Politics,Social and Economic Change,and Crime: Exploring the Impact of Contextual Effects on Offending Trajectories *

Outlining the Life Course Perspective

As many have come to recognize, ⁴ the life course perspective has had a dramatic impact on thinking within criminology, especially since the early 1990s. Indeed, a very large part of contemporary criminology’s theoretical apparatus is derived from the work of life course scholars. ⁵ A key aim of the life course perspective is to draw links between macro-level social history and social structures and the lives of individuals and communities. The life course perspective aims to explore “pathways through the age differentiated life-span,” and it is “manifested in expectations and options that impinge on decision processes and the course of events that give shape to life stages, transitions and turning points.” ⁶ Two concepts central to the perspective are the notions of “trajectory” and “transitions.” The first refers to a line of development over the life course (e.g., an employment career); the second refers to events (e.g., first job, promotion) that shape a trajectory. Robert Sampson and John Laub describe the perspective as focusing on “the duration, timing and ordering of major social events and their consequences for later social development.” ⁷ As Elder, Modell, and Parkes note, rapid social change has the ability to rearrange the timing and sequence of events in the transition to adulthood. ⁸

Glen Elder and Janet Giele also note the importance of locating people (particular individuals, subgroups, or cohorts) in specific communities and at specific historical moments. ⁹ Such thinking forces one to recognize that individuals do not exist in isolation; they are embedded in wider social, economic, and political contexts and relationships. Similarly, individuals are key constitutive parts of what is termed the “family-cycle,” whereby people move from being children to leaving the parental home, forming partnerships, rearing their own children, and surviving into old age. These events are not strictly age-defined, so the typicality of any form of such cycle has waxed and waned (as age of marriage, age at first childbirth, and rates of step-families have fluctuated). As individuals exert agency, the combination of people and wider social structures brings forth the possibility of what Elder and Giele refer to as a “loose-coupling” between age-graded life courses and individual choice. ¹⁰ Norms exist, but individuals are able to depart from them or to adapt them to suit their own circumstances. This highlights the extent to which variations in the timing and sequence of life course events may produce substantive differences in outcomes (or result from other differences).

The focus on wider social and economic structures in the work of Elder and others highlights the ways in which individuals’ lives are linked to one another. ¹¹ Events and long-term trajectories in the lives of parents in a family may alter the life courses of their offspring. As the individuals who make up families age, they form an aging social network, referred to as a “social convoy”: ¹² a group of interconnected people who move through time together. As Phyllis Moen and Elaine Hernandez note, ¹³ an individual’s resources, deficits in those resources, strains on them, increases in them, and so on, become drivers of transitions or turning points not just in the lives of the individuals themselves, but also in the lives of those people who are in some way related to them (either socially or biologically). For example, the loss of work for a parent on whom a family had relied affects not just the individual concerned but their dependents. As Elder remarks, “Each generation is bound to fateful decisions and events in the other’s life course.” ¹⁴ Similarly, the individual’s social network may be affected by the individual’s loss of work or divorce. Thus the concept of a social convoy can be extended from family members; school mates, coworkers, acquaintances, loosely engaged strangers, and so on are potential members of such convoys. Similarly, a lot of attention has been devoted to the idea of timings and to different types of “time.” Duane Alwin distinguishes between historical time (essentially the historical era in which one lives) and biographical time (the life course of the individual concerned) and shows how they may interact. ¹⁵ Meanwhile, Elder introduces the concept of social timing (the duration, incidence, and sequence of age-related expectations and beliefs). ¹⁶

Critiquing Life Course Criminology

One of the debates haunting life course studies is the need for answers to pressing social challenges now. Immediate answers are not always possible. Implicit in this critique is the idea that the social and cultural processes, economic forces, and political ideas that shape the experiences of one generation may not be the same as those which will affect future generations. In short, societies change, and as they change, different drivers to such changes may come to have different impacts over time. Some cultural forces may die out and cease to have any purchase on individuals’ life courses, whereas others may grow in importance. Such thinking, of course, recognizes the problem of age, period, and cohort effects. ¹⁷ The institutions that shape lives may change or cease to exist, yet few criminologists have recognized that the changes in such institutions have been dramatic in some societies. Michael Benson argues that those “working in the life course tradition have not yet devoted as much effort to understanding the contextual effects on trajectories in crime as they have to studying the parameters of careers at the individual-level.” ¹⁸ Since Benson’s critique, others have started to explore the role of neighborhood factors. ¹⁹ However, even that misses the macro-level processes Benson’s critique identifies. Similarly, the educational sociologist Dennis Gleeson argued that behavioralist explanations that focus on individual psychologies and family characteristics pathologize, aiding the creation of stereotypes and simultaneously overlooking the macro-level political, social, and economic processes at play. ²⁰ This critique suggests that additional variables may be needed to comprehend the causes of offending more fully.

