Lost in Categorisation? Employment Subsidies – Bringing the Beneficiaries Back In

Abstract

Employment subsidies are important active labour market policy (ALMP) tools, suited to a variety of labour market challenges. This paper engages with recent ALMP categorisation debates by appraising Cronert’s (2019) recent typology of employment subsidies. It uses empirical material to assess the typology’s explanatory power and produce insights to inform further typological development. The illustrative case of the British ‘Wage Incentive’ (2012-2014) is used to assess the typology’s analytical purchase. Cronert’s typology helpfully identifies key distinctions in the distributional profiles of employment subsidies, but further understanding of the category is impeded by the practice of defining them as demand-side interventions. The paper argues for a reappraisal of their supply-side characteristics, maintaining that the (potential) worker should be included in the analysis, and that employment subsidies’ relationship with training and job creation should be acknowledged. It proposes a redefinition of employment subsidies reflecting their real-world use, and suggests a framework for further exploring varieties of employment subsidy design from the perspective of beneficiaries.

Introduction

Welfare reforms over recent decades have moved away from ‘passive’ forms of support towards an ‘activation paradigm’ characterised by conditional benefits and increasing demands upon claimants (Weishaupt, 2011). While some identify ‘workfarist’ convergences (Peck, 2001), others emphasise the significance of distinctive national trajectories (Clasen and Clegg, 2006). Accounts of such developments invariably consider the tools brought to bear on labour market problems, and wider strategies of activation. Active Labour Market Policies (ALMPs) are accounted for at varying levels of abstraction, in terms of both policy tool mixes and strategic orientations. Specific measures are generally considered in terms of policy mechanisms, such as the categories used by OECD databases (Martin and Grubb, 2001). Ideal-type binaries often describe wider strategies (e.g. supply-side/demand-side’, ‘work-first’/‘human capital development’. Sometimes these are combined, generating more detailed typologies. Bonoli (2010), for example, considers ALMPs in terms of political-economic impact, reflecting on their relative ‘pro-market employment orientation’ and human capital investment. While such approaches highlight pertinent features of ALMPs, relevant features may be overlooked in categorisation, which to some extent explains ongoing efforts to account for them.

With a simple core mechanism and inherent configurability, employment subsidies represent important policymaking tools. Since COVID-19, their use has proliferated (OECD, 2020), reinforcing a need to account for them effectively. There are however some difficulties associated with this task, including longstanding inconsistencies in the use of subsidy-related terminology (Hamermesh, 1978) and a decline in the depth and precision of accounts over time. Early treatments were rigorous in their discussion of design features, recognising a broad range of options available to policymakers (Haveman and Christainsen, 1978; Hamermesh, 1978; Haveman and Palmer, 1982). Since then accounts have generally been less developed. Many treat employment subsidies as a uniform category (Cronert, 2019), or in purely economic terms. Even the more detailed treatments neglect design features that were highlighted in earlier accounts (EPPO, 2014; ILO; 2015). Furthermore, although a few acknowledge supply-side effects (Meadows, 2006; ILO,2015), employment subsidies are overwhelmingly treated as demand-side interventions in terms of policy-levers in being understood primarily as payments to employers. Recent work has sought to advance more nuanced understandings of them: Cronert’s (2019) typology accounts for employment subsidies in terms of their support for labour market insiders and outsiders and their use for cyclical or structural labour market challenges. However, this too perpetuates demand-side characterisations. Supply-side aspects are neglected and with that, the position and experience of the beneficiaries. This paper engages with ALMP categorisation debates and contributes to this analytical program by using empirical material to test the explanatory limits of Cronert’s framework, and producing insights to inform further typological development.

First, it appraises Cronert’s model, considering whether additional dimensions of subsidy design are significant and should be accounted for. The model is evaluated through the illustrative case of the UK Wage Incentive Scheme (WIS), a youth-targeted employment subsidy (operational 2012-2014), and Dinan’s (2019) typology of activation incentives. WIS has, to date, received limited attention in the literature but the case usefully illustrates several complexities of real-world employment subsidy use. While Cronert’s typology is useful, it neglects significant subsidy design features, and his definition overlooks aspects of their use in practice. Its second contribution is to re-define employment subsidies to reflect their real-world use, their relationship to job creation and training, and their supply-side levers and effects. Demand-side definitions encourage us to overlook the (potential) workers that employment subsidies seek to mobilise, and the conditions under which they do so. Acknowledging this gap creates an opportunity to explore within-category differences beyond the terms of Cronert’s typology. The final contribution is to propose three ways in which beneficiaries can be brought into the centre of the analysis to explore the category further. The paper begins by outlining its methods and discussing employment subsidy use and effects. WIS is then contextualised and considered through Cronert’s model. Gaps within that account are considered through Dinan’s typology. The strengths and limitations of Cronert’s typology are then discussed, and a re-definition of employment subsidies is proposed. The paper concludes by considering options for further exploring employment subsidies from the perspective of (potential) workers.

