Ontological contradictions in the UK’s Universal Credit reforms

ABSTRACT The Universal Credit reforms of the 2010s were a crucial turning point in the UK’s social security system. The reforms have been widely criticized in the literature for placing too much responsibility on welfare recipients, for using cultural explanations of poverty, and for prioritizing incentive-based solutions. This article argues that these common points of criticism actually point to demonstrable contradictions in the formation of Universal Credit, contradictions that are problematic regardless of the strength of the aforementioned criticisms. The focus is on “ontological contradictions”, which derive from fundamental assumptions about how individual agents relate to their material and ideational contexts. To make this argument, a critical realist framework is developed in a dialogue with existing poststructuralist approaches.


Introduction
This article looks back at the creation of the UK's Universal Credit system and considers the philosophical foundation on which it was originally built.Universal Credit (UC) is the UK's main system for distributing in-work benefits, and was created, as its name implies, to be a single payment to recipients, drawing together many existing benefit schemes.The controversy surrounding UC did not arise from its attempt to unify existing benefit payments but from the way in which the new system was designed.This design was the product of the Conservative Party's reappraisal of its social policy and so called "modernization" between 2005 and 2010, and its time in office from 2010 (McEnhill and Taylor-Gooby 2018;McKay and Rowlingson 2016), which focused heavily on notions of "responsibility", "incentives", and a "broken society" (Hayton 2017).Critics have focused specifically on UC's draconian conditionality, its approach to payments, and its insufficient levels of support (Dwyer and Wright 2014;Millar and Bennett 2017).This article argues that assumptions underpinning the initial design of the UC system contain fundamental and demonstrable contradictions.
The analysis builds on a strand of the social security literature that focuses on underlying policy assumptions (e.g.Millar and Bennett 2017;Monaghan and Ingold 2019), and specifically on "ontology" (e.g.Dobson 2015 andWhitworth 2016).Ontology concerns assumptions, controversies, and theories about the nature of reality, especially the questions of whether there is a reality external to our perception of it, how individuals relate to their social environment, and whether "social structure" exists and if so what it is.Ontology matters both theoretically, as an access point into debates on social science, and empirically, as a conceptual tool in qualitative analysis.The concern with ontology has a long history in the social sciences, especially in the philosophy of social science (van Orman Quine 1976), political analysis (Hay 2002;Marsh and Furlong 2002), economics (Lawson 1997;Sum and Jessop 2013), public administration (Raadschelders 2011;Stout 2012), gender studies (Butler 2002;Witt 2010), disability studies (Hughes 2007;Pfeiffer 2002), dialectics (Dixon, Woodward, and Jones 2008), anthropology (Holbraad and Pedersen 2017;Paleček and Risjord 2013), and more generally in social theory (Bourdieu 1977;Foucault 1966).Many of the authors cited here have led explicit "ontological turns" in their disciplines, especially in anthropology and economics.
Theoretically, this article advocates a critical realist approach to ontology, which is seen to differ from the poststructuralist perspective that dominates the critical social policy literature, but the article also argues that empirical findings from poststructuralist and realist studies can inform and support one another in spite of their philosophical differences.From a poststructuralist perspective, ontology is considered in terms of a hegemonic "causal essentialism" that constructs the reality of social security from politicians to street-level bureaucrats and claimants.A critical realist perspective represents at least three key differences: firstly, a broader understanding of "causality" and "essentialism" is taken, demonstrating that these concepts are not in themselves problematic for social security policy; secondly, rather than identifying a hegemonic neoliberal ontology, problems are instead identified within the fractured and contradictory ontologies within public policy; thirdly, it is argued that policy contradictions form "situational logics" that have causal consequences for agents.
Following from this theoretical discussion, the article presents empirical findings on the underlying assumptions of the Universal Credit reforms.It finds that the development and justification for these reforms was wrought with ontological contradictions that undermine the integrity of the system and undermine its capacity to deliver intended interventions.That contradictions exist within government policy is not in itself surprising.However, the value of ontological analysis lies in the specific location and identification of those contradictions, which are held to have causal consequences through the "situational logics" they create.One consequence of contradictions at this fundamental level is that policies do not justify the moral responsibility they place on individual claimants, a finding that adds further weight to arguments made against UC in the existing literature (Millar and Bennett 2017).

Ontology, social policy, and critical realism
Ontology is the philosophy of being; it is a field of philosophy that is concerned with the basic nature of reality.Social ontology focuses on "the basic nature and structure of social reality" where "social reality comprises all those phenomena whose existence depends necessarily on us" (Lawson 2019, 11).A central question of social ontology is the relationship between subjects (thinking, experiencing beings) and objects (the reality within which they think, experience, and be).Many other issues derive from this central question, including the differences between agency and structure, ideas and institutions, minds and bodies, emotions and rationality, simplicity and complexity, and change and stability.As in most ontological debates, the very distinctions between such concepts are themselves sites of contest.These questions about ontology are an unavoidable part of social philosophy and have tended to arise in most philosophical traditions across the globe and throughout history.In the West, they became particularly prominent in debates about phenomenology in the early twentieth century, and have been brought to the fore by poststructuralist and realist scholars in the years since.
