THATCHERITE IDEOLOGY, HOUSING TENURE AND CRIME: THE SOCIO-SPATIAL CONSEQUENCES OF THE RIGHT TO BUY FOR DOMESTIC PROPERTY CRIME

However one views it, the changes to housing tenure in the 1980s were pronounced and have had enduring effects in terms of the housing market. In this paper, we throw light on the relationship between housing tenure and the experience of property crime in and around what might be referred to as domestic environments (i.e. people’s homes). In so doing, we explore the ideological positions which both of the (then) main political parties had adopted towards housing during the 1970s (during the build up to the sale of council housing) and the ways in which the legal framework surrounding housing was modified in order to effect these ideas at, quite literally, ‘street level’. Using the General Household Survey, the British Crime Survey and the British Social Attitudes Survey, we examine the general relationship between housing tenure and crime and explore how these unfolded both in terms of time (i.e. an historical analysis) and social space (i.e. in terms of the socio-spatial location of these crimes).


Introduction
Our aim in this article is to explore the degree to which the changes produced in the provision of housing in England and Wales produced changes in the socio-spatial 1 distribution of economic need and social deprivation, which in turn changed the distribution of crime. It is our contention that the changes in the distribution of crime were, in part, shaped by changes in other aspects of government action elsewhere, especially in terms of the management of the economy and the housing sector, and were accordingly lagged . As Stewart and Burridge (1989: 78) argue, 'during periods of crisis, law is instrumental in the uneven and unequal distribution of national resources, and that while the normative form of many laws pretends towards national equality, the individual rights that are the vehicle for recent housing policy effectively convey spatially-defined uneven benefits', or, less prosaically, the law acts as a system for producing unequal outcomes despite its claims to produce national-defined equality. Whilst there *Stephen Farrall, School of Law, Sheffield University, Bartolome House, Winter Street, Sheffield S3 7ND, UK; s.farrall@sheffield. ac.uk; Colin Hay, Centre d'estudes europeenes, 28 rue de Saints-Peres, Sciences Po, Paris; Colin Hay, Centre d'estudes europeenes, 28 rue de Saints-Peres, Sciences Po, Paris; Will Jennings, Department of Politics and International Relations, University of Southampton, Southampton, SO17 1BJ; Emily Gray, School of Law, Sheffield University, Bartolome House, Winter Street, Sheffield S3 7ND, UK. have been numerous studies of who bought their homes under the right to buy scheme and what some of the social consequences have been, few have sought to piece together empirically the legislation, the changing social nature of housing provision which this produced and the impacts which this has had on the distribution of crime.
We also seek to contribute to the understanding of the structurally driven processes of victimization. In so doing, we turn to the work of Sandra Walklate (1992;Mawby and Walklate 1994;Walklate 1996) who proposed a 'critical' approach to exploring victimization. Her proposed agenda calls for empirically based research, an integration of quantitative and qualitative methodologies, and comparative and longitudinal studies which are able to explore those social processes which 'go on behind peoples' backs' (Walklate 1996). Such an endeavour ultimately seeks to understand and explore the processes associated with victimization in such a way that their socio-economic and cultural contexts were acknowledged (Walklate 1992;1996). Whilst many have explored the processes which drive victimization using mixed methodologies, few, to our knowledge, have used longitudinal data to explore the ways in which the socio-spatial distribution of crime has changed, nor simultaneously linked this to legislative changes.
Our article starts with a review of the ideological importance of housing in the wider Thatcherite agenda and the ways in which central government's commitment to local authority housing developed in the period between 1945 and 1979. Following this, we explore the wider institutional and legal frameworks which shaped the approach taken to the legislation during the 1980s and review key aspects of this legislation from 1977 to 1989. Having briefly summarized the extent to which housing budgets were cut, we assess the degree to which council house sales were evenly distributed spatially and socially. We next explore the concepts of polarization and residualization before making a brief historical assessment of the notion of the 'problem estate'. Following this, we use various data sets [such as the General Household Survey (GHS), The British Crime Survey (BCS) and the British Social Attitudes Survey (BSAS)] to explore the extent to which social renters were (1) increasingly drawn from the more disadvantaged sections of society throughout the 1980s and (2) increasingly more likely to be victimized in or around their homes.

The Ideological Background
Housing has always been considered an early target of Thatcherite policy initiatives because of the electoral support it was expected to garner. Whilst there are good reasons to see the sales of council housing as being in tune with an older, 'pre-Thatcherite' element within the Conservative Party (see Davies 2013, which traces the ideological preferences for home ownership going back to 1945), it was nevertheless the 1979-83 Thatcher government which took the first serious strides towards changing housing tenure in the United Kingdom. The sale of council housing, however, was particularly attractive to the Conservatives in 1979 as it spoke to several policy ambitions simultaneously; it was an attack on collective consumption and ownership; it dissolved the interests of council tenants and councils themselves; it disrupted the council's status as a landlord; it transferred the public subsidy from local control to private bodies (former tenants, banks and building societies) and, of course, it was expected to expand home ownership and individual choice (Stewart and Burridge 1989: 75;Hay 1992). It also represented a 'weak point' in the fabric of the welfare state (Cole and Furbey 1994: 181).
