Nice work if you can get it: Labour market pathways of Belgian service voucher workers

Introduction

Several European countries have started to use tax money to boost the demand for domestic services such as house cleaning or ironing (Carbonnier and Morel, 2015). Next to improving the work–life balance for people already at work, an essential and recurring motive has been to create more ‘decent’ jobs as an alternative to precarious, informal work or no job at all (OECD, 2021). However, some people worry that such policies may also effectively institutionalize socio-economic inequalities (Gavanas, 2010; Hellgren, 2015; Shire, 2015; Van Hooren, 2018). In this article, we examine the case of the heavily subsidized, tightly regulated and yet also extremely popular service voucher scheme (SVS) in Belgium. It is the largest subsidized domestic work scheme in Europe by far (OECD, 2021). We ask: does it provide stable and secure employment to individuals outside or at the fringes of the labour market? The Belgian case is exciting because the domestic work jobs created by the scheme are of the highest quality in terms of working conditions and social protection (Ramos Martín, 2018).

We contribute to the literature by analysing empirically how domestic service workers fare in and outside subsidized domestic service work, in our case, the SVS. Despite the growth in subsidized domestic service work in Europe, little is known about the employment pathways of people employed under such schemes (Barbiano di Belgiojoso and Ortensi, 2019; OECD, 2021). Most research has analysed the issue using a qualitative (policy analysis) approach (Carbonnier and Morel, 2015; Jokela, 2017). For Belgium, research has identified positive employment effects of the SVS for both low- and high-skilled women and the unemployed (Raz-Yurovich and Marx, 2018, 2019). A recent study showed that the scheme is no stepping stone toward non-subsidized employment (Leduc and Tojerow, 2020). However, we still lack empirical research that offers a detailed picture of the labour market pathways that lead up to participation in subsidized domestic service work and enables us to understand how differences in these pathways contribute to divergence in employment outcomes. Therefore, in this article, we employ a sequence analysis approach to investigate how participation in the SVS is embedded in the broader context of labour market pathways.

Subsidized domestic service work: A way out of or into precariousness?

The Great Recession and the COVID19 crisis highlighted many existing vulnerabilities in advanced economies. For a long time, rising inequality and ‘dualization’ – the emergence or deepening of new insider–outsider divides – were thought to be inevitable by-products of changes that brought greater prosperity. Since the early 1990s, extensive research has demonstrated that labour markets in advanced economies are subject to polarizing forces. Researchers witnessed the gradual growth in both high and low paying jobs in some labour markets, with the gradual disappearance of mid-level jobs (Autor et al., 2003; Goos, 2013; Palier, 2019; Howell and Kalleberg, 2019). Goos and Manning (2007) pointed to the development of many highly paid ‘lovely jobs’ on the one hand, taken up by the highly educated, whose productivity was boosted by the digital revolution, and on the other hand, the development of low-paid ‘lousy jobs’ in the service sector such as retail, catering, construction, personal and domestic care.

A growing site of low-quality and poorly paid ‘lousy service jobs’ in Europe occurs in the domestic services sector (Carbonnier and Morel, 2015). Domestic work progressively disappeared in most European countries during the 20th century but reappeared over recent decades, much of it in the informal sector (Carbonnier and Morel, 2015). The growth of domestic services is said to stem from, among others, the rise in inequality, the rise in (undocumented) migration, unemployment growth, as well as new social needs (Bettio et al., 2006; Cancedda, 2001). It can also be understood against the rise of the post-industrial service economy, where ‘huge areas of servicing are labour intensive and low-skilled’ (Esping-Andersen, 1999: 96). Another explanation is increased female labour force participation (Leschke, 2015). With domestic service work on the rise, however, the sector never left the realm of women’s ‘natural’ domain, nor did its commodification and marketization increase its cultural or monetary value (Lutz, 2008; Farris, 2018). Paid domestic work enabled middle-class women to start performing clean well-paid employment outside the home while relegating the ‘dirty work’ to domestic workers (Anderson, 2000). This ‘growing commodification of reproductive labour’ (Abrantes, 2014) in contemporary affluent countries is disproportionately staffed by migrant women, thus reproducing systems of gender, class and race domination (Romero, 1992), with migrant women said to face a ‘double disadvantage’, both by gender as well as migrant status (Altman and Pannell, 2012; Jokela, 2019). Domestic work stands at the intersection of varying gender, care, and migration regimes, resulting in varying policies and risks of precariousness in different countries/welfare states (Lutz, 2008; Kilkey et al., 2010).

