The end of political and technological limits

The political limits to the growth of conditionality are fast disappearing. Gone is the aversion to scrutinising recipients’ behaviour under the quasi-Titmuss paradigm, which long held behavioural conditionality at bay (Deacon, Reference Deacon2002). Dwyer and Wright identify a ‘new ubiquitous conditionality’ drawing in not just the unemployed but lone parents, the ill and disabled, and those in low-paid work (Dwyer and Wright, Reference Dwyer and Wright2014: 33). Though the new conditionality is associated with the idea that paid employment is the pre-eminent obligation of citizens (Patrick, Reference Patrick2011: 276), governments are experimenting with using welfare conditionality to enforce a diverse range of other behaviours incumbent on citizens. Governments are even willing to use conditionality to shape behaviour, such as parenting, which has traditionally been regarded as private (Taylor et al., Reference Taylor, Gray and Stanton2016). Coercive obligations enjoy support from the electorate and are seen by politicians as a way of shoring up taxpayers’ support for the benefits system at a time when public sympathy for working-age people on low income is low (Henman, Reference Henman2011; King, Reference King1999: 304; Taylor-Gooby, Reference Taylor-Gooby2015; Deeming, Reference Deeming2015).

Technological developments have been important to the growth of behavioural conditionality (Henman, Reference Henman2011). This is an example of the capacity of digital technology to break administrative limits (Hood, Reference Hood2010: 529–30). Technological innovation and the endless documentation of the culture of audit have both enormously expanded the amount of behavioural information recorded and stored. Behavioural conditionality flourishes where behaviour is observable, measurable and verifiable. Networked electronic infrastructure enables cooperation and information sharing between government departments responsible for governing different areas of life. This information-sharing facilitates the linking of payments to behaviour not intrinsically related to the purposes of those payments (Henman, Reference Henman2011: 3). Through electronic networks governments can also receive regular information from non-government organisations well-placed to monitor compliance with behavioural standards such as schools, and providers of housing, health and employment services. This enables government to use its coercive power to reinforce the authority of a host of agents of governance (Watson, Reference Watson2015). It is striking that many of the behavioural conditions for receiving welfare payments concern the accessing of services. Because accessing services is verifiable and leaves a documentary trail, it lends itself to becoming a focus of eligibility requirements (Watts and Fitzpatrick, Reference Watts and Fitzpatrick2018: 37–40).

Welfare conditionality is being introduced to enforce an ever-increasing number of responsibilities. The links between the behavioural condition and the benefit programme are not automatic but orchestrated consequences, and could be either vivid to recipients or highly tenuous and contrived. Benefits have been made conditional on all manner of behaviours connected to preparing for employment, seeking employment and staying in employment, even having the right attitude to employment (Friedli and Stearn, Reference Friedli and Stearn2015). Drug testing now enables benefits to be made contingent on refraining from illicit drug use (Wincup, Reference Wincup2014; Pérez-Munoz, Reference Pérez-Munoz2017). Benefits have been made conditional on complying with expected parenting behaviour such as ensuring children are attending school and accessing health services (Taylor et al., Reference Taylor, Gray and Stanton2016). Welfare payments have been linked to attendance at classes for budgeting and parenting (O’Brien, Reference O’Brien2013: 740), and the accessing of therapeutic services – from drug and alcohol treatment (NCSL, 2017) to psychometric testing (Friedli and Stearn, Reference Friedli and Stearn2015). Behavioural conditionality has aimed to shape citizens’ reproductive decisions (O’Brien, Reference O’Brien2013: 740) thereby assisting the state to meet population goals (both more children and fewer children).

Conditionality could in theory be put to an almost infinite array of purposes. Traditional civic duties such as voting, jury duty and national service could be tied to welfare payments, as could compliance with the routine administration that makes the state possible: such as filling in census forms or tax returns. Welfare payments could be linked to the duties of the good neighbour or the good tenant: such as maintaining residential property, caring for shared spaces, not making excessive noise, and cooperating in waste management. In the UK the possibility of using conditionality to encourage claimants to seek treatment for obesity and other health conditions has been mooted (Black, Reference Black2016: ch 2). Cancer screening could easily be linked to benefits, and as personal monitoring technologies develop we could imagine the potential to link eligibility for benefits to all manner of health promoting behaviours from smoking cessation to daily physical exercise to taking prescription medicine as directed. The capacity to use monitoring devices in this way could be increased by introducing benefit sanctions for people who do not comply with efforts to monitor them (see Lupton, Reference Lupton2014). We could even imagine benefit sanctions being applied to people who fail to follow authorities’ directions in times of civil unrest, natural disaster, epidemics or other emergencies.