A further limitation of criminal-careers research is that, for many very good reasons, studies are often based in one location. The Cambridge Study in Delinquency Development, the Pittsburgh Youth Study, the Rochester Youth Study, the Edinburgh Study of Youth Transitions, and the Peterborough Adolescent Development Study, for example, were based on data from respondents recruited from one place (a whole city or a part or parts of a city). Although in some cases such cities are large and heterogeneous, such designs do not allow for changes affecting different parts of the same country in different ways. Even if the locations where the data were collected are not monocultural at the local level, they are spatially invariant at the macro level; hence the impacts of social and economic changes that do not fall evenly within a country are not available for analysis. One of our aims is to explore the impact of changing social and economic processes on the offending careers of one generation of UK citizens in a way that incorporates macro-level variables that may be causal antecedents of offending and in so doing avoids some of the pitfalls of behavioralist explanations.

In addition, and despite the undoubted quality of much of the research into criminal careers, it remains the case that much quantitative research has tended to tackle causal processes of offending in a largely individualized manner. It ranges from a near-total emphasis on individual-level processes to individual-institutional interactions, ²¹ although more ecological models do exist. ²² This observation led Robert Sampson to note that society and the idea of social change was one of the key elements missing from current research on criminal careers. ²³ Similar observations about life course research have been made by those working outside of criminology. For example, Karl Ulrich Mayer noted that the “unravelling of the impacts of institutional contexts and social processes . . . on life courses has hardly begun,” adding that “we know next to nothing about how the internal dynamics of life-courses and the interaction of developmental and social components of the life-course vary and how they are shaped by the macro contexts of institutions and social policies.” ²⁴

Thus, although life course criminology has meticulously focused on proximal institutions (families, schools, employers, and communities), those institutional arrangements, the discourses and policies that surround and flow from more distal institutions (political parties, governance structures), ²⁵ and the ideas they promulgate (discourses about “the family,” ideological stances on education, economic policies, and so on), have not received very much attention at all. In short, the current approaches adopted by life course criminologists tend to encourage the construction of “the offender” in individualistic terms. It is an offender’s decision making, marriages (or lack thereof), skills, and employment chances that are often found to account for their onset, maintenance, and desistance. ²⁶ This body of scholarship, as technically sophisticated as it is, has failed to engage with literatures suggesting that political decision making (such as economic or welfare policies) shapes the lives and life courses of citizens. This lacuna in criminological research means that the field reproduces constructions and discourses relating to offenders that are inherently pathologizing and that fail to consider the ways in which macro-level political and economic processes and policies may shape individual life courses and engagement in crime. The wider political, economic, and social processes that are implicated in the onset of offender careers and that sustain involvement in crime are therefore routinely excluded from criminological theorizing. Herein we seek to encourage the consideration of the more “distal” influences on criminal careers, such as national-level social and political changes, and the economic and social policies shaping (and shaped by) these. We aim to explore the impact of economic change (and the political ideologies that preceded these transformations and the social consequences that followed them) on offending trajectories using data from the United Kingdom. ²⁷

Inserting Political, Social, and Economic Change into Life Course Criminology

Benson is one author who has attempted to theorize how social and economic change may alter environments in ways that might affect engagement in crime over the life course. ²⁸ He notes that very few criminologists have explored the ways in which the state (and hence changes in those things the state can influence, such as taxation or welfare policies) can shape criminal careers. ²⁹ Benson argues that as well as decisions over which behaviors are criminalized, states can shape offending careers through policies relating to transportation systems, housing, and economic decisions. ³⁰ In exploring the concentration of poverty among the United States’ black population, Benson draws on Wilson’s work to show how decisions made by politicians, state officials, and private individuals have helped to leave the United States’ black population with higher levels of family disruption and residential instability than is the norm in the United States. ³¹ Those processes took several decades to emerge and were underscored by changes in the economy, which saw a long period of economic slowdown after 1974, with the loss of many manufacturing jobs. ³² Such jobs were key to ensuring informal social control among lower-class males, since they offered a chance for relatively highly paid employment without requiring high levels of education. As jobs started to be shed, inner-city areas started to spiral into decay and crime, and the fear of crime began to rise. ³³ Benson theorizes that such changes would have affected the lives of young men (especially) and women who lived in such communities.

Another criminologist who has theorized the ways in which criminal careers may be shaped by wider social and economic factors is John Hagan, who develops a theory of crime and capitalization based on a detailed review of the changes in the US economy since the end of World War II. ³⁴ He notes, like Benson, how economic restructuring, rises in economic inequality, residential segregation, and the concentration of poverty have left some US inner cities with few meaningful employment careers on which to build law-abiding lifestyles. Some communities have, instead of legitimate incomes, started to rely on deviant activities (drug sales, prostitution, and other illegal services) as a means to secure an income. Illegal activities become entrenched over time, as few people have links to legitimate employers and crime becomes embedded in communities’ daily routines. Such activities therefore also start to shape the lives of the men and women who live in those communities.