While there exist many definitions and approaches to precariousness, also outside of the focus on domestic service work, it generally refers to ‘employment under conditions that create uncertainty and insecurity for individuals in the labour market’ (Jokela, 2019: 34; Vosko et al., 2009). Kalleberg (2018: 241) specifies that the uncertainty, instability, and insecurity associated with precarious work also entails that ‘employees bear the risks of work (as opposed to businesses or the government) and receive limited social benefits and statutory entitlements’. The consequences of precarious work go beyond the quality and quantity of the job and extend into other realms, such as people’s well-being, their ability to form families, and broader community disintegration and disinvestment.

An essential aspect of the growth of domestic services has been the active promotion and subsidization of domestic and care services in several European countries (Hellgren, 2015; Marx and Vandelanoote, 2015; Morel, 2015). Such policies are seen as a vehicle for ensuring service sector job growth to the benefit of the low-skilled without having to accept higher wage inequality. In other words, subsidized domestic services emerged because good jobs at decent wages for low-skilled people became structurally scarcer in advanced economies (Raz-Yurovich and Marx, 2018). By subsidising domestic service work, EU countries aim at solving the tension between the need for more jobs for low-skilled people and the need to improve the qualities of these jobs while at the same time supporting the productive potential of the high-skilled and responding to novel social needs in care and the household.

Critical literature, however, points out that subsidized domestic and care work in Europe has so many side effects that it tends to ‘reinforce or even foster socio-economic inequalities between the low-skilled and the high-skilled, between migrants and non-migrants’ (Morel, 2015: 188). Rather than aiming to up-skill and promote activities with higher added value, the design of subsidy programmes reduces labour costs and de-regulates the labour market, contributing to a ‘dualization’ of the labour market. Jokela (2019) indicates that, despite promoting domestic work to provide job opportunities for low-skilled women across affluent countries, these women are at higher risk of precarious employment, working in non-standard employment settings with high job instability and low wages. Hence, we find a ‘subsidy paradox’ whereby several countries introduced policies subsidising domestic service work, which produces precariousness for workers.

Even in countries with specific regulations regarding paid domestic labour, workers in that sector occupy a precarious position in the labour market. In the Netherlands, domestic workers became excluded from employee social insurance coverage such as unemployment and pensions (Van Walsum, 2011). In Sweden, only the largest domestic service companies have signed collective agreements, and only their workers have access to the full benefits and protections of the formal labour market (Gavanas, 2010). Many – especially immigrants – are exposed to small entrepreneurs and subcontracting chains and work informally for meagre wages (Hellgren, 2015). Germany similarly attracted criticism for increased precariousness due to the extensive use of the mini-job system by domestic workers (Prosser, 2015). With tax credits mainly geared to supporting families (and professional working women) and to promoting the formalization of employment in the domestic services sector, women working in domestic services in Germany are relegated to low-wage, temporary work with little social protection (Jaehrling and Weinkopf, 2018; Shire, 2015). Precariousness is, however, not only associated with regulated markets for domestic services in Continental Europe. The UK, for example, showed an alternative process of formalization of domestic services through private-sector domestic service agencies that marketized and professionalized domestic services. Similar to what happens on the Continent, research for the UK indicates that domestic workers’ services continue to be treated as low status with low pay (Farris, 2018).