Methods

This paper draws upon documentary analysis conducted for doctoral research about the Youth Contract. Documents reveal insights about how social structures work (McCulloch, 2004) and are often naturally occurring, relevant, and available (Silverman, 2011). ‘Fast policy’ contexts however present access challenges as ongoing revisions displace older versions of documents. The Wayback Machine (WayBack Machine, 2020) potentially offsets this issue in archiving defunct webpages. To date it has been used within IT, archival science, and law (Arora et al., 2016), but it is also useful for researching policy development. For this paper, several different document types and versions were identified and analysed to build a comprehensive understanding of the Wage Incentive Scheme’s (WIS) design and operation. These relate to internal/external subsidy use, employer-facing marketing, internal guidance (DWP, 2013d), memos (DWP, 2014b), terms and conditions (DWP, 2012a, 2012b), ‘Freedom of Information’ requests (DWP, 2014c), evaluations (Jordan et al., 2013; Coleman et al., 2014), statistics (DWP, 2014a, 2015), and evidence gathered for policy scrutiny (WPSC, 2012). With this understanding, WIS is then used as an illustrative case to appraise Cronert’s (2019) typology, and explore options for further typological development.

The analysis proceeds by outlining Cronert’s definition and typology of employment subsidies, which outlines four key distinctions within the category. Accounting for WIS through this framework suggests a ‘borderline’ case, resulting from its inconclusive position in relation to one of the typology’s analytical dimensions. A comparison of WIS to predecessor subsidies indicates significant differences between them. As Cronert’s framework does not capture such distinctions, it is suggested that further development of the typology would be beneficial. Dinan’s (2019) typology of activation incentives provides for the analysis of ALMPs more generally, it is used in the subsequent discussion to produce a more systematic and multidimensional account of WIS, including features distinguishing it from predecessors. The account indicates that some limitations of Cronert’s typology stem from a narrow definition. The subsequent discussion proposes a redefinition of employment subsidies, acknowledging their real-world usage. The redefinition also provides for the use of three additional subsidy design features to build upon Cronert’s typology, with an understanding of scheme beneficiary experiences at the centre of the analysis.

Employment Subsidies

Employment subsidies are typically portrayed as demand-side interventions in policy-lever terms, with wage rates altered to affect employment (Hamermesh, 1993). Cronert (2019) additionally notes that they are publicly-funded, while hiring and firing decisions remain with employers. Some also acknowledge a supply-side to subsidies (sometimes through reference to ALMPs generally) in terms of their quantitative effects i.e. in terms of impacts upon the effective size or composition of the labour supply (Martin and Grubb, 2001; Haveman and Christainsen, 1978). The recognition of qualitative effects on the supply-side are much less common (and not referred to as such): at best, this involves references to jobseekers having altered perceptions of their chances within the jobs market (ILO, 2015). They are however generally not considered as supply-side in terms of policy levers, unless they are also understood to include payments made to employees. Employment subsidies have been used with varying success throughout Europe over recent decades (EPPO, 2014) and reservations remain about their effectiveness and value-for-money, especially in the context of larger firms and slack labour markets (Boone and van Ours, 2004; Welters and Muysken, 2006). In Britain, both retention subsidies (for preventing redundancies) and recruitment subsidies (for new hires) were used into the 1990s. The latter then dominated until the 2020 Coronavirus Job Retention Scheme. While subsidy use in Britain has undergone regular changes, other states have successfully integrated them into day-to-day practice; a subsidy for long-term unemployed Danes has been used since the late 1970s (EPPO, 2014).

Retention subsidies aim to prevent redundancies while recruitment subsidies alter hiring decisions or bring them forward. The marginal-stock variant instead supports additional job creation – because it is paid “on increments to the employment level in a firm” (Haveman and Christainsen, 1978, p.268), it is often ignored in recent accounts of employment subsidies (e.g. EPPO, 2014; ILO, 2015; Cronert, 2019). Recruitment and marginal-stock subsidies compensate employers for additional ‘running in’ costs associated with target groups (Ingold and Stuart, 2015). They are generally advocated for equity reasons such as promoting labour market attachment amongst disadvantaged groups or in response to recession (OECD, 2009). Employment subsidies are a diverse ALMP category with their effects “crucially determined by their design features” (Clasen et al., 2016, p.30). Terminological consensus about sub-categories is lacking, but agreement about the significance of design features to some extent prevails. This includes: duration/rate of subsidisation; payment type (instalments/back-loading/up-front); rate (fixed/variable); targeting (geography/industry/cohort); alongside the primary policy objectives (e.g. redundancy prevention, additional role creation, modification of hiring practices) (Hamermesh, 1978; Haveman and Palmer, 1982; Standing, 1999).

