Automating activation in Australia: a critical policy discourse analysis of the new employment services model

ABSTRACT

This article critically analyses the policy discourse that accompanied the introduction of the ‘New Employment Services Model’ (NESM) into the Australian social security system. Critical policy discourse analysis of documents presenting reforms to employment services provision shows that the NESM embeds a logic of activation that construes job seekers both as passive units requiring external motivation (through ‘employment support’ and ‘mutual obligation’), and as active agents responsible for their own situation. This contradictory logic results in inconsistencies between the espoused aims of the reforms and the experience of those subject to the policy mechanisms. Analysis of departmental outcomes data illustrates the inconsistent operational consequences of the conflation of ‘targeted compliance’ and ‘support.’ Due to the technological automation of activation processes, the real discursive innovation of the NESM is one of ‘digital depoliticization,’ whereby the moralizing assumptions that inform punitive activation are embedded and therefore concealed in the policy technology. The operational effect is to subordinate the achievement of employment to compliance with mutual obligation requirements as a structured aim of employment services provision.

KEYWORDS:

• Activation
• social security
• workfare
• employment services

Introduction

Over four decades, the Australian social security system has incorporated strategies of surveillance and punishment for people deemed only conditionally deserving of state support (Carney 2020a). These strategies have been institutionalized through ‘mutual obligation’ requirements and recently intensified by technological development (Casey 2022b, 2022a; Staines et al. 2021). Developments include the implementation of the Targeted Compliance Framework (TCF) to monitor job seeker participation in mutual obligation activities, and the New Employment Services Model (NESM), a predominantly online platform to support job seekers to engage with employment services. These two policy mechanisms are connected, with the Targeted Compliance Framework now embedded within the New Employment Services Model. Except for Casey (2022b, 2022a), O’Sullivan, McGann, and Considine (2021) and Considine et al. (2022), few studies have explored these developments in depth, and yet, taken together, they constitute a new welfare regime. Considine et al. coined the term ‘machine bureaucracy’ to describe ‘a distinct “variety of digitalization” that comprises an attempt to not just re-orientate the discretion previously held by frontline workers but to automate it.’ (2022, 531).

This article complements Casey’s analyses of the operational implications of the shift to digital services and further elaborates digitalization ‘trade-offs’ developed in Considine et al. (2022). We adapt the critical policy discourse analysis approach developed by Mulderrig, Montesano Montessori, and Farrelly (2019) and Hyatt (2013) to analyze the discursive processes by which policy changes were rationalized, and by which normative assumptions about employment and mutual obligation were embedded in the policies in operation. Two research questions guided this analysis; firstly, a policy question of what aims are furthered by the introduction of the NESM, that is, what innovation does the NESM represent, and secondly, a critical policy analysis question of how the policy in operation affects the job seeker experience.

The materials analyzed are policy documents that introduced the TCF and NESM, providing detail of policy aims and operation, and departmental reporting on outcomes. The first part of the article introduces the policy changes and situates them in the context of prior research on welfare conditionality and activation. The second part outlines the critical policy discourse analysis approach and selection of texts for analysis. The critical policy discourse analysis then highlights contradictions between activation policy aims, as expressed in the language used in the documents, and the experiences of jobseekers as reported in TCF outcomes. In doing so, the analysis demonstrates how the activation discourse provides authority to the TCF and NESM, which are constructed as objective, apolitical and mundane technologies, yet which incorporate normative assumptions of job seekers and their need for activation. The effect of these changes adds a digital dimension to the strategies of discursive depoliticization described by Wood and Flinders (2014), and by Busso (2017) in relation to social policy. Depoliticization displaces the issue of concern from the political sphere to a different ‘arena,’ for example from the sphere of political debate to the arena of individual choice. Wood and Flinders note the implication is ‘not so much that the issue of concern has become any less political, in the sense of its impact on individuals or society, but that it has been transferred to a less obviously politicized arena’ (Wood and Flinders 2014, 155), in the following case, to the digital enclosure of the new employment services model.

Background: the targeted compliance framework and new employment services model

In Australia, the Social Security Act 1991, and the Social Security (Administration) Act 1999 and related guidelines generate assessment and classification terminology that determine eligibility for income support payments and access to employment services. The discourse of eligibility has been central to recent policy change. The 2017–18 Federal Budget introduced a ‘targeted compliance framework’ (TCF) as a central policy technology for ‘A New Era of Welfare,’ that would ‘strengthen mutual obligation requirements and provide more incentives to help people move from welfare to work’ (Cash, Tehan, and Keenan 2018). This was complemented by expansion of private employment services’ administrative power in assessments of eligibility, especially in relation to identification of ‘non-compliant’ behavior (Commonwealth of Australia 2019).

