Introduction
There is a growing enthusiasm on the part of governments in many countries to incorporate insights garnered by the behavioural sciences into the design of policy and programmes. These disciplines – which include cognitive and social psychology, behavioural economics and behavioural finance – have produced robust evidence over recent decades that humans differ in predictable ways from the rational decision-makers assumed by neo-classical economics (Ariely, Reference Ariely2009; Kahneman, Reference Kahneman2011; Shafir, Reference Shafir2013; Tversky and Kahneman, Reference Tversky and Kahneman1974). Much has already been written about the potential of the behavioural sciences to spur innovation in retirement savings, organ donation, charitable giving, health promotion, environmental policy, and crime prevention (Dolan et al., Reference Dolan, Hallsworth, Halpern, King and Vlaev2012; Le Grand and New, Reference Le Grand and New2015; Marteau et al., Reference Marteau, Ogilvie, Roland, Suhrcke and Kelly2011; Oliver, Reference Oliver2013b; Shafir, Reference Shafir2013; Thaler and Sunstein, Reference Thaler and Sunstein2008). This article argues that insights from the behavioural sciences may yet change the trajectory of welfare policy. It seeks to stimulate the debate over the possible ways that empirical findings about judgment and decision-making could be used to argue for stronger income support for the unemployed and underemployed as well as greater public provision of services.Footnote 1 Richard Thaler and Cass Sunstein have gained much attention for their argument that policy makers should use the insights emerging from the behavioural sciences to nudge citizens towards the behaviour they seek, thereby reducing the need for coercion. This approach they call libertarian paternalism (Thaler and Sunstein, Reference Thaler and Sunstein2008). Despite the ethical concerns which have been raised about libertarian paternalism (discussed below), it is at present the pre-eminent way of understanding the policy implications of behavioural insights. Its dominance is largely due to the success of the Behavioural Insights Team – or ‘Nudge Unit’ — that David Cameron's government established within the Cabinet Office in 2010. This model has since spread internationally, inspiring similar developments in Australia, Canada, Denmark, France, Germany, Saudi Arabia, Singapore, and the United States (Bell, Reference Bell2013; Bennhold, Reference Bennhold2013; Rutter, Reference Rutter2015). Recently influential actors in international development policy have embraced the behavioural insights movement (World Bank, 2015; Rutter, Reference Rutter2016). Some scholars have connected the enthusiasm for nudge with the new paternalism promoted by the influential New Right thinker Lawrence Mead (Goodwin, Reference Goodwin2012: 85; King and Ross, Reference King, Ross, Castles, Leibfried, Lewis, Obinger and Pierson2010). As nudge spreads further afield coercive paternalism is sometimes mistakenly labelled nudge politics, generating misunderstanding and making it harder to discuss the pros and cons of behaviourally informed policy (Bielefeld, Reference Bielefeld2014; Davis, Reference Davis2015). To pave the way for a debate about the progressive possibilities of applying behavioural insights to social policy, the present article makes clear the significant contrasts between libertarian paternalism and new paternalism.
In recent decades there has been a trend in the US, UK, Europe, Canada, Australia and New Zealand to reframe citizenship entitlements to financial assistance as privileges which may be revoked if behavioural conditions are not met. Since the 1980s behavioural conditions are being specified more precisely and sanctions for non-compliance are becoming harsher (Watts et al., Reference Watts, Fitzpatrick, Bramley and Watkins2014: 3). Increasingly, policies are emphasising the personal responsibility of the poor for their circumstances. Sociological explanations for unemployment and poverty have lost ground to explanations focused on individuals’ failings. Across a wide range of social policy areas, the impulse has been to present citizens with an array of incentives and disincentives to reform their behaviour rather than to seek understanding of how the wider social environment shapes that behaviour (Henman, Reference Henman2011: 11–12). This article's central contention is that behavioural insights provide no justification for continuing down this route. Guy Standing (Reference Standing2011a) has called behavioural economics and behavioural conditionality ‘twin trends’. He perceives behavioural economics to be antithetical to a rights-based approach to social welfare. David Halpern – one of the leading actors behind the creation of the behavioural insights team in the UK – saw behavioural economics as compatible with a political philosophy of personal responsibility (Halpern et al., Reference Halpern, Bates, Mulgan, Aldridge, Beales and Heathfield2004). The present article by contrast argues that insights from the behavioural sciences undermine the logic of new paternalist reforms and provide reasons for a shift back towards a more social democratic approach to welfare provision.
This article proceeds by firstly summarizing the paternalist rationale for welfare conditionality articulated by Mead and highlighting his assumptions about the psychology of the poor. It then summarises some of the key findings of the behavioural sciences in recent decades and introduces the concept of libertarian paternalism. It argues that, though cognitive psychology and behavioural finance are disciplines that focus on individual behaviour, what they have found highlights the importance of social contexts, social cues and social norms. Particular attention is paid to the emerging behavioural research into how poverty affects decision-making. The article then employs what behavioural scientists now know about the decision-making of those with very tight budgets to mount a critique of Mead's arguments for paternalism.