A longstanding design challenge involves a trade-off between targeting and take-up (Martin and Grubb, 2001). A desire for appropriate subsidy use may produce strict rules which, if over-specified, may undermine employer engagement through perceptions of administrative burdens or cohort stigmatisation (Katz, 1996). If rules are relaxed, other side-effects emerge: Deadweight occurs where subsidised behaviours would happen anyway. Displacement occurs at scale over the longer-term, where subsidies confer a competitive advantage to participating firms, at the expense of roles elsewhere. Substitution occurs when beneficiaries replace un-subsidised workers. (Calmfors, 1994). Substitution can however also be an explicit policy objective, “…since shuffling the queue of job-seekers is in part what the schemes are intended to do”. (Fay, 1996, p.20). On this issue, Kettner and Rebien’s (2007) distinction between types of substitution is instructive: direct substitution involves the dismissal of existing employees; delayed substitution occurs after beneficiaries have been inducted; indirect substitution takes place as existing vacancies are filled (Moczall, 2013). Of these, indirect substitution is the one likely to be a desired policy effect, with subsidies modifying labour demand and (to the extent that they are effective) the composition of the remaining labour supply (though demand-side accounts often neglect this). Deadweight risks persist whenever employment subsidies are used, and displacement risks emerge at scale unless specific (non-market) activities are subsidised. It remains difficult to prevent delayed substitution given the time-lag involved but direct/indirect substitution may be countered through marginal-stock configurations and rules preventing subsidy use in roles vacated by redundancy.

Background: the Wage Incentive Scheme (WIS)

WIS was active from April 2012 to August 2014. It should be understood in the context of the UK Conservative-Liberal Democrat Coalition (2010-2015), austerity politics and welfare reform. Building on New Labour’s contractual welfare reforms, the Coalition introduced a tougher jobseeker regime: the 2012 Welfare Reform Act limited payments to claimants through a ‘benefit cap’ and established a ‘claimant commitment’ based on the notion that “looking for work should be a full-time job” (DWP, 2013c). Disqualifications for non-compliance were extended from a maximum of 26-weeks to three years. Unpaid work experience placements were increasingly used, and the quasi-market in public employment services (PES) expanded. Provision for long-term unemployed people was contracted-out to external delivery organisations under the Work Programme. Providers were paid ‘by outcomes’ through a differential payment model where claimants were segmented based upon (perceptions of) their proximity to the labour market. Fiscal consolidation was an immediate coalition priority. In May 2010 two subsidised employment programmes – the Six Month Offer and (youth-targeted) Future Jobs Fund (FJF) – were scrapped (H.M. Treasury, 2010). FJF’s cancellation, before evaluation, was for many premature (WPSC, 2010). In the following year, concerns about this decision seemed justified as youth unemployment hit one million shortly after the London riots (ONS, 2011) ¹ .

WIS was the centrepiece of the Youth Contract (YC), package of support, which also included work experience and apprenticeship placement funding. It sought to ‘level the playing field’ for unemployed young people, offering £2,275 to organisations employing eligible 18-24-year-olds for a sustained 26-week period. Eligibility was initially restricted to Work Programme clients (unemployed for over nine months), but from July 2012 it became available to Jobcentre Plus (JCP) clients (unemployed for six months or longer) in unemployment ‘hotspots’; by December it was rolled out across the entire JCP network. A £374 million budget made it the largest component of YC; over three years this aimed to fund 160,000 subsidies. To date, WIS has received limited discussion in the literature. Most accounts derive from official evaluations (Jordan et al., 2013; Coleman et al., 2014), which, alongside statistical releases (DWP, 2013a; 2013b; 2014a; 2015), form the basis for most statements about WIS. Evaluation focused on accounting for process (e.g. marketing/administration) and exploring employer and adviser perspectives. There were however significant omissions e.g. part-time work, job quality, and beneficiary (worker) experiences. WIS was widely regarded as ineffective. When performance statistics were published, ‘job starts’ were included, which presented a better impression of performance (DWP, 2013a). The Department for Work and Pensions (DWP) however acknowledged that job starts were based on unreliable administrative data (DWP, 2013b). Final figures showed payments made for 36,470 individuals, of which 27,750 were for 26-week outcomes (17 per cent of the target). This meant that 24 per cent of payments saw beneficiaries drop out before 26 weeks ² . ‘Job starts’ implied stronger performance, with up to 116,140 job transitions brokered via WIS – interestingly, this expanded substantially after the subsidy’s introduction to Jobcentre Plus (Figure 1). This 26-week/job-start gap likely reflected a combination of inappropriate referrals, inadequate in-work support, and employer reluctance to claim (Coleman et al., 2014, p.38).

Figure 1. Wage Incentives: Monthly Job Starts and Sustained Employment Payments (DWP, 2015)

In the literature, discussion of WIS is generally linked to other interventions or employer engagement (Ingold and Stuart, 2015). It is often regarded as a failure or irrelevance, with success framed in terms of employment outcomes (Crowley and Cominetti, 2014; Gregg, 2015; Maguire, 2015; Sage, 2016). Together, such accounts convey an image of low employer take-up and high levels of deadweight, essentially also implying that it was expensive (EPPO, 2014; Tassinari et al., 2016). While that characterisation is largely accurate, it overlooks a broader range of policy effects and WIS’s inherent multi-dimensionality. In using WIS as an illustrative case to evaluate Cronert’s typology, this paper also presents a fuller, contextualised account of WIS.