The 2021–22 Federal Budget framed the NESM as a new ‘digitally driven’ approach to employment services provision (Commonwealth of Australia 2021c), and budget allocations committed the Government to ‘strengthening mutual obligations to provide better support for job seekers in their search for work.’ (Commonwealth of Australia 2021b, 17). The Social Security Legislation Amendment (Streamlined Participation Requirements and Other Measures) Act 2022 (‘Streamlined participation’) then reoriented provisions of the Social Security Act to ‘support the New Employment Services Model by streamlining and modernizing social security law’ (Commonwealth of Australia 2021a).

A dramatic increase in demand on the social security system precipitated by the pandemic moved elements of online services from pilot phase to dominant mode of service delivery prior to their originally scheduled start date of mid-2022 (ANAO 2020a). For those classified as both ‘job ready’ and ‘digitally literate,’ ‘digital self-servicing’ (O’Dwyer 2019) became their primary mode of engagement with both employment services and mutual obligation requirements. Engagement is specified in a ‘job plan’ and takes the form of recording mutual obligation activities and responding to system generated communications. Differential support is offered according to assessments of job seeker job-readiness and ‘digital literacy,’ with face-to-face employment services reserved for job seekers with additional needs (DJSB 2018a).

The assemblage of digital and face-to-face private employment services provision was launched in July 2022 under the official heading ‘Workforce Australia,’ immediately receiving adverse media coverage that drew attention to individual job seekers being required to leave paid employment to attend to compliance activities, and their referral by employment service providers to employment support programs run by the referring provider (Henriques-Gomes 2022c, 2022a, 2022b, 2022d; Young and Hermant 2022). The incoming Federal government subsequently established a House Select Committee on Workforce Australia Employment Services due to inquire into and report on ‘the implementation, performance and appropriateness of Workforce Australia Employment Services’ by the end of September 2023 (Commonwealth of Australia 2022a).

Workforce Australia represents the automation and integration of social security with ‘work first’ approaches to social protection. Rationalization emphasized efficiencies to be generated by digital systems, reflecting the international ‘digital turn’ in the governance of welfare (Alston, 2019; Carney 2020b). Alston (2019) warned that in the development of digital welfare states, a preoccupation with use of digital systems for surveillance and fraud prevention would subvert the potential for such systems to realize the right to social protection. The Australian approach is illustrative, combining a seven-year ‘back end’ project to redevelop the information technology systems used to administer social security payments (ANAO 2020b), with a ‘front end’ discursive reframing of the objectives of the social security system that foregrounded surveillance and fraud protection.

Considine et al. (2022, 520) used the phrase ‘machine bureaucracies’ to describe the qualitatively different form of digitalized welfare represented by the New Employment Services Model. Previously, digital applications were used to ‘augment the jobseeker–advisor relationship’ whereas the NESM proposed ‘to remove the human element of service provision for all claimants other than those identified as having the most complex employment barriers.’ Carney’s (2020a, b) analyses of automation highlighted the mixed experience of digitalization of the Australian welfare system, including ‘… poor or rushed Al designs, poor understanding of client characteristics and personal vulnerabilities, and inadequate understanding of the dynamics within contracted-out social security services settings’ (2020a, 24). In relation to the automated compliance process that is currently the subject of a Royal Commission of Inquiry referred to as ‘Robodebt’ (Commonwealth of Australia 2023), Carney commented that while ‘gravely bungled’ implementation does not constitute an argument for retention of human administered systems, ‘[p]oorly designed Al systems however put at risk values of procedural fairness, the rule of law, and government accountability’ (Carney 2020a, 50).

Welfare, mutual obligation and activation

Australia’s social welfare system was premised on the equation of work and wellbeing, as captured in the phrase ‘wage earner’s welfare state.’ The origins of this phrase lay in Castles’ international comparative studies of different systems, and his identification of unique characteristics of an Australian ‘social contract’. That contract had the features of non-discretionary benefits, bolstered by industrial relations regulations that sought to ensure the connection between income and livelihood (Castles 2001, 539).

Subsequent policy shifts in the late 1990s, including changes in the industrial relations system, erosion of universal health care provisions, and the introduction of ‘mutual obligation’ requirements for beneficiaries of state support caused Castles to revise his view and sketch the outline of a new social contract underpinning the Australian settlement. This new social contract reinstated and sharpened a distinction between ‘deserving’ and ‘undeserving,’ requiring applicants to demonstrate their eligibility for increasingly conditional support. Examples of programs operationalizing this rationality include ‘work for the dole’, income management programs, such as the cashless debit card, drug testing of welfare recipients, compulsory training, and intensification of job search requirements (McKeever and Walsh 2020; Peterie et al. 2019; Mendes 2020a, b). Benefits were reframed as a ‘wage-like payment’ to be earned through completion of mutual obligation requirements (McKeever and Walsh 2020). The reconceptualization of welfare as conditional supported the imposition of mandatory requirements, monitoring and sanctions, enacted on citizens by government to change behaviors understood as ‘causing’ unemployment (Parsell et al. 2020; Peterie et al. 2019; Carter and Whitworth 2017).