Nevertheless, we can find no other studies that take seriously the idea that political decision making (around the management of the economy, the allocation of resources, and the extent, e.g., tolerated levels, of economic inequality) may in some way provoke or be associated with offending behaviors at the individual level. Accordingly, our aim is to incorporate an understanding of the role of political processes into the statistical modeling of individual criminal careers, such that current criminological debates about criminal careers pay greater attention to the role of politics (especially with regards to the management of the economy) in the evolution of such careers. In this way, our research adds to the literature on the penal-welfare nexus, in that we explore the ways changes in the social and economic policies pursued by the Thatcher-led governments led to increases in crime, which in turn led to “toughened” criminal justice policies under Prime Minister Tony Blair. ³⁵ We approach the social and economic changes brought about or augmented by the Thatcher and Major governments from 1979 to 1997 as being distal events, which radically altered the social and economic circumstances for many UK citizens (such as regionally varying unemployment). These distal events in turn shaped proximal events and processes that led to some individuals’ becoming involved in crime.

The rest of our article unfolds along the following lines. The next section reviews what is known about the relationship between the economy and crime rates, focusing on the UK experience. It also serves as an overview of the economic history of the United Kingdom in the 1970s. We then outline the data on which we rely and our analytic strategy. The following section presents two near-identical path diagrams (one that does not include a measure of economic change at the county level and one that does). Finding that economic restructuring does help to explain offending careers, we then control for the extent of economic restructuring at the county level to assess the geographical impact of economic restructuring on offending careers. Finally, we reflect on the findings we have produced and locate them within wider literatures on both offending careers and economic, social, and political change.

The United Kingdom in the 1970s, the 1980s New Right, and the Radical Reshaping of Society

There is an extensive and long-standing literature on the relationship between economic conditions and crime, focusing in particular on effects of the business cycle and unemployment rates. ³⁶ Most studies test the economy-crime link in the context of evidence from the United States or the United Kingdom. ³⁷ Economic hardship interacts with effects of other social and criminal justice policies. For example, Christopher Hale and Dina Sabbagh find that the effect of increases in police manpower in Britain during the 1980s was nullified by the sizable increases in unemployment experienced at that same time. ³⁸

The literature suggests that lagging and coincident indicators of economic conditions, such as the unemployment rate, income inequality, GDP, and consumer sentiment, contribute to increased rates of property crime through either the production of need and hence criminal motivation or the relative availability of “stealable” goods and commodities. ³⁹ It is, however, rare for analyses of the economy-crime link to be embedded in theories about wider social, economic, and political changes. That is to say, the existing studies are conceived as a test of the effect of economic variables on rates of crime, rather than a reflection on either the historical backdrop to changes in the crime rate or the theoretical foundations of the economy-crime link.

Let us now turn to a consideration of the United Kingdom’s economic fortunes between the 1970s and the 1990s. Throughout much of the 1970s, the United Kingdom faced considerable economic difficulties. The inflation rate, which started at around 7 percent in 1972, reached 24 percent in 1975; ⁴⁰ it stood at over 13 percent in 1979. Unemployment rates fared little better (despite a drop in 1973–74 to 3 percent); they stabilized at around 5 percent for the period from 1976 to 1979. ⁴¹ The breakdown of the relationship between unemployment and inflation led British governments to retreat from Keynesian fiscal policies and focus, to varying degrees, on monetarist policies and the welfare retrenchment associated with these. ⁴² In addition, the United Kingdom borrowed $3.9 billion from the IMF in 1976, resulting in cuts in education budgets. ⁴³ Declining real wages from 1975 to 1978 and widespread strikes culminated in the so-called Winter of Discontent in 1978–79. ⁴⁴

In May 1979, the United Kingdom elected its first “radical right” government. Space precludes a full discussion of the impact of the Thatcher government’s policies on UK society, ⁴⁵ but it is important to highlight some of early economic policies associated with this government. One of the first things it did was to increase interest rates, which had the unintended consequence of immediately weakening the United Kingdom’s manufacturing sector and producing a sharp fall in manufacturing output between 1979 and 1981. ⁴⁶ Indeed, the economy experienced negative growth for much of the early 1980s. ⁴⁷ As the early Thatcher governments focused on reducing inflation over the goal of full employment, unemployment increased in the 1980s (Fig. 1). So damaging were their policies that the Conservatives abandoned their monetarist ideals (their monetarist phase ran between 1981 and 1984). However, the UK economy’s troubles persisted for many years, with widespread economic disruption and unemployment. During the mid-1980s there was a year-long miners’ strike, which the National Union of Mineworkers lost, resulting in the closure of many mines and the loss of tens of thousands of jobs. The abandonment of monetarism was followed by a focus on privatization and financial deregulation (1983–86), which contributed to the so-called Lawson Boom of 1986–88. The economic and social turbulence, however, was not evenly distributed across the United Kingdom. The communities most heavily affected were those most reliant on heavy industry and manufacturing, which were located in the Midlands and the North of England, South Wales, and central Scotland (all places once associated with mining or steel production and manufacturing). Accordingly, the unemployment rate rose dramatically, reaching almost 12 percent by the mid-1980s, from around 2 percent in the mid-1970s. From the mid-1980s, inflation also rose, and successive budgets reduced personal taxation. The economic restructuring, almost always associated with processes of deindustrialization (which had started in the late-1960s), ⁴⁸ was consistently associated with rising unemployment, which led to increasing social and political polarization. ⁴⁹

