by Adam Perkins
Based on the discovery that individuals with the employment-resistant personality profile are over-represented amongst the welfare-claiming sector of the population (for example, Vaughn et al., 2010), the association shown in Figure 9.1 provides circumstantial support for the notion that the more employment-resistant an individual is, the more willing they are to have extra children in order to garner extra welfare benefits. But this finding could also be an artefact of the greater material needs of larger families, leading to greater reliance on welfare benefits for reasons unconnected to personality.
The finding that troubled families on average contain 2.5 children now becomes useful to the welfare trait theory because family size was not an admission criterion for the Troubled Families Programme, but antisocial behaviour was. Moreover, whilst not all troubled families are workless, 83 per cent of them contain at least one unemployed adult. This means that the troubled families can be conceptualised as representing a subset of workless households in which the members possess particularly employment-resistant personality profiles. In line with the notion that employment-resistant personality characteristics engender a willingness to have extra children in order to increase welfare income, we would expect the troubled families to contain more children than ordinary workless families. And they do, as shown in Figure 9.2.
Figure 9.1 Bar chart showing the average number of children under the age of 16 in working, mixed and workless households, April–June 2013, UK
Source: Labour Force Survey Household Dataset.
Second, there was a significant association between adult offending and youth offending in the troubled families: 45 per cent of families with an adult involved in antisocial behaviour also contained a young person involved in antisocial behaviour, compared to 20 per cent of families with no antisocial adults. Finally, once again in line with the notion that antisocial personality characteristics engender a willingness to have extra children in order to increase welfare income, the report found that there is a statistically significant association between a family having more than three children and there being an adult family member with a recent criminal conviction.
Figure 9.2 Bar chart showing the average number of children under the age of 16 in working, mixed, workless and troubled households, April–June 2013, UK
Source: Labour Force Survey Household Dataset and Department for Communities and Local Government.
None of these findings provide causal proof for the welfare trait theory, but their patterning is consistent with it, since it suggests that troubled-family status is primarily a product of employment-resistant personality characteristics in the parents/adults. Additional findings in the report reinforce the impression. For example, it shows that 29 per cent of troubled families were experiencing domestic violence, a rate that is approximately four times higher than the rate of domestic violence in the nation as a whole (seven per cent). Moreover, 23 per cent of the troubled families contained at least one child identified as being in need, due to abuse or neglect, and six per cent of the families had a child in the care of the local authorities. To put that latter number in perspective, government figures show that as of 31 March 2014, there was a total of 68,840 children in care in local authority care in the whole of England. In the UK at that time there were approximately four and half million families with dependent children: as a very rough estimate, we can therefore say that approximately one per cent of families in England have a child in care. As a result, we can see that rates of child neglect in troubled families are approximately six times higher than average.
The report also found that on average police were called to the households of the troubled families five times in the previous six months. One family had 90 police callouts in the previous six months and 21 families had more than 30 callouts in the same time frame. The extremely high number of police callouts generated by the troubled families is just one example of the high costs that troubled families impose on the public purse. These costs are detailed in a separate report entitled ‘The Fiscal Case for Working with Troubled Families’ that was published in February 2013 by the Department for Communities and Local Government. It concluded that the overall cost of troubled families in England to the UK government was estimated to be around £9 billion per year. Given the previous finding that there are 111,574 troubled families in England, this allows us to estimate that each troubled family costs the taxpayer approximately £81,000 per year.
This calculation is particularly interesting in the context of this book because in Chapter 8 we saw that on average each person in the UK costs the public purse approximately £1,900 per year. In 2010 there were an average of 2.3 people per UK household and so we can see that per year the average UK household costs the public purse £4,370, approximately 18 times less than the cost to the public purse of each troubled family.
However, a third government report (The National Evaluation of the Troubled Families Programme) found that the average troubled family contains more individuals than the average household (4.2 people). If we divide £81,000 by 4.2, we can see that each member of a troubled family on average costs the public purse £19,285 per year or approximately ten times more than the cost to the public purse of the average person in the UK (£1,900). From these calculations, we can see that the per-person estimate in Chapter 8 of the costs to the taxpayer of each individual with the employment-resistant personality was overly conservative: I estimated that the cost to the taxpayer of each employment-resistant individual was £12,000 per year. However, as I mentioned in Chapter 8, my estimate did not include the cost to society of paying for the care of children neglected by employment-resistant parents, whereas these new government figures do. For example, Figures 9.3 and 9.4 from ‘The Fiscal Case for Working with Troubled Families’ use pie charts to show how the £9 billion is spent.