In the study of social policy, these debates have tended to be informed by poststructuralist perspectives, especially the work of Foucault (Mckee 2009) and Laclau and Moofe (Howarth 2010).The key manoeuvre of a poststructuralist relational approach is the shift away from the idea that the social policy arena consists of individual entities causally interacting, and a shift towards the idea that it consists of contingent and fluid processes in which objects and subjects are in a constant state of becoming, to the extent that they never actually are anything.This approach to ontology is characteristic of the tradition of "critical social policy" and the wider philosophical movement of poststructuralism, within which "an emphasis on heterogeneity and contingency offers a refreshing skepticism about the full range of 'things' usually associated with policy, including policy itself" (Bacchi and Goodwin 2016, 4).
In her landmark article, "Power, Agency, Relationality, and Welfare Practice", Rachel Dobson (2015) builds on these traditions, putting ontology at the heart of social policy analysis in two important ways.Firstly, she questions the ontological assumptions of existing academic research into social policy, suggesting that the dominant approach of "essentialism" misunderstands the relationship between the subjects and objects of welfare practice.Secondly, she explains that the essentialist approach also dominates in policy making itself, functioning as an ontological foundation for "individualising, pathologising and moralizing" approaches to poverty (Dobson 2015, 689).Dobson (2015) calls for social policy research to move away from an ontological position in which welfare objects, such as "welfare organisations" and "the welfare state", are theorized as fixed and discreet entities that each have their own fundamental essences.The essentialist approach that Dobson critiques considers welfare objects as separate from one another and separate from welfare subjects, so that each are isolated and only connected by simple causal relations.Welfare subjects, especially policy practitioners and welfare recipients, are thus cast in one of two simplistic roles: they either resiliently resist the causal powers of welfare objects or they succumb to those powers as vulnerable victims (Dobson 2015).As an illustration of this point, Julian Le Grand's (2003) influential four-part framework for public policy analysis embodies this essentialism by distinguishing between altruistic "knights", self-serving "knaves", victim "pawns", and free-choosing "queens".Dobson rejects such essentialist approaches, using an ontological argument that society is far too changeable and interconnected for an "objects-with-essences perspective".Her argument is bolstered with the claim that essentialism leads to pernicious social policy where welfare subjects, theoretically severed from their structured social environment, are blamed as carriers (or worse as perpetrators) of poverty.
The empirical analysis of this article (laid out below) does indeed identify assumptions in government policy that tend towards this problematic separation of welfare subjects and welfare objects.The empirical analysis also shows that one consequence of the detachment of object and subject is that certain policies exaggerate individual responsibility, just as Dobson describes, while others treat individuals merely as objects in simple causal models.However, it does not follow that these problematic policy assumptions are necessarily the result of an "essentialist" and "causal" ontology.Indeed, Newton (1998) argues that poststructuralist approaches themselves struggle to avoid certain implicit types of essentialism, not least because of the need to assert some universal features of the human condition, including basic material needs and the capacity to suffer, features that are crucial in the study of social policy.From a critical realist perspective (see Sayer 2000 andCollier 1994 for an introduction), it is possible to argue that welfare objects and subjects are entities with unique causal powers, while at the same time rejecting the strict separation of object and subject that leads to the problematic ontology that Dobson rightly criticizes.As O'Mahoney (2012, 735) puts it "from a CR perspective, it is entirely plausible to reject these incorrect and sometimes damaging forms of essentialism … but to accept the forms of essentialism that allow a coherent and consistent discussion of discourse, subjectivity and resistance".
Let us take the claim common to Dobson (2015), Hunter (2015), and Bacchi and Goodwin (2016), that we cannot speak of welfare in terms of "entities" or "things" with fixed essences, because the constant flux of society means that objects, and indeed subjects, are in a constant process of being made and unmade.In a weak sense, this claim is important but uncontroversial; few would claim that society is entirely static and few would claim that the entities within it are eternally unchanging.However, it does not follow that we therefore abandon the notion of these entities having essences.All essences are temporally limited, but they can be said to exist as long as they hold a particular form.For example, an institution of social policy, such as the UK's "Department of Work and Pensions", can be said to exist as an entity with an essence because it has a particular set of causal powers.Some of these powers belong to the organization as a whole and some to the individual roles within the organization, in the form of rights and responsibilities, but together they make up a real entity that has unique causal powers (Lawson 2019).At some point, a government will no doubt decide to rename the Department of Work and Pensions, or merge it with another department, or significantly change its structure and remit, but the inevitability of its future reform does not provide sufficient grounds for rejecting its present existence as an entity.Analysing welfare institutions as causally efficacious entities is crucial if we are to understand cause and effect in the welfare system, and thereby attribute blame, identify priorities for reform, or to acknowledge that individuals within the system have some degree of agency.Hunter (2003) seeks to make space for agency within the legacy of a restrictive poststructuralist framework, emphasizing the capacity of the welfare recipient to contribute to the formation of their own identity in a struggle with the makers and implementors of policy.Dobson's (2020) empirical analysis builds closely on these ideas, exploring how welfare practitioners constantly negotiate their (incomplete and unique) identities through institutional space.Mckee (2009, 481) goes a step further, arguing that a poststructuralist approach to social policy "disregards empirical reality, downplays the role of the state, neglects social difference, inadequately theorizes resistance, and sanitizes politics out of the policy process".She instead seeks to relocate Foucault-inspired approaches to social policy analysis within a realist ontology, using the work of Stenson (2008) to argue for an approach that is able to account for the day-to-day reality of social policy implementation as well as the macro-discursive context.While McKee's "realist governmentality" is closer to a critical realist approach than Hunter and Dobson's relational ontology, all these approaches are attempts to correct a poststructuralist tendency to underplay the agency of welfare recipients and street level bureaucrats (Lipsky 1980).Rather than rework poststructuralist theory, critical realism offers a different ontological foundation for theorizing the basic nature of social reality, one that builds-in the causal capacity of human agency.Taking a critical realist approach does not mean rejecting the rich and insightful analysis of poststructuralist research on social policy.It instead means taking a different ontological starting point, and thereby mobilizing different theoretical frameworks to address similar empirical problems.At a more practical level, much of the findings of poststructuralist and critical realist research may be mutually reinforcing and may offer similar lessons for how to improve the wellbeing and agency of welfare recipients.
