by Adam Perkins
The British welfare state evolved over centuries but began in its modern, cradle-to-grave form in the late 1940s (Timmins, 2001). This means that if the welfare trait theory is correct, in the UK we should find an upswing in criminal violence that began in the late 1950s/early 1960s, as the first generation of personality-damaged welfare babies reach the age of criminal responsibility. Figure 9.5 plots homicide rates in England and Wales for each census year from 1901 to 2011 and appears to support this hypothesis by showing that homicide rates in England and Wales began a steep upswing in 1961which continued to the mid-1990s before dipping down again. This recent dip in homicide seems to counter my theory, since the welfare state was still in place. But it can explained in part by modern medicine’s increased capacity to save the lives of assault victims (for example, studies in the USA have shown that homicide rates in the late 1990s would have been up to three times higher if victims of assault were treated with the medical technology of 1960; Harris et al., 2002).
Figure 9.5 Homicides in England and Wales between 1901 and 2011
Source: Office of National Statistics.
The rise in homicide rate over time that is shown in Figure 9.5 is backed up by statistical testing: there is a statistically significant positive correlation between homicide and year, which means that 40 per cent of the variance in homicide rates is accounted for by the year. Since most other nations in the Western world show a similar trend in homicide (for example, Pinker, 2011) and they also tended to implement modern welfare states in the aftermath of the Second World War, evidence of this type appears to back up the welfare trait theory.
However, whilst this correlation is reassuring, since the absence of statistically significant correlations that fit the welfare trait theory would set alarm bells ringing about its validity, the presence of such correlations does not provide conclusive proof that the welfare trait theory is valid, because correlation does not imply causation. The blunt truth is that, if we go fishing in large datasets, we can find statistically significant correlations that support any theory we wish to dream up but are, in reality, spurious. Or, stated another way, visually impressive and statistically significant correlations can occur by coincidence. This is perhaps best illustrated by visiting the website that mines US government datasets for spurious correlations in order to show just how misleading correlations can be (www.tylervigen.com). In Figure 9.6, I have plotted some US government data which show there is a statistically significant positive correlation in the USA between the annual per-capita consumption of cheese and the number of people who die each year after becoming tangled in their bed sheets. This correlation is extremely strong, explaining an impressive 89.7% of the variance between the two variables, yet is completely spurious because there is no sensible causal mechanism that can explain why eating cheese should cause death by tangled bed sheets.
Figure 9.6 The relationship in the USA from 2000 to 2009 between per-capita consumption of cheese and deaths due to tangled bed sheets
Source: Centers for Disease Control and Prevention and US Census.
In contrast, the statistically significant positive correlation between levels of cigarette smoking and rates of lung cancer that was first demonstrated in the 1930s (for example, Müller, 1939) is not spurious because there is a sensible causal link between the two variables, namely the carcinogenic properties of cigarette smoke. What these considerations mean is that the use of correlations in science is a two-stage process: first we need to use statistical techniques to prove that a correlation exists between two variables. Second we need to use our common sense to determine whether there is a causal mechanism that can reasonably explain the statistical connection between the variables or whether that statistically significant correlation is merely a coincidence. In the case of this book, we have a sensible causal mechanism which can explain the statistically significant correlation in Figure 9.5 between the onset of the welfare state in the late 1940s and the upsurge in homicide in the early 1960s, namely the welfare trait theory. But that is still not sufficient, since if we try to corroborate the welfare trait theory with correlations that we then corroborate via the same theory, we risk falling into the trap of circularity.
To make sure the welfare trait theory is not guilty of circularity, we need to take a broader view of the scientific literature. Most importantly, since few theories ever emerge in total isolation, if the welfare trait theory is valid, the scientific literature should already contain theories created by other scholars that point in approximately the same direction. And it does. For example, as was mentioned in Chapter 4, the economic historian Greg Clark has already published a book – A Farewell to Alms – that argues cogently for the notion that differential reproduction can shape social norms via personality changes, much as I argue in this book.
But Clark’s book concerned the origins of the Industrial Revolution and does not address the welfare state, nor the apparent shift towards a greater frequency of violent crime that occurred in the nations of the Western world around 1960. In regard to this, we are fortunate to have at our disposal the work of the criminologist Manuel Eisner, who is a leading authority on the causes of the upswing in criminal violence that hit the Western world around the year 1960. After surveying a wide range of data, Eisner concluded that this upswing was caused by a shift in the Lebensführung of the Western world (Eisner, 2008). This is a German term popularised by Max Weber that represents the way of life that a society regards as good or correct. Lebensführung has been translated into English as lifestyle, but Eisner finds this an inadequate representation, maintaining that ‘Lebensführung or conduct of life refers to a much wider cultural script encompassing work, politics, beliefs, education, and individual character. These models of conduct of life become reinforced and stabilized through institutions such as schools, families, the church, and bureaucracies’ (Eisner, 2008, p. 290).