Employment ‘activation’ as a form of welfare conditionality

For workfare states, employment in paid work is understood as a central obligation of citizenship, with the state and network of employment services providers positioned as enforcers of that obligation (Bengtsson 2014; Mestan 2014; Whitworth 2016; Yeatman 2000; Mendes 2020b). This leads to the development of activation programs which, in theory, are designed to facilitate peoples’ transition from welfare to paid employment (Carter and Whitworth 2017; Mestan 2014). In practice, incentives, particularly the threat of financial sanctions, compel people to engage in ‘work-like’ practices (Peterie et al. 2019). According to Bengtsson, activation programs reflect the post-Keynesian neoliberalisation of economies, and the shift in focus from full employment to ‘full employability’ (Bengtsson 2014, 54). Raffass (2017, 362) argued that consequently, employability, rather than actual employment, becomes the fundamental goal of activation:

The internationally codified human right to work receives no acknowledgement within the paradigm of the activating state… Since employment is not guaranteed as of right and, statistically, there are not enough jobs to be held by every person wanting a job, what is being enforced is not even the duty to work, but merely a duty to activate.

As Bengtsson (2014, 66) argued, and similarly, McDonald, Marston, and Buckley (2003, 500–501) and Peterie et al. (2019, 802) in the Australian context, the focus of conditionality on policing workfare practices, shifts the emphasis from job seeking toward enforcing standardized job seeking activities, regardless of their value to the jobseeker. Activation programs combine capability development through training programs with threats of penalties for noncompliance affecting eligibility for entitlements (Bengtsson 2014). Jobseekers are understood as either lacking required skills, or ‘work-shy’ (Peterie et al. 2019; Mendes 2020b). The ‘activation turn’ in workfare economies was underpinned by the normative assumption that all adults should be active participants in labor markets, or at least ‘job ready,’ that is, capable of quickly transitioning into labor markets should an opportunity arise (OECD 2012; Bengtsson 2014; O’Sullivan, McGann, and Considine 2021). Over decades this assumption has supported the development of a specialist industry of employment support services, and a corresponding professionalization of job seeking (Sharone 2007), in the Australian context through the development of a ‘quasi-market’ in employment services provision (O’Sullivan, McGann, and Considine 2021).

A substantial body of international literature has appraised different approaches to active participation. These consist of systematic meta-analysis of evaluations (Vooren et al. 2019) and critique of the coercive elements of activation programs (Peterie et al. 2019; Whitworth and Carter 2018; Carter and Whitworth 2017; Raffass 2016, 2017). Vooren et al. (2019) developed an econometric meta-analysis of active labor market program evaluations to appraise their relative effectiveness in terms of the likelihood of employment. Their overall finding was that ‘the average effects … are relatively small’ (139) and that ‘such small average effects highlight the need of thorough cost–benefit analyses’ (140), pointing to the significant investment such programs require.

The efficacy of both employment services and mutual obligation processes are questionable (Commonwealth of Australia 2019; O’Halloran, Farnworth, and Thomacos 2020), especially in relation to people considered ‘disadvantaged’ in relation to the labor market (O’Sullivan, McGann, and Considine 2021, 215). Mutual obligation requirements in Australia have been found to provide few tangible benefits for job seekers in terms of gaining meaningful, lasting employment (Commonwealth of Australia 2019; Gerards and Welters 2021). Casey (2022a, 2) equated ‘work first’ outcome incentives with Raffass’ concept of punitive activation: ‘the threat of sanctions for non-fulfillment of (often punishingly) extensive job-search requirements or refusing job offers, mandatory unpaid work and the psycho-compulsion involved in mandatory skills building and motivational training’ (Raffass 2017, 350). Peterie et al. (2019, 796) argued that consequently, jobseekers experience stigmatization and blame through their participation in employment services and internalize ‘dominant discourses that frame unemployment as a personal failure’.

The fusion of active participation into ‘activation’ effectively reframes agency, with organizations and processes credited with the power to activate job seekers, and job seekers held to account for the limited agency they are granted in their efforts to maintain a level of social security, regardless of their structural likelihood of employment. Busso noted, “active policies can limit individuals’ autonomy when they are included in a framework of conditionality that requires that users comply with a contract in order to access any other benefit.” (Busso 2017, 436). A further systemic implication reinforced by Demazière is that ‘everything serves to keep the problem of unemployment and its solutions at a strictly individual level’ (Demazière 2021, 195), maintaining the premise that individual compliance with the ‘duty to activate’ will result in employment, regardless of evidence to the contrary and broader labor market conditions. Busso (2017) identified individualized intervention as one of three strategies used in the depoliticization of social policy. The other two are ‘the shift of emphasis from justice to effectiveness’ (422) and development of measures of policy effectiveness that foreclose on discussion (political debate) of policy goals, and the absorption of civil society into governance structures, muting potential dissent. All three strategies are evident in the discursive rationalization of the new employment services model.