WIS as ‘outsider subsidy’: the limits of Cronert’s typology

Cronert critiques treatments of employment subsidies as a uniform ALMP category, and seeks a “parsimonious typology” reflecting their versatility, and use “…as means to rather different distributional ends” (Cronert, 2019, p.17). He defines them as “…demand-side labour market interventions that provide economic incentives to employers to increase the employment and/or earnings of selected groups of persons experiencing difficulties in the labour market” (Cronert, 2019, p.3). Following convention, he distinguishes them from training, direct job creation, and supply-side wage subsidies which “…affect workers’ job search behaviour rather than employers’ recruitment behaviour” (Cronert, 2019, p.3). The typology shows how subsidies address specific labour market challenges, but there are limits to its descriptive power. Distilling a subsidy into a single “underlying dimensionality” may be difficult, and Cronert acknowledges the likelihood of messy borderline cases (Cronert, 2019, p. 15). There is therefore a tension between the diversity he acknowledges and the parsimonious typology he develops to account for it.

The typology emerges from a ‘multiple correspondence analysis’ of subsidy design characteristics within an EU database. Cronert acknowledges limitations to this, e.g. incomplete data on job eligibility criteria (Cronert, 2019, p.9). Clusters are identified with shared characteristics used as the basis for proposing two dimensions. A cyclical/structural dimension, relating to labour market contexts and challenges, emerges from Cronert’s reading of programme duration and targeting: cyclically-oriented subsidies are viewed as shorter, broadly-targeted initiatives, while structurally-oriented subsides have longer periods of subsidisation and focused targeting. An insider/outsider dimension sees ‘insiders’ targeted by subsidies that maintain existing employment (retention schemes) or increase the employment of disabled people ³ (Cronert, 2019, pp. 9-15). Four employment subsidy types emerge from the analysis: Structural-outsider programmes of medium-to-long duration (typically targeted at the long-term unemployed); longer-term Structural-disabled subsidies; Cyclical-general (cyclical-outsider) programmes of short duration; and Cyclical-insider (retention) subsidies (Cronert, 2019, pp. 13-15). How well does this typology describe WIS?

The ‘outsider’ credentials of WIS are unambiguous – it neither maintained existing roles nor supported disabled people into work. Its place within the structural-cyclical dimension is less clear. Drawing attention to broad targeting and short duration, Cronert exemplifies cyclical-general subsidies using WIS’s immediate predecessor the Six Month Offer (SMO) (Cronert, 2019, p. 14). SMO was more broadly targeted than its own predecessor the New Deal Employment Option (NDEO), which operated within a context of growth; but the successor (WIS) was more targeted than SMO in a context of higher unemployment. Could this reflect a ‘borderline’ case leaning towards structural objectives, or an ideological aversion to spending? Answers cannot be found at the level of subsidy duration because each of the three subsidies above lasted for six months. The design features emphasised by Cronert do not readily distinguish these subsidies; other factors are therefore worth considering.

Timing presents one option: WIS was launched as youth unemployment rose, did that make it cyclically-targeted? It is hard to say. Cyclical and structural factors can be interdependent: a crisis may compound longer-term structural trends, or be used as “…an opportunity to promote structural reforms” (Scarpetta et al., 2010, p.26; Gregg, 2015). The length of subsidisation, timing, and targeting of a subsidy may then resemble a cyclically-targeted intervention while other design features do not. A key issue may be the extent to which it stimulates demand as post-crisis youth unemployment reflects reduced labour demand and how “youth workers tend to play a role of a ‘buffer’ to absorb macroeconomic shocks, through wider fluctuations in their unemployment rates” (Jimeno and Rodriguez-Palenzuela, 2003, p.6). A couple of options present themselves for exploring this further. The amount of subsidisation, or hours subsidised could be considered. WIS offered £2,275 over six months (roughly half the minimum wage at the time), more than the £1,000 offered for SMO but it is notable that SMO was only one part of a “Golden Hello” package and coupled with a £1,500 training subsidy (Harari, 2010). Looking at NDEO (adjusted for inflation) a similar payment is found, again with training attached (Jarvis, 1997). WIS was then more distinguished from its predecessors in lacking training support than in terms of the funding provided. Cronert follows a well-established tendency of bracketing off employment subsidies from training programmes, and it is true that these are different policies. The convention, however, risks overlooking an important aspect of employment subsidy design: because it is sometimes viewed as more successful in combination with training (Katz, 1996).