It is useful to consider three advantages of taking a critical realist approach to social policy analysis by way of establishing the analytical framework used in this article.Firstly, developing critical realist approaches to social policy opens an important disciplinary dialogue with poststructuralism that can provide a broader arena to debate ontological questions and constructively challenge the theoretical assumptions of all perspectives.This advantage has already been demonstrated to some extent in the above discussion, but it is a longer-term project beyond the scope of any single article.
Secondly, and more immediately relevant, critical realist social theory addresses ontological questions by imposing analytical dualisms between concepts that are usually deliberately conflated in poststructuralist accounts.There is therefore an insistence on an explicit distinction between agency and structure, between the unique causal powers of individual creativity and intention on the one hand and the unique causal powers of their structured social context on the other.There is a similarly explicit distinction between the material and ideational aspects of social reality, between (i) the existence of underlying material conditions, including the natural world, human-made technology and the distribution of resources, and (ii) the existence of ideational and cultural phenomena, including languages, thoughts, beliefs, discourses, and ideologies.One advantage of clearly distinguishing between these basic aspects of social reality is that it becomes possible to analyse the causal relations between them (Archer 1995).As McAnulla (2006, 133) puts it "a critical realist view can support an ontology in which the world is stratified with agents, discourses and social structures constituting distinct but related layers of reality".This layered approach enables a wider lens through which to understand individual action, especially of political leaders and senior policymakers (James 2021).These concepts can function not just as an underlying theory, but also as a lens or framework with which to study policy assumptions.Therefore, in order to understand policy assumptions and identify contradictions, we can consider how any particular text constructs agential, structural, ideational, and material dimensions.While this framework could also be deployed from a poststructuralist perspective, it would stand in tension with its underlying ontology and broader aims to deconstruct analytical dualisms.
The third advantage of taking a critical realist perspective is to theorize the causal capacities of assumptions and contradictions themselves, distinct from the causal capacities of the agents that believe/use them.Again, using the words of McAnulla (2006, 128), "critical realism, in stressing the importance of a stratified ontology is able to accommodate the existence of social forms which may not be directly observable, but can nonetheless be causally significant".Although this article focuses on exposing the assumptions rather than mapping their consequences, it is an important part of the argument that those assumptions do have causal capacities, in that they provide reasons to act (Fairclough, Jessop, and Sayer 2015), but also in that they produce situational logics within which agents act.Margaret Archer (1996) argues that there is a basic structure to ideas in so far as they are related by either necessity or contingency, and related by either contradiction or complementarity.This produces what Archer calls "situational logics", where any two ideas can be necessarily contradictory, contingently contradictory, necessarily complementary, or contingently complimentary, with each situational logic exerting a different force on the agents involved.Those agents include policy makers, policy implementors, and welfare recipients, and the situational logics will have different consequences depending on how any given individual relates to the welfare system.For policy makers, this is more likely to entail exploiting complementarities or avoiding contradictions, while welfare recipients are more likely to be experiencing the material consequences of policy contradictions, such as trying to negotiate competing demands that the system places on them.
Therefore, the distinctions made in critical realism between the agential, structural, ideational, and material aspects of social reality, and the approach to understanding the causal capacity of ideas in their relations of necessary/contingent contradiction and complementarity, provides a powerful framework for the analysis of ontological assumptions within social policy.Before considering the practical methods of this approach and the findings of the analysis, it is first important to consider the existing literature on Universal Credit.