Behavioural research offers some support for Mead's observation that people on very low incomes often make decisions which are not in their own best interests. However, behavioural insights lead us away from the moralising interventions advocated by Mead for several important reasons. Firstly, unlike Mead, behavioural research does not ‘other’ people living in poverty. It locates the explanation for counter-productive behaviours in the cognitive biases humans share, rich and poor alike. People experiencing financial hardship make bad choices not because they are aberrant, but because they are human. Behaviour research reveals that counter-productive behaviour is to be expected when people are placed in circumstances hostile to their well-being. Rather than locate the deficit within individuals, it locates it within the context of scarcity in which some individuals are forced to live. This points not to a need to remedy the character of individuals – rather it suggests the importance of remedying social conditions. A scientifically rigorous understanding of human decision-making provides a powerful impetus to reject the moral lens through which we have been asked to view poverty in recent decades. Policies which aim at helping welfare recipients by hassling them appear likely to do more harm than good.
New paternalism
‘New paternalism’ is a term coined by the highly influential US political scientist Lawrence Mead. In a series of books (Mead, Reference Mead1986, Reference Mead1992, Reference Mead and Mead1997) he has argued that welfare perversely entrenches the poverty it is intended to alleviate. Like Charles Murray, he argued that a major cause of unemployment is the unwillingness of the low skilled to apply themselves regularly to the low waged, low status jobs which are available to them. However, in contrast to Murray's argument that welfare gives people incentives to behave in ways which will keep them poor, Mead stressed that it was not rational economic calculation that prevented the long-term unemployed from seeking work, but their inability to ‘advance their own self-interest’ (Mead, Reference Mead and Saunders2000: 47). He posited that an underclass existed in the US who suffered ‘functioning problems such as difficulties in getting through school, obeying the law, working, and keeping their families together’ (Mead, Reference Mead1986: 22). Central to their difficulties was an ‘inability to commit steadily to any task’ (Mead, Reference Mead1986: 50). He held that these problems were ‘traceable to an unstable family life, marked by absent fathers, erratic parenting, and low self-esteem and aspiration' (Mead, Reference Mead1986: 22). Mead believed that the permissiveness of social policy exacerbated the failure of the poor to take responsibility for getting themselves out of poverty. He wrote that ‘government must persuade them to blame themselves’ (Mead, Reference Mead1986: 10; italics in original).
Mead argued that rather than cutting welfare, the best way to deal with welfare dependency was through programmes which provide personal supervision and direction. Paternalistic programmes should provide the poor with a combination of ‘help and hassle’ (Mead, Reference Mead and Mead1997: 24). The standards of behaviour expected by society needed to be ‘clear’ and ‘forcefully communicated to the poor’ (Mead, Reference Mead1986: 247). Income support payments should no longer be seen as an entitlement but should be tied to behavioural requirements. The capacity to deny individuals their payments then becomes a lever which can be used to ensure compliance (Mead, Reference Mead and Mead1997: 5). The wishes of the welfare recipient are treated as unimportant. A central assumption is that ‘enforcing society's interest in good behaviour is deemed to serve the individual's interest as well’ (Mead, Reference Mead and Mead1997: 4). Thus, this approach is not merely regulatory but paternalistic – it finds its rationale in the presumed well-being of welfare recipients (Mead, Reference Mead and Mead1997: 22).
The paternalism observed and advocated by Mead is labelled new paternalism to distinguish it from the paternalism of the eighteenth and nineteenth centuries with which it shows much continuity (Mead, Reference Mead and Mead1997: 7–8). It differs in that it involves supervision of welfare recipients' lives by governments rather than churches or charities, and because supervisors are guided by rules instead of enjoying the discretion to apply their own moral judgements. It further differs from older forms of paternalism in that the targets of supervision do not reside in a custodial institution.
New paternalism is directly at odds with the hitherto dominant social policy paradigm in the UK, led by Titmuss, which strove to avoid making distinctions between the deserving and undeserving poor. Mead's philosophy provided the theoretical underpinnings for welfare reform in the US under Clinton and subsequently in the UK under Blair and Australia under Howard (Deacon, Reference Deacon2000; King and Wickham-Jones, Reference King and Wickham-Jones1999: 65; Saunders, Reference Saunders and Saunders2000: 23). Characteristic features of these welfare-to-work reforms include their focus on the behaviour of the unemployed as the explanation for their unemployment, and the linking of desired behaviours (such as training and accepting jobs) to eligibility for benefits. Paternalism is not the only rationale given for welfare conditionality. Policy makers have also drawn upon economic efficiency, social control and the principle of reciprocity to offer justifications for conditionality (Paz-Fuchs, Reference Paz-Fuchs2008; Watts et al., Reference Watts, Fitzpatrick, Bramley and Watkins2014: 15). This article focuses on paternalism and its relationship to the behavioural sciences because the mistaken view that behavioural economics and new paternalism are inherently linked constitutes a barrier to understanding behavioural insights as a resource for advocates of a rights-based approach to social welfare.