As these pie charts show, approximately 41.5 per cent of the £19,285 per year that each person in a troubled family costs the taxpayer per year relates to child protection. From this we can see that my estimate of £12,000 as the annual cost of each employment-resistant individual is approximately £8,003 too low. If we add this latter figure to my estimate, it ends up as £20,003. At a mere £718 pounds more than the UK government estimate, this calculation suggests there is close agreement between these two estimates, providing reassurance that employment-resistant individuals are much more costly to society than average citizens (about ten times more costly).
Figure 9.3 Targeted annual expenditure on troubled families
Source: The Fiscal Case for Working with Troubled Families.
Figure 9.4 Reactive annual expenditure on troubled families
Source: The Fiscal Case for Working with Troubled Families.
As a footnote to these calculations, it should be noted that in an interview with the Sunday Times in August 2014, Louise Casey, the UK government’s leader of the Troubled Families Programme, gave an update on the prevalence and costs of troubled families. She estimated that, in addition to the 111,574 ‘first tier’ of troubled families, there are 400,000 ‘second-tier’ troubled families in the UK whose problems are less severe but still significant. She estimated that the total cost to the taxpayer of all 500,000 first- and second-tier troubled families was in the region of £30 billion per year. If we refer back to Chapter 8, the number of employment-resistant individuals in the UK who exist as a result of flawed welfare policy was estimated at 98,040. At the initial per-person annual cost estimate of £12,000, this produced an extra annual cost to the public purse of approximately £1.2 billion or one per cent of welfare spending. We can now see that sum was too low: the new estimate is that employment-resistant individuals in the UK who exist as a result of welfare policy cost the public purse approximately £1.9 billion (98,040 × £19,285) or 1.5 per cent of welfare spending.
The data gathered by the Troubled Families Programme are reassuring for the validity of the welfare trait theory, yet they are cross-sectional and so do not shed light on causation. The clea
rest corroborating evidence of all for my theory would be data showing that the implementation of the welfare state is followed by an upsurge in a behaviour that is more common in people who possess the employment-resistant personality profile, yet is not connected to the welfare state. One such behaviour is criminality, because it is not necessarily connected to the world of work, nor is it necessarily motivated by economic gain, yet it is more common amongst employment-resistant individuals. Therefore, a convincing clue that the welfare state really is increasing the frequency of mis-developed, employment-resistant personality profiles would be the presence of an upsurge in crime following the introduction of the welfare state.
Chapter 2 introduced the research on problem families in Sheffield that was conducted by W. L. Tonge and colleagues (Tonge et al., 1975, 1981). Amongst other things, we saw evidence that the adults of the 33 problem families had personality profiles that were significantly less conscientious and agreeable than those of the adults from 33 comparison families, who were matched for location and socio-economic background. In line with the claim that low levels of conscientiousness and agreeableness constitute the employment-resistant personality profile, we also saw that the problem-family adults had significantly worse work records than the adults in the comparison families. Moreover, in Chapter 8 we saw that the problem-family adults averaged approximately seven convictions per head versus one conviction per head amongst the adults in the comparison families.
Viewed as a whole, these findings by Tonge et al. (1975) provide background support for an association between criminality, unemployment and the employment-resistant personality profile. They also fit the data gathered by the Troubled Families Programme. But they also suggest that we need to be more specific regarding which types of criminality are important. For example, Tonge and colleagues found that high levels of criminality were also displayed by some participants whose personality profiles were well-adjusted.
One of our subjects with a lengthy list of criminal convictions due to his activities as a burglar was an exemplary husband and father to his family. We were impressed by his attitude of social responsibility when he explained why he burgled only commercial premises, which he expected to be adequately insured. He avoided all private houses, even large ones ‘because you never know how poor they are’. Consistency of purpose and coherence of meaning must be lacking if the diagnosis of personality disorder is to be sustained.
(Tonge et al., 1975, p. 26)
Based on this evidence, it is necessary to narrow down the type of criminality associated with the employment-resistant personality profile. Fortunately, Tonge and colleagues itemised the types of criminality perpetrated by the Sheffield families, showing that the biggest difference in type of conviction between the two family types was in criminal violence. More specifically, 8 of the 33 men in the problem families had convictions for violence, yet only 1 of the 33 men in the comparison families had a conviction of this type, a ratio of 8:1. In contrast, 20 of the 33 in the problem families had convictions for crimes of dishonesty but so did 7 of the 33 men in the comparison families, a ratio of almost 3:1. These data show that criminal violence was especially common amongst the men of the problem families, suggesting that there is indeed a link between violent criminality and the employment-resistant personality profile.
These results point to the existence of a link between employment-resistance and violent crime. The existence of this link is further supported by homicide statistics. For example, in 2007, the UK government began recording the employment status of individuals convicted of homicide in England and Wales, classifying them as employed, unemployed, retired, student and other. Table 9.1 summarises these data from 2007 until 8 November 2013 (the latest available figures), showing that unemployed individuals perpetrate more killings than individuals in all the other four employment categories combined. Moreover, given that the unemployment rate in England and Wales was approximately seven per cent during this time period, we can see that unemployed people are approximately nine times more common amongst convicted killers than amongst the general population as a whole.