Ontological assumptions and Universal Credit
In this section, the literature on Universal Credit (UC) is considered in terms of its current engagement with ontological assumptions.This section does not therefore represent a comprehensive engagement with the vast and varied literature on Universal Credit; rather it identifies the parts of this literature that engage with the central questions of social ontology.Although there are few explicit mentions of "ontology" in the literature, existing research does implicitly engage with ontological questions.This not only further justifies the focus of this article, but it also allows for its situation among existing research on UC.Morris (2019, 287) argues that the Cameron-led governments' classification and stratification of benefit claimants according to desert is ultimately "justified by notions of dependency as a behavioural choice".The normative assumptions of desert within social policy-making can be seen to rely on more fundamental ontological assumptions relating to the capacity of the individual to determine their own behaviour (i.e. the structure-agency issue).Therefore, although Morris (2019, 287) does not refer to ontology explicitly, she does effectively offer an assessment of the Coalition's welfare ontology: "we find a welfare approach based on individual achievement and responsibility".Fletcher and Wright (2018) argue that the conditionality and surveillance that characterize the British welfare system post-2010 means that it is best understood as an "authoritarian" system.This would seem to stand in tension to Morris's view of a system based on "behavioural choice" and "individual responsibility".And in turn, both perspectives stand in contrast to the analysis of Hickman et al. (2017), who outlines a system in which the state seeks to actively and intensively create responsible benefit claimants.These three different visions of the underlying assumptions in the social security system could be respectively summarized as "the assumption that agency is supreme", "the assumption that agency needs suppressing", and "the assumption that agency needs creating".As will be seen in the following section, these differing views are not the product of academic disagreements, but are instead the product of contradictions within government policy itself.
In other parts of the literature, this concern with the underlying assumptions of UC is taken further.Millar and Bennett (2017, 176) argue that a "commitment to independence is directly contradicted by the increased control inherent in the Universal Credit design".On this basis, it could be said that they identify two ontological assumptions: firstly, there is the assumption that people are independent agents, who are free to make choices and take responsibility for those choices; secondly, there is the assumption that individuals need re-socialising so as to make them into independent agents.These two positions stand in clear tension, and the tension is comparable to the one identified above between Morris's (2019) assessment of individualism, and Fletcher and Wright's (2018) assessment of authoritarianism.
Wiggan (2012) applies Fairclough's critical discourse analysis to policy documents in order to identify the "stories" told about welfare.He finds that these policy documents "construct the persistence of poverty and unemployment as originating in the poor choices and behaviour of individuals" (400), and that "the cause of poverty and unemployment is portrayed as originating solely in a culture of out of work benefit dependency" (389), and finally that "the reader is led to believe that it is the marginal financial value of paid work that is the most salient factor in whether or not an individual enters the labour market" (392).This again reflects three different explanations of poverty that stand in tension with one another.
Whitworth (2016) explicitly argues that these tensions exist at an ontological level, identifying a tension between a neoliberal "rational" conception of the subject and a hard paternalist conception of the subject as "unable or unwilling to operate effectively within the required framework of choices and responsibilities" (416).The recognition that we are dealing with ontology allows Whitworth to specify the tensions between the two assumptions, and demonstrate how those assumptions contradict.This also links to older debates in the social security literature that link to early formations of neoliberal approaches.Page (2015, 139) and Williams (2015, 82) both suggest that the Universal Credit reforms were influenced by Lawrence Mead (1997), an American scholar who advocates a "paternalist" approach in guiding people back into work.Whitworth and Carter (2014, 105) also focus on Mead's paternalism as one of the twin pillars of Coalition welfare reform (the other being neoliberalism), explaining that paternalism "emphasises the state's role as persuading, encouraging, tutoring and if necessary compelling the poor to act as 'good citizens'".This stands in contrast to the work of authors like Charles Murray, who focused more directly on the personal responsibility welfare recipients.
This brief snapshot of the existing literature exposes at least three strands ontological assumptions associated with the Universal Credit reforms and the broader neoliberal tradition with which it is associated: the first focuses on the responsibility and therefore agency of the individual; the second focuses on the incentive frameworks within which rational actors operate; and the third focuses on the cultural context of poverty.In the following section, the analysis traces these strands in government policy documents and specifies their role in the Universal Credit reforms.

Ontological contradictions in the Universal Credit reforms
The roots of the post-2010 social security reforms can be traced back to the outputs of Iain Duncan Smith's think tank, the Centre for Social Justice, and to two landmark Conservative Party reports, Breakdown Britain (SJPG 2006) and Breakthrough Britain (SJPG 2007).Produced as part of the Conservative Party's policy review, these reports were the foundation of the party's "social justice agenda", which aimed to explain poverty as the product of five "pathways": family breakdown, educational failure, worklessness, addiction, and indebtedness (SJPG 2007).
This agenda can be traced through the sixteen texts that are considered in the current analysis.These texts were selected on the basis of an in-depth literature review on the Conservative Party and social policy during this period (Newman 2010).The analysis applied a combination of Archer's (1996) morphogenetic approach and Fairclough's (2003) critical discourse analysis, and built on an in-depth exploration of the structure-agency and material-ideational issues (Newman 2019a(Newman , 2019b(Newman , 2020)).A close analysis of each document revealed particular sentences and particular discourses that relied in some notable way on ontological assumptions.Any phrase, sentence, or paragraph that, in itself or as part of a discourse, entailed an ontological claim about the structure-agency or material-ideational questions was extracted and listed within a particular theme.Many different themes were identified, some examples being "cultures of worklessness", "the family", and "material rationalism".
