Only a small proportion of Outmovers (3.5 per cent) wanted to move back to their previous area, and 8 per cent wanted to move on to another area altogether.

In terms of attitude and process we examined for the Outmover group the influence on residential satisfaction of desire to move, choice in the process of moving and perceived distance moved. Table 3 shows that desire to move had no significant effect upon post-move dwelling satisfaction, with those who did or did not want to move being equally satisfied with their dwelling afterwards. However, those who did not want to move beforehand were significantly less satisfied with their neighbourhood after the move, with fewer being ‘very satisfied’ and more being ‘dissatisfied’ compared with the group who had wanted to move.

Table 3. Residential satisfaction by desire to move, Outmovers.

	Desire to move
	Housing satisfaction		Neighbourhood satisfaction
	Wanted to move^a	Did not want to move^b	Wanted to move^a	Did not want to move^b
Very satisfied	31.2	24.4	25.9	11.8
Fairly satisfied	45.7	55.8	62.6	67.1
Neither	7.2	3.5	3.6	7.1
Fairly dissatisfied	8.0	8.1	2.2	9.4
Very dissatisfied	8.0	8.1	5.8	4.7
N	138	86	139	85
P	0.519	0.015
Note: Table shows number of individuals satisfied with home or neighbourhood within each desire to move group (column per centages).
^a Includes those who said they moved due to demolition and wanted to move beforehand, plus those who said they moved for a non-demolition reason.
^b Includes those who said they moved due to demolition but did not want to move beforehand or could not recall whether they desired to move.

As Table 4 shows, post-move satisfaction with both the home and neighbourhood is associated with the resident's perceptions of choice within the process. Those Outmovers who felt they had choice of area, neighbourhood, and fixtures and fittings were more satisfied afterwards than those who felt they had no choice in these matters. Furthermore, those who perceived that they had ‘a lot’ of choice were extremely highly satisfied after the move.

Table 4. Residential satisfaction by perceived choice, Outmovers.

	Area	Home	Fixtures and fittings
	% satisfied with neighbourhood	% satisfied with home	% satisfied with home
Degree of choice
A lot	96.9	91.7	96.7
Some	91.2	87.0	83.7
None	74.8	68.6	71.4
N	222	223	219
P	0.001	0.001	0.006

When we look at neighbourhood satisfaction according to people's perceptions of change and distance, we see that those who consider that they still live in the same neighbourhood as before are the most satisfied, and those Outmovers who live ‘a long way’ from their previous neighbourhood are the least likely to be satisfied, although a majority still are satisfied (Table 5).

Table 5. Neighbourhood satisfaction by perception of neighbourhood change, Outmovers.

	% Outmovers	% Satisfied (row %)
Perceived current location
Part of same neighbourhood as before	35.0	93.6
In adjoining or nearby neighbourhood	26.0	89.7
A long way from previous neighbourhood	39.0	74.7
N	223	223

4.5.3 Social outcomes

On four out of five items, Outmovers reported engaging in neighbourly behaviours more often than Remainers, notwithstanding their shorter length of residence (up to 3 years), and for three of the four items, the difference was large and statistically significant (see Table 6). In the multiple regression analysis, controlling for personal and household characteristics, built form and length of residence, Outmover status still had a significant, positive effect upon neighbourliness, though this effect was reduced for female Outmovers and Outmovers in two parent households (see Table A1).

Table 6. Levels of neighbourliness.

	Remainers (%)	Outmovers (%)	p
Know ‘many or ‘most’ people in neighbourhood	25.1	30.4	0.067
Speak to neighbours most days of week	35.9	28.5	0.004
Visit neighbours in their home^a	28.6	57.1	< 0.0001
Borrow things and exchange favours^a	12.0	46.0	< 0.0001
Stop and talk to people in the neighbourhood^a	51.3	65.6	< 0.0001
N (minimum)	669	221
^a Those who answered ‘a great deal’ or ‘a fair amount’.