In keeping with Hay (1992: 43), we see the development of Thatcherite policies as pragmatic as much as ideological. With regards to housing, as was the case with other privatizations of, for example, the gas supply, telecommunications and other public services, one key aim of the changes initiated during the 1980s was to restructure the state and the ways in which it interacted with the citizenry. Such a restructuring, naturally, made a reversal of the potential Thatcherite settlement all the harder (Hay 1992: 45). The provision of council housing increased hugely during the 1950s (Dunleavy 1981;Hay 1992: 48). However, the Conservative government continued to promote private housing over council housing until 1964 when it was replaced by a Labour government which was committed to the expansion of council housing (Hay 1992: 49). The public's support of highrise flats was garnered on the basis of arguments linking their development with notions such as 'progress' and 'technology', and as such can be read as the physical manifestation of the logic of modernity-that the city could be transformed socially and in terms of economic productivity by changing it's very physical nature (Dunleavy 1981;Pawley 1966;Hay 1992: 50). Although high-rise flats were not initially popular with those for whom they were intended, the support for high-rise living amongst the working class was produced on the basis of arguments that they represented an improvement on the 'slums' which had existed previously, and that they were 'exciting' and 'new'.
Following Thatcher's election as leader of the Conservative Party in 1975, housing started to form a key plank in the articulation of other policy ideas. In this respect, it is hardly surprising that the idea of selling council houses to their tenants was presented in the 1979 Conservative Party's election manifesto under the heading of 'Helping the Family'-another key element in Thatcherite ideology. In this way, the Thatcherite approach to housing must be seen as part of a much wider ideological package of moral authoritarianism which incorporated various 'traditional values' associated with the idea of the 'nuclear' family (Hay 1992: 54;Hayes 1994).
Once in office, the Conservative government's strategy was to move away from the idea that the role of governments was to produce a consensus in which the hopes and aspirations of citizens could be accommodated. This was characterized by Thatcher in her speech to a Conservative Party rally in Cardiff during the 1979 general election (16 April) in which she argued that Moses had not asked his 'brothers' for a 'consensus'. This approach required what might be thought of as a 'political triage' in which some sectional interests were accommodated and others were downplayed, marginalized or just ignored. This 'split-nation' strategy as some have termed it (Loney 1986;Rentoul 1987;Walker and Walker 1987;Jessop et al. 1988) ran not along two lines as Jessop et al. suggest (the 'haves' and the 'have nots'), but rather along three lines, viz, the 'haves', the 'have nots' and the 'might haves' (Hay 1992: 55). In housing terms, the 'haves' were those who already owned their homes; the 'have nots' were those who did not, whilst the 'might haves' were those who currently did not own their own homes, but who might be tempted to do so if offered enough opportunities and incentives to do so. Against this backdrop, the stage was set for a radical recasting of the legal instruments surrounding housing.

The Wider Institutional and Legal Frameworks
As Murie (1985: 173) notes, a government keen to develop private housing in place of public-owned housing, as Thatcher's was, did not need to 'create' market-based organizations, ideas and mechanisms (such as lenders, estate agents and the concepts of 'property', 'mortgages' and 'domestic space') from scratch, since these institutions already existed. Similarly, the incoming Conservative government also inherited a ready-made mechanism for controlling (and hence reducing) new capital expenditure in the form of Housing Strategies and Investment Programmes (Housing Plans in Scotland), introduced by the previous government in order to introduce cash limits (Murie 1985: 173).
In examining the ways in which crime and disorder were redistributed during the 1980s, one needs to explore the context in which the legal framework surrounding housing provision and the housing market itself in England and Wales changed following the 1980 Housing Act. Undeniably housing policy played a crucial role in the elections of 1979 and 1983 (Monk and Kleinman 1989: 121) with a flurry of policy innovation unleashed after 1979 (Stewart and Burridge 1989: 66). Table 1 summarizes the key Acts of Parliament which affected housing provision in the United Kingdom (and in particular England and Wales) between 1977 and 1989.