The above is sensational and partly speculative. However, this long list shows the arbitrariness of singling out one form of deviation from social expectations as a basis on which to punish people. In myriad ways our behaviour intentionally or unintentionally affects those around us. Hence our duties to our fellow citizens are considerable. Yet the enforcement of these duties through an ever-growing regime of surveillance and punishment is a dystopian vision.

For the sake of balance, some remaining practical limits to conditionality are worth acknowledging, such as the potential reluctance of non-governmental authorities to cooperate in the enforcement of conditions. Professionals’ ethical duties to their clients might constitute a barrier to their full cooperation with the punitive ambitions of the state. The increasing density of conditions may hit a natural limit: at some point many recipients will have no more money left that the state can threaten to take away. There is also evidence that some highly vulnerable individuals are choosing to disengage from the social security system altogether rather than continue to be subject to conditionality (Jordan and Fowkes, Reference Jordan and Fowkes2016; Taylor et al., Reference Taylor, Gray and Stanton2016; Fletcher and Flint, Reference Fletcher and Flint2018). Nevertheless, policymakers do seem to be in the thrall of the promise of incentives to fix problems. More debate about the ethical concerns posed by welfare conditionality is needed to prevent them unreflectively reaching for new conditions in response to ever more social problems.

Incentives and intrinsic motivation

Welfare conditionality amounts to using the social security system as a means of administering financial incentives; so, to understand the ethics of this behaviour change tool, it is useful to reflect on the ethics of incentives. The use of financial incentives in the pursuit of policy goals is widespread. The growth in welfare conditionality should be seen in the context of the pervasive assumption by policy makers that self-interest is the key to individual motivation (see for example Murray, Reference Murray1984). This questionable assumption is also responsible for propelling market norms into spheres hitherto governed by nonmarket norms. From this assumption it follows that individuals’ behaviour can be deliberately shaped by government through offering incentives or disincentives that they will factor into their decisions in predictable ways (Le Grand, Reference Le Grand1997). Yet behavioural economists now realise that focusing narrowly on material gain as the driver of human behaviour overlooks the strong influence that our desires to be particular kinds of people (a good mother, a good neighbour, a trustworthy person) have on the choices we make (Bowles, Reference Bowles2016). Though recent interest among policymakers in the behavioural sciences is sometimes linked by scholars to the growth in welfare conditionality (Curchin, Reference Curchin2017), these disciplines in fact offer a resource for critiquing the reductionist view of human motivation underpinning conditionality.

Behavioural scientists distinguish between two types of motivation: intrinsic and extrinsic. While intrinsic motivation derives from features of the task itself, extrinsic motivation derives from rewards outside of the task (Deci, Reference Deci1971). While people feel intrinsically motivated to perform tasks which are fun or offer a sense of mastery or appeal to their curiosity or sense of fairness, extrinsic motivation takes the form of financial incentives, material rewards, praise and threats of punishment (often referred to as sticks and carrots). Though neo-classical economics suggests that multiple motivations to do the same task deriving from different sources should reinforce each other, a vast literature from behavioural science attests to the complex relationship between intrinsic and extrinsic motivation (Bowles, Reference Bowles2008). In fields such as educational psychology and organisational psychology it is now acknowledged that attempts to elicit desired behaviour through extrinsic motivators – such as praise, incentives and rewards – can backfire because of their unanticipated effects on intrinsic motivations such as public-spiritedness or creativity (Frey and Jegen, Reference Frey and Jegen2001; Titmuss, Reference Titmuss1970). Hence treating people as solely motivated by personal economic gain can become a self-fulfilling prophecy (Sandel, Reference Sandel2012).