Outmovers were also more likely than Remainers to report having two of three forms of social support available to them, namely practical and financial support (see Table 7). In multiple regression analysis of the social support index, Outmover status had a large, though not statistically significant, positive effect on the score. However, the effect of Outmover status on social support was significantly reduced for Outmovers in family households (see Table A2).

Table 7. Levels of available social support.

	Remainers (%)	Outmovers (%)	p
Practical: to go to shops if unwell	66.7	82.1	0.0000
Financial: to lend money for a few days	54.3	68.6	0.0000
Emotional: to give advice and support in crisis	63.0	67.0	0.309
N (minimum)	597	188
Note: Percentages are of those respondents who had recourse to one or more people to provide that type of support.

The two-thirds of Outmovers who said they no longer lived in the same neighbourhood as previously were asked whether their ‘closest neighbours’ from where they lived before still lived ‘very nearby’ to them5: a quarter (26.0 per cent) said yes, two-in-five (42.5 per cent) said no and around a third (31.5 per cent) did not know. The first two groups of respondents were then asked how they felt about this outcome. As Table 8 shows, although those who kept their close neighbours from before were the happiest group, the vast majority of those Outmovers who had ‘lost’ their previously close neighbours did not mind either way about this. In fact, two-thirds of the entire Outmover sample who had moved to another neighbourhood were indifferent to the issue of retention of previous neighbours.

Table 8. Feelings about retention or loss of ‘closest’ neighbours, Outmovers.

Whether ‘closest’ neighbours were retained	Feelings about retention of neighbours (row %)
	Happy about it	Not happy about it	Do not mind either way	N
Yes	47.4	0.0	52.6	38
No	8.2	13.1	78.7	61
p = 0.0001

4.5.4 Psychosocial benefits of home and neighbourhood

In terms of psychosocial benefits of the home, the majority of both study groups derived a range of autonomy-related benefits, though slightly more of the Outmover group did so (Table 9). However, with regard to status-related benefits, although the majority of Outmovers derived these benefits from their house and neighbourhood, only half or fewer of the Remainers did so.

Table 9. Psychosocial benefits of home.

	Remainers (%)	Outmovers (%)	p
Autonomy items
Have privacy in my home	65.6	74.1	0.019
Feel in control of my home	62.1	68.8	0.072
Feel safe in my home	64.7	77.7	< 0.0001
Can get away from it all in my home	63.1	72.8	0.008
Can do what I want in my home	65.6	71.4	0.110
Status items
My home makes me feel I'm doing well	49.9	68.3	< 0.0001
Most people would like a home like mine	39.4	65.6	< 0.0001
My home expresses my personality and values	49.7	65.2	< 0.0001
Living in this neighbourhood makes me feel I'm doing well in my life	32.0	70.0	< 0.0001
N	678	224
Note: Table shows number of individuals who responded ‘agree’ or ‘strongly agree’.

In regression analysis, the Outmover class did not have a significantly different autonomy index score from Remainers after controlling for other factors. Higher autonomy scores were reported for single parent families living in flats and houses (rather than in high-rise flats)—which, of course, is an outcome delivered through relocation, and for two-parent families living in an area for 3–5 years (see Table A3). The Outmover class remained significant in the regression analysis of the status index, being positively associated with the score (see Table A4). As with autonomy, higher status scores were reported for single-parent families in flats and houses, and for women who had lived in an area for 1–2 years. For both autonomy and status, single parents in the Outmover group had lower scores.

4.5.5 Sense of community

Half the Outmovers (53 per cent) felt they had moved to an area where the feeling of community was ‘better’ than in their previous location, whereas one-in-seven (14 per cent) felt it was ‘worse’. Outmovers reported a stronger sense of enjoyment, belonging and inclusion in their neighbourhoods and communities than did Remainers (see Table 10), and this was true even when we compared Outmovers only with those Remainers who had lived in their homes for less than 5 years (not shown). All three community outcomes were higher for those Outmovers who had ‘a lot’ of choice about the area they moved to, and for those whose ‘closest’ neighbours still lived ‘very nearby’.6

Table 10. Sense of community.