According to Eisner, prior to the mid-twentieth century in the Western world there was ‘an emphasis on self-control as an ideal of personality; domesticity and familialism as guidelines for private life; and respectability as the yardstick for public appearance’ (Eisner, 2008, p. 303). Eisner believes that the upswing in criminal violence that began in the late 1950s/early 1960s was a result of a shift away from these values of self-control, domestic duty and respectability that he viewed as causing the reductions in crime witnessed in the previous centuries. Furthermore, Eisner identifies a causal factor which links the welfare trait theory to the shift in the Lebensführung of the Western world. That factor is socialisation.
As you may recall, in Chapter 8 we saw how a person’s scores on conscientiousness and agreeableness represent the effectiveness of their socialisation as a child and young adult. The work of David Lykken has shown us that a person’s level of socialisation is determined by a combination of nature and nurture: the higher a baby’s dose of genes for being difficult to socialise, the better the quality of parenting they require in order to grow into a socially well-adjusted and economically productive adult (Lykken, 1998). If we now return to Eisner’s work, we can see that he blames the shift in Lebensführung towards greater criminal violence that occurred around 1960 partially on a change in patterns of socialisation. For example, he wrote that:
the major shifts in levels of interpersonal criminal violence over the past 160 years were associated with broad changes, across Europe, in shared cultural models of what constitutes a desirable and good ‘conduct of life’. These are said to influence levels of interpersonal violence through their effects on patterns of socialization as well as by affecting expectations about adequate interaction in daily situations, especially in public space.
(Eisner, 2008, p. 290)
So what we need now to tie the work of Eisner to the theme of this book is a specific mechanism to explain how a welfare state could alter patterns of socialisation. This need brings us full circle because, as I mentioned in Chapter 1, my answer is that the welfare state alters patterns of socialisation by
boosting the number of children who are born into disadvantaged households and who suffer personality mis-development as a result. This difference in reproduction is crucial to my argument because, as we have already seen, the personality profiles of welfare claimants are on average less well socialised than employed citizens (for example, Vaughn et al., 2010) and lack of socialisation impairs parenting competence (Cleckley, 1988).
The favoured reproductive status of the welfare claimants therefore means that their offspring will not only be less well socialised on average than offspring of employed citizens, but also as every generation passes they will be proportionately more numerous. As they proliferated in the years following the introduction of the welfare state, I argue that these inadequately socialised young people spread a culture of impulsive, irresponsible behaviour, culminating in the soaring crime rates of the 1960s, 1970s and 1980s. Eventually, by the 1990s, the nations of the Western world realised that the easy-going ‘Dixon of Dock Green’ law-enforcement practices that had been adequate up until the 1950s were no longer fit for purpose. As a result, they hardened their policing procedures, and gradually (along with advances in trauma medicine) helped to stem the rise in homicides (as shown in Figure 9.5).
As a caveat, it is important to acknowledge that Eisner (2008) argued that the sheer rapidity of the change in criminality indicates that it has a strong cultural component that does not map directly onto the concept of personality traits as a product of nature and nurture. In response, I suggest that the focus of the welfare trait theory on differences in reproduction and child socialisation between claimants and employed citizens should not be allowed to overshadow the purely environmental effect of the welfare state on the behaviour patterns of adults that has been demonstrated by economic studies. As mentioned in the introduction, these studies show that when the welfare state becomes more generous, the personality profile of the population becomes more willing to break societal norms (Heckman, 2008).
Applied to Eisner’s demonstration of a shift in the Lebensführung of the Western world towards greater criminal violence from 1960 onwards, I suggest that these economic studies point to two rather simple explanations as to how the implementation of the welfare state can also temporarily raise the level of criminal violence in adults whose personality profiles are already formed: first, I suggest a role for social learning – learning from others – which allows an attitude or behaviour to spread quickly through a population, as a form of local tradition. Anecdotal examples of such an effect are readily found in human populations, such as the rapid spread of football hooliganism amongst developed nations in the 1970s and then its equally rapid decline in the 1990s. The power of social learning to cause rapid changes in behaviour has been proven formally via studies in chimpanzees that show it provides ‘a high speed “second inheritance” system that interacts with genetic inheritance to enrich behavioural evolution’ (Whiten, 2014, p. 178).
Second, I suggest that the welfare state also alters behaviour by an even simpler mechanism, namely that, by providing claimants with an income without requiring work in return, it gives them the option of sleeping late in the day and staying awake at night. In the case of claimants who happen to possess pro-social personality characteristics, this change would not affect crime: they might spend their extra time by cultivating an innocuous hobby. But in individuals who possess antisocial, predatory urges, the welfare state opens a Pandora’s box of opportunities for nocturnal criminality by allowing them to sleep in late during the day and thus have plenty of energy to hunt for victims under the protection of darkness. This idea receives circumstantial support from Eisner’s work, since he found that the upsurge in criminal violence that took place in the Western world from the 1960s to the 1990s was not a result of increased infanticide or domestic violence (both of which seemed to decline in frequency), but was primarily driven by a rise in violence inflicted on young men in public places by other young men who were usually strangers (Eisner, 2008).