Behavioural insights
In recent decades researchers working at the intersection of economics and psychology have demonstrated the wide gulf between the assumptions about decision-making incorporated into neo-classical economic models and the decision-making processes of actual humans. Homo economicus – or economic man – is capable of processing all available information and chooses only options that maximize his expected utility. The decision-making of homo sapiens, however, is greatly affected by our limited cognitive abilities, limited attention and limited self-control. Through experiments in the laboratory and the field, as well as survey studies, researchers have demonstrated that failures of judgement and faulty choices are common (Le Grand and New, Reference Le Grand and New2015: chap 5). It is not merely that we must often make our decisions on the basis of inadequate information. Experiments have shown that people – including beneficiaries of higher education at elite institutions – often assess probabilities in irrational ways (Tversky and Kahneman, Reference Tversky and Kahneman1974). We tend to receive new information in ways that reaffirm our existing beliefs (Rabin and Schrag, Reference Rabin and Schrag1999). Our decisions are shaped by our difficulty in imagining the unfamiliar and our assessments of events’ probabilities are biased by whether we can vividly imagine those events (Lichtenstein et al., Reference Lichtenstein, Slovic, Fischhoff, Layman and Combs1978). We do not evaluate our options by rationally calculating their costs and benefits and thus do not always respond to incentives in the ways expected by economic theory (Kahneman and Tversky, Reference Kahneman and Tversky1979).
Research has shown that how a decision is framed makes a difference to which option is chosen (Tversky and Kahneman, Reference Tversky and Kahneman1981). The features that are most salient in the moment of making a decision carry disproportionate weight (Kahneman and Thaler, Reference Kahneman and Thaler2006). Our actions are more strongly swayed by the possibility of a loss than the possibility of a gain of the same magnitude (Kahneman and Tversky, Reference Kahneman and Tversky1979), and far more by the prospect of gains (and losses) today than future gains (and losses) (Laibson, Reference Laibson1997). We have difficulty finding the self-control necessary to make sacrifices in the present for the sake of our future well-being (Laibson, Reference Laibson1997). Self-control is a limited resource that can become depleted with use (Baumeister et al., Reference Baumeister, Bratslavsky, Muraven and Tice1998). We are given to procrastination and we have a tendency to continue the status quo (Samuelson and Zeckhauser, Reference Samuelson and Zeckhauser1988; Ariely and Wertenbroch, Reference Ariely and Wertenbroch2002). Most people are unrealistically overconfident about their own abilities (Svenson, Reference Svenson1981; Cooper et al., Reference Cooper, Woo and Dunkelberg1988). The behavioural sciences have amassed evidence for the impact on decisions of seemingly trivial features of the immediate decision-making environment – due to effects such as anchoring (i.e. heavy reliance on an initial piece of information when making judgements, even when that piece of information is irrelevant) (Kahneman, Reference Kahneman2011: chap 11) and priming (i.e. influence by subconscious cues) (Bargh, Reference Bargh2006). Many decisions are made not through careful, effortful deliberation but via a much quicker and easier cognitive process that is impulsive, associative, and more emotional (Kahneman, Reference Kahneman2011).
The inductive approach of the behavioural sciences has also revealed the naivety of conventional economics’ assumptions about motivation. Though their methodology is individualist, the behavioural sciences reveal humans to be social creatures. By focusing on individual behaviour they illuminate our extreme susceptibility to social cues and the importance of phenomena such as social norms (Darley and Latané, Reference Darley and Latané1968; Thaler and Sunstein, Reference Thaler and Sunstein2008: chap 3; Burke and Peyton Young, 2011). For example, the research demonstrates that people feel compelled to reciprocate helpful actions (Cialdini, Reference Cialdini1993: chap 2), we are highly sensitive to unfairness and are motivated to punish it even when this comes at a personal cost (Henrich et al., Reference Henrich, Boyd, Bowles, Camerer, Fehr, Gintis and McElreath2001). We feel a powerful need to maintain consistency with our publicly expressed commitments (Cialdini, Reference Cialdini1993: chap 3), and many people resist encroachments on their autonomy even when this negatively impacts their welfare (Brehm and Brehm, Reference Brehm and Brehm1981).
Libertarian paternalism
Sunstein argues that what is now known about human error ‘cuts at some of the foundations of the Harm Principle’ (Sunstein, Reference Sunstein2014: 5). The predictable flaws in judgment and decision-making provide an opening for paternalistic government interventions which aim to ‘make it more likely that people will promote their own ends, as they themselves understand them' (Sunstein, Reference Sunstein2014: 19; italics in original). Thaler and Sunstein have argued that policy makers should use the empirical findings concerning judgement and decision-making to influence citizens’ behaviour. They advocate the uses of ‘nudges’. Nudges are minor interventions in the decision-making context that leave the choice set unchanged yet by cleverly exploiting knowledge of our cognitive biases – such as the importance of defaults – increase the likelihood that people will choose the ‘better’ option. Nudges supplement – and to some extent replace – policy makers’ existing array of tools for changing behaviour: economic incentives, information provision and compulsion. Nudges are typically minor, inexpensive interventions such as sending reminders, changing the graphic design and wording of communication, changing the timing of programmes, and changing defaults.