Theoretical approach, materials, and methods

This study is a critical policy discourse analysis (CPDA) of documents and speeches introducing the Targeted Compliance Framework and New Employment Services Model. Critical policy discourse analysis recognizes the role policy discourses play in the constitution of problems and responses. Analysis proceeds from the premise that social processes and discourses are mutually constitutive, and that social change is both accompanied and precipitated by discursive framing. Mulderrig, Montesano Montessori, and Farrelly (2019, 6) noted the language of policy ‘plays a role in the ongoing negotiation and co-constitution of subjectivities and practices of the wider socio-economic order, and for this reason warrants critical scrutiny.’ This perspective informs ‘explanatory critique’ that illuminates the mechanisms by which policy discourses maintain and enhance relations of power (Smith-Carrier and Lawlor 2017). For Hyatt, analysis of language reveals naturalized power relations whereby ‘ … language acts as a social control agent, through which members of society are conditioned to accept conventions and practices that may not be in their best interests. These language practices are represented as “common sense”, inevitable and beyond challenge’ (Hyatt 2013, 839–840). Mulderrig, Montesano Montessori, and Farrelly (2019, 8), described CPDA as:

…a tool in strategies of resistance. It provides a theoretically informed language of explanatory critique, capable of showing not only how, but why, the language and logic of neoliberalism comes to dominate and colonise even those voices which ostensibly stand in opposition to it.

Prior research on social welfare policy has identified and analyzed how discursive constructions of welfare subject positions are used to reinforce and legitimate policy settings (Marston 2006; McDonald, Marston, and Buckley 2003), or to effect change by reframing public understandings of the causes of social problems (Marston 2006, 2008; Farrelly, Mulderrig, and Montessori 2019). Discursive constructions of welfare subjects are also used to enforce social and moral distinctions influencing entitlement to state support and to maintain hegemonic constructions of ‘the reality’ of unemployment and employment (Marston 2008; Whitworth 2016). More recently, Staines and Smith explored the new discursive potential of ‘big data’ to justify ‘the increased rationalization and gradual withdrawal of state support, both in the form of social security payments and state-funded social services’ (Staines and Smith 2022, 15).

CPDA Framework for analyzing policy documents

Hyatt proposed CPDA to study policy aims and impacts, through analysis of ‘policy drivers,’ defined as ‘expressions of intended aims or goals of a policy,’ ‘levers,’ understood as mechanisms, which shape or direct change, such as the use of performance targets or audit, and ‘warrant’ to describe how policy actions are justified and given authority (Hyatt 2013, 838). Hyatt described three modes of legitimation, understood as ‘the process by which policies are justified to their audience by attachment to dominant norms and values’ (840):

• Authorisation – reference to authority as unimpeachable, beyond challenge;

• Rationalisation – claimed value and usefulness of a social action;

• Moral evaluation – connection with values systems which are claimed to be desirable.

Hyatt emphasized lexico-grammatical construction as an important element of discourse analysis in the study of policy, especially in relation to questions of power and responsibility. A dominant strategy is the use of the passive voice to intentionally remove the agent from the process, thus reducing clarity around who is taking the actions in question. For a study of policy frameworks and technologies, this approach has added resonance as it can help convey where responsibility and power lie in the implementation of the policy.

In this article, CPDA identifies contradictory logics informing discourses of employment and social support in political announcements, legislation, and operational documents introducing and rationalizing the TCF and the NESM. Elements of Hyatt’s CPDA framework informs the analysis of 29 documents describing the aims, implementation and outcomes of the TCF and NESM (see Table 1) over a five-year period, commencing with announcements inaugurating the ‘new era of welfare,’ and culminating in the introduction of Streamlined Participation legislation (2017–2021). The documents were identified through keyword web searches (‘Targeted Compliance Framework’ and ‘New Employment Services Model’) of the Federal Parliamentary website (for political statements) and Departmental websites (for advice to claimants). The analysis of documents highlights the discursive processes of rationalization and legitimation both in political discourse and in the operating procedures of the NESM.

Table 1. Sources for analysis.

Source	Document
Department of Education, Skills and Employment	DESE (2018, 2019, 2020d, 2020f, 2020e, 2020a, 2020b, 2020c, 2021b, 2021a, 2021d, 2021c, 2021f, 2021g, 2021e)
Department of Jobs and Small Business	DJSB (2018b, 2018a, 2019)
Department of Social Services	DSS (2020)
Commonwealth of Australia	Commonwealth of Australia (2019, 2021a, 2021e, 2021b, 2021c)
Australian National Audit Office	ANAO (2020a, 2020b)
Media Releases, Speeches issued by Government Ministers	Cash, Tehan, and Keenan (2018), Porter (2016), Morrison (2015)

The analysis concentrates on legitimation strategies, political warrant and identification of drivers and levers. Analysis of key announcements and related information explaining the NESM and amendments to the Social Security Act focus on explanatory critique of the assumptions about employment that are supposed to give activation logic its force, guided by the questions: What aims does the NESM further? What is the innovation that it introduces?