With thousands of words of extracts from each policy document, a further layer of analysis was needed.Themes were reorganized so that some were abandoned, and some were grouped together under broader umbrella themes.The final major part of the analysis was to write detailed summaries of the ontological assumptions, contradictions, and complementarities that existed within (and in some cases between) different themes.This also entailed more detailed analysis of a handful of key sentences that effectively state ontological assumptions explicitly; many of these sentences are included as extracts in the analysis below.In these policy documents, three main strands of ontological assumptions were identified, each representing a position in the structure-agency and material-ideational controversies, and each apparent in the existing literature on Universal Credit and welfare ideology.
The first of the three strands can be labelled the "life-course model", a name derived from the policy documents themselves, which all state that their central strategy "follows a life-cycle structure" (DWP 2012a, 13).This accords with what Wiggan (2012) calls "a culture of out of work benefit dependency".This is especially prevalent when the texts are developing or implicating the "pathways to poverty" explanation.In the life-course model, individuals are held to be the product of the institutions of their life-course (family, community, education, etc.), within which they are conditioned and socialized by the dominant values and attitudes.Poverty is therefore a consequence of these lifecourse institutions rather than the agency of the individuals.Simply put, the lifecourse model entails a structuralist position in the structure-agency debate and an ideational position in the material-ideational debate.It is structuralist because individuals are assumed to be the product of the institutions of their life-course rather than products of their own choices, and ideational because this process of socialization is assumed to occur through the intergenerational transmission of values and attitudes.
The second strand of ontological assumptions can be labelled the "rationalist model" because there are key sections and policies that are fundamentally based on a rational understanding of the individual.This links to Whitworth's (2016) identification of a rational conception of the neoliberal subject.This is perhaps the simplest of the three models, as it assumes individuals always competently seek material gain.This approach was central to the Conservative Party's 2010 manifesto, which proposed a system of financial incentives for (amongst other things) rail franchises, the NHS, university graduates, energy policy, recycling, and water conservation (Conservative Party 2010).Poverty is therefore explained by the "perverse incentives" of the legacy benefits system (SJPG 2007, v. 2, p. 5).The "rationalist model" is structuralist because it assumes that individual actions are produced by the incentives built into their social context.The focus on the functioning of financial incentives demonstrates that the rationalist model takes a materialist position on the material-ideational issue.
The third strand can be labelled the "responsibility model", because, following the centrality of "responsibility" to Cameron's modernization of the Conservative party (Conservative Party 2006), the concept retained a significant position during the party's time in office (Atkins 2015).This is similar to Morris's (2019) identification of "a welfare approach based on individual achievement and responsibility".The notion of responsibility is not just a communicative tool, it also entails a significant commitment to a certain ontological position.By emphasizing the need for greater responsibility in any given situation, a commitment is made to an agency-centred ontology.Throughout the documents, there is a call for people to take greater responsibility, implying that both the ability and failure to exercise responsibility are within the power of human agents.When we consider the various ways that the word "responsibility" is used in the texts, it is possible to see a consideration of both material and ideational factors."Responsibility" is simultaneously used to mean, (a) individual agency, (b) the moral duty of individuals to take a particular course of action, and (c) the remit of activity for which they are legally or practically accountable.The notion of a moral duty to a course of action clearly evokes a set of beliefs about right and wrong, suggesting an important ontological place for the causal significance of ideational factors.The notion of a remit more subtly evokes the significance of materiality, in the sense that certain responsibilities derive from material situations, such as parenthood, unemployment, and addiction.
In order to demonstrate the persistence of the three ontological models and the contradictions between them, six elements of the Coalition Government's social security reforms are considered (see Table 1).The first three policies shown in Table 1 are based coherently on one of the three ontological models, but it should be noted that this "coherence" only refers to the internal coherence of the particular policy.Because each policy is based on a different ontological model, contradictions exist between them, reflecting inconsistencies in the wider reform programme.The other three policies (numbered 4-6 in Table 1) contain internal contradictions in that they depend upon a problematic combination of ontological positions.There are therefore two layers of ontological incoherence in these policy documents: there is ontological incoherence between policies, what we might call "contingent contradictions", and there is ontological incoherence within particular policies, what we can call "necessary contradictions".The remainder of this article will discuss the six policies shown in Table 1, demonstrating the ontological underpinnings of each with reference to the relevant policy documents.