These data are for England and Wales, but it should be noted that the general trend for unemployed individuals to be over-represented amongst convicted killers has also been found in other nations of the Western world. For example, Mucchielli (2004) profiled individuals who were convicted of murder in the department of Yvelines in France over the previous ten years. This study is interesting because it not only profiled the offenders at the time of their conviction but also attempted to trace their prior life histories. The relevant finding was that of the 122 murderers in the study, only 48 (39 per cent) were in employment at the time of the offence and of these employed murderers, only ten had consistent work records in the years prior to the offence.
Table 9.1 Employment status of individuals convicted of homicide in England and Wales from 2007/2008 to 2012
Note: Figures as at 8 November 2013. Figures are subject to revision as cases are dealt with by the police and the courts, or as further information becomes available. The data for more recent years is incomplete, as cases are still going through the courts. Excludes 373 suspects whose employment status was unknown. Percentages may not sum due to rounding.
These homicide data from the UK and France show that there is a significant association between criminal violence and unemployment, to the extent that unemployed people are approximately ten times more common amongst convicted killers than amongst the general population. However, because homicide is sometimes a financially acquisitive crime, some commentators might attempt to explain away this association between homicide and unemployment as a product of poverty rather than personality.
This objection is implausible because millions of people endure poverty yet are never convicted of a violent crime. Also, as we have already seen, there is evidence to show that antisocial personality disorder is associated with a high probability of both violent criminality (Hodgins, 2007; Jüriloo et al., 2013) and a poor work record (Cleckley, 1988). Nevertheless, it may be the case that on occasion an unemployed person might be driven by poverty to attempt an armed robbery and end up accidentally killing rather than merely robbing the victim. Moreover, homicide statistics do not provide data on the personality characteristics of the killers, preventing us from being sure that the causal link between violent crime and unemployment is personality.
Both these concerns are addressed by a study that profiled 336 UK males convicted of domestic violence (Gilchrist et al., 2003). This study is especially important to the topic of this book because Gilchrist and colleagues not only recorded the employment status of the offenders but also profiled their personalities. Moreover, unlike other forms of violent crime, such as mugging or armed robbery, the perpetrator of domestic violence does not benefit financially from his crime. This means that any links between proneness to domestic violence and unemployment are more likely to be a direct product of an antisocial personality profile rather than the side effect of a desire to acquire money.
The results of this study provide clear evidence for the notion that antisocial personality characteristics underpin both criminal violence and unemployment. First, Gilchrist and colleagues found that 108 of the men (47 per cent of the sample) met criteria for antisocial personality disorder. Given that epidemiology studies suggest that only around one per cent of the general population meet these criteria (for example, Coid et al., 2009), we can see that individuals with antisocial personality disorder are almost 50 times more common in these offenders than amongst the general public. Second, in line with the finding that individuals with antisocial personality disorder are over-represented amongst the unemployed sector of the population (for example, Vaughn et al., 2010), Gilchrist and colleagues found that 60 per cent of their subjects were unemployed. As the rate of unemployment in the UK in 2003 was approximately five per cent, we can see that unemployed men are almost 12 times more common amongst this group of convicted abusers than amongst the g
eneral population.
Since I earlier defined the employment-resistant personality profile as a mild form of antisocial personality disorder, these data gathered by Gilchrist and colleagues suggest that a rise in the frequency of employment-resistant personalities due to the implementation of the welfare state will indeed trigger a rise in violent crime. As I mentioned in the introduction to this chapter, we should find that the onset of a welfare state is followed some 10–15 years later by a rise in criminal violence, as the first generation of personality-damaged welfare babies grow old enough to begin swelling government statistics for violent crime. This evidence has special importance because no other personality theory to my knowledge can explain changes in criminal violence via this pattern of timing in relation to the onset of the welfare state.
There are many indices of criminal violence, but homicide statistics are regarded by criminologists as the most reliable index of trends in criminal violence, because killing another human has been regarded as a serious crime for millennia and so is more likely to show up in historical crime statistics than other types of violence. As Steven Pinker documented so brilliantly in his recent book The Better Angels of Our Nature (Pinker, 2011), homicide rates have been declining across the Western world for centuries – except for a mysterious upswing that began a decade or so after the end of the Second World War and continued until the 1990s. The causes of this upswing in homicide across the Western world has not yet been conclusively explained, opening the door to the possibility that it might be related to the welfare state, since the same era saw the widespread implementation of welfare legislation in advanced nations. Bearing this in mind, we will now look in more detail at homicide trends before and after the implementation of the welfare state in the UK, whilst acknowledging that the introduction of the welfare state is just one event that happened in this period – others may also have an influence.