Results

The targeted compliance framework: discursive analysis of aims and operation

The introduction of the TCF is significant not as a shift in overall policy direction but for its intensification of existing settings and broader reframing of ‘social protection.’ The Social Services Legislation Amendment (Welfare Reform) Bill 2017 merged five allowances and two pensions into the ‘Jobseeker payment’ or ‘age pension’ depending on individual circumstance and age. This nominal change in the terminology – from (making) allowances, to ‘participation payments,’ - normalized the idea that income security payments are received for participation in mutual obligation activities, that is, they are conditional and contingent on participation rather than as a citizen right. This shift was reinforced by political statements, for example, ‘at the end of the day the best form of welfare is a job and creating jobs and getting people in a position to take one is even more important than system reforms itself’ (Morrison 2015).

Design of the TCF is informed by concern with surveillance and fraud protection. In a guideline for employment service providers, DESE (2018) stated the aim of the TCF is:

to ensure only those job seekers who are persistently and wilfully non-compliant incur financial penalties while providing protections for the most vulnerable. It is designed to encourage job seekers to engage with their employment services provider … take personal responsibility for managing and meeting their Mutual Obligation Requirements, and actively look for work.

This statement outlines the key driver of the TCF; to identify job seekers who appear to have obtained welfare payments without meeting their obligations. A key lever, ‘to encourage job seekers to engage with their employment services provider,’ is articulated through threat of financial penalty. This statement exemplifies the TCF focus on ensuring job seeker compliance with mutual obligation requirements through engagement with employment agencies, rather than on supporting people to find work. Both policy drivers and levers assume job seekers in Australia are disinterested in finding work and actively avoid their obligations. The universal scope of the TCF is justified on the basis that it will both address the behavioral transgressions of a few and ensure that compliant, responsible job seekers are protected. This is associated with the political warrant constructed in these statements, which justify the TCF as an essential measure to address fraud, but also to protect the concept of citizenship as entailing work, responsibility and duty.

This framing is also evident in Ministerial communications, which presented the TCF via analogy with a driver’s license:

A new compliance framework based on a demerit points system – similar to how a driver’s licence operates – will target welfare recipients who are not serious about finding work, while providing greater protections for vulnerable job seekers.

(Cash, Tehan, and Keenan 2018)

The emphasis on personal responsibility reinforces the construct of job seekers as ‘not serious,’ requiring behavioral management through the TCF.

Advice to job seekers is direct: ‘You need to know how to meet your mutual obligation requirements to avoid your payment being put on hold or financial penalties applied’ (DESE 2020d). The passive expression is opaque, indicating that penalties will be applied, though not specifying who makes this determination. To some degree, this lack of clarity underpins the authorization of the TCF itself, as the framework, with its detailed rules and processes, evaluates job seeker performance alongside expectations and sanctions applied digitally.

Explanations draw on discursive strategies associated more with game rules than penalties for noncompliance with road rules. The following advice to claimants illustrates the framing in terms of ‘zones’ and ‘demerits’:

In the Warning Zone, if you get 3 demerits you will have a Capability Interview with your Provider or, for online job seekers, with the Digital Services Contact Centre. If you get 5 demerits you will have a Capability Assessment with Centrelink … If Centrelink decides at a Capability Assessment you can meet your requirements, you will move to the Penalty Zone where you may lose money if you do not meet your requirements.

(DESE 2020d, emphasis original)

In this procedural account job seekers are managed through threat of payment cancellation, while the concept of the ‘penalty zone’ itself is a discursive construct implying that a person who is subject to penalty is guilty of behavioral transgressions. As in this analogy, a person who has accumulated the requisite number of demerits will potentially lose the freedoms afforded them through their payment.

The official discourse distinguishes between a minority of job seekers who are not taking responsibility for ‘managing and meeting’ their requirements, and job seekers genuinely seeking work (DJSB 2018b, 3). Penalties incorporated into the TCF are justified by repeated reference to willfully non-compliant job seekers (DESE 2020f, 2020e; DJSB 2019, 2018b). Conversely, the TCF is also claimed to protect ‘the majority of job seekers who do the right thing’ (DESE 2021a; DJSB 2019, 14), who are seen to be ‘genuine’ in their commitment, and ‘vulnerable’ job seekers (DESE 2021a, 2020f, 9, 2020e, 1; DJSB 2018b, 2019):

The TCF is fair, working as intended, deterring deliberate and wilful non-compliance while providing protection for vulnerable job seekers/participants. It supports those who are genuinely trying to meet their obligations and only results in financial penalties for the small number of job seekers/participants who persistently and deliberately do not meet their Mutual Obligation Requirements.