In Thaler and Sunstein's view nudges are more appealing than alternative methods for changing citizens’ behaviour because they are non-coercive. They argue that the possibility of deliberately influencing choices without closing any options off makes possible a libertarian form of paternalism: non-coercive benevolent action by government motivated by predictable flaws in citizens’ judgment. They claim this approach balances the imperative for government intervention in a variety of policy spheres with appropriate respect for freedom of choice thereby offering a viable ‘Third Way’, capable of being embraced by political leaders of both the left and right (Thaler and Sunstein, Reference Thaler and Sunstein2008: 252–53).
The features which make libertarian paternalism widely politically palatable have become reasons to criticise it. Nudge provides governments with a way of responding to public problems without increasing regulation, without reducing inequality, and without increasing public spendingFootnote 2 – in other words without threatening the interests of the powerful (Jones et al., Reference Jones, Pykett and Whitehead2013; Mackenzie et al., Reference Mackenzie, Collins, Connolly, Doyle and McCartney2016: 2–3). Nudge has become associated with promoters of austerity, and critics argue that governments ideologically disposed to shrinking the public sector have used nudge as a substitute for the more transformative policies that are required to adequately address environmental, health and economic crises. Some scholars have further argued that libertarian paternalism serves to prop up neoliberalism by resolving the contradiction between the rationality of the calculating, choice-making individuals presupposed by conventional economics, and the irrationality of actually existing human beings (Jones et al., Reference Jones, Pykett and Whitehead2013; Davies, Reference Davies2014). The idea is that by providing nudges governments induce citizens to consume and invest in the ways they would if they were rational choice makers, thereby concealing the inherent self-destructiveness of a society governed by market logic.
The psychology of poverty
While nudge largely ignores the socio-economic determinants of behaviour (Bonell et al., Reference Bonell, McKee, Fletcher, Haines and Wilkinson2011), there are some behavioural scientists who shine a spotlight on the relationship between poverty and behaviour. When the well-off make errors of judgment or give in to temptations, they usually suffer no life-changing consequences. Having more than they need affords them room to make some sub-optimal decisions without precipitating a crisis. However, the precarity of the poor means that errors of judgment on their part can have serious ramifications (Bertrand et al., Reference Bertrand, Mullainathan and Shafir2004). Importantly, behavioural scientists have demonstrated that the causal relationship between poverty and flawed decision-making does not just run one way (Spears, Reference Spears2011). The brain development of children who are raised in poverty is negatively impacted by the chronic stress they experience (Evans and Schamberg, Reference Evans and Schamberg2009; Blair et al., Reference Blair, Granger, Willoughby, Mills-Koonce, Cox, Greenberg, Kivlighan and Fortunato2011). Moreover, the experience of poverty has also been shown to have immediate and reversible effects on decision-making capabilities. Normal human psychological limitations – our limited cognitive abilities, limited attention and limited self-control – interact with the environment of poverty to produce the behaviour that advocates of harsh welfare regimes attribute to permissive welfare.
Behavioural scientists offer a number of explanations for this phenomenon. The everyday purchasing decisions people on tighter budgets must make are more cognitively demanding than the everyday financial decisions of the better-off because they involve difficult trade-offs. When low incomes are uncertain and unstable this exacerbates the cognitive effort required to live within a budget. Being constantly preoccupied with money problems leaves people on inadequate incomes drained of the cognitive resources needed for other mentally demanding tasks. In addition people who are poorer have less access than the wealthy to technologies and services which minimise the need to draw upon our mental resources (Banerjee and Duflo, Reference Banerjee and Duflo2011: 68–70). Further, features of the environments people on low incomes often live in – such as lots of background noise and overcrowding – may rob them of available cognitive resources (Bronzaft, Reference Bronzaft1981; Mullainathan and Shafir, Reference Shafir2013: 42). People who are poorer have been shown to have lower-quality sleep and this also has an impact on cognitive performance (Patel et al., Reference Patel, Grandner, Xie, Branas and Gooneratne2010).
Mullainathan and Shafir argue that ‘scarcity’, by which they mean the experience of ‘having less than you feel you need’ (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 4), impacts on decision-making capacity because ‘the mind orients automatically, powerfully, toward unfulfilled needs’ (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 7). Scarcity creates a distinctive mindset characterised by a narrow and involuntary focus on immediate pressing needs. This mindset can make it hard to process information, pay attention, make good decisions, plan for the future, and resist temptations. As an involuntary, ever-present experience of deprivation, the experience of living in poverty exemplifies the psychology of scarcity.Footnote 3 The daily challenge of balancing a very tight budget robs people of the ability to give their full attention to other tasks. Keeping to a tight budget means exercising self-control frequently to resist temptations. On the basis of a randomised control trial, Spears (Reference Spears2011) argues that making difficult budgeting decisions depletes a person's reserves of self-control and therefore makes them more impatient than they otherwise would be. This finding has important implications for a variety of activities that demand self-control such as parenting, quitting smoking and losing weight (Mullainathan and Shafir, Reference Shafir2013: 159).