One danger with incentives is that by crowding out intrinsic motivation they can inadvertently ‘create the conditions for their own perpetuation’ (Grant, Reference Grant2012: 119). For tasks for which people have little intrinsic motivation, incentives are appropriate since there isn’t anything to undermine (Grant, Reference Grant2012: 119). But some tasks linked to welfare payments are the kind for which we would expect people to have some intrinsic motivation. For example, the desire to improve one’s circumstances provides intrinsic motivation for job search and training. The susceptibility of intrinsic motivation to disruption by external motivation seems to be behind the resistance of some voluntary agencies in the United Kingdom to workfare. Concerned that forcing unemployed people to perform tasks normally performed by volunteers undermines the meaning of volunteering, hundreds of organisations have signed an agreement to ‘keep volunteering voluntary’.Footnote ¹

As Michael Sandel argues ‘To decide whether to rely on financial incentives, we need to ask whether those incentives will corrupt attitudes and norms worth protecting’ (Sandel, Reference Sandel2012: 91). This moral assessment would need to be made on a case-by-case basis, exploring the intrinsic motivations people have for performing the desired behaviours to which welfare payments are now attached in realms from job-seeking to parenting to health prevention. In what follows I take the case of childhood immunisation and try to draw out the ethical implications of providing incentives through the social security system.

The illiberal justification for ‘No Jab, No Pay’

In response to falling immunisation rates particularly among young children (under the age of six) the federal government in Australia initiated a suite of new measures in 1997. The Seven Point Plan included financial incentives for both general practitioners and parents (Ward et al., Reference Ward, Hull and Leask2013). In 1998 a one-off means-tested payment – the Maternity Immunisation Allowance (MIA) – was introduced to reward timely immunisation, and child care subsidies were made contingent on children being fully immunised. By 2000 the rate of children fully immunised rose substantially, but because multiple strategies were implemented at once it is difficult to be sure how much of this change was attributable to the financial incentives for parents (Bond et al., Reference Bond, Davie, Carlin, Lester and Nolan2002; Lawrence et al., Reference Lawrence, MacIntyre, Hull and McIntyre2004). In the years since 2000, there have been modifications to the names and design of family payments and childcare subsidies, but they continue to be linked to immunisation status (Ward et al., Reference Ward, Hull and Leask2013). Initially the federal legislation making payments to parents contingent on immunisation included provision for philosophical exemptions as well as exemptions on medical grounds. In 2015 amending legislation was passed to remove philosophical exemptions. It is noteworthy that in Australia the number of parents who are strongly opposed to vaccinations is very small (Pearce et al., Reference Pearce, Marshall, Bedford and Lynch2015).

In a speech to parliament on the removal of conscientious objection the Australian Minister for Social Services said: ‘The choice made by some families not to vaccinate their children is not supported by public policy or medical research and advice to the government, nor should such action be supported by taxpayers in the form of family payments’ (Morrison, Reference Morrison2015). The authoritarian implication of this claim is that, unless every aspect of one’s behaviour is approved of by government, one has no entitlement to any cash transfer. The suggestion that providing a citizen with cash transfers amounts to endorsing every aspect of their behaviour is especially troubling because many other obligations of citizens could potentially be linked to benefits. No strong case has been made that the obligation to have one’s child immunised has a greater claim than others. This logic widens the scope for punishment by government enormously.

Unlike workfare, childhood immunisation conditionality is not justified by reference to a failure by recipients to competently pursue their own best interests. The claim is never made that the parents who resist immunisation are incapable of acting autonomously (cf. Mead, Reference Mead1986, Reference Mead and Mead1997). Rather the policy is justified by reference to the communitarian idea that ‘responsibilities are the flip side of rights’ (Albrechtsen, Reference Albrechtsen2015: 21). Families have a duty to contribute to the important public health goal of preventing contagious childhood diseases. Parents who refuse to vaccinate their children are seen as free riders – their families benefit from herd immunity without taking the small risks involved in contributing to it (Salmon et al., Reference Salmon, Teret, MacIntyre, Salisbury, Burgess and Halsey2006: 440). Writing in support of the removal of the conscientious objection exemption, the conservative commentator Janet Albrechtsen asked rhetorically, ‘Why should taxpayer funds be directed to those who cannot exercise the most basic civic-responsibility?’ (Albrechtsen, Reference Albrechtsen2015). Discernible here is the idea that popular support for the welfare system depends on the public being persuaded that recipients are meeting their obligations (Deacon, Reference Deacon2002: 110).

While the communitarian argument that people have a duty to vaccinate their children is strong, the argument that the state should enforce this duty through the welfare system is weak. Goodin argues convincingly that enforcing strict conditionality is not the most desirable structure of mutuality for the welfare state (Goodin, Reference Goodin2002). Reimaging welfare as a deal along the lines of ‘ordinary commercial contracts, where one party’s breach cancels the contract for both’ (Goodin, Reference Goodin2002: 585) corrupts the idea of the welfare state. The incidence of parents defaulting on their responsibility doesn’t relieve the government of its responsibility to them, because the source of the government’s responsibility is elsewhere (in its obligation to prevent poverty, for example). Nor is it the case that parents’ receipt of subsidised childcare is the source of their responsibility to vaccinate their children – parents who do not use childcare still have this responsibility. It flows from their duty as parents to ensure preventative healthcare for their own children and their duty as members of community to contribute to public health goals (such as herd immunity). The obligations between citizens and the state in this case are mutual but independent.