Pathways to poverty and the National Citizens Service
From the start of the Conservative Party policy review in 2006, through to the 2010 general election, the notion of "the pathways to poverty" retained a central importance.Although the phrase "pathways to poverty" never became a policy itself, disappearing once the Coalition took office, its underlying life-course ontology persisted.For example, the crucial DWP (2012a) text Social Justice: Transforming Lives, was effectively structured according to the five pathways, and retained an insistence on many of the same causal mechanisms, such as "intergenerational worklessness".In the life-course model, the individual's cultural context is held to be the dominant causal force.This model therefore underpins the commitment to "prevention throughout a person's life, with carefully designed interventions to stop people falling off track and into difficult circumstances" (DWP 2012a, 1).The centrality of this model to the DWP's overall approach is perhaps best demonstrated by their prediction that "it will take many years to see the impact of our reforms work their way through, as today's children reach adulthood" (DWP 2012a, 35).
In terms of the explanation of poverty offered in the documents, the life-course model can be seen to dominate in a number of ways: it is the most commonly discussed, it underpins central statements of aims and values, and it underpins the overall structuring of the documents.However, despite this dominance in the discourse, there are very few policies that actually build from the same ontological position.The attempts to re-socialize people into the world of work (see the below sub-section on The Work Programme), and the attempts to encourage couples to form and sustain marriages, are policies influenced by the life-course model, but both are ultimately delivered according to the rationalist model, using systems of incentives as their mechanism for change.Therefore, the Conservative Party and the Conservative-led DWP developed a complex explanation of poverty based on a life-course ontology, but in all their most significant measures to tackle poverty, they reverted to rationalist or responsibility ontologies (see below).One of the few policies that actually sought to reform the institutions of the life-course by changing cultural norms was the National Citizens Service (NCS).
Established in 2011, the NCS is "a nationwide initiative to recruit teenagers for a summer of activities in the hope of producing 'socially responsible' young people" (Bulley and Sokhi-Bulley 2014, 453).According to the programme's own website, 500,000 people took part during the 2010s.Having explained poverty as the cause of persistently problematic cultures within the institutions of the family, the school, the community, and the benefit system, it seems a logical step for the government to seek to create a positive life-course institution that socializes citizens among a culture of "leadership, teamwork and communication" (Gov.uk 2019).From the ontological viewpoint of the life-course model, there is not a choice between individual freedom and government control, but there is instead only the choice between the negative and positive cultures of life-course institutions, within which the individual is socialized and ultimately produced.Although the NCS conforms closely to a life-course ontology, it is an exception; all the other main features of the social security reforms were based, in full or in part, on other ontological positions.

The Universal Credit incentives framework
By bringing multiple benefits together, UC allows an individual's benefit payments to be steadily withdrawn as they increase their earnings through work, meaning that the system is able to guarantee claimants "a minimum financial gain of 35 pence for every £1 extra take home pay" (DWP 2014, 35).This primary concern with a standardized application of incentives demonstrates an assumption that financial motivation is the key causal force behind individual decision making (DWP 2010a, 19).Overall, the UC incentive framework relies fundamentally on a rationalist model of ontological assumptions, with the attempt to "improve work incentives" held to be crucial for the government to put their "principles into practice" (DWP 2010a, 18)."[A]t its heart, Universal Credit drives behaviour change among jobseekers by helping them understand that Universal Credit rewards work and showing that employment is the best route to financial independence" (DWP 2015, 36).The amount of theoretical weight put on these central reforms to improve incentives is demonstrated in the assumption that they will change individual behaviour on a mass scale, putting "an additional 300,000 households in work, once the impact of Universal Credit is fully realized" (DWP 2014, 31).
Within discussions of incentives, it is the rationalist model that is almost exclusively implied or deployed.The responsibility model is noticeable in its absence from expositions or explorations of incentivisation.In the few instances where the word "choice" appears, it is proceeded by either "rational" (DWP 2010b, 42) or "logical" (DWP 2010a, 13).The collocations "rational choice" and "logical choice" transfer the theoretical causal power from the individual decision-maker to the choice-framework within which the individual operates.The dominance of incentive systems over the agency of individuals is further apparent in the discussion of benefit fraud: "fraud is always wrong, but we must recognize that the benefits system is making matters worse by pushing valuable work, and the aspiration that this can engender, underground" (DWP 2010a, 13).Although this quote begins with the statement that "fraud is always wrong", it is simultaneously praising the "valuable work" and the work "aspiration" that leads people to commit fraud.Fraud in the benefit system is therefore assumed to occur not because of individual failures of moral responsibility but because "working legitimately is not a rational choice for many poor people to make" (DWP 2010a, 13).This leads us to the conclusion that within the policy mechanism of incentivisation, the tension between the rationalist model and the responsibility model is overcome by a simple denial of the moral and causal agency of individuals.

The Claimant Commitment
A policy that demonstrated the presence and importance of the responsibility model in the UC reforms was the "Claimant Commitment", which was "rolled out nationwide, for all those making new claims to Jobseeker's Allowance" (DWP 2014, 8).The claimant commitment is a document signed by benefit recipients that lays out their responsibilities (DWP 2014, 24), and the signing is presented as an attempt to "encourage and promote personal responsibility" (DWP 2014, 8).If this were understood as an attempt to increase personal responsibility through re-socialization, the Claimant Commitment could represent a complementarity between the responsibility model and the life-course model.However, where a genuine complementarity would recognize those individuals who are capable of "exercising personal responsibility" and those who are not, on the basis of their life experience, the policies presented under the heading "promoting personal responsibility" suggest instead that the Claimant Commitment document is a list of expectations uniformly signed by all jobseekers.