Importantly, behavioural scientists have demonstrated that the scarcity mindset can be temporarily created in the laboratory by placing subjects in artificial situations which induce the feeling of not having enough (Shah et al., Reference Shah, Mullainathan and Shafir2012). As a consequence, experimental subjects from middle-class backgrounds suddenly display the short-term thinking and irrational decision-making often attributed to a culture of poverty. The mindset engendered by scarcity encourages people to make short-sighted decisions which exacerbate their financial difficulties: scarcity perpetuates itself (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 14). Mullainathan and Shafir believe that the psychology of scarcity is an important explanation for why poor people do things that entrench their own poverty.
Commonalities between libertarian paternalism and new paternalism
That libertarian paternalism and new paternalism are sometimes associated with each other is understandable. Their commonalities are worth acknowledging before I clarify what sets them apart. Like Mead, libertarian paternalists argue that state interventions aimed at influencing particular behaviours which are not illegal and do not directly harm others are justifiable because their benefits outweigh their harms (Mead, Reference Mead and Mead1997).Footnote 4 As Sunstein writes ‘The unifying theme of paternalistic approaches, however diverse, is that government does not believe that people's choices will promote their welfare, and it is taking steps to influence or alter people's choices for their own good’ (Sunstein Reference Sunstein2014: 54, italics in original). Both libertarian paternalism and new paternalism ground their reasoning on the claim that people cannot be relied upon to act in their own best interests or in accordance with their own values. Like behavioural economists, Mead makes much of seemingly counterproductive behaviour suggestive of a gap between the choices people make in day-to-day life and their actual values (Mead, Reference Mead and Mead1997: 24). Both libertarian paternalists and new paternalists explicitly challenge ‘the unreality of economic assumptions’ about psychology (Mead, Reference Mead and Mead1997: 28), although the former do so from a more sophisticated evidence base. This leads them to insist that the conventional tools of policy – such as incentives and information provision – are insufficient to achieve policy objectives as they unrealistically presume the rationality of their targets.
Although Thaler and Sunstein, unlike Mead, claim to value liberty highly, both libertarian paternalism and the new paternalism have been criticised for showing too little respect for the value of autonomy. While nudges which support people to make well-informed, conscious decisions can be considered empowering, nudges which harness subconscious biases to promote desired behaviour are manipulative and potentially disempowering (Hausman and Welsh, Reference Hausman and Welsh2010; Jones et al., Reference Jones, Pykett and Whitehead2013: 52). Doubts have been raised about the ethics and political legitimacy of using nudges in the latter category, especially if they are not subject to adequate democratic scrutiny and deliberation (Jones et al., Reference Jones, Pykett and Whitehead2013: 46).
The nudge agenda and welfare conditionality are also linked by the controversial premise that policy makers know what is best (Standing, Reference Standing2011a). Libertarian paternalism and new paternalism find justification for government interventions in the values of the people whose behaviour the inventions aim to change, but critics are concerned about the authority with which government claims to know what these values are (Bielefeld, Reference Bielefeld2014). The possibility of policy makers using libertarian paternalism as a cover, to impose the majority's values on a minority who do not share them, seems to be at the heart of Standing's objection to nudge (Standing, Reference Standing2011b: 265). Sunstein and Thaler present libertarian paternalism as a form of ‘means’ paternalism rather than ‘ends’ paternalism, and therefore as compatible with value pluralism (Sunstein, Reference Sunstein2014: 17, 63). However, Sunstein and Thaler could be accused of erroneously assuming that everyone shares some of their values (Saint-Paul, Reference Saint-Paul2011), and thereby subtly importing ‘ends’ paternalism into their proposals. Mead's claim to be merely closing ‘the gap between intention and action’ (Mead, Reference Mead and Mead1997: 28) is even more contentious. His tendency to assume individuals’ interests converge with societal interests is more marked and his writing displays a blindness to the possibility that, with respect to paid work and marriage, some people hold values very different to his own.
Using behavioural science to critique Mead
Despite their common vulnerability to some ethical objections, it needs to be acknowledged that new paternalism and libertarian paternalism have contrasting psychological foundations which make them significantly different. The first important contrast is that the behavioural sciences are more concerned with observing and explaining human behaviour than with moralising. Experimental techniques have revealed that prosocial behaviour has its roots more in the vagaries of decision-making contexts than in virtuous personalities (Darley and Batson, Reference Darley and Batson1973). The tendency to give too much weight to actors' characters when assessing causal responsibility for events is in fact shown by the behavioural sciences to be a predictable cognitive quirk – the fundamental attribution error (Jones and Harris, Reference Jones and Harris1967). One striking feature of nudge is the situationist psychology underpinning it (McTernan, Reference McTernan2014: 9). Thaler and Sunstein's work is infused with the conviction that behaviour is far more influenced by context than people realise.