(DESE 2020f, 2)

Questions around genuineness of job searches and job applications are repeatedly raised using phrases such as ‘willful noncompliance’ intimating a deliberate practice associated with deception, which further justifies policing of all job seekers. While it is acknowledged that not all job seekers engage in this practice, all job seekers are nonetheless viewed with suspicion.

Naturalization of job seeker ‘failures’

The TCF penalizes job seekers for a range of what is described as ‘failures.’ As the DESE (2020e, 6–7) explained, incurring a suspension and ‘Moving into the Warning Zone helps to reinforce the importance of meeting Mutual Obligation Requirements and discourages Participants from committing any further Mutual Obligation Failures.’ In addition to ‘Mutual Obligation Failure’ (always capitalized), the TCF formalized a nomenclature of additional failures job seekers can ‘commit,’ replacing prior penalty-attracting failures (of failure to attend an interview or connect/reconnect with an employment service), with ‘Unemployment Failure,’ and ‘Work Refusal Failure,’ each of which attract combinations of demerits, movement between zones, penalties and ‘post-cancellation nonpayment periods,’ if the jobseeker is not able to offer a ‘Reasonable Excuse.’ To illustrate, five ‘Mutual Obligation Failures’ within six months results in a person entering the ‘Penalty Zone,’ wherein further failures ‘will result in person being determined to have “persistently committed mutual obligation failures” and will result in them facing escalating financial penalties’ (Commonwealth of Australia 2022b), culminating in payment cancellation and nonpayment periods of up to four weeks. Payment cancellation and the four-week preclusion is also incurred by a person in any zone, who commits a ‘Work Refusal’ or ‘Unemployment Failure’ (DESE 2020f, 9). The technical explanation of the TCF framework concurrently develops a new language of job seeker ‘failure,’ which job seekers demonstrate by insufficient participation in mutual obligation processes. This language naturalizes the assumption of job seekers’ unwillingness to engage with employment agencies, while rationalizing punitive strategies.

Of all the potential misdemeanors, ‘work refusal failures’ incur greatest penalties. A work refusal failure occurs when a job seeker rejects a potential employment opportunity, regardless of the type of work and whether this fits with the job seeker’s aspirations. The construction of job seekers as ‘work shy’ fundamentally misunderstands this population. Demographic data on payment recipients for March 2022 (for example) showed that 19% of 855,870 jobseeker recipients had earned over AU$250 from employment in the prior fortnight (DSS 2022). It should be noted that earnings exceeding that amount trigger dramatic reductions in income security payments: for every dollar over AU$150/fortnight, payment reduces by ‘50 cents for each dollar between $150 and $256 then 60 cents for each dollar over $256’ (Services Australia 2022a). The income test for eligibility (for a single person with no dependants, for example) provides for earnings of up to $626.75/week ($32,591/annum) before the payment reduces to zero (Services Australia 2022b). The coexistence of income security payments and income earned through employment demonstrates the reality that unemployment/employment is not an either/or state for many claimants. The system logic however is founded on a construction of claimant employment status as exclusively unemployed and seeking work, not as partially employed, or available only for work that is not in conflict with family commitments (Yeatman 2000).

How does the TCF in operation affect the job seeker experience?

As Hyatt (2013) noted, CPDA also seeks to understand policy impacts. This section draws on data which demonstrates the impact of the TCF in operation in relation to the vulnerable groups that it claims to protect, to answer the second research question of how the policy in operation affects the job seeker experience. The following tables present disaggregated compliance data provided to the Senate Education and Employment Committee estimates hearings (DESE 2020e, 2020a).

As indicated in Table 2, on 30 June and 31 December 2019, over 200,000 job seekers were classified as being either in the ‘warning zone’ or ‘penalty zone’, meaning that they had registered at least one active demerit for a breach of their mutual obligation requirements (demerits remain active for six months) and had received a payment suspension (DESE 2020e, 2020a, 2020b). Further, Table 3 suggests that job seekers who are part of ‘vulnerable populations’ are not protected through this system, with specific groups either over-represented in the warning or penalty zones or experiencing payment cancellations.

Table 2. Number of TCF job seekers across all programs, by TCF zone, on 30 June and 31 December 2019 (DESE 2020f, 2019).

Date	Total TCF job seekers (all zones)	Green zone	Warning zone	Penalty zone
31December 2019	583,429	346,445 (59%)	228,391 (39%)	8,593 (1.5%)
30 June 2019	579,189	352,310 (61%)	219,806 (38%)	7,073 (1.2%)

Table 3. Number of Jobactive participants canceled off payment, per population, per TCF zone (DESE 2020a, 2020b).