The first key legislative change we need to document actually relates to the prior Labour administration: the 1977 Homeless Persons Act. This act was, in many respects, another key plank in the welfare state, and although many or the larger urban local authorities had already developed strategies for homelessness. The Act was intended to extend housing provision to those individuals and families hitherto unable to secure their own accommodation (Atkinson and Durden 1990: 118;Ginsburg 1997: 141) and was therefore responding to trends which were emerging prior to 1977. The Act extended the definition of homelessness, thereby increasing the numbers of those deemed homeless and placed new responsibilities to house those with greater social needs (Murie 1989: 213;Atkinson and Durden 1990: 118) and created a 'radically new housing market' (Bottoms et al. 1992: 140). Ginsburg (1997: 141) reports that the numbers of people presenting themselves as homeless increased every year from 1979 to 1991; in 1979, 57,200 were accepted a homeless in England, peaking at 151,720 in 1991. As such, throughout the 1980s, there was a growing social polarization between those in council housing and those living in homes they owned (the private rented sector had Local Govt and Housing Act Prevented LAs from using rates to subsidise rents been in decline for many years; Murie 1989: 214) as those in council stock were increasingly drawn from those 'in need'. However, in addition to the 1977 Act, there had been a series of attempts to reduce public expenditure on housing since 1976 (Murie 1989: 213). Thus, when the Conservative government was elected to office in 1979, it inherited a system in which housing expenditure was already being cut and local authority house-building was already in decline (Murie 1989: 213). Housing had been a particular focus of the Conservative election manifesto in 1979, being given more space than social security, health, welfare or education. The key aim, however, had been the introduction of the right to buy one's council house for existing tenants. This was introduced by the 1980 Housing Act.
The key change, in addition to the introduction of the 'right to buy', was the change in the levels of expenditure made with regards to housing. The right to buy proved to be particularly popular with the skilled working class, who, as Stewart and Burridge (1989: 70) note, might otherwise have moved to different locations and become owneroccupiers anyway. During the passing of the Act, amendments were introduced by the House of Lords, resulting in a number of concessions on the part of the government, one of which was that accommodation designed or adapted for the elderly ought to be excluded from the right to buy provisions. Whilst understandable in terms of trying to prevent the exposure of elderly people to the processes of marketization being unleashed, the amendment had the effect of 'trapping' elderly people in some accommodation in some estates and may have contributed, albeit in a small way, to the processes of residualization (see below). Together, the 1982 Social Security and Housing Benefit Act combined the existing systems of rent and rebates with allowances, and supplementary benefits assistance with rents, transferring the administrative burden of this welfare provision from central to local government (Hay 1992: 57). These Acts effectively placed local authorities in the situation of needing to explain and implement the very cuts they had fought against (Cole and Furbey 1994: 198). Subsequently, in the first four years of its operation, the 1982 Act effectively removed around 1 million households from eligibility (Cole and Furbey 1994: 198). Later, the 1986 Social Security Act removed a further £450 million from housing benefit expenditure (Cole and Furbey 1994: 198) and in the same year, the 1986 Housing and Planning Act laid down the measures by which publicly owned housing could be transferred en masse to private owners (Hay 1992: 58). In effect, the 'might haves' were now amongst the 'haves' and the 'have nots' and their interests could be jettisoned. The 1986 Buildings Societies Act was a response to the fact that many building societies were reluctant to release money for council house sales (Murie 1985: 185) and that building societies were being hampered by the existing legislation (Boddy 1989: 92-3) and was the first comprehensive legislation on building societies since 1874. The 1986 Housing and Planning Act extended the discounts available. As a rejoinder, the 1996 Housing Act withdrew the right of homeless families to tenancy in the social rented sector, thereby pushing more of them towards the private rented sector (Ginsburg 1997: 141). However, as Blandy and Hunter (2012) demonstrate, whilst there were a number of legal challenges made to the acts (most notably that by Norwich City Council), the courts tended to reinforce the 'responsiblizing' aspects of the new legislation. In this respect, they argue, the courts acted in a more neo-liberal manner than central government, which had attempted to limit purchaser's expose to risk and responsibility (Blandy and Hunter 2012: 25).
Data from the BCS and the BSAS (Table 2) suggests growing proportions of owners (including those buying with mortgages) and decreasing proportions of people living in the social rented sector. The 1982 BCS sample, e.g., consisted of some 61 per cent of owners/people buying with a mortgage and 30 per cent social renters [slightly under-representing owners; Pawson and Wilcox (2013: 122, table 17a) put these at 67 per cent]. The 1983 BSAS sample consisted of 66 per cent of owners/those buying with a mortgage and 27 per cent social renters. By 1992, the BCS consisted of some 68 per cent of owners/those buying and 22 per cent social renters, whilst the BSAS of 1993 had 71 per cent owners/buying with a mortgage and 23 per cent social renters (so both under represented owners and over-represented those in social housing). Indeed, both datasets therefore appear to track the decline in the social rented sector (which would always have been dominated by local authority landlords) and the increase in those people owning or buying their homes (see Table 2).