Conclusion
The welfare trait theory dovetails with the work of criminologists, who have found that there was an upswing in criminal violence in the Western world from about 1960 to 1990 that was caused by a reduction in the value placed upon self-control, duty and respectability. This reduction in solid citizen values corresponds to the pattern of personality change predicted by the welfare trait theory. This is of course a retrospective interpretation and does not prove the welfare trait theory, but it is what we would expect to see if the welfare trait theory is correct.
10
What Next?
In the first chapter I listed three alternative narratives for the future of the welfare state, which were as follows:
1. The welfare state should be retained without change.
2. The welfare state should be abolished.
3. The welfare state should be amended to take account of personality.
As this book demonstrates, I believe the balance of the evidence favours the third narrative. In this final chapter, I will therefore make some basic suggestions as to how the welfare state could be amended so that it looks after the unemployed but does not warp the personality profile of the population towards greater employment-resistance. Richard Dawkins provided a pointer back in 1976 when he showed, using the rubric of evolutionary biology, that reproduction must be disconnected from subsistence if the welfare state is to be sustainable in the long run. Dawkins’ disconnection suggestion is crucial because, as we have seen during the course of this book, it will reduce the capacity of the welfare state to shift the average personality profile of the population towards employment-resistance. It achieves this not only by preventing the proliferation of children carrying a larger than average dose of the genes for being difficult to socialise, but also, more importantly, by reducing the number of genetically average children who are swayed towards employment-resistance by being born into disadvantaged households, where they are at a higher than average risk of personality mis-development.
But the big question in policy terms is how to accomplish this disconnection? As we saw in Chapter 6, in 1976 Richard Dawkins suggested in The Selfish Gene that welfare benefits should be conditional upon claimants utilising contraception. Dawkins’ suggestion would no doubt be effective but, since 1976, we have discovered that welfare claimants voluntarily limit their reproduction if the welfare state no longer makes it worth their while financially to have extra children (for example, Argys et al., 2000). This discovery means that I do not advise mandatory contraception. Instead I suggest that disconnection of reproduction from subsistence should be achieved by adjusting welfare generosity until the average number of children born to claimants matches the average number of children born to non-claimants.
In the UK, the effects on reproduction of such an amendment to the welfare state can be monitored because, as we saw in Table 4.1, the government already records the number of children born into each of three levels of welfare dependence. Figure 10.1 shows these data as a bar chart. When thinking about these data, the crucial variable to bear in mind is the average number of children in working households (the left-hand column in the bar chart), because this provides an approximate baseline measure of the average level of reproduction that can be sustained by citizens who are not reliant on the welfare state for their survival.
Figure 10.1 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.
The data shown in Figure 10.1 indicate that the level of welfare generosity that no longer endangers the personality profile of a population is whichever level causes mixed and workless households to have the same number of children, on average, as working households – what could be dubbed ‘the point of reproductive equilibrium’ (indicated by the dashed line). Therefore, my specific policy suggestion is that welfare generosity should be gradually reduced until mixed and workless households have, on ave
rage, the same number of children as working households. The importance of reducing generosity gradually is that it allows the effect of welfare generosity on reproduction to be tracked in a systematic manner and also means that claimants will have plenty of time to get the message that the government will no longer allow children to be used as tools to gain extra welfare benefits.
At first glance this policy suggestion might seem absurd as we have already seen that the risk of developing the employment-resistant personality profile is increased by childhood disadvantage (Heckman et al., 2013). Reduced welfare generosity would therefore seem likely to increase disadvantage amongst the children of claimants and thereby proliferate the employment-resistant personality profile, with dire prospects for all concerned. However, we saw evidence in Chapter 4 that employment-resistant individuals have a tendency to spend their welfare benefits on unnecessary purchases such as electronic gadgets and luxury chocolates, instead of using the money to improve the lives of their children (Tonge et al., 1975). This discovery suggests that fluctuations in the generosity of benefits do not make much difference to the quality of life for the children of claimants: they tend to be neglected regardless of household income.
More generally, reduced welfare generosity might seem likely to harm society by increasing inequality. However, as noted by James Heckman in his 2013 lecture at the British Academy, increased redistribution of wealth via an expanded welfare state (as happened in the USA in the mid- to late 1960s and early 1970s during the so-called War on Poverty) does not cause significant improvements in long-term measures of inequality such as intergenerational mobility. Based on the personality data summarised in this book, I suggest this is likely to be at least partly because the same unconscientious attitude that increases an individual’s risk of ending up unemployed and claiming welfare will mean that increased welfare generosity won’t produce lasting improvements in their circumstances, because that extra money won’t be managed conscientiously.