The precariousness associated with national subsidy systems for domestic work is paradoxical. The paradox stems in part from a clash of motives behind the design of domestic work subsidies. A system geared to support families attain a better work–life balance (Germany) or stimulate entrepreneurship in domestic workers (the Netherlands) does not focus or is not interested in the working conditions and social protection of domestic workers. Keeping these various subsidy programmes in mind, we study the Belgian case. The question behind our research is whether domestic service schemes and subsidies can at all ‘get things right’. We take Belgium as a case study as its scheme’s subsidy levels are extraordinarily high, and the scheme offers elaborate social protection for workers. In the following section, we analyse the design of the SVS, which will serve as a background to our empirical research questions.

The Belgian service voucher scheme

The SVS needs to be understood in the context of the Belgian labour market. Belgium has one of the most regulated and institutionalized labour markets in the OECD, with extensive collective wage bargaining and comparatively tight legal restrictions on hiring and firing practices, temporary contracts and other forms of non-standard employment. As a result, job quality for the low-educated in Belgium is relatively high compared to other OECD countries. However, the flip side is a strongly segmented labour market characterized by a persistent insider–outsider divide. As low-skilled work is relatively expensive and strongly regulated, many jobs come with strictly defined educational requirements. So, whereas Belgium does not exhibit flexibilization at the margin, the downside is low employment opportunities for those with few skills or with foreign qualifications that are not (yet) recognized (OECD, 2020). Immigrants also face more persistent barriers in the Belgian labour market than elsewhere in Europe, especially immigrant women who come from outside the EU (OECD/EU, 2018). A subsidized employment scheme such as the SVS aims to remedy these low job opportunities for people at the bottom of the labour market.

The SVS emerged in 2004 to create extra jobs for people with no or little formal qualifications, reduce informal sector domestic work and support people’s work–life balance (Raz-Yurovich and Marx, 2018). The scheme has become spectacularly popular: the number of SVS workers grew tenfold in less than 10 years, from about 15,000 in 2004 to more than 170,000 in 2019 (Figure S1 in appendix). The scheme gradually became a significant employer of female workers: nearly 8% of all working women were employed under the SVS. Almost all the employees are female (97%). There are about a million service voucher users in an adult population of around 8 million, with clients being predominantly two-adult households in which both adults are working full-time. The second group of users, which is expanding, includes individuals above 65. Most users are high-skilled and relatively high-up in the income distribution (Raz-Yurovich and Marx, 2019).

The architecture of the SVS exhibits features that may be regarded as a serious attempt to prevent the precariousness of the domestic service jobs it aims to create. First, the SVS is among the most heavily subsidized schemes. A family is entitled to 800 vouchers at the cost of €9 per voucher per year. One voucher pays for 1 hour of domestic work and is used to pay for a clearly defined set of domestic activities such as cleaning, ironing, preparing food, shopping, and transporting persons with reduced mobility. Users buy the vouchers from a private contractor called Sodexo. Each user is eligible for a 20% personal income tax credit, which reduces the actual consumer cost per voucher to €7.2. The voucher is subsidized in two ways. The government subsidizes Sodexo €14.4 per voucher, and the consumer can deduct 20% of the voucher’s price from his or her income tax. This deduction costs the government €1.8 per voucher. Adding up both elements, the government subsidizes €16.2 per voucher or 69% of the total voucher cost (€23.4).

The second important feature of the SVS is its elaborate labour regulations. Domestic workers under the scheme are all formally employed. No worker needs to negotiate her wage as the wages and working conditions (pay, holidays) are set in collective labour agreements that are generally binding and apply to all workers, unionized or not. The level of protection granted to workers in Belgium is, by all international and even European standards, among the highest (OECD, 2021). SVS workers receive a minimum gross wage of €11.6 per hour worked, paid by the service sector company employing the worker, which is higher than the collective minimum wage (€9.7). This wage can be higher depending on the contract terms negotiated between the employer (not the client) and the employee. Other social-legal protections stipulated in the scheme regulations are that workers automatically receive a contract of unlimited duration after three consecutive months of employment with the same company. Additionally, workers are guaranteed a minimum of three subsequent working hours per assignment and 10 h per week. The wage and working hours must remain stable for the entire duration of the employment contract (Mousaid et al., 2017).