A related difference is that, while new paternalism focuses on the behavioural deficiencies of the poor, the behavioural sciences understand all people, not just those in poverty, to be irrational in the way they respond to incentives. This is underlined by the different pronouns that dominate the writing of different advocates for paternalism: while Mead focuses on ‘their’ behaviour, the libertarian paternalists write about ‘our’ behaviour. Mead claims that overly generous welfare provision causes departures from rationality. By contrast, the behavioural sciences hold that departures from rationality are normal, predictable and largely adaptive. A further difference is the temporal dimension of their explanations for counterproductive behaviour. Mead attributes such behaviour to long term socialisation, holding that dysfunction is socially transmitted from parents to children. Meanwhile, behavioural researchers have shown through experiments that short-term thinking and impulsiveness can be induced by the immediate context (Spears, Reference Spears2011; Shah et al., Reference Shah, Mullainathan and Shafir2012). Any of us could quickly fall into behaving like the poor if our income was severely cut. Rather than encourage self-satisfaction with our virtuous behaviour, this approach encourages empathy – there but for the grace of God go I. What we know about the psychology of scarcity makes the distance between the poor and the comfortably well off seem far more tenuous and vulnerable. The empathy that Mullainathan and Shafir promote is important in light of anti-welfare populism and the stigmatisation of poverty (Hoggett et al., Reference Hoggett, Wilkinson and Beedell2013).
These very different ways of understanding the causes of counterproductive behaviour suggest very different social policy responses. Mead's moralistic way suggests a thoroughgoing attack on the culture of poverty and what he sees as the indulgent treatment of the poor by the state. The concepts of punishment and reward are central to Mead's understanding of how behaviour can be rectified. Mead advocates supervision which will prevent people from acting against their presumed best interest (Mead, Reference Mead and Mead1997: 5). This denies them the right to make the wrong choice. By contrast, libertarian paternalism does not forbid any options. People remain free to make suboptimal choices if they wish.
According to Mead, ‘Behavior stems in the first instance from the individual, and there is no way to change it unless at least some responsibility is imputed to the individual’ (Mead, Reference Mead1986: 46). Mead emphasises that the poor should be taught to blame themselves for their poverty and to internalise the consequences of their poor decisions. The finding that behaviour is highly responsive to seemingly trivial features of the context challenges this assertion. Moreover, it makes the goal of increasing personal accountability for behaviour appear unnecessary. Instead it suggests that altering decision-making contexts is a more fruitful way of encouraging desired behaviour. Libertarian paternalism has a tendency to shift responsibility to third parties such as private sector companies who contribute to the choice architecture implicated in problematic choices e.g. through regulation specifying how information about their products is to be conveyed, how choices are presented to customers and so on (Thaler and Sunstein, Reference Thaler and Sunstein2008: 191–2).
The empirical findings from the behavioural sciences should change our expectations of what the poor should be capable of if only they had more resolve, greater self-discipline and stronger incentives. Policies which focus on bolstering resolve by increasing the consequences of failure look counterproductive. Mullainathan and Shafir argue that policies which imposes incentives for desired behaviour on the assumption that the poor are insufficiently motivated to help themselves miss the mark. Providing greater incentives (for example, to find paid employment) does not necessarily work: ‘Psychological biases often persist despite more extreme consequences’ (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 83). When income support is tied to compliance with a list of conditions, it punishes most those who are suffering most. Qualitative studies reveal that benefit claimants often do not comprehend the complicated rules that apply to them and that some are being punished for lack of understanding rather than deliberate non-compliance (Henman, Reference Henman2011: 14; Watts et al., Reference Watts, Fitzpatrick, Bramley and Watkins2014: 7). Understanding the array of behavioural expectations and the consequences attached to non-compliance increases cognitive load (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 173). Successfully negotiating a complex set of behavioural conditions requires more mental energy than people surviving on low incomes can spare (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 173). Threatening to cut benefits in response to non-compliance worsens uncertainty and income instability.
Policy aimed at attempting to induce superhuman vigilance among people living in poverty appears misguided when seen in this light. Mead asserts the corrective potential of short-term hardship. Research into poverty has made behavioural economists doubt the wisdom of this. Bertrand et al. (Reference Bertrand, Mullainathan and Shafir2004: 422) for example, criticise the excessive ‘hassle’ involved in applying for welfare benefits. While Mead asserts more stress is required to activate people (to counteract what he presumes is the passivity induced by welfare receipt), the behavioural research suggests that this would be counterproductive –‘the adversarial tone' of welfare programmes may exacerbate hopelessness and reduce compliance with behavioural requirements (Bertrand et al., Reference Bertrand, Mullainathan and Shafir2004: 422).