	Remainers (%)	Outmovers (%)	p
Enjoy living here	71.1	82.1	0.001
Feel I belong to this neighbourhood	59.6	73.2	< 0.0001
Feel part of the community	54.3	69.2	< 0.0001
N	678	224
Note: Table shows number of individuals responding ‘a great deal’ or ‘a fair amount’.

In the regression analysis of the sense of community index, Outmover status remained significant, with a large, positive effect upon the index score, though this was substantially reduced for Outmovers in a house, which might reflect a micro-locational effect to do with where houses were built for Outmovers, given that house itself had a positive main effect (see Table A5).

4.5.6 Trust

Between two and three times as many Outmovers as Remainers reported feelings of safety at night, reliance on informal social control by neighbours, and expectations of honesty in neighbours (see Table 11). Similar magnitudes of difference existed when we compared only British citizens' responses in the two samples. Perceptions of the reliance and honesty of neighbours were higher among those Outmovers who had ‘some’ or ‘a lot’ of choice about the area they moved to. Perceptions of honesty were higher among those Outmovers who considered that they still lived in the same neighbourhood as before.

Table 11. Local trust.

	Remainers (%)	Outmovers (%)	p
Safety: feel safe walking alone in neighbourhood after dark	24.5	64.3	< 0.0001
Reliance: likely someone would intervene in local harassment incident	16.3	33.4	< 0.0001
Honesty: likely a lost wallet would be returned intact	6.6	22.8	< 0.0001
N	678	224
Note: Table shows number of individuals responding ‘safe’ or ‘very safe’, and ‘agree’ or ‘strongly agree’.

In the regression analysis of the Trust Index, Outmover status had a significant and positive effect upon the score (Table A6). Trust was substantially lower for families living in an area for more than a year, and for all those living in flats for more than a year.

5 Discussion

Let us now summarise our findings in relation to our four research questions, before considering their implications for how we study and talk about restructuring and relocation.

We did not find evidence that relocation is a ‘forced’ or imposed event for most of the people involved. Of course, in one sense, relocation is imposed, because without a public policy intervention it would not have happened for the residents in question. However, when we take the residents' perspective into account, we find that most Outmovers (around three-in-five) had either been wanting to move anyway or considered that they had moved for reasons unconnected to the restructuring process. This is not to say that relocation is problem-free, because we also found sizeable minorities of those relocated experiencing problems of upheaval, cost and uncertainty during the course of moving, particularly two-parent families with dependent children.

Relocation has delivered better residential outcomes for those who move, particularly in terms of dwelling quality and the neighbourhood physical and social environments, though less so for the neighbourhood service environment. Most of those who moved considered that they had ‘bettered’ their residential conditions, though again less so in neighbourhood than in dwelling terms. This reflects the fact that the positive effects of relocation are limited in this case because most people have re-located to areas which are still very deprived in national and local terms, so their residential worlds have not been ‘transformed’ by the move (see GoWell, 2011). Nevertheless, within 3 years, a high degree of settlement had been achieved by those relocated, with most Outmovers happy to stay in their new location and very few seeking to return to their original location.

Perceptions of physical displacement among Outmovers were mixed. On the one hand, more people considered that they had remained in the same neighbourhood than the reality of movement across identifiable estate boundaries would have suggested. On the other hand, more people also considered that they had moved ‘a long way’ from their previous neighbourhood than the actual distances moved might have suggested.