OPEN IN VIEWER

Figure 1.

National Unemployment Rate (Percentage), 1970–2006.

The liberalization of the United Kingdom’s financial markets and changes to the tax regime from the 1980s contributed to rising inequality (Fig. 2) as the Thatcher government abolished higher rates of taxation and increased indirect taxation, which had a greater marginal effect on the less well off.

OPEN IN VIEWER

Figure 2.

Income Inequality (Gini Coefficient, after Housing Costs), 1970–2006.

These trends in the postwar British economy correspond to long-term patterns of change in the rate of property crime (Fig. 3), compiled from the Home Office’s recorded statistics for England and Wales, 1961–2006. The steepest rises in the rate of property crime occurred during the 1980s and early 1990s, also coinciding with sharp increases in levels of unemployment and inequality. This is consistent with the findings generated by statistical analyses of the link between economic conditions and crime. ⁵⁰

OPEN IN VIEWER

Figure 3.

Property Crime per Capita.

Statistical modeling suggests that the economic policies initiated by the Thatcher governments forced several state-owned industrial companies (e.g., British Steel, British Coal, and car manufacturers, some of which were state-owned) to make employees redundant. ⁵¹ The eventual loss by the National Union of Miners of the year-long miners’ strike resulted in substantial job losses. As these job losses cascaded through the economy, so jobs in sectors that depended on mining (such as railways and engineering) were also lost. In short, the economic restructuring that started in the 1960s reached a zenith during the 1980s. From the mid- to late 1980s the United Kingdom started the transition to a postindustrial nation. The official document Social Trends for 2007 reports that:

over the last 25 years the UK economy has experienced structural change. . . . The extraction and production industries, made up of agriculture and fishing, energy and water, manufacturing, and construction showed a combined fall of 43 percent from 8.2m jobs in 1981 to 4.7m jobs in 2006. Manufacturing alone accounted for 81 percent of this decline, with the number of employee jobs in this sector nearly halving from 5.9m in 1981 to 3m in 2006. ⁵²

As modeling has shown, ⁵³ the resulting increase in unemployment was associated with an increase in property crime. Such policies also resulted in declines in the real term value of welfare benefits (which did not keep up with inflation), increases in unemployment, economic inequality, and crime, a toughening of the criminal justice system, ⁵⁴ and increases in penal populism. ⁵⁵

Nevertheless, there are deficiencies with this previous work. The first is the obvious question of the ecological fallacy. Are the national-level findings corroborated by individual- and region-level data? In addition to this, national-level modeling is unable to assess the degree to which the effects of the economic policies pursued fell unevenly in the United Kingdom. The United Kingdom’s industrial geography is an uneven one, with the industrial base spatially clustered. ⁵⁶ Herein we develop both of these matters via the use of longitudinal, individual-level data, analyzed in such a way as to enable us to explore the geographical effects of economic policies. As well as speaking to debates in criminology, our article therefore makes contributions to political geography and wider understandings of macro-level economic policies on the lives of citizens.

Research Design

Recall that our key aim is to explore processes of political change and offending trajectories at the individual level. Although no data set could ever be perfect for this, the British Cohort Study (hereafter BSC70) makes an extremely good vehicle with which to study the impact of dramatic social, economic, and policy change on a cohort of people. The cohort was born in one week of April 1970 and grew up in the 1980s (during which they would have experienced changes in economic, social welfare, housing, and schooling policies). The BCS70 is large enough for us to explore the unfolding of life courses over several years. In all, 16,135 babies were born and recruited into the BCS70 (98 percent of all births in that week). Although the births are limited to one week in 1970, there are no reasons why this cohort ought to be considered unique or nonrepresentative in any way. As such, the cohort has been repeatedly used as if it were a nationally representative sample. The BCS70 allows us to explore the social and economic changes of the 1980s because of the regular timing of the follow-up interviews (at ages five, ten, sixteen, twenty-six, and thirty). Mothers were interviewed in 1970, providing us some background data on the social environments in which the cohort would spend its formative years. That was also one year before the United Kingdom’s 1971 census was undertaken. In 1975 mothers were reinterviewed and asked questions about their children’s behavior. In 1980 the children were interviewed and questions relating to crime were first fielded. That interview preceded the United Kingdom’s 1981 census by one year. In 1986 (when the cohort was sixteen) the survey questions were expanded to include contact with the police and convictions in court. Those topics were revisited in 1996, when the cohort was twenty-six, and again in 2000. Teachers and head teachers were also interviewed while the children were at school (in 1975, 1980, and 1986). The survey regularly fielded questions on cohort members’ social and economic circumstances (type of housing, neighborhood characteristics, schooling and employment experiences, household composition, home leaving, homelessness, relationship formation, marriage and child rearing, peer relations, and medical experiences), as well as social attitudes, political affiliation, alcohol consumption, and psychological well-being. The BCS70 cohort’s geographical location was also recorded at each interview, allowing us to undertake county-level analyses. ⁵⁷