Applying behavioural insights to social policy: beyond nudge
One way of using the findings from behavioural research on poverty is to reform existing welfare-to-work policies to make them more effective. The behavioural sciences could provide insight into which of the interventions advocated by Mead and currently implemented in various countries are likely to help (and which to merely harm). Mullainathan and Shafir call for anti-poverty programmes to be designed with the psychology of scarcity in mind (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 168–176). They suggest ways of tinkering with the incentives to make them more effective (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 170). They also have suggestions for improving the support provided to welfare recipients. For example, they advocate designing training programmes so that participants who slip up can recover from setbacks. Behavioural finance has generated some techniques and tools for supporting individuals to make better decisions under adverse circumstances and tricks to make saving easier (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 208, 214). It also provides a rationale for making tools which wealthy people use to protect themselves against their own mistakes more widely available to people on lower incomes. For example, we should give poor people access to financial tools such as savings accounts.
Many of these proposals are consistent with libertarian paternalism in that they treat the consequences of poor choice-making as given and concentrate on reducing the likelihood of poor choices being made.Footnote 5 But behavioural insights could instead inspire an alternative political philosophy that emphasises the state's responsibility to ameliorate the consequences of poor choices. Though the work of the behavioural insights team within the Cabinet Office has focused on identifying small tweaks and quick fixes in the immediate context of decision-making, this is not the only possible policy response to behavioural insights. Nothing in behavioural insights necessitates policy makers focusing on the smaller, immediate context rather than the bigger societal context.
In some policy areas the lesson from the behavioural sciences is that nudges are inadequate and stronger government responses are required (Lunn, Reference Lunn2012: 436; Sunstein, Reference Sunstein2014: 164). Some behavioural scientists are now calling for a second generation of behaviourally informed policy which aims to ‘more aggressively address the underlying causes of problems’ (Bhargava and Loewenstein, Reference Bhargava and Loewenstein2015: 397). Many scholars have argued that behavioural insights provide a rationale for regulation to protect consumers from manipulative behaviour by private sector companies (Ariely, Reference Ariely2009: 290; Bertrand et al., Reference Bertrand, Mullainathan and Shafir2006: 20; Bonell et al., Reference Bonell, McKee, Fletcher, Haines and Wilkinson2011; Goodwin, Reference Goodwin2012; Leggett, Reference Leggett2014; Oliver, Reference Oliver2013a: 696, 698). Gilles Saint-Paul (Reference Saint-Paul2011) agrees there is no inherent link between behavioural insights and small government. Writing from a libertarian position, he warns that the behavioural sciences could spur a dystopian expansion of government intervention into citizens’ private lives.
Behavioural insights can be used to mount a case for a more generous, less punitive approach to income support. The starting point is to acknowledge that the expectations made of the poor which drive new paternalist welfare reform are unreasonable. The nature of the cognitive machinery humans share imposes limits on how self-disciplined and conscientious we can reasonably expect people living in adverse circumstances to be. The behavioural sciences show us the unremarkableness of making mistakes concerning money especially among citizens in advanced capitalist societies. Behavioural scientists have contended that high levels of consumer debt are a predictable weakness of a financial system which provides instantaneous access to credit, rather than a sign of the moral failings of the indebted (Laibson, Reference Laibson1997). That large numbers of people in capitalist societies will be burdened by worries about money is entirely foreseeable if we take the view, as behavioural economists George Akerlof and Robert Shiller (Reference Akerlof and Shiller2015) do, that it is in the nature of capitalism that firms bombard consumers with opportunities to make foolish decisions, and that many firms confuse, manipulate and deceive their potential customers.Footnote 6
Akerlof and Shiller reason that, since personal financial insecurity is an affliction no one would choose, its commonness in rich countries at a moment in world history of such high productivity is a signal that poorly regulated markets are creating flawed choice-making contexts (2015: xiii). They argue that the challenge in budgeting effectively, in an environment where firms use great ingenuity to systematically prey upon consumers’ weakest moments, is often underestimated (Akerlof and Shiller, Reference Akerlof and Shiller2015: 153). In such an environment, the expectation that everyone will succeed in saving adequately for retirement becomes fanciful (even when the issue of inequality in earnings is left aside). For this reason some behavioural scientists now consider that nudging people into better retirement savings plans is not enough (Bhargava and Loewenstein, Reference Bhargava and Loewenstein2015: 400). Guaranteed income support from government remains necessary to see people through the years when they are not earning due to retirement, unemployment or disability (Akerlof and Shiller, Reference Akerlof and Shiller2015: 154).