Rather than finding that relocation results in social displacement, we found evidence to the contrary, namely that moving appears to have stimulated neighbourliness and social support for those relocated, or rather between them and their new co-residents, though again we note that the positive effects were reduced for some families. Relocation did, in most cases, sever connections between residents and their previous close-by neighbours, but this did not necessarily result in ‘disruption’ because most of those who lost their proximately close neighbours were indifferent to this outcome of moving, and as we have seen, levels of neighbourliness were relatively high for Outmovers, despite this consequence. However, Outmovers who kept their neighbours were happier about this outcome than others, and their sense of community was higher. So, we might conclude that in most cases (though we should not assume in every case) retention of neighbours is better for people, but losing neighbours is by no means as detrimental as often described.

We also examined psychosocial outcomes, an often ignored aspect of displacement, and found that relocation was positively associated with higher status-related benefits of the home and neighbourhood, higher sense of community and greater trust in co-residents, though not with higher autonomy-related benefits. However, some of the positive psychosocial effects were lessened in the case of single-parent families (status and trust) and two-parent families (trust).

Finally, we found that aspects of attitude and process mattered for perceived outcomes from moving. A prior desire to move was positively linked with post-relocation neighbourhood satisfaction. The larger the perceived distance moved, the lower the post-move neighbourhood satisfaction for Outmovers, and the lower their perceptions of the honesty of their neighbours. The widest-ranging effects, however, came from choice. Those Outmovers who felt that they had been given choice in the process of moving, particularly choice of area to live in, had higher post-move residential satisfaction, stronger sense of community and greater trust in their co-residents.

6 Conclusion

In a study of relocation resulting from the restructuring of social housing areas in a deprived city in the north of the UK, we have not found evidence to support the notion that relocation can be characterised as ‘forced’, nor that it generally results in ‘displacement’ for those who relocate, at least not in the short-to-medium term. This has implications for how we study and talk about restructuring processes. Given that displacement is considered the core component of gentrification by many researchers in the field, we would argue that on the basis of both a theoretical consideration of what relocation and displacement might entail, and the findings presented here, the restructuring of social housing areas should not a priori be considered to constitute ‘state-led gentrification’, at least not in terms of the consequences for those who move out.7 In a sense, the state-led gentrification label unnecessarily prioritises, in a pejorative manner, the ‘mixed-communities’ policy objective over and above the objective of improving residential conditions for residents.

Thus, there is a need to be careful in our use of language in describing processes of urban change, whether state-instigated or not. Terms such as ‘displacement’ and ‘gentrification’ are emotive, hold strongly negative connotations for many writers and readers, and assume known policy intentions and effects. To use such terms as a widespread descriptor of restructuring processes which are variable and uncertain in development, and whose outcomes are unknown, runs the risk of becoming a ‘hegemonic discourse’ as much as the ‘neo-liberal’ gentrification and mixed-income policy themes against which they are counter-posed (Lees & Ley, 2008).

Furthermore, the argument for a more ‘critical take’ on policy-led gentrification (Slater, 2006) requires a more objective and considered approach to the study of restructuring processes and their effects, especially pertaining to issues of prior conditions, resident attitudes, processes and outcomes. First, there is a need for more robust and independent studies of restructuring which do not adopt easy characterisations of the kinds of communities undergoing change; as Gibson and Langstaff pointed out a long time ago, we should ‘dispose’ of the myth ‘that every older housing area contains a cohesive community implacably opposed to clearance and rehousing’ (1982, p. 45). Nor should we make assumptions about the preferences, experiences and feelings of those who relocate; residents may like change more, or respond to it better, than many commentators expect.

Second, evaluations of public policy interventions should entail investigation of the processes of implementation (Hill & Hupe, 2002). Mechanisms of information, advice, assistance, support and choice offered to residents in restructuring and relocation situations vary a great deal and are likely to be influential in the resident experience. Third, the outcomes of relocation merit a broader consideration than simply physical distancing (although distance is an important factor as demonstrated here). As we have shown, terms such as ‘displacement’ require unpacking so that their implications become clear and their various dimensions identifiable; this opens the way to a more nuanced examination of whether displacement occurs, what forms it takes and to what extent, and for whom.