Whether a subsidy is configured to stimulate additional role creation (and directly reduce unemployment) is a second factor to consider. Evaluation of WIS indicated some additional job creation (19 per cent of surveyed employers) but also noted this was not the explicit objective of the scheme (Coleman et al., 2014, p.68). WIS was designed to modify demand for labour through indirect substitution – additional role creation was a secondary effect not hard-wired into its design. This stands in contrast to marginal-stock subsidies such as the 2020 Kickstart scheme which require employers to evidence that subsidised roles are “new jobs”, explicitly ruling out direct, indirect and delayed substitution ⁴ (DWP, 2020). Marginal-stock subsidies are absent from Cronert’s discussion of employment subsidies, which is strange as they are overtly cyclically-oriented in directly reducing unemployment (interestingly, Cronert (2017) argues elsewhere that partisanship in ALMP use is seen in a tendency to use (or not use) them to reduce unemployment). Marginal-stock subsidies may pose a challenge for neat categorisation in blurring conceptual boundaries between employment subsidies and job creation; this is revisited later in the discussion.

In summary, Cronert’s model classifies WIS as an ‘outsider’ subsidy. It resembles a cyclically-oriented case, but not in terms of the targeting/duration that Cronert emphasises and it did not directly reduce unemployment. There are therefore grounds for considering it a ‘borderline’ case. Cronert’s typology captures key aspects of the Wage Incentive’s design but it was more than a ‘borderline’ structural/cyclical subsidy for outsiders. Dinan’s (2019) activation incentive typology provides a tool for exploring this in greater depth.

Reading WIS through activation incentives

With inherent adaptability and inconsistent spending implications, Dinan (2019) finds ALMPs challenging to compare on the basis of cost. A typology is instead proposed based upon the incentive mechanisms that influence target populations’ behaviour. Dinan’s model considers whether policies engage workers or the firm (supply/demand-side) against four incentive mechanisms types. The first two, positive and negative financial incentives, concern whether actors stand to gain from participation or lose from non-participation. The third and fourth are “human capital incentives” which “…invest in an individual’s capacities to increase skills or qualifications”. These are differentiated according to intervention intensity: organizational human capital incentives are “…soft skills that relate to organizational skills, such as the ability to “market” oneself”, concrete human capital incentives are “…hard skills associated with concrete knowledge, such as training and vocational training and education”. (Dinan, 2019, p.8). Together these elements produce an eightfold typology, now used to structure a deeper analysis of WIS.

Dinan’s first category, demand-side negative financial incentives, does not apply to WIS. The second, increased labour search incentives, concerns supply-side penalties, e.g. sanctions claimants receive for non-compliance with job-search conditions. For Dinan this “most resembles workfare-ism” in seeking to counter ‘incentive traps’ and ‘dependence’ (Dinan, 2019, pp. 8-11). This does not apply directly to WIS as such features relate to the benefit system rather than the subsidy, yet close links between the two complicate our understanding of WIS’s operation. Lødemel and Trickey’s (2000) ‘form-based’ definition of workfare, considers how ‘offers’ like work placements are connected to the threat of penalties, constituting ‘throffers’. These guide claimant behaviours use a combination of ‘carrots’ and ‘sticks’, with disqualifications for benefits as the ultimate sanction. WIS and other YC measures were conceived of in such terms (providing the ‘Youth Contract’ with its contractual element). This was outlined in the strategy document Building Engagement, Building Futures which also argued that YC opportunities justified stricter governance of jobseekers.

“As we are providing more support and more opportunities for young people, we also expect more in return. Those failing to engage positively with the Youth Contract will be considered for Mandatory Work Activity ⁵ . Those who drop out of a Work Experience place or a subsidised (or other) job without good reason will lose their benefits.” (H.M. Government, 2011, p.8).

In isolation, WIS cannot be characterised as an “increased labour search incentive”, but it was also more than a ‘throffer’ component. ‘Positive engagement’ with WIS was underwritten by the threat of sanction, but YC ‘opportunities’ including WIS were also used to legitimate the intensification of conditionality. Considering the British New Deals, Peck (2001) noted that workfarist interventions were ‘fair weather’ policies, reliant upon favourable labour market contexts for their continued legitimacy and success. The case of WIS however suggests that employment subsidies could provide a tool for policy-makers to ‘double down’ on their commitment to a stricter regime in the absence of a supportive labour market (and use the crisis as an opportunity to promote structural reform). The offer of (largely unrealised) subsidised placements replaced job vacancies as ‘carrots’ legitimating greater use of ‘the stick’. Indeed, sanctioning against 18-24-year-olds was sustained at unusually high levels during WIS’s lifetime [Figure 2 and Figure 3]. Dinan’s typology helps to acknowledge subsidies’ roles in such supply-side strategies but this is difficult to capture under Cronert’s model. WIS is by no means unique in this respect. In the 1980s UK policymakers believed young people to be ‘priced out’ of the labour market and so employers were subsidised to pay young people at lower rates than older colleagues under the Young Workers Scheme (Finn, 1987, p142). Such context may therefore be invaluable for appreciating how an employment subsidy actually works.