One of the clearest statements of an undiluted responsibility model is made using the words of a Jobcentre employee: "the onus is firmly on the claimant to take responsibility for themselves" (DWP 2014, 8).Using the claimant commitment to place moral responsibility on claimants ontologically contradicts the detailed life-course explanations of institutional conditioning and socialization identified above.The life-course model places the causal-burden with the institutions in which an individual has lived, but the claimant commitment, underpinned by the responsibility model, places the causal and therefore the moral burden of responsibility firmly with the individual.

The Work Programme
Though discontinued in 2017, the Work Programme aimed to provide "support for people who are long-term unemployedor are at most risk of becoming so" (DWP 2012b, 2).It was held to be "central to the Coalition Government's ambitious programme of welfare reform" … "along with the Universal Credit benefit reforms" (DWP 2012b, 2).Two guiding principles of the Work Programme were "personalized support" (DWP 2010b, 39), and "long-term contact" (DWP 2012b, 10).The idea that each individual is the product of a unique life-course is reflected in the notion of "personalized support", and the idea that behavioural change requires a long-term change to the cultural context of the individual is reflected in the notion of "long-term contact".Furthermore, in the two policy documents that focus on the Work Programme, there are concerns about "households where no one has ever worked" (DWP 2012c, 2), which are underpinned by assumptions about intergenerational cultural transmission, and explicit claims that "long-term unemployment is damaging to individuals and communities" (DWP 2012c, 2); this in turn relies on assumptions about the reinforcement of negative attitudes towards work within communities.These various assumptions would suggest that the Work Programme was an attempt to maintain an ideational lifecourse explanation of poverty and to offer a policy solution that coherently maintained the same ontological position.
However, the actual delivery of the Work Programme relied exclusively on financial incentives, and therefore on an underlying rationalist ontology: "the programme design combines strong long-term incentives with freedom for service providers to innovate" (DWP 2012b, 2).A payment-by-results system meant that the government had no control over how the policy operated and therefore had no way of ensuring that service delivery provided solutions to the specific problems identified in the life-course explanation of poverty.The Work Programme operated on the rationalist assumption that the incentivisation of "providers" ensures that they deliver required outcomes.The policy was therefore only able to realize its ideationally motivated aim of "long-term contact" by paying providers significantly more for supporting participants over the longer term.A clear contradiction arises in relation to the crucial concept of "personalized support", because there was nothing within the Work Programme incentive framework that was intended to ensure or even encourage personalized support.If the "government is providing freedom for providers to personalize support for the individual" (DWP 2012b, 2: 8), then the government is equally providing "freedom" for Work Programme providers to provide "one-size-fits-all services" (DWP 2012b, 2: 8).Therefore, although the life-course model remained a significant part of the justification of The Work Programme, it was largely replaced by the rationalist model in the actual delivery.

Direct monthly payments
As part of the UC reforms, it was decided that benefits would be paid on a monthly basis in order to "replicate the world of work" (DWP 2014, 8).The primary aim of this policy was to "encouraging individuals to become self-reliant, financially independent and productive (in work)" (DWP 2014, 42).Similarly, UC payments also sought to "promote personal responsibility … by paying the housing element directly to the household, which is then responsible for paying the rent to their landlord" (DWP 2014, 9).The life-course justification for these policies involves delivering "life change for people, setting them on a path from dependence to independence" (DWP 2014, 3).This highlights the government's own role in "delivering life change for people" and in turn the implication that individuals do not deliver their own life change.However, this policy is ultimately underpinned by the assumptions of the responsibility model.A clear example of this can be seen when the government claims to "promote personal responsibility" but ultimately "expects the majority of tenants to manage their finances, including their own housing costs" (DWP 2015, 32).A subtle but crucial difference must be recognized here, because personal responsibility is therefore not created through this policy but is instead expected by the policy.Therefore, the policy itself does not intend to reform the attitudes of individuals or to help them develop their capacity for agency but is simply a reform that allows and expects individuals to use their pre-existing causal and moral agency in the organization of their own affairs.

Conditionality
Conditionality is a principal and policy that pre-existed the Coalition Government, but it is one that the Coalition embraced as a central feature of its social security policy.The policy documents explain that the requirements of jobseekers would no longer be three work-search activities per fortnight, but instead would entail full-time hours in order to find work.These increased compulsory expectations are accompanied by an increased use and severity of sanctions, which initially entailed a withdrawal of benefits for periods of up to 3 years.Therefore, although incentivisation is described as the central aim of UC (DWP 2010b), it is the conditionality and sanction system that clearly represents a more significant change to the lives of claimants.The underlying ontological assumptions of "conditionality" fluctuate significantly between a responsibility justification and a rationalist justification.Although the two models were both used by the government to provide a justification for increased conditionality, they ultimately contradict at an ontological level.