The Distribution of Sales
By 1990, over 1.5 million council houses had been sold (Malpass and Murie 1990: 96). Within the country, however, the distribution of sales was not even. Murie (1989: 220) reports that sales of council housing were higher in the South and East and lowest in the North and Inner London. Atkinson and Durden (1994: 186) offer further detail showing that, until at least the mid-1980s, most sales took place in the south-east and adjacent areas. By June 1988, none of the 15 local authorities which had sold more than 30 per cent of their housing stock were in the North, Yorkshire and Humberside, the West Midlands, the North West or Inner London (all were in the South or East). Yet even within regions variations existed; sales were highest in those districts which had fewest flats and where homeownership rates were already high (suggesting that not all accommodation was equally desired and that there was a 'follow-your-neighbour' effect to house purchases). Atkinson and Durden (1990: 121) note that the majority of sales were to those living in houses (as opposed to flats). McNabb and Wass (1999: 671) report that single people were the least likely to buy their homes, and this decreased with the presence and number of children. Cole and Furbey (1994: 198) note that the sales of council houses were concentrated amongst prosperous (rather than economically depressed) areas, middle-aged tenants with adult children (rather than the elderly or younger residents) and the skilled working class (as opposed to welfare claimants). In terms of the stock which was purchased, as Murie (1997: 28) notes, it was the 'estates with which contain[ed] the best quality, most spacious, traditionally-built houses with gardens tend to be those which were most popular and in highest demand'. As Forrest and Murie (1988: 80) note, 'the erosion of local authority discretion in the provision of housing has heightened the geographical divisions in council house ownership between north and south, inner city and suburb, tower block and terrace'. In other words, national-level laws played out in different ways in different places.

Residualization and Polarization
Residualization refers to the long-term trend for council housing to (1) become less associated with and used by affluent members of society; (2) cater increasingly for lower income groups (such as the elderly, those not working and some ethnic minority groups); and (3) accordingly to cater less and less for the affluent working class and lower middle class (Murie 1997: 26).
As such, residualization is associated with a shrinking of the numbers of people relying on council housing, the concentration of entrenched social and economic problems amongst those housed in council-owned accommodation and the political marginalization of council tenants in discourses and debates about housing (Cole and Furbey 1994: 198). 2 Prior to Thatcher's election, both owner-occupation and council housing had been rising (at the expense of the private rented sector). However, the increases in the reductions of the price with which tenants could buy their council homes and the rises in the mortgage interest relief rates meant that the public sector moved closer towards catering for the poorest and most disadvantaged households (Monk and Kleinman 1989: 127). In addition to this, the requirement for councils to house some of the most-needy households (a requirement of the 1977 Homeless Persons Act) meant that this process was further enhanced. In line with Ginsberg (see above), Kemp (1992: 77, table 5.6) reports that in 1978, the number of households being accepted for housing on the basis of homelessness was just over 53,000, whilst by 1986, this had doubled to almost 103,000 and increased to over 126,000 by 1989. Williams (1992: 177) quotes even higher figures, noting that 'probably as many were turned away'. The causes of this dramatic growth are complex; social and demographic changes would have accounted for some of it (divorce rates were increasing, see David 2014: 180 andHill andWalker 2014: 86-8) and people were living for longer too. In addition, unemployment would have forced some either to relocate, leaving existing homes or to move, possibly making dependents homeless. Nevertheless, Kemp (1992: 78) is not alone in attributing the rise in homelessness to government policies, albeit unintended ones, such as changes to the rules governing social security eligibility. Williams (1992: 187) goes as far as to suggest that by the early 1990s, the main route into council housing was via homelessness. The better housing stock in the better areas became privatized, and hence (typically) mixed-tenure estates emerged in suburban areas, whilst inner city areas underwent a process of ghettoization (Stewart and Burridge 1989: 76). Williams (1992: 182) argues that 'All the evidence shows growing concentrations of people on the lowest incomes without employment and reliant on state benefits' in the local authority housing sector. Monk and Kleinman (1989: 129) argue that this has led to the emergence of a new underclass of housing with around 20-30 per cent of the homeless and badly housed who can expect little improvement in the situations. In this respect, as Cole and Furbey (1994: 208) argue, during the period from 1979 until the early 1990s, housing tenure became increasingly polarized between an owner-occupation sector catering for the affluent and/or aspiring skilled working class and middle class (the 'haves') and a council-provided sector catering for the poor, the unemployed, the elderly and ethnic minority households (the 'have nots'-the 'might haves' having migrated into the 'haves'). These trends are confirmed by our analyses of the BCS, BSAS and GHS. 3 Tables 3 and 4 report on various markers of residualization from the BCS and BSAS. We see, e.g., that the BCS owners are likely to experience low levels of unemployment (between 3 Further details about the dataset we have collated can be found in Jennings et al. (2015). The BCS was first conducted in 1982 and was commissioned by the UK government to measure the 'dark figure' of unreported crime incidents. The survey moved to an annual basis from 2001. As well as collecting information on victimization and fear of crime, it collects information on a range of attitudes towards the criminal justice system, causes of crime and demographic information about respondents. Our collation of the data collected from 1982 until 2012 resulted in a dataset of almost 600,000 respondents. The BSAS began in 1983. It is based on an annual random probability, face-to-face survey of Britons. The series was designed to act as a counterpart to other large-scale government surveys such as the Labour Force Survey or the General Household Survey, which provide data on behavioural actions and tangible 'facts'. It has been conducted every year since 1983, except in 1988 and 1992. Our collation of the data collected from 1982 until 2012 resulted in a dataset of almost 90,000 respondents. The General Household Survey started in 1971, and interviews about 10,000 people each year.