Analytic Strategy and Results

We adopt an analytic strategy building on that proposed by Glen Elder and Lisa Pellerin for linking structures and human lives. ⁵⁸ We explore how involvement with key social institutions and the organizations associated with them can initiate and maintain (or in some cases alter) the trajectories of offending careers. Because much offending occurs in early adolescence, but for some individuals is likely to be maintained into adulthood, our model measures offending twice: at age sixteen and in adulthood, up to age thirty. We first build a “baseline” structural equation model including key individual-level variables (e.g., relationships with key social institutions such as families of origin, schools, employers, and marital relationships). ⁵⁹ This model (outlined in Fig. 4A) allows us to identify the causal structure of offending and the social, institutional, and organizational backgrounds against which this takes place. This represents a fairly standard model of offending over the life course and fits the data well (Fig. 4B). Next, we rerun the model (outlined in Figs. 5A and 5B), but include a measure of economic restructuring. Let us commence by outlining Figure 4A.

OPEN IN VIEWER

Figure 4A.

Baseline Model of Offending.

OPEN IN VIEWER

Figure 4B.

Baseline Model of Offending (with Estimates).

OPEN IN VIEWER

Figure 5A.

Expanded Model of Offending.

OPEN IN VIEWER

Figure 5B.

Expanded Model of Offending (with Estimates).

Figure 4A specifies a regression path from being Disciplined at School (as reported by the child’s teacher in 1980) to feelings of School Alienation (as reported by the child in 1986). ⁶⁰ The first of these variables relates to more serious forms of school punishment, such as being suspended or excluded, being caned, being given another form of corporal punishment, or having a report on the child’s behavior sent home to the parents. The questions recorded how often the teachers had used those forms of punishment in general (rather than for the child in question) and were coded as “never,” “rarely,” “occasionally,” and “often.” Although this measure does not provide us with data on the child’s experiences, it does provide us with an insight into the approach to discipline in the schools where the child was located while at school (an important contextual variable associated with school outcomes). ⁶¹ A child who sees directly or learns of severe forms of punishment secondhand will be affected by the general punishment milieu. Higher scores indicate greater experiences of the more serious forms of punishment. School Alienation was measured by seven items that focused on the child’s feelings about school (and dealt with the extent to which the child felt that “school is largely a waste of time”; that they were told to be “quiet in classroom and get on with work”; that they thought “homework is a bore”; that they found “it difficult to keep my mind on work”; that they “never take work seriously”; that they did “not like school”; and that they believed making plans was “pointless; take things as they come”). The children could say whether they found these statements “very true,” “partly true,” or “not true at all.” These were factor analyzed (KMO value of .836 and producing one factor with an Eigen value of 2.924). Higher scores on the School Alienation measure indicated higher levels of alienation. From School Alienation we specify three further paths: to Employed (at 26), to Offending (10 to 16), and to Offending (16 to 30). ⁶² These were all asked of the cohort members. The first of these variables is a binary of whether or not the cohort member was working (full time or part time, and including those studying, temporarily sick and off work, in training or looking after the home) or were unemployed (including those on long-term sick leave). The first offending measure relates to the number of times they had been cautioned at a police station between ten and sixteen. In order to measure the same sort of experiences between sixteen and thirty, we used data reported at age thirty relating to experiences of being arrested and taken to a police station going back to sixteen. We specify a path from Employed (at 26) to Offending (16 to 30), and one from Offending (10 to 16) to Offending (16 to 30). The upper half of the path diagram models forms of formal social control (school and employment); the lower half models familial processes. We specify a path from being on the “At Risk” Register (at 10) (measured by school nurses or health visitors and based on school health records), to the child’s report of the Quality of Relationship with Their Parents (at 16), which is a sum of their answers to questions about their parents “treating me like a child”; “not understand[ing] me/my motives”; and “being too strict, bossy and having too many rules” (children could agree or disagree with each statement). This we also regress on to Offending (10 to 16), and a further self-report of the Quality of Relationship with Their Partner (at 30) (a single item) and from this, finally, to Offending (16 to 30). ⁶³