A third feature of the SVS additionally shields workers from potentially abusive clients through the ‘three-party’ setup: SVS workers are not employed by individual consumers but by SVS companies. These companies have, in turn, contracts with clients. That has advantages for all parties. Consumers are not dependent on a particular worker, ensuring continuity of service. Workers are not dependent on the wishes and whims of individual consumers either, placing them in a less ‘subservient’ position compared to their counterparts in many other countries. The fact that a company employs them also brings continuity of work for the workers. The employing company has legal responsibilities towards the worker. Critical research, however, indicated that the three-party setup does not always shield women from exploitative relationships, with some companies violating workers’ social rights (Michielsen, 2018). Research showed that especially immigrant workers are also more vulnerable to exploitation by customers, among others, due to their more uncertain residency status (Mousaid et al., 2017).

A fourth aspect of the SVS architecture that renders the labour market position of workers more secure and stable is the workers’ full integration into the social security system. Workers accrue social rights for unemployment, pensions, sickness and disability. It may not be surprising then that researchers found evidence that SVS workers experience a relatively high level of job satisfaction (Michielsen, 2018). Finally, the scheme offers stability and security by not setting any time limits to the use of the scheme by workers. Any worker can be part of the scheme for an indefinite period.

Research questions

The overarching research question is whether the SVS offers a way out of exclusion or precariousness. Kalleberg’s (2018) definition of precarious work indicates the first main feature: the quality of the work itself in which ‘employees bear the risks of work and receive limited social benefits and statutory entitlements’. As we have seen, the Belgian legislation tackles this aspect of precariousness; the SVS performs relatively well in this regard. A second important feature of precarious work is ‘uncertain, unstable and insecure’ work. This aspect is what we examine in the remainder of this article. We want to know if the SVS enables individuals who find themselves at the margins of the labour market in a situation of instability or uncertainty to escape precariousness. We examine whether the SVS can help individuals escape unstable or uncertain work or come out of a situation of exclusion from the labour market due to inactivity or long-term unemployment. Hence our research question is: To what extent does the SVS provide stable and secure employment to its original target group, that is, low-skilled individuals who were previously long-term non-employed or in unstable employment?

For the target group – low-skilled women outside or at the fringes of the labour market – the SVS may be easy to access as it does not require high qualifications. Entering the scheme also allows these workers to gain access to social security benefits, meaning that those with occasional or no employment spells would have a relatively high incentive to enter the scheme and stay inside the scheme. The same is expected regarding those active in the informal economy, yet we cannot trace them due to the nature of informal activities. We expect most women in long-term unemployment and inactivity (including informal work) to find stable employment in the SVS scheme. However, SVS companies face fierce competition and intense pressure to hire the most productive and reliable workers. Such people are more easily found among those with work experience than those with longstanding joblessness. Workers who get attracted by the SVS from another paid job can compare the scheme’s benefits with alternative employment options and may have a lower incentive to stay inside the scheme. We expect these women not to be attracted to stay in the scheme for a long time.

At the same time, there has been an increasing inflow of immigrants in the scheme. Some immigrants might enter the SVS relatively early in their integration trajectory, and their profile may not match the original target group. We expect that SVS jobs could act as short-term stepping stones into regular employment for some of these immigrant workers. However, a study for Flanders demonstrates that options for immigrant women in low-skilled work are much more limited than for men as typically feminine jobs – notably in the care sector – often require high language qualifications (Van Hoof et al., 2020). Such gender segregation in the labour market may force immigrant women to enter the SVS and trap workers with limited language proficiency inside the scheme.

Data and methods

Findings

What pathways do SVS women take?

A first finding indicates that the scheme attracts women with more vulnerable labour market profiles than non-SVS workers (Tables S3–S4 in appendix). We create a typology of career pathways using sequence analysis to explore this further, focusing exclusively on women who entered the SVS between 2004 and 2013. Figure 1 shows the clusters using state distribution plots, providing a good overview of the quarterly distribution of labour market states. Table S5 in appendix labels the clusters from the resulting typology and presents quantitative measures on their size, most typical sequence and labour market state composition. Figure 2 presents the evolution of the average daily wage by pathway type over time.