Demographic	Green zone	Warning zone	Penalty zone	Total number job seekers canceled off payment	Total participants in jobactive program per demographic, at 31 August 2019
Indigenous	4,377	7,147	325	11,849	52,335
Single Parents	2,894	2,941	46	5,881	48,907
Disability	5,936	7,576	208	13,720	89,753
Homelessness	3,897	6,203	253	10,353	45,301
CALD	5,130	8,914	204	14,248	68,679
Ex-offenders	5,572	9,232	413	15,217	55,018
Total	38,179	64,829	1,472	104,480	388,018

While the TCF is rationalized on the basis that it will only punish those who are ‘willfully non-compliant,’ this data shows that large numbers of job seekers have been affected through suspension or cancellation, particularly those who are Indigenous, live with a disability and/or experience homelessness. The data reinforce the disproportionate consequences of the TCF in operation.

The new employment services model

We now analyze the development of the New Employment Services Model, in which the Targeted Compliance Framework has been embedded. The Commonwealth announced its new approach to employment services as responding to the need to provide personalized assistance relevant to a changing employment landscape. The Employment Services Expert Advisory Panel recommended a new ‘digital’ approach, equating ‘digital’ with ‘personal powerful, and efficient’ support (DJSB 2018a):

The future system will cost no more than it does today. Funding will be invested in smarter ways … It is smarter to use data and analytics to inform what we do. It is smarter to automate compliance and monitor administration.

(DJSB 2018a, 4)

Equating digitization and automation with smartness is a significant discursive framing, aiming not only to authorize to digital processes, but also to situate authority in the automation of compliance and monitoring administration. Power is displaced to the digital systems and processes underpinning the TCF and NESM, which are constructed as disinterested tools of measurement and audit, rather than being underpinned by ideological constructs of welfare and job seekers. Claims of smartness in terms of resource efficiency justify this approach on economic grounds.

Online job seeker responsibilities

Claimants are required to engage with the automated system to establish their eligibility for support and to agree to become a ‘jobseeker.’ Significant numbers were projected to make use of digital services, ‘a core service that includes online tools and jobs matching’ (DESE 2021d); however, evidence of their efficacy is mixed (DESE 2021f, 2021e, 67–77). Making a claim requires claimants’ agreement in advance that they understand and accept conditions on receipt of payments, for example:

I agree to search for work by contacting eight employers per month. I agree to report and provide evidence of these jobsearch contacts … I am aware that I need to look for any suitable work, not just work that I would like to do.

(Commonwealth of Australia 2021e, 7)

This contractual frame offers minimal choice to the claimant, drawing on the framing of work as a citizen responsibility and continuing the construct of ‘work refusal’ as an unacceptable behavior. In place of the opportunity to engage in work, work-related practices, through mutual obligation activities, fulfil that responsibility.

Specific obligations are agreed to by claimants in the application process, documented in an employment pathway plan (described in communication to job seekers as a ‘job plan’), which must be agreed to before income support payments commence (Commonwealth of Australia 2021a; DESE 2020c). To illustrate, under ‘agree to job plan’ there is a lengthy restatement of agreement to seven clauses, starting with: ‘I declare that I understand the mutual obligation requirements in my Job Plan, and that I know I must meet them to get my income support payment’ (Commonwealth of Australia 2021e, 8).

In other words, claimant agency is limited to the choice to not agree to a job plan and to receive no income support. This implies that choice and freedom are earned through participation in paid work. Through their ‘failure’ to do so, job seekers are constructed as lesser citizens who have not earned this freedom. Thus, the NESM rationalizes curbing job seeker choice regarding employment opportunities.

At the same time, the second reading of the Social Security Legislation Amendment (Streamlined Participation Requirements and Other Measures) Bill 2021 framed automation as enhancing choice: ‘This bill will allow jobseekers to choose their own job plan requirements, within departmental guidelines, and to manage their job plans online’ (Commonwealth of Australia 2021a).

Choice in ‘job plan requirements’ refers to the ‘points-based activation system’ introduced to give effect to the expert advisory panel’s recommendation for more flexible mutual obligation requirements (DJSB 2018a). Points-based activation allocates points to activities using ‘work first’ weightings, e.g. job application = 5 points, job interview = 10 points, volunteering = 5 points/10 hours (DESE 2021g). Compliance is demonstrated through the achievement of the required number of points, however claimants are advised ‘Job search activities that are directly linked to employment such as applying for jobs and interviews will remain the focus’ (DESE 2021d). Choice is centered around the ways that job seekers might demonstrate compliance but is not afforded in relation to the type of work that job seekers can accept.

Targeting and ‘tailored servicing’

Explanations of the purpose of targeting have varied as system imperatives changed. Currently, it is intended to differentiate job seekers requiring ‘tailored servicing,’ by employment services providers from those who can self-service through ‘personalized service options’ (DESE 2021d, 1). It is claimed to ‘ensure that resources are directed to those who are most in need’ and ‘also essential for ensuring that employment services operate within the budget parameters set by the Government.’ (DSS 2020, 3)

While the preliminary automated assessment described as a ‘job seeker snapshot’ is framed as a mechanism to ensure that each job seeker is provided the appropriate level of support, it also constitutes the reduction of job seekers to suites of characteristics, which can be effectively acted upon by employment services providers, and for the digitally literate, by automated tools:

… assessments will use data analytics techniques to help target support and personalise services, such as providing job seekers with training and job search recommendations. Ongoing assessments and checks will also be used to identify when job seekers’ servicing needs change over time, and trigger interventions to ensure job seekers at risk of becoming disengaged receive appropriate support.