Figure 2. JSA Sanctions ("Decision to apply a sanction (adverse)" 18-24-year-olds (Stat-Xplore, 2019)

Figure 3. 18-24-years-olds: Average Monthly Adverse Sanctions as % of 18-24 Claimant Count (authors own calculation from ONS, 2019; Stat X-plore, 2019)

The third category subsidized employment (demand-side positive financial incentives) clearly applies to WIS. It refers to interventions reducing labour costs, including subsidies as understood by Cronert (direct cash transfers) as well as reduced social contributions for employers and sub-legal minimum wages (Cronert, 2019, p.10). For Dinan, employment subsidies and direct job creation share the same incentive mechanism while Cronert separates the two (Cronert, 2019, p.3). It is worth reflecting on the reasoning behind this. Cronert references Bonoli’s treatment of ALMPs (Bonoli, 2010) according to relative pro-market employment orientation (whether roles are created within the public/private sector). For Cronert, job creation schemes are distinct because they “…are not intended to promote demand-driven market employment but rather to keep people occupied in non-market jobs of community benefit, usually in the public or non-profit sector.” (Cronert, 2019, p.3). Bonoli similarly suggests that direct job creation is “…often used to re-create an entitlement to unemployment insurance rather than to increase the chances of landing an unsubsidized job.” (Bonoli, 2010, p.440). While direct job creation and employment subsidies are different ALMPs, sustaining their separation on such grounds may pose challenges, under circumstances where a significant grey area emerges between them. WIS illustrates some of the issues involved.

Cronert acknowledges that employment subsidies are used across private, public and non-profit sectors: earlier we also saw that subsidies can be used in a disciplinary sense. They sometimes also create additional jobs directly – if ‘marginal-stock’ subsidies create additional private sector roles, these are no more ‘market driven’ than additional public sector roles. If roles are created in the public sector, the subsidy begins to resemble direct job creation. WIS was not designed as a marginal-stock subsidy, but it was sometimes modified via local top-up subsidies to create new roles. Bradford and Nottingham, for example, included “additional role” and “community benefit” conditions of the kind Cronert associates with direct job creation (Leeds City Region, 2013, p.9; LGA, 2014, p. 35, 38). We also cannot assume that subsidies necessarily increase beneficiaries’ chances of landing unsubsidised roles any more than direct job creation does; as either could be augmented with support to promote subsequent labour market progression.

The fourth category, fiscal incentives (supply-side positive financial incentives), does not apply to WIS as payments were made to employers rather than workers. For Dinan, this includes low-earnings supplements and re-employment bonuses (paid to jobseekers for moving into employment). These are objectively different instruments, supporting income security and employment transitions respectively. Some however consider re-employment bonuses alongside employment subsidies because both may exert downward pressure on wages, the former by reducing employer wage bills, the latter by effectively lowering workers’ reservation wage (Huo, 2006; Vlandas, 2013, p.7). In Britain re-employment bonuses have been less common than subsidies paid to employers but a national ‘back-to-work bonus’ operated in the 1990s and local variants have been used more recently (Waite, 2014).

Category five, administrative services (demand-side organisational human capital incentives), “…create and reinforce links between businesses and local and regional governments” and “simplify the matching process between workers and labour” (Cronert, 2019, p.10). While a ‘matching approach’ to ALMP is typically associated with the work of public employment services in addressing supply-demand information asymmetries, for Bredgaard (2018) this also applies to wage subsidies because (subject to ‘moderators’ such as willingness/motivation) they create links between jobseekers and employers. They also establish links between employment services and employers – WIS was used for employer engagement by Jobcentre Plus and Work Programme providers to ‘market a wider range of services’ (Coleman et al., 2014, p.40). It was originally designed as an exclusive ‘sales tool’ for The Work Programme (WPSC, 2012, Ev. 124). WIS’s design reflected this: the back-loaded payment mirrored the outcome payment schedule for providers; payments of WIS often entailed payments to providers. WIS was then also a subsidy to Work Programme providers – by facilitating employer engagement it effectively lowered the threshold for securing job outcomes. WIS’s introduction was timely for providers: the Coalition’s reforms were fiercely contested in 2011/12, and some anticipated the Work Programme’s imminent collapse (Butler, 2012). In particular, media criticism of unpaid placements and an associated boycott campaign undermined employer engagement (Grover and Piggott, 2013). WIS engaged employers without the controversy of unpaid placements, some suggested it had “driven business” to providers (WPSC, 2012, Ev.61).

Category six, employment services are supply-side measures which “attempt to reduce barriers to the labour market by enhancing individuals’ capacity to market their skills” (Dinan, 2019, p.10). This applies to WIS, and challenges understandings of employment subsidies as demand-side interventions in terms of their levers. How subsidies are presented to employers is significant. If advisers contact employers with the offer, it functions in clear demand-side terms. If a claimant instead uses it to market themselves to prospective employers, it assumes supply-side characteristics as the voucher becomes the lever, altering beneficiaries’ job-search behaviours. This happened with WIS: practices viewed as effective in SMO were re-adopted (Adams et al., 2011). Vouchers (Figure 4) were issued as a tool for claimants to ‘sell themselves’ with. They were referred to in face-to-face interactions, on CVs and in applications (Jordan et al, 2013). Voucher ‘self-marketing’ receives mixed reviews in the literature. For WIS it was viewed as successful for some beneficiaries (Coleman et al., 2014, p.24) though stigmatisation risks also emerge from the practice (Burtless, 1985). Vouchers represent an inexpensive modification, which explains why they are a common adaptation; but the practice also implicitly challenges exclusively demand-side definitions of employment subsidies.