OPEN IN VIEWER

Figure 1.

Typology of labour market pathways (eight-cluster solution): state distribution plots. Note: this graph presents the time-dependent distribution of the state variable by cluster.

OPEN IN VIEWER

Figure 2.

Average daily gross wage by pathway (cluster) type over time (in years around SVS entry).

Cluster 1: Redirection

The largest cluster, accounting for 28% of the sample, groups together pathways dominated by a transition from a long spell of employment to a long and stable spell of SVS employment. On average, the daily wage in the pre-SVS period decreases sharply prior to entry into the scheme, indicating that some women in this cluster experienced a short non-employment spell before entering the SVS. Subsequently, the average daily wage in the scheme reaches consistently higher and steadier levels than in the pre-SVS period. Notably, the large size of this cluster indicates that the quality of social protection, pay and job stability provided by the SVS compares well with jobs in the regular labour market. Figure S2 in appendix shows that around 27% of the women were doing non-manual jobs in the pre-SVS period, and 23% had a full-time contract. More than half enter the SVS from other sectors, such as manufacturing, trade, catering, other services and education, suggesting that a significant share makes the transition to subsidized employment deliberately, perhaps because of the flexibility in terms of part-time work. Most women who take up SVS also work part-time, showing that women working in the scheme indeed ‘double up’ domestic duties, both in their own home and in their middle-class employer’s home. Domestic service work remains relegated to women’s traditional realm, be it now partly through commodification to the advantage of middle-class women.

Cluster 2: Inactivity exit

Another 17% of women use the scheme as an exit plan from inactivity. The second largest cluster is dominated by transitions from persistent inactivity to long and stable SVS employment – although in some cases with switches back and forth to inactivity. We see a sharp increase in the average wage, which remains relatively constant in the long run. There is, therefore, evidence that the SVS serves as an entry gate into the labour market for the long-term inactive. Moreover, as inactivity also covers informal work, it is plausible that a share of the long-term inactive women in this cluster was actually doing undeclared domestic work.

Cluster 3: Break

A short-term work experience in the scheme, which is preceded and subsequently followed by other types of employment, is also one of the most common career patterns, as it concerns about 15% of women workers. The SVS creates a temporary break in regular employment careers in this cluster. Moreover, these trajectories were more volatile than those in the Redirection cluster, and daily wages prior to SVS were comparatively lower on average. We can assume these women utilized the SVS job as a career stabilizer and a wage booster. Most of all, it is a tool to transition to another job outside of the scheme subsequently. This strategy seems to work well since women secure less volatile and better-paid jobs in their post-SVS careers (Figure S3 in appendix).

Cluster 4: Unemployment exit

This cluster amounts to 13% of our sample and gathers women who develop a stable SVS career after long periods of unemployment and a boost in average wages. In contrast to previous clusters, the average wage after SVS entry progressively declines, reflecting volatile careers. In sum, the scheme appears less effective as a bridge to subsidized employment for women entering volatile unemployment situations than those who have been long-term inactive or employed, possibly due to previous informal domestic work experience. Still, the Unemployment exit trajectory provides a relatively stable job and wage to a large proportion of previously unemployed women.

Cluster 5: Integration

The scheme is often the first entry point to the Belgian labour market for recent immigrants, as it secures regular employment while being easy to access. Before their SVS entry, some immigrants experience periods of inactivity, possibly resulting from their participation in compulsory integration programmes. Immigrants who fare best in terms of a stable SVS career enter the scheme shortly after arrival. At the same time, this relatively quick entry into the scheme is associated with steady income levels over time. This pathway is followed by 10% of SVS workers.