(DESE 2021d, 2)

The use of the phrase ‘servicing needs’ reiterates the mechanistic construction of a job seeker as comprised of parts, which can be accurately assessed and acted on. This is framed as being in the best interests of job seekers, though ultimately, it further shifts power toward the governing processes and systems. The repeated use of passive voice, such as ‘[o]ngoing assessments and checks will also be used’, reinforces the deference of power toward systems in the evaluation, support and punishment of job seekers. This reinforces the authority of the systems underpinning the NESM and TCF, removing people as agents in this process.

By distancing people from analysis and decision making, the model is portrayed as guiding judgment, based on the rational calculation of the algorithm. If a job seeker is non-compliant according to the model, then they are framed as being unambiguously non-compliant. These framings are embedded in the automated system logic of the NESM that constructs claimants as jobseekers ‘whose primary goal is to gain and maintain employment’ (Commonwealth of Australia 2021e, 8).

Discussion and conclusion

The analysis interrogated the discursive logic of activation informing the targeted compliance framework and the new employment services model. Changes to both employment support services and social security provisions, were ostensibly introduced as modernization necessary to meet the requirements of contemporary employment challenges. A series of assumptions underpin automation, starting with an assumed connection between activation and employment outcomes, achieved through enforced participation of job seekers.

The activation strategies incorporated into the model are informed by limited understandings of agency and causality, constructing both job seekers and employment services according to machine metaphors, as requiring external force for motivation. The analysis presented here illustrates how the rationalization of recent activation reforms rests on a dominant political discourse of people seeking state support as agents of their own situation, whereas in the new employment services model job seekers are framed as passive problems requiring various game-like strategies of compliance enforcement. This contradiction is manifest not only at the level of the ‘welfare subject’ positioned as requiring activation, but also in the equation of provision of support with registration on a website, the assumption that activation will produce ‘suitable work,’ and that suspension of income security payments motivates compliant behavior, that is, the conflation of compulsion and support in punitive activation (Raffass 2017).

The analysis of announcements of policy and operational changes suggests a shift in system logic. Whereas prior system settings may have supported workfare assumptions of active participation, the NESM logic substitutes compliance and incorporation into digital systems as the dominant operating objectives of activation. Strategies and system reforms thus prioritize ‘getting people in a position to take one’ (Morrison 2015), while effecting their depoliticization through both discursive and digital sequestration.

As Wood and Flinders noted, ‘ … depoliticization occurs when the debate surrounding an issue becomes technocratic, managerial, or disciplined toward a single goal, and hence changed in content’ (Wood and Flinders 2014, 161). This is evident in the classification process through which income security claimants are constructed as jobseekers, the coercive behavioralist measures underpinning targeted compliance, and impact of this approach on populations already marginalized in Australian society.

While proponents of the NESM claim that it is an innovative approach to employment service provision, the innovative element of the digital turn in employment services is framed in terms of potential savings and efficiencies. O’Sullivan, McGann, and Considine (2021, 5) observed ‘The narrative has always been that … the introduction of privatized service delivery, makes services more responsive and individualized. That narrative has seemingly been abandoned. In its place is a focus on efficiency and ease.’ Carney (2020a) cautioned against ‘[e]mbedding replacement Al within legacy systems constructed around human decision-making without system redesign,’ (28) especially in relation to ‘a segment of society with among the highest incidence of people at risk of being harmed by any design deficiencies.’ (44). In the NESM digitalization has encoded normative assumptions about the nature of work, employment services support and the causes of unemployment (for example, the belief that without mutual obligation requirements job seekers would fail to look for work) into the machine bureaucracy. Further, the ambivalence implicit in the dual role of employment support and mutual obligation enforcement structured into employment services that was the subject of concern in consultations on the new approach (Commonwealth of Australia 2019), is now structured into the NESM. Political representations of the NESM promoted it as responding to the need to provide tailored employment services to support job seekers, however, the design features of the NESM indicate it is more concerned to automate compliance with obligations.

The analysis presented here supports the conclusion that the real discursive innovation of the NESM is that the moralizing assumptions informing punitive activation have been embedded and therefore concealed in the policy technology. Critical policy discourse analysis reveals how this image is discursively constructed, as a form of ‘digital depoliticization’ that shifts the processes by which people seeking state support are conceived and treated from the political to the technological arena, concealing their operational impact, individualizing the experience of the social security system, and generating further questions for the right to social protection.

Disclosure statement

No potential conflict of interest was reported by the authors.