Figure 4. Wage Incentive "self-marketing" voucher

Dinan’s final categories – company training and upskilling – are more intensive forms of human capital development. Neither applies to WIS but, as noted earlier, training was linked to WIS’s predecessors and subsidy success is sometimes connected to complementary training (EPPO, 2014). Subsidies may be augmented with additional training funds as in NDEO or SMO, and designs may stipulate conditions relating to skills investment (Adams et al., 2011). Training programmes and employment subsidies are clearly different ALMPs but it is worth acknowledging that subsidy designs may vary by the extent to which they are augmented by training. The implications of this reading of WIS through ‘activation incentives’ are discussed in the next section.

Discussion: redefining employment subsidies/bringing the beneficiaries back in

Cronert’s typology highlights key distributional profiles of employment subsidies but its usefulness and descriptive power is restricted to outlining basic policy intent. International studies often consider comparable economic contexts – during downturns, the typology distinguishes between retention and hiring subsidies; during growth, it distinguishes between disabled and other beneficiaries (Cronert, 2019). These are relevant considerations, but much more can be said about employment subsidies than is captured under the framework. The four sub-groups Cronert identifies are no more uniform categories than employment subsidies are themselves. Reading of WIS through Dinan’s (2019) activation incentives demonstrates the potential complexities of employment subsidies, yet potentially produces an unwieldy amount of information. Can Cronert’s frugal typology be built upon to elucidate further distinctions? Before this is considered, Cronert’s definition of employment subsidies should be revisited. Viewed through Dinan’s typology, the case of WIS suggests issues with its reference points.

Cronert’s definition partially explores employment subsidies in negative terms; they are not non-selective tax reductions, tax credits, supply-side subsidies, training, job-search, or job creation. His positive definition describes them as publicly-funded with hiring/firing decisions made by other actors; They are “…demand-side labour market interventions that provide economic incentives to employers to increase the employment and/or earnings of selected groups of persons experiencing difficulties in the labour market” (Cronert, 2019, p.3). The definition prompts a number of questions, but an initial point is noteworthy. Employment subsidies can also be used to reduce ‘beneficiary’ earnings – the Young Workers Scheme (mentioned above) did this in the 1980s in assuming young people to be ‘priced out’ of the labour market (Finn, 1987). Cronert’s reference to increased earnings seems unnecessary and certainly secondary to the goal of preventing unemployment. Retention subsidies are employment subsidies by virtue of securing continued employment for existing workers. Increased earnings stem from that goal but are not integral to the definition. Employment subsidies promote transitions to employment, or prevent transitions to unemployment.

While they require demand-side buy-in, definitions of employment subsidies as demand-side are problematic because they also work on the supply-side. Firstly, as noted in some existing accounts, in terms of quantitative supply-side effects i.e. they change the size or composition of the labour supply (Haveman and Christainsen, 1978). Second, subsidies sometimes assume a role in (qualitative) supply-side strategies (i.e. altering workers’ motivations or making them cheaper) e.g. when coupled with sanctions as ‘throffers’ (Lødemel and Trickey, 2000). While they are separate to a jobseeker regime they may support or legitimate it. Such connections can be crucial for understanding employment subsidies’ operation, yet they are lost within de-contextualised demand-side accounts. Finally, there are situations where employment subsidies directly assume supply-side characteristics in terms of policy levers: when ‘self-marketing’ vouchers are issued to beneficiaries. Demand-side definitions overlook these realities and, with them, the beneficiaries at the heart of the intervention. Employment subsidies rely on demand-side buy-in but they are not exclusively demand-side: (potential) workers and their experiences are important considerations too.

Issues also arise from Cronert’s negative definition, specifically that employment subsidies are not supply-side subsidies, job-creation or training programmes. In acknowledging that they sometimes work on the supply-side, their relationship to other supply-side subsidies should be reconsidered. Some are clearly not employment subsidies: childcare allowances and mobility grants remove labour market barriers rather than funding employment transitions. Reemployment bonuses however change job-search behaviours as recruitment subsidies do when ‘self-marketing’ vouchers are used. The key difference is that recruitment subsidies pay employees through employers while reemployment bonuses pay employees directly. Similar mechanisms are at work – a reemployment bonus is also a publicly funded employment transition, where hiring/firing decisions are in the hands of other employers. Both allow employers to hire beneficiaries at lower rates (than is legal or would normally be acceptable to employees). Their impacts on the wider labour market may also be similar, potentially in promoting an expansion of low-waged work (Vlandas, 2013). Cronert’s exclusion of re-employment bonuses seems based upon understanding them as supply-side and employment subsidies as distinctly demand-side. With the complicating issue of self-marketing, the residual ground for maintaining the separation appears to be the slightly different question of whether they are paid through an employer. If this is sufficient for maintaining the separation, then the label ‘employer subsidy’ is perhaps more appropriate.