Cluster 6: Return to inactivity

In contrast to the previous pathway types, clusters 6–8 show pathways in which subsidized employment does not result in a stable employment career. In cluster 6, women enter an SVS job after being previously inactive and transition back to inactivity after a relatively short experience in the scheme (over 7% of the sample). SVS employment does not have lasting beneficial career effects for women in this cluster. Similarly, the average daily wage increases significantly after the women enter the scheme but quickly falls back to earlier levels.

Cluster 7: Return to unemployment

In this cluster, women enter the scheme from precarious employment careers, which are interrupted by relatively long periods spent in non-employment. Some women in this cluster might be using the SVS job experience to accumulate the rights for unemployment benefits. For others, the scheme might be sending women into a spiral of non-employment. Cluster 7 is the only cluster where women’s employment and wage situation reach worse levels 4 years after their SVS entry than before. In this regard, this cluster has the least desirable outcome. It affects above 6% of women with SVS experience.

Cluster 8: Challenged integration

The last career pattern is the second one typical for recent immigrants (only 3% of the sample), for whom the SVS is only a short-lived solution. After being inactive, their short experience in the scheme is followed by non-employment and results in an unstable career with low average wages after the SVS exit (just slightly higher than in the Inactivity and Unemployment exit clusters). Compared to the more successful Integration pathway (Cluster 5), immigrants who end up trapped in non-employment enter the scheme from more extended periods of inactivity, probably partly due to the lack of more suitable employment options pushing them to enter the SVS as a last resort.

Who follows which pathway?

We now outline who is more likely to follow each of the above pathways. Table 1 presents average marginal effects derived from a multinomial logistic regression. Marginal effects measure the effect of the covariate on the independent likelihood of being observed in a pathway type at specified covariate values. Here, we estimate average marginal effects (AMEs), which represent the effect averaged across the values of the other covariates in the model.

OPEN IN VIEWER

Table 1.

Factors predicting pathway type (average marginal effects).

	Redirection	Inactivity exit	Break	Unemployment exit	Integration	Return to inactivity	Return to unemployment	Challenged integration
Age (35–49)
18–24	−0.104***	−0.074***	0.065***	−0.004	0.014*	0.021**	0.056***	0.026***
25–34	−0.059***	−0.025**	0.015*	0.006	0.007	0.012*	0.034***	0.010**
50–59	−0.021	0.050*	−0.035*	0.014	−0.018	0.036	−0.012	−0.014*
Household type (Couple with children)
Couple without children	0.048***	−0.056***	0.005	−0.022**	0.041***	−0.033***	−0.008	0.023***
Single parent	−0.047***	−0.141***	−0.003	0.132***	−0.002	−0.021**	0.086***	−0.005
Single	−0.021	−0.135***	0.013	0.077***	0.037***	−0.025**	0.046***	0.008
Other	−0.089***	−0.057**	0.038	−0.014	0.099***	0.004	−0.008	0.027***
Highest level of education (low-skilled)
Medium-skilled	0.049***	−0.035***	0.042***	−0.013*	0.002	−0.024***	−0.025***	0.004
High-skilled	−0.014	−0.009	0.086***	−0.048***	0.010	−0.013	−0.028**	0.015**
Immigrant background (Belgian-born)
EU14	−0.114***	0.069***	−0.084***	−0.024*	0.129***	−0.006	−0.032***	0.062***
EU12	−0.246***	0.115***	−0.165***	−0.137***	0.471***	−0.019*	−0.074***	0.055***
Non-EU Europe	−0.217***	0.167***	−0.101***	−0.092***	0.159***	0.049**	−0.046***	0.083***
North Africa	−0.245***	0.125***	−0.107***	−0.033*	0.121***	0.071***	−0.015	0.083***
Sub-Saharan Africa	−0.198***	0.134***	−0.120***	−0.056***	0.116***	0.064***	−0.045***	0.104***
Asia	−0.174***	0.155***	−0.072***	−0.101***	0.141***	0.036*	−0.041**	0.056***
Latin America	−0.177***	0.043	−0.120***	−0.067***	0.304***	−0.013	−0.065***	0.096***

*p < .05; **p < .01; ***p < .001. AMEs for women with a Turkish and North American background are not reported because of too small sample sizes.