Not only does our growing understanding of the limits of self-control, cognitive resources and attention suggest the necessity of alleviating material hardship through a socially provided safety net, the finding that choices are greatly influenced by environmental factors also gives us reason to think that a socially guaranteed safety net is fair. By undermining the idea that individuals should personally bear responsibility for their poverty, behavioural insights offer support for the whole social democratic project of socialising risk. This includes, but goes beyond, tax and transfer policy. This entails, for example, more forgiving education systems which provide people, who failed or dropped out while young, opportunities to re-enter. It would mean greater limits on the extent to which the logic of the market governs access to the various goods and services the economy produces; in other words, moving away from the principle of ‘user pays’ in healthcare, education, childcare, housing and so on. The dominant social policy paradigm influenced by the New Right gives too much attention to individual responsibility at the expense of the societal responsibility to remedy the structures that make life for the worst off so fraught. A shift in the focus of social policy to address the real problem – precarity – could have a follow-on consequence of reducing self-destructive behaviour because income stability is ‘psychologically transformative’ (Mullainathan and Shafir, Reference Mullainathan and Shafir2013: 178). Adequate public provision of services could similarly have beneficial effects on the everyday decision-making capacity of people experiencing hardship. For example, Mullainathan and Shafir argue that it is worth publicly investing in better access to quality childcare for low-income parents because, when childcare is reliable, it enables parents to stop worrying about their childcare arrangements, direct their mental resources to other tasks (2013: 176–7).
Conclusion
Robert Goodin has written that both sides of the debate over welfare reform agree that the long term unemployed
are depressed and dispirited and (for the most part) passive and quiescent. How best to interpret that phenomenon is the question. Where the conservative critic sees a failure of will, the welfare dependent sees an absence of realistic alternatives. What is the more apt characterization of the situation is, of course, largely a matter of empirical investigation (Goodin, Reference Goodin, Schmidtz and Goodin1998: 125).
Empirical observation by behavioural sciences has revealed that neither is entirely apt. Rather it has generated a third explanation: humans are predictably irrational in ways that make it unlikely that people on inadequate incomes will be able to consistently muster the attention, deliberative powers and willpower to seize the very limited opportunities for a better life available to them. This insight has the potential to move welfare reform beyond the current stalemate. Behavioural research observes the same symptom, but it reveals the flaws of Mead's diagnosis. In addition to mounting moral arguments about the proper limits to government intervention in people's lives, or about the unfairness of imposing greater scrutiny and compliance costs on the poor than on other citizens, those who want to resist the harsh conditionality in welfare systems around the world can use the behavioural sciences to question the efficacy of these policies.
Behavioural insights have the potential to make a valuable contribution to the broader debate about how much weight can be given to individual behaviour in explanations of poverty. The findings by behavioural scientists who study poverty suggest that there are circumstances in which humans are unlikely to flourish no matter how motivated they are. Ironically studies of individuals (often observed in a lab, abstracted from social context) have given policy makers good reason to look at the causal importance of the social environment. Though nudge ‘neglects how that environment is itself a historical product, and one shot through with the traces of previous decisions, contestations and relations of power’ (Leggett, Reference Leggett2014: 11), I share Patrick Brown's hope that the concept of choice architecture could be broadened beyond the immediate context of decision-making. In this way, policy makers’ current enthusiasm for behavioural insights could provide a starting point for a conversation about how behaviour is shaped by social structure (Brown, Reference Brown2012: 308).
By offering a defence of the value of the behavioural sciences to social policy I do not mean to overlook the inherent limitations of these disciplines. Research into poverty by behavioural scientists affords a central role to scientific experts and gives little voice to people living in poverty. Their methodology cannot offer the same deep insight into subjectivity that qualitative research promises. Though I have argued that the behavioural sciences provide impetus for paying attention to the social circumstances that produce precarity, other disciplines such as history, political economy, political science and sociology are necessary for understanding how these circumstances came about and what maintains them.
Nonetheless, if we understand social structures as the cumulative product of decisions made by interdependent agents, the behavioural sciences can assist in this task. It is sometimes assumed that ‘behaviour’ is code for the problematic behaviour of people marginalised by the present economic system (Rhys et al., 2013: 51; Sampath, 2015), yet there is no reason why behavioural scientists should not focus their inquiry on the problematic decision-making of the wealthy and powerful. We can learn how choice architecture affects decisions to racially discriminate, evade tax, hoard opportunities within privileged social networks, or pollute the environment. We could study how receiving large campaign donations from corporations affects legislators' votes. Studies could also shed light on ways decision-making by the powerful and privileged could be scaffolded to encourage more socially responsible decisions.
I have argued that insights about decision-making under scarcity provide justification for welfare state interventions aimed at providing a stronger safety net, not just subtle alterations to choice architecture. The present article affirms and extends Oliver's argument that the equation in some people's minds of behavioural insights with nudges have led to an under-appreciation of the full implications of behavioural insights for reshaping policy agendas. Oliver is right that ‘behavioural economics could be used for more appropriate and profound purposes' than those encapsulated by the nudge agenda (Oliver, Reference Oliver2013a: 692). In the social policy arena, behavioural insights could be used to support policies which are more ethical and more empowering for the precarious citizens whose behavior – indeed, whose lives – they seek to transform.