In setting out to study restructuring processes along these lines, we would argue that more attention therefore needs to be given to issues of time and context. Rather than simply drawing parallels with what happened in earlier periods of ‘urban renewal’ or ‘slum clearance’ from the 1950s and 1960s, one should also consider any differences between then and now which might affect restructuring outcomes. Again, language can get in the way here, if terms such as ‘working class community’ are readily used to describe those people living in areas of restructuring; the term itself implies a degree of close connectivity and engagement between residents, which may no longer exist in some places. Just as communities today may be different from those of half a century ago, so might the processes of relocation itself (with sometimes a more resident-focused approach), and in some contexts with different expectations and greater rights for those affected.

We would not argue that our findings on the effects of relocation are generalisable to all restructured social housing areas, because context is fundamentally important. Key elements of the context we have studied are that the communities involved are very deprived, have been residentially unstable for some time, are culturally very diverse and the housing stock consists of a high proportion of poor-quality, high-rise buildings. These factors will not all be present in every other place undergoing restructuring, but they influence residents' perceptions of change and the nature and degree of outcomes attained through relocation, and thus merit consideration.

Furthermore, Glasgow is a very deprived, peripheral city quite unlike, for example, London or other towns and cities in the south-east UK. To study restructuring processes and their effects entails taking into account the pre-existing context in terms of the type of communities and neighbourhoods to be restructured and the socio-economic and housing market circumstances of the town or city involved. What is feasible and what policy attempts to achieve in terms of housing restructuring and social change, and the mechanisms involved in doing so (for example, the extent of private sector developer involvement), will vary according to the inter- and intra-city context. Furthermore, what residents think and feel about relocation will reflect elements of both their origin and destination locations, and the contrast between the two, which the broader local context also influences.

Lastly, any study of the effects of restructuring and relocation is a product of its timing within and after the process of change. Outcomes are unlikely to be static, and what is found at one time may not be what is found earlier or later in the process. Restructuring of large social housing estates can take a decade or more to complete, so that the contrasting views of both Remainers and Outmovers will change over that period and beyond, necessitating the reconsideration of the relative merits of relocation over time. This dynamic perspective on restructuring and relocation has yet to be well incorporated into studies of urban restructuring.

Notes

¹ We note, however, that one recent analysis reported that the intensity of urban renewal activity in a city (in terms of federal spending per population) is positively associated with higher subsequent levels of property values, incomes and population in central cities than would otherwise have occurred—in the authors’ words: ‘…a far less dismal legacy than is commonly portrayed’ (Collins & Shester, 2011).

² We do not cover issues of functional displacement in this article as the survey did not adequately measure relevant items.

³ This is a reference to the Scottish Government's Warm Deal Programme which funds insulation works, improvements to partial central heating systems and other energy efficiency measures in public housing or housing occupied by those on low incomes. It typically improves the NHER (energy efficiency) rating of a property from 3.2 to 6.4 (on a scale of 0–10) (Scottish Government, 2005).

⁴ See www.gowellonline.com for further details.

⁵ We acknowledge that the term ‘closest’ could have been interpreted in two ways by respondents, as meaning emotional closeness or physical proximity. However, given the phrasing of the second part of the question, we believe that most people would have understood that the intention of this question was to ask about proximity.

⁶ In all cases, p < 0.01, except for the effect of choice or area upon feelings of inclusion in the community where p = 0.056.

⁷ Although not the subject of this paper, we would also caution against assuming that other in situ effects of restructuring also constitute gentrification, in terms of the other elements proposed by Warde (1991), such as impacts upon the built environment, property values, local services and culture. We also note that Nevin (2010) has refuted claims for gentrification through Housing Market Renewal in Liverpool on grounds of house prices and occupier incomes for new-build housing.