By conceiving reliance on the social security system – rather than living as a member of a community – as the source of duties, advocates of welfare conditionality are expanding the domain where wealth buys autonomy. This preservation of choice (for some) seems to be part of welfare conditionality’s appeal as a neo-liberal public health strategy: ‘Parents are still free to exercise their misguided choice not to vaccinate – only they won’t be supported by taxpayer funds when making a choice that endangers the community’ (Albrechtsen, Reference Albrechtsen2015: 21). Very wealthy parents who don’t want to vaccinate their children have the luxury of declining the payments for which they are eligible (the childcare rebate is not means-tested). In effect their wealth buys them a level of freedom to act on their principles that is unavailable to those on lower incomes. This absolves wealthy families from the duty to contribute to herd immunity, while compelling poorer families to do so.

What’s wrong with incentives for immunisation?

Let us now think about the complex relationship between intrinsic motivation and incentives in the context of immunisation conditionality. The threat of a benefit sanction is an extrinsic motivation for taking one’s child to be vaccinated. The desire to prevent one’s child from suffering diseases such as measles is an intrinsic motivation.

An understanding of the interaction of intrinsic and extrinsic sources of motivation can help explain why offering incentives can be insulting (Grant, Reference Grant2012: 56). To people who are already intrinsically motivated – for example, to have their child vaccinated – it can imply distrust and condescension. Why, a parent might think, does the government not trust me to be motivated to vaccinate my child? Does the government think I am lacking sufficient concern for my child’s health and the health of the other children in my neighbourhood? It also insults the parent who wants to vaccinate but is late due to practical barriers – such as lack of transport to clinics or family crises demanding attention – by implying that motivation rather than the family’s practical circumstances are the obstacle. This is important because research in Australia has shown that the majority of children who miss out on immunisation do so because of practical barriers experienced by their families, rather than parental concerns about the safety and efficacy of vaccines (Bond et al., Reference Bond, Nolan and Lester1999; Pearce et al., Reference Pearce, Marshall, Bedford and Lynch2015; Fielding et al., Reference Fielding, Bolam and Danchin2017).

Conditionality comes from a mindset that individualises problems rather than looks to their deeper social and structural sources. It appeals to governments seeking quick and simple fixes to behaviour that has complex root causes (Henman, Reference Henman2011: 12–13). When applied as a measure to increase use of government-funded services it deflects attention from aspects of the services that discourage people from using them. Conditionality offers an alternative to listening to potential users to understand the barriers they face in accessing the services and working with them to find ways of removing those barriers. The Australian education reformer Chris Sarra has criticised a programme that made Aboriginal parents’ welfare payments contingent on the school attendance of their children for deflecting attention from the responsibility of teachers and principals to foster positive relationships with Aboriginal school children and their families: in other words, to build their intrinsic motivation to attend school (Sarra, Reference Sarra2012: 314–15).

When conditions relate to accessing services, and people have no reasonable alternative but to comply, this may impact on the quality of services. Conditionality can deny clients the opportunity to vote with their feet when they perceive the services on offer are poor quality, inappropriate or even dangerous. Hence it removes an important check on the power that service providers have over clients. Welfare conditionality is particularly concerning when it displaces alternative ways of motivating desired behaviour such as investing in legitimacy and fairness. Relying on incentives to raise immunisation rates means missing the opportunity to increase the intrinsic motivation of the small number of parents reluctant to immunise (by, for example, building their trust in immunisation providers). The childhood immunisation strategy in the United Kingdom provides a useful contrast to Australia. There, comparable immunisation rates have been achieved by spending more resources on understanding parents’ perceptions of vaccine risks and feeding this understanding into improved public education about immunisation, instead of directing financial incentives to parents (Salmon et al., Reference Salmon, Teret, MacIntyre, Salisbury, Burgess and Halsey2006: 441).