The model fitted the data well (NFI = .956, CFI = .969, RMSEA = .011) and explained about 27 percent of the observed variance, with seven of the twelve paths statistically significant (see Fig. 4B for the standardized regression coefficients, which are marked in bold if they are statistically significant at least at the <.05 level). Four of the five nonstatistically significant paths relate to familial processes; all of the key explanatory variables associated with offending were associated with school or employment. Being on the “At Risk” Register (at 10) was the only familial process variable that was statistically significantly related to offending (for offending between ages ten and sixteen). Schools that tended to use more severe forms of discipline tended to have more “alienated” children at age sixteen. School Alienation (at 16) was related to Employment (at 26), such that those who were more alienated from school were less likely to be employed at 26, and Offending (10 to 16) and Offending (16 to 30) such that the more alienated were more likely to have been arrested by the police. As one might imagine, being on the “At Risk” Register (at 10) was also associated with Offending (10 to 16). Being employed at twenty-six was negatively associated with Offending (16 to 30), suggesting that employment acted as a suppressant factor on offending. There was a high degree of stability in terms of offending between childhood (ages ten to sixteen) and adulthood (ages sixteen to thirty). Quality of relationships (with parents and partners) was not a key part of the model; but it may reflect the instability of such measures, which may fluctuate, making measurement less stable. In short, the baseline model we have fitted replicates earlier models in this tradition, with the theoretical expectations being in line with the findings of others. ⁶⁴

The second step of our analytic strategy sees us repeat this model, this time including a variable that captures the degree of economic restructuring experienced by the community in which each BSC70 cohort member was living at age sixteen (Fig. 5A). Our measure of Economic Change (1971–81) uses data from the 1971 and 1981 censuses and is the sum of (a) the proportion of the economically active population employed in mining in 1971 and (b) the proportion of economically active unemployed males in 1981.

These proportions were summed for each county and applied to each cohort member based on the county where they were living in 1986 at age sixteen. We chose data for those working in coal mining in 1971, as they were a good barometer of industrial activity at that time. Mining was co-located with steel production and processing in South Wales, South Yorkshire, Central Belt Scotland, and Teesside; ship building (in and around Glasgow in particular); and the maintenance of locomotives and railway distribution in centers in Derby, Doncaster, Nottingham, Sheffield, York, and Central Belt Scotland. In 1970 there were approximately 290,000 people (mainly men) working in 293 mines. By 1986, following the year-long miners’ strike, there were roughly 91,000 working in 110 mines. ⁶⁵ By summing this with the proportion of unemployed males in 1981, we created a single item measure of areas that were heavily industrial at t1 (1971) and at t2 (1981) were experiencing high levels of male unemployment, in most cases due to job losses among industrial and related employers, which had not recovered after ten years. Our measure therefore describes the economic trajectories of industrial areas over ten years.

This variable measures change over time, since there was a significant and rapid loss of employment in industries associated with mining and male employees were among those most affected—hence, our use of two measures at two points of time. Although there were other social changes taking place alongside these processes (such as the greater inclusion of females in the labor market), for many individual households such processes were in part a response to the loss of traditional forms of (male) employment. Many such communities lived and worked closely together such that local state housing estates (“council houses”) were dominated by families who derived their household incomes from the same employer (or interdependent employers), meaning that when coal production declined or ceased altogether in one community, the livelihoods of whole estates were affected.

The expanded model is identical to the baseline model, with the exception of the inclusion of the Economic Change variable described above. This we locate to the left-hand side of the model (implying temporal and causal precedence), and from it we specify paths to five of the variables in the original model: Disciplined at School, School Alienation, Offending (10 to 16), “At Risk” Register (at 10), and Quality of Relationship with Their Partner (at 30). Of these five new paths, three were statistically significant: those leading to Disciplined at School, School Alienation, and Offending (10 to 16). The model (Fig. 5B) fitted the data well (NFI = .942, CFI = .958, and the RMSEA was .012), and again explained about 27 percent of the observed variance. Of the statistically significant paths specified, the model indicates that greater levels of Economic Change were associated with schools that reported using more severe forms of discipline, Disciplined at School (at 10). This suggests that as economic change took place, schools used more severe discipline measures more frequently. It could be that the children themselves were less well behaved (and hence the teachers responded more punitively), or it could be that economic change (independent of its effect on children’s behaviors) increased the use of severe school discipline measures. Similarly, in the expanded model, local Economic Change was associated with higher scores on the School Alienation (at 16) measure (suggesting that economic change was associated with increases in feelings of school alienation), and with Offending (16 to 30), such that people living in areas that experienced greater economic change were more likely to offend when aged ten to sixteen. The measure of the types of discipline meted out at school, Disciplined at School (at 10), is associated with alienated at school, School Alienation (at 16). Being on the “At Risk” Register (at 10) is also associated with Offending (10 to 16).