A second issue concerns the separation of employment subsidies from direct job creation (e.g. Cronert, 2019; Bonoli, 2010). Employment subsidies are not public works programmes, but they are sometimes configured in ways that approximate them. They are used in the public, private and non-profit sectors and may create additional employment. They can be used disciplinarily, for community benefit, and are more or less conducive to subsequent (unsubsidised) employment progression. Under certain configurations, they strongly resemble job creation; the extent to which they do this accounts for significant variation within the category. If beneficiaries are positioned at the centre of the analysis, this becomes a design consideration with distributional impacts: hiring subsidies can intensify competition between unemployed people by ‘re-ordering the exit queue’, promoting a re-composition of the labour supply, and exerting downward pressure on wages. ‘Marginal-stock’ subsidies instead alleviate such pressures, increasing demand relative to supply and reducing beneficiaries’ competition with other candidates. Finally, while employment subsidies are not training schemes, their relationship to training is sometimes a significant design feature. The success of structurally-oriented subsidies is sometimes linked to associated training components (EPPO, 2014) and many designs require skills investment (or commitment to progression/employability). Considered together, such ambiguities suggest a re-definition of employment subsidies is in order.

Employment subsidies are interventions that are active on both the supply and demand-side. Public funding is used to promote employment of specific groups while leaving hiring and firing decisions with employers. They may support transitions into employment or prevent transitions to unemployment. While a distinct ALMP category, they sometimes incorporate elements of training and occasionally resemble certain forms of job creation.

This re-definition, derived from the discussion above, acknowledges real-world employment subsidy use whilst remaining compatible with Cronert’s typology. It also opens additional reference points up for further exploration of the category, with beneficiaries at the centre of the analysis. Three options emerge for considering differences between Cronert’s employment subsidy sub-categories. First, how does the subsidy position beneficiaries in relation to the rest of the labour market? Does it create new roles for them or change their place in a queue? (Fay, 1996) Is retention supported by providing employers with additional funds, or is additional work-sharing encouraged? Put differently, is a subsidy designed to expand or to modify labour demand? Recruitment subsidies primarily have a compositional effect through ‘queue shuffling’ unless marginal-stock conditions are present to directly expand demand. Similarly, subsidies encouraging job-shares are more expansive than simple retention incentives.

Second, what does it do for beneficiaries? Do they simply receive work experience? Are their skills developed? Is progression or retention promoted? Are job quality controls in place? Are retained workers re-skilled for progression elsewhere? These are important questions. Third, what does it expect them to do? The subsidy’s connection to conditionality is relevant. Is the subsidy passive or ‘activating’? Minimal expectations of beneficiaries are likely to be that they conduct themselves appropriately and complete subsidised placements (in Britain beneficiaries would likely be disqualified from benefits if they left work voluntarily or were dismissed for misconduct). Alongside this, positive or negative incentives may be positioned against an (activating) expectation that claimants ‘sell themselves’, e.g. through vouchers or re-employment bonuses. The paper cannot properly operationalise these questions here, but they present avenues for exploring employment subsidy variety from the perspective of their primary beneficiaries. Given the importance of demand-side buy-in, a complementary endeavour would be to explore similar questions from employer perspectives, e.g. considering whether subsidies are delivered with complementary candidate screening, training or retention and progression support (Sissons and Green, 2017).

Conclusion

This paper has sought to contribute to ALMP categorisation debates by testing the analytical purchase of Cronert’s framework with empirical material, and producing insights to inform that analytical program through further typological development. The case of WIS illustrates the complexities of employment subsidy use within the contemporary policy landscape. Cronert’s (2019) model helpfully outlines different distributional profiles of subsidies but this is also its limit, and his definition of employment subsidies overlooks aspects of their real-world use. This paper has proposed a reformulated definition to address this issue, and suggested avenues for further exploration of the category with beneficiaries at the centre of the analysis. Many authors rightly emphasise the significance of employers for ALMP delivery, noting that suspicion of employment services and negative assumptions about beneficiaries often hold interventions back (Ingold and Stuart, 2015). Bredgaard (2018) notes that the role of employers is often implicit or assumed when ALMPs are treated as supply-side interventions. It has correspondingly been suggested here that supply-side aspects are similarly overlooked in examining employment subsidies.

Subsidy designs cannot guarantee outcomes but they do structure participant experiences. This is relevant to policy success because, like other ALMPs, subsidy designs communicate what employment services do, and signal the implicit value of workers to employers (Ingold and Stuart, 2015). This in turn has implications for which employers participate, and how they engage. Training or progression conditions may be burdensome for some but they also signal that workers are worthy of investment, meaning employers are more likely to engage on that basis. If subsidies instead construct (potential) workers as cheap, disposable, and lacking intrinsic motivation, employers are more likely to treat them as such. Employment subsidies require demand-side buy-in to work, but it is difficult to appreciate which employers will engage (including why and how they engage) without reflecting on the supply-side and bringing beneficiaries back into the discussion.