For vaccine hesitant parents, conditionality does not attempt to address the absence of intrinsic motivation so much as bypass it. Using incentives is problematic when accompanied by the abandonment of persuasion and other efforts to build intrinsic motivation. This matters in part because democratic society depends on discussion and justification to legitimate government policy (Gutmann and Thompson, Reference Gutmann and Thompson1996). Demonising people while simultaneously designing incentives for them is hazardous because it is easy to overlook the prosocial motivations of the people demonised. Parents who are reluctant to immunise their children are a good example.

The current approach of taking away entitlements treats resistant immunisers as beyond the pale, as so unreasonable as to be unpersuadable. This characterisation contrasts with the picture that emerges from Stuart Blume’s review of qualitative research into parents who were reluctant to comply with the health authorities’ vaccination schedules (Blume, Reference Blume2006). This research helps explain why in developed countries hesitant immunisers are most prevalent among the segment of the population with the most educational qualifications (Fielding et al., Reference Fielding, Bolam and Danchin2017). Blume argues that these parents were not the mindless followers of an anti-vaccination movement. Rather they were people who valued autonomy highly, and sought to make an active decision about which vaccinations were in the best interests of their child based on their own research into the relevant diseases and vaccination risks. They did not trust health authorities to make this decision for them partly because they feared that the commercial interests of pharmaceutical companies were having an undue influence on public policy. This is consistent with the findings of recent research undertaken in Australia with parents who are reluctant to immunise their children (Attwell et al., Reference Attwell, Leask, Meyer, Rokkas and Ward2017).

These parents were not failing to comply because of an absence of personal responsibility. Rather, their conflict with health authorities stems from their refusal to abdicate responsibility for making an informed decision. Hesitant immunisers found both the written information from health authorities and the attitudes of health practitioners patronising: they perceived them as more oriented to inducing compliance with vaccination schedules than helping parents make up their own minds. Taking “parents’ grievances and beliefs seriously and, in rational debate, looking to design vaccination programmes that better respect their agency and their competence” would be a possible alternative approach to inducing compliance through financial incentives (Blume, Reference Blume2006: 640; Attwell et al., Reference Attwell, Leask, Meyer, Rokkas and Ward2017).

Parents who are strongly opposed to immunisation but not in a financial position to forgo immunisation-linked payments will experience ‘No Jab, No Pay’ as an undue influence on their decision about their child’s healthcare. In her study of the ethics of incentives, Ruth Grant writes:

Clearly the small group of parents that has a strong principled aversion to having their children immunised are being given a financial incentive to vaccinate their children against their better judgement. Immunisation conditionality asks parents to override their concern for their children’s health for the sake of the money. Another way of putting this is that these parents are being bribed by the government. Bribes are morally objectionable because they ‘bypass persuasion and substitute an external reason for an intrinsic one’ (Sandel, Reference Sandel2012: 59). By offering incentives for immunisation the government is trying to substitute its decision for that of the child’s parents. This is a form of undue influence because the decision to vaccinate a child is not the government’s to make. Medical practitioners are still legally required to obtain the consent of the patient (or the patients’ parents) for immunisation.

It is important to stress that payments or services that support parents to act on their best intentions (for example, providing immunisation services in convenient locations) are not bribery. The distinguishing feature of bribes is that they induce people to act against their better judgement. Rewards and incentives can also reinforce people’s desire to act in accordance with their sense of their obligations. Much comes down to the design of the incentives (Bowles, Reference Bowles2016). The Conditional Cash Transfers that have been effective in the developing world can be understood as supporting the intrinsic motivations of families to send their children to school by providing the cash that enables them to attend. There is much scope for rethinking the framing of welfare conditionality so that it recognises and bolsters existing motivation. This would involve a move away from the current punitive orientation, and careful analysis of how incentives can be used in conjunction with other tools of government such as exhortation and capacity building (Schneider and Ingram, Reference Schneider and Ingram1990).

The provision for conscientious objection to immunisation in place in Australia for more than a decade can be interpreted as a response to this problem of undue influence (bribery). The provision for conscientious objection had minimal impact on immunisation coverage rates. The process of seeking an exemption was onerous enough that it forced ambivalent parents to consider whether they were sufficiently convinced by anti-vaccination arguments to identify as a conscientious objector. Importantly it also brought parents into contact with immunisation experts, creating an opportunity for experts to listen to parents’ concerns about vaccinating their child and to potentially allay their fears (Leask and Danchin, Reference Leask and Danchin2017). Public health experts fear that the removal of the provision for conscientious objection has taken away a reliable source of information on the prevalence of vaccine hesitancy. Such information is useful in informing strategies to address the different barriers to childhood immunisation (Fielding et al., Reference Fielding, Bolam and Danchin2017: 257). The provision for conscientious objection was a clever way of balancing the legitimate goal of promoting immunisation with the value of autonomy. Its removal, which enjoyed bipartisan support, is a policy shift in an illiberal direction.