2 and 4 per cent), whilst for the social renters, this is between 6 and 11 per cent. These track general trends in the unemployment rate, but the increases in the rates are higher for those in the social rented sector and slower to fall. Unsurprisingly, therefore, more of those in social rented housing experience low levels of income. Both the BCS and BSAS suggest that over time, the percentage of owners in the lowest income bracket declines, whilst the same figure for social renters either plateaus around 40-50 per cent (BCS) or increases (BSAS). The BCS asked its fieldwork team to assess the state of local housing in the area adjacent to the homes in which they complete interviews. Between 1984 and 1994, this hovered around 2 per cent for owners, whilst for social renters, it remained around the 9 per cent level (despite the work of Housing Associations towards the end of this period). Penetration into the owners market by ethnic minorities ran at around 4-5 per cent using both the BCS and BSAS. However, amongst social renters, both data sets suggest that more of the people living in this tenure were made up of ethnic minorities (rising from 4 to 5 per cent in the early 1980s to around 10 per cent by the late 1990s). Table 4 deals with welfare recipients in the two social groups we are most interested in (owners and social renters) and again supports the general notion that levels of residualization have increased amongst the social renters. For example, whilst 40 per cent of owners were in receipt of incapacity benefit in 1986, the same figure for social renters was 67 per cent, and whilst the owners slowly decreased between 1986 and 1991, for the social renters, there was very little meaningful change in the percentage claiming incapacity benefits. Whilst a very small number of owners were receiving single parent benefits (2-3 per cent) for social renters, this was rising (from 6 per cent to around 10 per cent). Similarly, owners saw declines in housing benefits, whilst social renters (always higher anyway) saw increases. A similar trend is observable for unemployment; although both groups saw declining percentages of people claiming unemployment benefit.
The GHS also asked about economic activity and tenure. Looking at economic activity by tenure since 1972 until the mid-1990s (Tables 5 and 6), one sees a distinct shift from social renters and owners/those buying with a mortgage having similar rates of unemployment to quite distinctly higher rates for social renters. It is clear, therefore, by all of the measures, we have explored, and using three of the most respected data sources on these topics, that the social renters have slowly become part of a residualized sector in society, earning less than owners, relying more heavily on specific welfare schemes, working less, being more likely to come from ethnic minorities and living in areas with both a high turnover of residents and a greater percentage of housing stock in poor conditions.

Crime and the 'Problem Estate': An Historical View
The council housing estates which provided homes for the affluent working class in the interwar and immediately post-war periods were not associated with crime, disorder or social disorganization at all (Murie 1997). At this stage, it was the declining, inner city, privately rented households which were associated with crime, high levels of turnover and low levels of social control (Murie 1997). Such areas developed a reputation for being 'poor'-along a number of indices: in terms of their physical structure, in terms of the other residents and in terms of their prospects for improvement (Bottoms et al. 1992: 132;Murie 1997: 24). Council housing did not make up a very large proportion of the areas which were deemed (officially or informally) to be 'problem estates', although as Baldwin and Bottoms (1976: 163) show, the age of an estate was associated with variations in the rate of resident offenders. As such, the social rented housing sector was an extremely complex one. Generally, however, council estates saw very high levels of employment and very low levels of disorder. In some cases, by dint of many residents sharing the same employer or working in very highly related and integrated sectors of the economy, the relationships which fostered appropriate levels of informal social control in the workplace, in schools and in trade unions often cascaded into relationships between neighbours-who often worked together in order to maintain and produce a well-ordered environment in which rules were obeyed and authority respected (Bottoms et al. 1992: 134;Murie 1997). In this respect, it is important to bear in mind that the association between council housing and crime is (1) a relatively recent one and (2) not, therefore, an automatic one.  III: unemployment, 1972-94 (GHS, even years) 197219741976197819801982198419901994 Owners 2  1  2  2  3  4  4  4  3  3  5  4  Social renters  3  2  5  5  6  11  11  11  8  8  11  11 All figures are percentages and unweighted. Page 10 of 18 Having reviewed the available evidence on the relationship between crime and housing tenure, Murie (1997: 26) concludes that '… the spatial distribution of crime and victimisation owes a considerable amount to the spatial concentration of council housing' and, in particular, the changes in the nature of the social groups who inhabit such housing. Murie explains the changes in the spatial concentration of crime via processes outlined above (chiefly residualization, polarization and privatization). As Murie (1997) notes, housing policies, the processes of allocation and social stigma all contribute to the incidence and spatial distribution of crime. For example, not all council estates were seen as equally 'good', and those applicants who were felt to be 'less deserving' were allocated accommodation on such estates (Murie 1997: 28). Similarly discriminatory practices helped to produce some estates in which there were concentrations of vulnerable people-but in the main the concentrations and their associations with crime were the exception, not the rule. However, whilst these trends can be identified going back over a number of years, the pace of such changes has quickened over the past 20 years.