We find two clusters dominated by employment: the Redirection cluster (28%) and the Break cluster (15%) capture almost half of the women in the sample. Table 1 shows that – all other things equal – being medium-skilled and being Belgian-born increases the likelihood of belonging to the clusters dominated by employment. The Break pathway additionally attracts high-skilled women to a greater extent. Being young enhances the likelihood of having a transitional SVS job, whereas being older and living as a couple (with children) increases the probability of staying inside the scheme.

Patterns where long-term inactive women enter subsidized employment, are found across two clusters. The Inactivity exit cluster and the Return to Inactivity cluster account for around 24% of the sample, yet the ‘unstable’ cluster is much smaller in size (7%) than the ‘stable’ cluster (17%). Women with partners, low-skilled and immigrant women, are more likely to follow these clusters. However, EU immigrant women only have a higher probability of following the Inactivity exit pathway, whereas non-EU immigrant women are more likely to follow the Return to Inactivity pathway. We find that the probability of the Inactivity exit pattern is higher among women aged above 25, while younger women are more likely to follow the Return to Inactivity pattern.

As for unemployment dominated clusters, again, we find that the ‘unstable’ Return to Unemployment cluster (7%) is much smaller in size compared to the ‘stable’ Unemployment exit cluster (13%). Being (a) single (mother) and low-skilled increases the chances of being in these clusters, whereas being an immigrant decreases it. Age is not associated with the chances of unemployment exit, but younger women do have a higher likelihood to return to unemployment after a brief period in subsidized work.

Finally, the Integration (10%) and the Challenged integration (3%) clusters are particularly important for recent immigrants. We find that being young and high-skilled increases workers’ probability of following these cluster pathways. Interestingly, EU12 and Latin American immigrant women have a much higher probability of following the ‘stable’ Integration pathway, while non-EU immigrant groups are more likely to be in the Challenged integration cluster.

Discussion and conclusion: A model to be emulated?

Belgium’s SVS came into existence with the noblest of objectives: to offer the prospect of a decent job to women who might otherwise be condemned to a life at the fringes of the labour market or outside of the world of (formal) work altogether. The SVS can be seen as an attempt to rise above the tension between the desire to increase the quantity of the jobs for people at the fringes of the labour market and the need to steer clear of creating low-quality, precarious jobs. This article looks at the labour market pathways of SVS workers and asks: Does the scheme effectively offer a way out of exclusion or precariousness?

We find that a significant share (around 44%) of the workers enter the SVS from the fringes of the labour market, either from long-term inactivity, long-term unemployment, or unstable employment. Going back to Kalleberg’s (2018) definition of precariousness, the scheme offers work that is ‘certain, stable and secure’ to women who were envisaged to benefit from it. Seven out of ten vulnerable workers secure a stable career in subsidized employment in the 4 years following entry. Nevertheless, not all of these workers improve on their situation, as roughly one-third of vulnerable workers end up back at the margins of the labour market. A significant finding is that young women were more likely to follow trajectories where SVS employment does not offer a way out of exclusion or precariousness.

Our analysis also shows that people enter the SVS from much less precarious situations, such as regular employment, often even from continuous employment trajectories. Medium-skilled and Belgian-born workers are most likely to develop such pathways. The fact that a considerable number of SVS workers (around 43%) come from steady non-SVS jobs is clearly at odds with the original objective of offering a stepping stone to those condemned to a life at the margins of the labour market. Most women take on part-time subsidized employment, reflecting the strategies lower-skilled women use to combine their paid domestic service work in middle-class women’s households with their continued unpaid domestic service and care work in their own homes. The commodification of domestic work thus reflects class, race, and gender divides, with subsidies mainly enabling middle-class women to relegate the dirty work to lower class migrant women (Farris, 2018; Romero, 1992). Next to the aim to create legitimate jobs for low-skilled women at the margins of the labour market, an essential motive for the SVS scheme was providing a better work–life combination. Our results indicate that the latter motive only benefits middle-class dual-earner families.