Our analyses suggest that economic change lay behind some of the (seemingly) individual-level processes associated with the commencement of criminal careers, such as weakened social bonds with key institutions such as schools and that economic change may (in some circumstances) encourage disengagement from schools. ⁶⁶ These relationships suggest that greater levels of economic restructuring are associated with greater use of punitive measures in schools in those areas that experienced greater levels of restructuring; the same was also true for alienation from school (namely, that children living in areas that experienced greater levels of restructuring felt more alienated from their schools). Economic Change was also directly related to Offending (10 to 16), such that those living in areas of greater economic change were likely to have been in trouble more often with the police at ages ten to sixteen than those living in areas with lower levels of economic change.

In order for readers to “locate” the parts of Britain most heavily affected by the economic changes of the early 1980s, Figure 6 provides a map of Britain showing the levels of economic restructuring using this measure plotted by county. The four rectangles mark four areas of Britain in which economic restructuring was heavily pronounced (namely, Central Belt Scotland, the North-East shoulder, Central Belt England, and the Welsh Valleys). ⁶⁷

OPEN IN VIEWER

Figure 6.

Disadvantage Index Score of Mining (1971) and Unemployment (1981).

The third and final stage in our analysis is to rerun the model in Figure 5A, this time controlling for the extent of economic change experienced in each area. We have thus far shown that Economic Change lay behind some of the individual- and school-level processes we have charted. However, what the model reported on in Figure 5B cannot do is to assess the extent to which the relationships between these variables were mediated by the extent of economic restructuring experienced. We reran the model reported in Figure 5B four times (Figs. 7A–7D), with these figures reporting first those in the lowest quartile of Economic Change (Fig. 7A), the next lowest (Fig. 7B), and so on, up to those who were living in the areas with the highest levels of Economic Change (Fig. 7D). This strategy allows us to explore the strength of the relationships between the variables in the model and to assess the relative impact of Economic Change in those areas with little change, as opposed to those areas with greater levels of change. Because we can identify which counties are included in each quartile of Economic Change, we are also able to comment on the geography of economic restructuring and the extent to which this strengthened the relationship between variables that account for offending in some areas more than in others.

OPEN IN VIEWER

Figure 7A.

Expanded Model of Offending: Lowest Level of Economic Change.

OPEN IN VIEWER

Figure 7B.

Expanded Model of Offending: Second Lowest Level of Economic Change.

OPEN IN VIEWER

Figure 7C.

Expanded Model of Offending: Second Highest Level of Economic Change.

OPEN IN VIEWER

Figure 7D.

Expanded Model of Offending: Highest Level of Economic Change.

The first thing to note is that for those who experienced the lowest levels of economic change (Fig. 7A), the Economic Change variable is not related to other variables. ⁶⁸ This suggests that in those areas with low levels of economic change, it was individual-level factors that explained offending (toward the right-hand side of the figure). The next thing to note is that the degree of Economic Change is only statistically significantly related to many of the other variables in the model for the two middle groups (Figs. 7B and 7C). However, the degree of Economic Change is related to being on the “At Risk” Register for those children in the first of the two groups of counties that experienced the highest levels of economic restructuring (Fig. 7C). It is also statistically significantly related to Offending (10 to 16) for those who experienced a low level of economic change (Fig. 7B) and those who experienced the highest levels (Fig. 7D). However, in Figure 7B it is a negative relationship, but for the highest group it is positive. This suggests that in places that experienced relatively low levels of economic change, children may have been encouraged into engaging with school (supporting earlier Scottish data on this matter). ⁶⁹ Hence the effects of economic change are mediated by the degree of change areas experienced.

In areas with relatively low levels of economic change (Fig. 7B), Economic Change is directly related to Disciplined at School, School Alienation, and Offending (10 to 16). However, the paths to the latter two of these are negative (suggesting that economic change reduced school alienation and offending). The model for the areas that experienced relatively high levels of Economic Change (Fig. 7C) suggests that economic change affects offending between ages ten and sixteen indirectly via school alienation and being on the at-risk register. So where economic change was higher it may have demoralized young people and placed them at greater risk of harm. In those areas that experienced the greatest levels of Economic Change (Fig. 7D), economic change was directly related to offending at ages ten to sixteen, suggesting that children in the areas most heavily affected by economic change were drawn directly into offending. Here the relationships with schooling and being “at risk” were not found. All of this suggests that the impact of economic restructuring is mediated by the degree of economic restructuring; the relationship is not simply linear.