Crime and the Right to Buy: Our Contribution
Our aims in this contribution are to explore the long-term social and socio-spatial effects of the 1980 Housing Act (and those Acts which followed it), and, in particular, the right to buy, on the spatial distribution of crime in English and Welsh cities. Our aim is to assess the extent to which those who remained in council-owned houses (and similar accommodation owned by other organizations such as housing associations) became at increased risk of property victimization, such as burglary, theft from the home, vandalism and car crime.
We explore the impact of the right to buy and related legislation on the socio-spatial distribution of crime using two principal data sets-namely, the BCS and the BSAS, and in so doing we focus on the data collected under the auspices of these data sets form 1982 (in the case of the BCS) and 1983 (in the case of the BSAS). Our main focus will be on crimes relating to domestic property crime in the year prior to the survey (namely, someone breaking into the respondent's home, someone trying to break in to steal or cause damage, anything being stolen from within the home or anything being stolen from outside the home, a definition very close to that used by Trickett et al. 1995). Ultimately, we show that the ideological basis of the right to buy legislation, as well as shifting the nature of housing tenure in England and Wales, is also implicated in the increase in property crime in some of our cities' poorer estates. As such, ideological changes, when enacted via legal systems can act to produce changes-albeit slowly-in the distribution of some forms of property victimization.

Crime prevention
Before we examine the relationship between housing tenure and victimization, let us look at the degree to which owners and social renters were protected from burglary victimization by, in this case, burglar alarms (we choose burglar alarms since the BCS has asked about these, with the exception of 1984, since 1982). This data (Table 6) suggest that whilst the two groups were roughly on a par with one another in 1982, this parity quickly disappeared, and by the turn of the century, three times as many owners and those buying with a mortgage had burglar alarms when compared to those in social renting. So, even though coverage increased, it did so unevenly.

Tenure and victimization
Let us now turn to explore the relationship between victimization (in or near to the homes of survey respondents) and tenure over time. One of the problems that we face in undertaking such an analysis is that the BCS did not start until 1982, and therefore the earliest self-report victimization data which it can provide us with is for the year 1981. As this came after the commencement of the right to buy, the BCS cannot provide with information about the relative distribution of victimization prior to the commencement of council house sales. However, the GHS fielded a series of questions about theft from and within the home in 1972, 1973 and 1980 which provide us with a 'baseline' picture of the relationship between tenure and victimization. Table 7 reports the data from the GHS for the questions about being burgled in the previous year for 1972, 1973 and 1980 (it was not asked in any other years prior to 1980). These questions suggest that housing tenure was not strongly associated with burglary in the years prior to the right to buy legislation coming into force. The data present a remarkably stable picture; about 2 per cent of owners and 3 per cent of renters had been burgled in the previous year (so 1971, 1972 and 1979). 4 Table 8 explores, by way of and Mann-Whitney's U-tests, the average number of such victimizations by owners/renters. 5 This suggests that the average number of domestic thefts for renters (which would have included those in the private rented sector) was higher than for those owning or buying their own homes. However, this was, even with just three data points, a relatively stable situation, with the mean difference (around 0.018) hardly changing at all. So whilst renters experienced on average more domestic thefts between 1971 and 1979, there appeared to be no underlying change in the relative risks in the eight or so years before the right to buy was introduced.  *** *** *** ***p < .000. 4 The GHS data distinguish between owners/buyers and renters (rather than those renting privately or from social landlords, although the vast majority would have been renting from social landlords, and in this era that would have meant local authorities). 5 Since our variable (the number of victimizations) is skewed, we report tests using Mann-Whitney's U. The Mann-Whitney test is a non-parametric test that allows two groups/conditions/treatments to be compared without making the assumption that values are normally distributed. The logic behind the Mann-Whitney test is to rank the data for each condition, and then see how different the two rank totals are from each other (Mann and Whitney 1947). Let us now turn to the BCS data for the experiences of owners and renters from 1981 (the year which the 1982 BCS would have asked about). Of course, the comparison with the earlier GHS is not perfect (since the two surveys ask slightly differently worded questions), but the data are sufficient for our purposes. Whilst the GHS data suggested little difference between the owners and renters, this picture changes when we look at the BCS data (for which we have combined into one measure items asking about entering the home and causing damage, entering the home to steal or cause damage, having items stolen from within the home and having items stolen from outside the home) (see Table 9).