Response to two possible objections

Finally I want to anticipate two possible objections to my argument. The first is that the harm prevented by immunisation conditionality outweighs the lack of respect it shows for autonomy. One might ask whether parents would rather be insulted by conditionality or watch their newborns die of whooping cough. Autonomy and public health are competing considerations, and sometimes autonomy should be traded off for important public health goals. I have not attempted to make a full ethical appraisal of childhood immunisation conditionality here. Such an appraisal would do well to consider the questions Watts and Fitzpatrick have put forth to guide the assessment of the ethical legitimacy of instances of welfare conditionality (Reference Watts and Fitzpatrick2018: 152). An important question is whether denying payments represents the ‘minimum level of “power over” people commensurate with achieving’ the legitimate goal of maintaining herd immunity to childhood diseases (Watts and Fitzpatrick, Reference Watts and Fitzpatrick2018: 152). This is an empirical question and it turns on the effectiveness of alternative ways of influencing parents who are reluctant to immunise their children. Even if an affirmative answer could be given to this question, it is noteworthy that in practice ‘No Jab, No Pay’ is not justified as a regrettable but proportionate use of power over vaccine hesitant parents. Rather, the concern to minimise the use of power over this demonised segment of the population is not there.

A second potential objection is that my argument places too much weight on autonomy. Some scholars argue autonomy is overrated and that the flaws and biases in human decision-making, as revealed by behavioural economics, justify coercive intervention by government (Conly, Reference Conly2013). Though I concede that people do sometimes make decisions that are not in their own best interests, there are multiple ways that governments can respond to flaws in citizens’ choice-making capacity that support rather than erode their autonomy (Thaler and Sunstein, Reference Thaler and Sunstein2008). In the case of childhood immunisation, some immunisation experts are calling for the funding of more sophisticated communication strategies so that vaccine-hesitant parents do not just receive generic information about the benefits of immunisation, but can have a dialogue about the relevance of immunisation to their families with a health professional they trust (Macintyre and Salmon, Reference Macintyre and Salmon2015; Fielding et al., Reference Fielding, Bolam and Danchin2017: 256). This approach supports parents’ decision making about their child’s healthcare, rather than overrides it. The general trend within social security systems towards disrespecting rather than scaffolding citizens’ autonomy deserves more public debate.

Conclusion

This article’s focus has not been on whether conditionality ‘works’, but on the ethics of pursuing behaviour change in this way. Policy attention has focused far more on whether people change their behaviour in response to government interventions than ‘why’. But ‘why’ is important because ‘[i]t makes an ethical difference, both for individual character and social norms, whether people are motivated by fear of punishment, desire for gain, generosity, rectitude, and so forth’ (Grant, Reference Grant2012: 53). I have argued that there are ethical costs to governments relying heavily on extrinsic motivation such as incentives to encourage desired behaviour instead of recognising and bolstering citizens’ intrinsic motivation by addressing legitimacy and fairness and working in partnership with them to remove practical barriers. This article has contributed to the literature on the ethics of welfare conditionality by drawing out the implications of thinking of the social security system as available to provide financial incentives for behaviours such as immunisation. The line of critique I have developed here has relevance not only to immunisation conditionality but to workfare and the many other forms of conditionality targeted to disadvantaged groups. The problems I have raised about coercion, undue influence and insult apply generally. Similarly the concerns about demonization, and the failure to engage recipients of cash transfers in dialogue that respects their autonomy, are not problems specific to this particular form of conditionality.

Other scholars have raised concerns about the unequal nature of citizenship created by welfare conditionality (Henman, Reference Henman2011). They correctly point out that conditionality leads to a fractured citizenship in which citizens on lower incomes are more vulnerable to punishment, intrusive monitoring, and onerous compliance costs than more affluent citizens. I have gone further by arguing that conditionality also corrupts the nature of citizenship through its lack of respect for the autonomy of citizens. The behavioural goals of governments are potentially endless and as the technical limits and even political limits to conditionality are surmounted, a growing proportion of the population is subjected to a controlling welfare regime that does not value them as self-determining adults. The effects on autonomy of each specific condition that is attached to welfare receipt might seem relatively small. However, the cumulative effects of many conditions pose a real threat to liberal democratic citizenship.