Appendix: Regression analyses

Regression analyses each included the following independent variables: sample (Outmover vs. Remainer), sex (male, female), household structure (adult, single-parent family, two-parent family, older person), built form (multi-storey flat, other flat, house) and length of residence in area ( < 1 year, 1–2 years, 3–5 years, 6–9 years, 10+ years). All two-way interactions were included in the initial models, the final models being determined by backward elimination of non-significant terms. Tables show intercept, sample effect, plus all other significant terms (p < 0.05) in final fitted models. Full versions of regression tables are available from authors on request.

Table A1. Associations with neighbourliness: regression analysis of neighbourliness index.

					95% Confidence interval
	B	Std. error	t	Sig.	Lower bound	Upper bound
Intercept	24.668	8.427	2.927	0.004	8.116	41.220
Time in area: 6–9 years	18.966	8.729	2.173	0.030	1.820	36.111
Time in area: 10+ years	18.750	8.482	2.211	0.027	2.091	35.408
Sample: Outmover	24.395	4.091	5.963	0.000	16.359	32.431
Sex: female × built form: flat	17.328	7.656	2.263	0.024	2.292	32.364
Sex: female × built form: house	12.769	4.724	2.703	0.007	3.490	22.048
Sex: female × sample: Outmover	− 12.004	5.224	-2.298	0.022	− 22.263	− 1.744
Hhd structure: older × built form: flat	− 28.892	11.331	− 2.550	0.011	− 51.148	− 6.637
Hhd structure: 2P family × sample: Outmover	− 23.151	9.198	− 2.517	0.012	− 41.216	− 5.086
Note: R ^2 = 0.217 (adjusted R ^2 = 0.171).

Table A2. Associations with social support: regression analysis of social support index.

					95% Confidence interval
	B	Std error	t	Sig.	Lower bound	Upper bound
Intercept	14.666	22.979	0.638	0.524	− 30.481	59.814
Hhd structure: 1P family	14.758	4.454	3.314	0.001	6.008	23.508
Hhd structure: older	− 12.166	4.955	− 2.455	0.14	− 21.900	− 2.431
Sample: Outmover	31.580	25.468	1.240	0.216	− 18.457	81.618
Hhd structure: 1P family × sample: Outmover	− 16.221	7.690	− 2.109	0.035	− 31.330	− 1.112
Hhd structure: 2P family × sample: Outmover	− 27.495	8.920	− 3.082	0.002	− 45.019	− 9.970
R^2 = 0.140 (adjusted R^2 = 0.114).

Table A3. Psychosocial benefits of autonomy: regression analysis of autonomy index.

					95% Confidence interval
	B	Std error	t	Sig.	Lower bound	Upper bound
Intercept	80.465	7.633	10.542	0.000	65.473	95.458
Hhd structure: 2P family	− 22.654	11.015	− 2.057	0.040	− 44.290	− 1.019
Sample: Outmover	.923	3.261	0.283	0.777	− 5.481	7.328
Sex: female × time in area: 1–2 years	21.384	10.634	2.011	0.45	0.497	42.271
Hhd structure: 1P family × built form: flat	22.805	7.577	3.010	0.003	7.923	37.686
Hhd structure: 1P family × built form: house	15.396	4.299	3.581	0.000	6.952	23.840
Hhd structure: 2P family × time in area: 3–5 years	29.595	11.270	2.626	0.009	7.460	51.730
Hhd structure: 1P family × sample: Outmover	− 26.459	6.919	− 3.824	0.000	− 40.048	− 12.869
Note: R ^2 = 0.206 (adjusted R ^2 = 0.158).

Table A4. Psychosocial benefits of status: regression analysis of status index.