This data suggest that both social renters and home owners saw increases in crimes around their homes during the 1980s. However, it does not suggest that there were huge variations in experiences; more social renters had experienced this sort of victimization, but their rates had not increased terribly much higher than those of owners. Table 10 suggests a slightly different picture, however; what the data demonstrates is that in 1982, social renters already had about twice the levels of burglary victimization than owners did (in line with the GHS proportions, see Table 8). Whilst the general trend for the social renters is upwards until 1996 (with a slight dip in 1984), the owners experienced a far 'shallower' increase (which plateaued from 1984 until 1992) and then a decline from a peak of 0.1724 in 1994. By the end of the data run in 2000, their average number of victimizations of this sort had declined to below 1982 levels. For social renters, the situation in 2000 had worsened considerably; in 1982, the average number of burglaries was 0.0891, a figure which had risen to 0.1685 by 2000 (in effect, a doubling).

Discussion
We have established that since 1980, the size of the social renting sector has (1) declined in size and (2) become residualized (i.e. narrowed the social base from which it draws). Whilst this narrative fits with the existing literature (e.g. Murie 1997;Kemp 2014), there are two potential ways in which these changes may have been related to higher and Table 9 Domestic property crimes by owners/mortgaged and social renters (BCS) 1982(BCS) 1984(BCS) 1988(BCS) 1992(BCS) 1994(BCS) 1996(BCS) 1998 Owners 8 8 9 9 11 10 9 Social renters 11 10 11 13 12 12 11 All figures are percentages and weighted for households. *** *** *** *** *** *** *** *** more concentrated levels of domestic property crime. The first of these relates to what may be thought of changes at the level of the individual household, whilst the second to changes in the social make-up of some council/social landlord estates. Let us take these in order. So, e.g., if more households in some of the more attractive estates with generally lower levels of crime buy their houses, then there is little redistribution of crime geographically-the changes which we have identified are simply the result of our 're-labelling' these household as 'owners'. This has the effect of making it appear that there was a change in crime levels associated with different tenures-but actually, there might be little (or no) change in where such crimes took place geographically, it is simply that we have shifted one set of social renters into a different category. Of course, over time, former-council renters who now owned their homes may choose to invest in additional security devices (such as burglar alarms). In any case, it does not assume that there was a process of what might be termed 'residential replacement' (i.e. a turnover in residents). When this residential replacement came, it too did little to alter crime rates, we suspect. This is because, when council house tenants who had bought their homes came to sell them, the general rise in house properties meant that they were (in the main) selling to those who could get mortgages and afford to buy these homes. Hence estates with low levels of crime tended to remain low crime estates. The second possibility (which operates at a different level of explanation) is that the 'worst' estates, which always had more crime and victimization, experienced an increase in property crime. This might have been because, over time, the concentration of households with social and economic problems increased as a result of the economic down turn and the loss of employment in manufacturing, coal mining and allied trades. Those who could leave did so leaving their council homes to be re-let by the council to those in greater need (residential replacement) and that this brought with it the sorts of social and economic needs associated with crime and disorder. Such processes would have been most keenly felt in those estates in which few had bought their council homes and where jobs, and the hopes which they had given people, were lost during processes of 'restructuring'.
Disentangling these two processes is, using the data available to us, not an easy task. However, some informed assessments can be made. It is unlikely, we think, that the right to buy led to dramatic increases in rates of residential replacement immediately. Those tenants who bought did so because they liked their homes and the areas they were in and were not speculating in the (then embryonic) housing market (Forrest and Murie 1990: 23). This suggests that the first explanation above is plausible. However, with the passage of time, there is a greater (and increasing) rate of residential replacement. The first of the right to buy homes start to be resold and in some areas the resale values of these put them beyond the reach of lower income households, tending to reinforce levels of 'exclusivity' (Forrest and Murie 1988: 172-3) and social control (and hence suppressing dramatic increases in crime-even if increases were experienced). In those areas in which fewer former-council houses were bought, the pressure to accommodate increasing numbers of those in social need meant that the estates which remained in council ownership saw increases in anti-social behaviour and crime. As this suggests, the second explanation starts to become more significant over time. Particularly notable here is that, using data from the 2000 BCS, Tseloni (2006) found that one of the strongest predictors of rates of property crime was share of social housing; social renting increased household crimes by at least 40 per cent compared with owner-occupancy, 'possibly because of entangled effects of social vulnerability and/or proximity to potential offenders' (Tseloni, 2006: 219;cf. Kershaw and Tseloni 2005).