Explained through the rationalist model, conditionality is a system of incentives and disincentives that guides the behaviour of claimants.The reforms in this regard are justified as "a new approach to conditionality that [aims] to incentivize people to enter work and progressincreasing hours and earnings until they move off benefits altogether" (DWP 2010a, 29).Although the notion of incentivisation appears in the policy documents primarily in relation to the rate of benefit entitlement, the change in these incentives is relatively marginal for most people.Conditionality sanctions, on the other hand, entail the full withdrawal of all support for a minimum of four weeks for jobseekers, increasing further where there is a "serious failure", including "failure to apply for a job" and "failure to accept a job offer" (DWP 2010b, 30).
However, the policy documents only occasionally present the conditionality regime in rationalist terms.Conditionality and sanctions are much more commonly justified in terms of the responsibility model, where benefits are considered "rewards" and sanctions are considered "punishments".Thus, the DWP argued that "those who fail to meet their responsibilities should face a sanction such as a benefit reduction" (DWP 2010a, 28).These rewards and punishments were justified on moral grounds, relating to the moral responsibilities of the individual, so that "changes to the existing conditionality and sanctions regime" aimed to "strengthen the link between people receiving benefits and meeting their responsibilities" (DWP 2010b, 26).Crucially, the primary causal force assumed here is individual agency; i.e. the moral power of an individual to overcome the temptation to follow the easiest or most rewarding course of action, and their causal power to enact that moral responsibility.Therefore, the contradiction between the rationalist justification for conditionality and the responsibility justification for conditionality relates to (1) whether the burden of responsibility is on the individual decision maker or the framework of incentives, and (2) whether benefit payments and sanctions therefore operate as a system of rewards and punishments for moral failures and achievements OR as a system of incentives and deterrents that have the causal power to drive individual behaviour in a particular direction.

Conclusion
This article has sought to build on the existing literature on the underlying assumptions of universal credit, and the wider literature that emphasizes the importance of ontology in social security studies.In contrast to the tendency towards poststructuralist analysis within this literature, the article has taken a critical realist stance and deployed a critical realist framework of analysis.This helps to widen the theoretical debate within the literature, but more practically it enables divergence from existing studies in two senses.The first is to endorse analytical dualisms on key ontological issues, so that structural, agential, material, and ideational aspects of social reality are considered to be distinguishable in order to analyse the interrelations between them.This has provided a useful framework for analysis, but has also allowed the article to focus on the ideational aspect (in the form of ontological assumptions) and conclude by theorizing its causal capacities.This is the second unique contribution of a critical realist analysis, to theorize contradicting policy assumptions as creating "situational logics" that exert a causal force on the agents that act within the welfare arena.While this article has not traced through this causal force to study its actual effect on agents, it has specified the nature of the situational logics produced by ontological contradicitons in the Universal Credit reforms.Future research could explore how these situational logics affect agents, using Archer's morphogenetic approach (Archer 1996;Newman 2020) and realist theories of the policy cycle (e.g.Croci, Laycock, and Chainey 2022).
Through an in-depth document analysis, three main ontological models were identified, each taking a different position in the "structure-agency" and "material-ideational" controversies: (1) the life-course model, which is structuralist and idealist, in that it assumes individuals are the product of cultural socialization within the institutions of the lifecourse (especially the family); (2) the rationalist model, which is structuralist and materialist in that it assumes that individual decisions can be explained in terms of their contextual framework of financial incentives; (3) the responsibility model, which is primarily agentialist and recognizes both material and ideational factorsthis model assumes that regardless of life-experiences and contextual incentives, individuals have a unique causal power over their lives and actions, which attaches to them a moral responsibility.
These models have all been shown to be present in the UK's social security reforms, underpinning key policies in various ways.The "National Citizens Service", the "UC incentives framework", and the "Claimant Commitment" were shown to have internally coherent ontological assumptions, associated respectively with the life-course, rationalist, and responsibility models.However, these policies contradict one another, with each informed by different ontological positions.They therefore create a situational logic of "contingent contradiction".In contrast the "Work Programme", "direct monthly payments", and "conditionality" were shown to contain internal ontological contradictions, meaning that they produce a situational logic of "necessary contradiction".The causal consequences cannot be "read off" from these situational logics, because outcomes depend on the agency of policy makers, policy implementors, and welfare recipients.However, it can be said that policies containing necessary contradictions will be harder to sustain and potentially less effective, while contingent contradictions will only have consequences when and where those contradicting policies interact with one another.
Therefore, although this article has not considered the wider implications of ontological incoherence, this is an avenue for future research.It is expected that a policy reform programme will have problematic consequences if it is based on an incoherent understanding of the relationship between structure and agency, and the ideational and material.One notable implication of the research presented in this article is the fundamental injustice of a government holding benefit claimants morally and causally accountable for their actions when key parts of that same government's policy programme fundamentally relied on the assumption that the actions of benefit claimants are determined by contextual factors.The logical demonstration of this injustice in this article is one of many reasons why social policy analysis must continue to consider the ontological assumptions underpinning social security policies.

Table 1 .
The six policies analysed in this section.