					95% Confidence interval
	B	Std error	t	Sig.	Lower bound	Upper bound
Intercept	61.708	8.210	7.516	0.000	45.579	77.836
Hhd structure: older	13.587	4.221	3.219	0.001	5.294	21.880
Built form: flat	11.253	5.456	2.062	0.040	0.534	21.971
Sample: Outmover	12.457	3.794	3.283	0.001	5.003	19.910
Sex: female × time in area: 1–2 years	24.384	11.963	2.038	0.042	0.884	47.884
Hhd structure: 1P family × built form: flat	23.809	9.792	2.432	0.015	4.574	43.044
Hhd structure: 1P family × built form: house	19.296	5.658	3.410	0.001	8.181	30.411
Hhd structure: 1P family × sample: Outmover	− 24.869	7.378	− 3.371	0.001	− 39.362	− 10.375
Note: R ^2 = 0.235 (adjusted R ^2 = 0.201).

Table A5. Sense of community: regression analysis of community index.

					95% Confidence interval
	B	Std error	t	Sig.	Lower bound	Upper bound
Intercept	52.540	11.111	4.729	0.000	30.717	74.363
Built form: house	9.209	2.885	3.192	0.001	3.542	14.876
Sample: Outmover	18.436	5.397	3.416	0.001	7.836	29.036
Built form: house × sample: Outmover	− 22.302	6.155	− 3.623	0.000	− 34.391	− 10.213
Note: R ^2 = 0.179 (adjusted R ^2 = 0.140).

Table A6. Local trust: regression analysis of trust index.

					95% Confidence interval
	B	Std error	t	Sig.	Lower bound	Upper bound
Intercept	− 22.265	30.741	− 0.724	0.469	− 82.689	38.160
Household structure: 1P family	72.319	26.838	2.695	0.007	19.567	125.071
Household structure: 2P family	47.156	21.187	2.226	0.027	5.509	88.802
Household structure: older	12.072	5.277	2.288	0.023	1.699	22.446
Built form: flat	116.596	36.023	3.237	0.001	45.790	187.402
Sample: Outmover	18.141	4.310	4.209	0.000	9.670	26.613
Sex: female × time in area: 1–2 years	49.468	19.497	2.537	0.012	11.144	87.792
Sex: female × time in area: 6–9 years	47.988	19.534	2.457	0.014	9.592	86.384
Hhd structure: 1P family × built form: house	− 16.164	6.891	− 2.346	0.019	− 29.708	− 2.620
Hhd structure: older × built form: flat	− 39.835	17.490	− 2.278	0.023	− 74.214	− 5.457
Hhd structure: older × built form: house	− 22.730	8.037	− 2.828	0.005	− 38.529	− 6.932
Hhd structure: 1P family × time in area: 1–2 years	− 93.866	27.631	− 3.397	0.001	− 148.179	− 39.554
Hhd structure: 1P family × time in area: 6–9 years	− 91.079	27.427	− 3.321	0.001	− 144.990	− 37.167
Hhd structure: 1P family × time in area:10+ years	− 71.998	26.996	− 2.667	0.008	− 125.062	− 18.935
Hhd structure: 2P family × time in area: 1–2 years	− 59.408	21.348	− 2.783	0.006	− 101.369	− 17.447
Hhd structure: 2P family × time in area: 6–9 years	− 56.861	21.712	− 2.619	0.009	− 99.538	− 14.184
Hhd structure: 2P family × time in area:10+ years	− 46.331	21.448	− 2.160	0.031	− 88.488	− 4.173
Built form: flat × time in area: 1–2 years	− 115.549	36.606	− 3.157	0.002	− 187.501	− 43.596
Built form: flat × time in area: 3–5 years	− 117.842	40.409	− 2.916	0.004	− 197.271	− 38.414
Built form: flat × time in area: 6–9 years	− 108.032	38.707	− 2.791	0.005	− 184.114	− 31.950
Built form: flat × time in area: 10+ years	− 97.216	36.898	− 2.635	0.009	− 169.743	− 24.690
Note: R ^2 = 0.281 (adjusted R ^2 = 0.213).