The relative pace at which these processes took hold will, of course, be shaped by local policies and economic fortunes. Objectively similar estates in, say, the south-east and the northeast may have experienced different processes which operated at different speeds due to the availability of employment locally. Additionally, the nature of the stock on council estates (high-rise, low-rise or houses-and again the age and size of these) will also have shaped long-term outcomes.

Conclusion
Our article has sought to contribute the understanding of the structurally driven processes of victimization. Herein, we have taken legislation relating to housing (specifically, the right to buy one's own council house-a key plank in Thatcherite ideology) as the key element in changing the structural processes which shaped the socio-spatial distribution of crime in England and Wales in the late 1970s. Whilst the structures we have explored have been legislative ones, changes in the law do not come from nowhere; they are produced by governments seeking to promote particular ideological agendas. As the processes associated with this restructuring took hold during the 1980s, so we see, using data from a range of respected sources (including government-sponsored surveys) increases in the disparities between owners and social renters in terms of their exposure to domestic property victimization. Elsewhere we have charted the ways in which government policies pursued in the 1980s were associated with rises in property crime Farrall and Jennings 2014). This article suggests that as well as producing a rise in property crime rates (which of course led to the much lauded crime drop), other aspects of the social policies pursued at this time, produced a social and (eventually) geographical concentration of crime amongst some social groups and in some areas of our towns and cities.
Our approach has been informed by the work of Sandra Walklate (1992;Mawby and Walklate 1994;Walklate 1996) who proposed that criminologists adopt a more 'critical' approach to exploring victimization. Using longitudinal data, we have been able to explore the legal and social processes which 'go on behind peoples' backs'. Whilst many of the people living in social housing may be acutely aware of the exaggerated risks they face, few have explored the ways in which political imperatives and legal apparatus operated together to produce such risks. As such, not only might one argue that governments help to shape the levels of crime a society will experience (by, e.g., their handling of the economy) so, it would appear, they can shape which social groups experience increases (or decreases) in their relative rates of victimization. Similar findings have been reported in other countries which underwent New Right-led restructuring; Currie (1990: 308) notes that cuts to the US housing budget in the 1980s 'meant that public housing did often become housing of last resort…' with a 'concentration of social pathology -drug-dealing, violence, gang warfare and family disruption' whilst Levitt (1999: 87) reports that in the mid-1970s, poorer households in the United States were burgled less than richer ones, but by the mid-1990s, this situation had reversed.
For Thatcher, the political requirement to build a social base for the ideas embodied in the-at that time-still embryonic project we now know as 'Thatcherism' meant that housing became a key policy arena. For various reasons, many of which we outline above, housing was targeted for ideological reasons (which supported the project of Thatcherism at an ideational level) and for what it might do for the electoral base (the assumption being that hone-owners would start to vote conservative). In these ways, the changes to housing laws enacted in the 1980s were not (as Blandy and Hunter 2012 note) a 'neutral' instrument, but one which could be used to reinforce various forms of inequalities-some of which, as we document above, related to exposure to victimization. One of the ironies in the tale of Thatcherism and crime was that council housing was not associated with crime prior to the right to buy legislation (Murie 1997), but slowly became virtually synonymous with it. This in turn, along with rising crime rates generally, and the consequences of social policy endeavours in other parts of society (such as the withdrawal of economic support from some sectors of the economy, the changes to the social security system, Care in the Community policies (Murie 1997) and the later increases in the exclusion of children from school, which was associated with anti-social behaviour, Timmins 2001: 566), led to a situation in which some of the key ingredients in the recipe for rising crime rates were brought together in temporally and spatially. For example, league tables of school exams were first published in 1992 (Timmins 2001: 519), which encouraged head teachers to exclude unruly children (in order to increase school examination performance). Accordingly, school exclusions rose throughout the 1990s (reaching a peak of 12,668 in 1996-97; DFeS 2001), resulting in increases in anti-social behaviour as such children continued to behaviour badly, only now in public areas, such as residential streets, rather than at school (Berridge et al. 2001;Timmins 2001: 566). This concentration, socially and spatially organized as it was, tended to produce feedback loops, whereby those who could leave such estates did so [Forrest and Murie (1990: 47) report that movers were more likely to be non-manual workers than those who bought their former-council homes and remained in them] and those who were left behind were increasingly marginalized from the political debates (the 'have nots'; Hay 1992). Such marginalization, however, also allowed for this social group to be presented (against the backdrop of worklessness, crime, fragmented families and 'failing' schools) as 'workshy' and as 'benefits scroungers'-an image which successive governments have not attempted to dispel. As such, legal instruments and political imperatives are heavily implicated in the concentration of crime (and hence victimization risk) and the magnification of inequalities in England and Wales in the 25 or 30 years since the introduction of the right to buy.

Funding
We would like to acknowledge the support of the Economic and Social Research Council (as award number ES/K006398/1 in enabling this research to have been conducted. 6