The attractiveness of safe and stable SVS jobs also raises the question of whether SVS jobs are not too attractive. One could argue that a job can never be ‘too good’, but if the consequence is that the prime target group is crowded out, then policymakers are facing a dilemma. There is anecdotal evidence that SVS companies prefer to hire workers with established employment histories. These people are regarded as more productive, reliable, and attuned to the rhythm and demands of formal employment.

Finally, the scheme plays an ambivalent role in the labour market integration process of female immigrant newcomers. Most of the immigrant newcomers enter SVS employment after spending a considerable amount of time in inactivity. This situation seems to correspond to the policy’s target group. However, a significant share of the newcomers – particularly migrants from EU12 countries – enter the scheme soon after arrival. These immigrants might have succeeded in landing a steady non-subsidized job if the SVS had not existed. A substantial share of these EU labour migrants is high-skilled, so it is debatable whether entering the scheme is the best option for them, even if it leads to stable and secure subsidized employment. There might be a risk that EU labour migrants crowd out more vulnerable groups by entering the scheme (see recent evidence for Sweden) (Rickne, 2021). At the same time, for a smaller group of primarily non-EU newcomers, entering subsidized employment as the first job in Belgium does not ensure effective integration into the labour market. Our findings suggest that – at least in the short-term – the scheme does not provide a stepping stone to other professions for female newcomers, which is especially problematic for those who are tertiary educated.

In conclusion, does the SVS offer an example to emulate across Europe? There can be little doubt that SVS jobs are of good quality in terms of pay, social benefits and labour protections. However, this brings problems. Many workers enter the highly subsidized scheme from regular non-subsidized jobs, which was never the policy’s intention. At the same time, many people stay in SVS jobs for a long time. Such career stability is favourable for those who experienced prior trajectories marked by instability, uncertainty, and precariousness. However, it is also true that many predominantly immigrant women with high education levels get trapped in domestic work for which they are effectively overqualified. In that sense, the scheme can be seen as a case of institutionalized second-tier employment. More than half of SVS workers have completed at least secondary education, and around 7% have obtained a higher education degree (but in the case of immigrants, these shares are considerably higher). That is again an indication that the scheme caters only to a limited extent to those envisaged initially as the leading target group: women with limited education and few opportunities in the regular labour market.

The three-party setup of the SVS, in which individual consumers do not act as direct employers of workers, has significant advantages. Consumers can outsource the hiring, monitoring, and control of workers to the SVS company while enjoying continuous service and having no administrative costs or worries. The three-party setup also has significant benefits for the SVS worker. The worker does not have to look for clientele herself, is ensured a steady flow of work, enjoys having her contractual and administrative requirements handled and receives full social insurance benefits. However, at the same time, this setup creates intense competitive pressures among SVS companies competing for business. This situation is good for consumers and favourable for workers fortunate enough to be healthy and productive. Yet, the productive demands this setup brings are less advantageous for the people targeted by the scheme in the first place: low-skilled women at the margins of the labour market.

The SVS can be seen as a case of policy overshooting. A strong case could be made for stricter entry criteria or lower subsidization levels. Given the high current subsidization level, it could be argued to restrict access to service jobs to those coming in from (extended) spells of unemployment or inactivity. One could even contemplate barring entry to those with advanced levels of education unless perhaps they have demonstrable difficulties obtaining a steady non-subsidized job, as in the case of immigrants. At the same time, it would make sense to provide continuous education and (language) training to lower-skilled (immigrant) SVS workers to enhance their chances for upward mobility. All this is unlikely to happen. The scheme has become extremely popular among Belgium’s affluent middle classes, who sway decisive electoral power in a highly competitive political landscape. Consumers care most about the quality and continuity of their services. They care less about the social objectives of the scheme, let alone that they are even aware of those. Any policy intervention jeopardising the continuous delivery of high-quality domestic services will be strongly resented. The SVS has, in effect, become the third rail of Belgian politics: touching it spells political trouble.

Supplemental Material