Discussion

Focusing on the United Kingdom’s economic troubles during the 1970s (to which Margaret Thatcher’s election can be read as a response) and the early years of the Thatcher governments themselves (1979–81), we have explored the extent to which macro-level economic changes can be used to understand the causal antecedents of offending over the life course. To our knowledge, few have ever attempted to locate offending careers empirically within wider macro-level structures or linked these to macroeconomic policy and political decision making. Our modeling found that while the addition of a variable capturing change over a ten-year period did not improve the percentage of variance explained when compared to a model that did not include such a variable (Figs. 4A–5B), the fact that three of the five additional paths incorporated were statistically significant suggests that incorporating such a measure has theoretical merit. ⁷⁰ Further modeling (Figs. 7A–7D) explored this finding in greater depth, assessing the extent to which the relationships between the variables were modified by the degree of economic restructuring experienced. Although the model fitted the data well, the role of economic restructuring was especially key in areas with greater levels of economic restructuring. Substantively this would suggest that the social and economic changes initiated (as either deliberate policy developments or unintended consequences) during and since the early 1980s have altered citizens’ engagement in crime, especially for those in counties that experienced the greatest levels of economic restructuring.

Our theoretical position has been that the relationship between increases in unemployment and crime in the early 1980s, especially in communities that relied on coal mining and heavy industry for their incomes, was affected most when monetarist policies were adopted in the United Kingdom in the early 1980s. Our article is concerned with the effects of changes in the political governance of the economy in the early 1980s on engagement in crime among young people living in the United Kingdom at that time and on their subsequent offending in adulthood. The dramatic increases in crime witnessed in the United Kingdom during the 1980s led to the development of a more punitive criminal justice system in the 1990s. ⁷¹ Such dramatic changes in unemployment (which stood at just over 4 percent in 1979 but increased to almost 8.5 percent in 1981) pushed many living in the United Kingdom’s industrial heartlands into unemployment and forced many families who depended on (typically) male breadwinners for their livelihoods into, or close to, poverty. That these job losses were clustered made the impact of such levels of unemployment on crime all the greater, since it encouraged the development of offending networks. ⁷² Hence the radicalism of the early phase of Thatcherite economic policies created quite profound and immediate economic problems (not simply unemployment but long-term, geographically concentrated, unemployment) that fell disproportionately on communities that had relied on manufacturing and mining for their livelihoods. Over time, and augmented by retrenchment in other social policies, ⁷³ the neighborhoods of those who had relied on heavy industry for their incomes became associated with crime and disorder. On the whole, economic change was associated with increases in levels of school alienation, which, following what we know from other studies ⁷⁴ as well as our own modeling, ⁷⁵ was associated with juvenile offending (ages ten to sixteen) as well as offending into adulthood. School failure, unemployment, and engagement in crime started to take on intra- and interrelationships familiar to criminal careers researchers. Interlocking social and economic policies relating to housing, schooling, and economic management started a set of processes that altered the social and economic geography of the United Kingdom, such that the greater levels of impoverishment started to coalesce in both regions of the United Kingdom and smaller geographies. Part of this story of change, then, relates to the uneven geographical impact of political decision making. When the Labour Party returned to power in 1997, their manifesto promised to tackle just such uneven deprivation (as part of their New Deal for Communities program). ⁷⁶ Much of this, of course, resonates with what we already know about the causal antecedents of offending: those with lower levels of engagement with schools and who are at risk are likely to start offending early in life. Similarly, early engagement in crime is likely to foster continued engagement in offending. Employment (especially after the age of twenty-five) is negatively associated with offending. ⁷⁷ What is novel, however, is our focus on the wider economic restructuring that may have entrenched such causal processes in the first place and the political ideology that shaped them.

Like Benson’s review of the political processes that may shape criminal careers, we found that the key processes that appear to have operated as causal antecedents took many years to lead to entrenched offending. The budget of 1981 (when the BCS70 children were eleven) increased levels of unemployment, which affected the lives of some of the cohorts’ families. Their peak age of conviction would not have been for around another nine to ten years (1990–91), and those who remained engaged in crime would have done so well into their late twenties (the late 1990s), as suggested by our modeling. The point at which many of the BSC70 would have become increasingly engaged in crime (around fifteen or sixteen) coincided with a dramatic increase in crime in England and Wales. The data we have at our disposal do not allow us to explore the micro-level processes that Hagan suggests took place in the United States. ⁷⁸ However, other studies of the relationships between neighborhoods and crime suggest that such processes may have been operating in the United Kingdom in some places too. ⁷⁹ More generally, our modeling suggests that other studies, wherever possible, ought to incorporate variables into their modeling that capture the role of political processes and the social and economic changes stemming from them. Our article delivers the substantive message that background structural causes of offending at the individual level may rest as much with a country’s politicians as they do with “street-level” actors but that, in periods when crime rises dramatically as a result, it may be subsequent governments who have to deal with the fallout. Politics shapes the economic policies adopted, which in turn affect the communities people live in, through the availability of work, the extent to which young people may regard their own futures as unworthy of investment (leading to alienation from school), and the extent to which offending (out of either frustration or economic need) may be an appropriate response. Given what we know about offending careers, and the difficulty of ceasing them once they have begun, ⁸⁰ offending pathways can easily become entrenched at the individual and community levels.