Introduction

In Asia, fee-charging agenciesFootnote 1 have long been the primary gatekeepers to temporary employment for millions of migrant workers. As such, they are significant actors in the global infrastructural governance of international migration (Goh et al, 2017; Xiang & Lindquist, 2018). In the past decade, policymakers have increasingly associated recruiters’ practice of charging fees to migrants with debt bondage and other poor employment conditions (UNDOC, 2015; ILO, 2015b). In response, ‘abusive recruitment’ has catapulted to the top of international policy agendas, largely in the slipstream of anti-trafficking advocacy. Recruiters, especially those in migrants’ origin countries, have become subject to new forms of policy governance and business codes of conduct. Most notably, the Global Compact on Safe, Orderly and Regular Migration (GCM) Objective 6 aims to ‘facilitate fair and ethical recruitment’ in the context of ‘safeguarding the conditions that ensure decent work’ (Assembly, 2018). From the ILO’s perspective, decent work is generally understood to include, at a minimum, the four fundamental rights to freedom of association, elimination of forced labour, abolition of child labour and the elimination of discrimination in respect of employment and occupation. However, to date, agencies’ responsibility for systemic and discriminatory racialised and gendered employment patterns remain largely ignored by policymakers. This paper argues that recruitment and discrimination should be on the agenda for advocating for decent work for migrants; the neglect of this issue is a substantial obstacle to the advancement of a rights-based and socially fair approach to the global regulation of worker migration.

Over the past two decades, structural discrimination has increasingly been recognised as a pervasive factor driving global inequality (World Bank, 2020). It traps people in low-paid, ‘informal’ economy jobs and is a recognised risk factor in forced and child labour (ILO, 1999). At the same time, structural discrimination, by class, gender, race and ethnicity, remains firmly entrenched in low-wage temporary labour migration, including for migrant domestic work (Truong, 1996; Lan, 2006; Parrenas, 2012; Piper et al., 2017). In part this is because, for states, retaining a distinction between the rights accorded to citizens and those to migrants (non-citizens) is an essential component to sovereignty (Ruhs, 2012). Regulations that grant differential rights to residence, work and welfare based on immigration status de facto systemically discriminate on the basis of national origin, and often additionally on the basis of sex, ethnic and social origin (Guild & Mantu, 2011).

Discriminatory regulation and employer hiring practices are well-documented in the literature on migrant domestic work in Asia (Constable, 1997; Lan, 2006). In particular, numerous actors, including the ILO, have challenged the governance practice of ‘banning’ female migration for overseas domestic work (ILO and GAATW, 2015). However, the specific role of placement agencies in brokering structural discrimination as a core part of their business—rather than as a passive response to employers’ hiring preferences—has thus far been relatively neglected by scholars as well as by policymakers.

This paper responds by drawing on data from a qualitative research study conducted with migrant domestic work agencies in Jordan, Lebanon, and Bangladesh between 2013 and 2015. At the time of the research, Bangladeshi women had recently become the most significant of all nationalities in this occupation in these two countries. The paper shows that placement agencies in Amman and Beirut recruited and supplied Bangladeshi women as the cheapest domestic workers to their clients. In so doing, agencies profited from engaging in racialised and gendered discriminatory recruitment and employment practices. This paper thus advances our understanding of agencies’ regulatory roles in transnational migrant labour markets (Jones, 2014; Shire, 2020) and as infrastructural governance actors (Goh et al, 2017; Xiang & Lindquist, 2018).

The argument unfolds in three stages. The first section embeds the paper in the existing literature on recruitment and discrimination. The following section presents the main relevant international norms on non-discrimination and the ILO’s Fair Recruitment Initiative, a flagship programme in the Organisation’s approach to the global governance of migration. The subsequent sections set out the empirical findings, evidencing agencies’ systemically discriminatory roles in sourcing women, marketing them to clients, and establishing their ‘financial value’ in the labour market. As agencies’ activities impacted on Bangladeshi women’s position in the labour market, including their pay and ability to organise (Cheng, 2003; Lan, 2006; Bakan & Stasiulis, 1995), the paper concludes with a challenge to international policymakers to tackle this important rights issue.

The discriminatory practices of the recruitment industry: the missing piece in the story

Domestic work placement agencies are a form of labour market intermediary, brokering employment for a fee (Coe et al., 2010). For employers in Asia, they serve as the ‘front office’ of extended global networks of actors enabling them to access workers often millions of miles away, with whom they lack a shared language and culture. For workers in Asia, they are often the ‘back office’ of extended local and transnational networks of informal, community or family-based networks which facilitate their mobility. Placement agencies, together with their subcontracted partners in migrants’ origin countries, arrange the end-to-end logistics of labour migration, facilitating travel, visas, work permits and accommodation, as well as negotiating the employment placement and contract. Activities are organised according to a strict division of labour: it is only the placement agencies—the topic of this article—which deal with the clients. Contracted recruitment agencies and/or brokers in migrant origin countries mobilise workers ready for placement up to the point of their departure to the country of employment (Deshingkar, 2019). In many countries in Asia, using the services of an agency is increasingly legally required for both employers and migrants alike and to not comply can render migration irregular (Jones et al., 2017). Even when their usage is not legally required, migrants in Asia are often still dependent on a range of legally licensed or informal brokers to organise the increasingly complex regulatory logistics of migrating. In this region, in both origin and destination countries alike, agencies are, therefore, both ubiquitous and structurally embedded in the political and economic governance of temporary migration (Goh et al., 2017; Guevarra, 2009; Rodriguez, 2010; Xiang & Lindquist, 2018).

Global recruitment industry representatives proudly claim that they provide a valuable service in matching un- or under-employed workers with unfilled employment opportunities (WEC, 2020). However, matching labour supply and demand is never a neutral activity (Peck & Theodore, 2001). Therefore, what agencies do and how they do it, matters. As mediators of migration and employment, agencies’ activities have consequences for individuals as well as wider societal impacts.

A growing and ethnographically rich migration studies literature conceptualises migrant ‘brokerage’, primarily in migrants’ origin countries (Deshingkar, 2019). These studies address the social embeddedness of brokerage into local cultural contexts, the importance of migrants’ own agency in interacting with brokers and the ambiguity of brokers’ actions (Awumbila et al, 2019). However, this research explores brokerage through the lens of migration, largely ignoring the role of intermediaries as transnational labour market actors. Indeed, the role of placement agencies as regulatory actors in gatekeeping entry to (destination country) labour markets has to date been comparatively neglected by scholars (Goh et al., 2017).

Research on the recruitment of migrant domestic workers has explored some aspects of how agencies behave in discriminatory ways. Scholars have analysed how recruiting agencies in migrants’ origin countries shape domestic workers’ ‘on the job’ performance, including of docility and other related subjectivities (Guevarra, 2009; Rodriguez, 2010). In destination countries, in Asia and beyond, multiple studies have shown how (childcare and domestic work) agencies pitch candidates to their (employer) clients using idealised and essentialised personality characteristics such as docility, obedience, industriousness and subservience (Pratt, 1999; Bakan & Stasiulis, 1995; Lan, 2006; Constable, 1997), the very same subjectivities shaped by agencies in migrants’ origin countries. These stereotypes—deemed by agencies to be those which are most sought after by employers—are commonly depicted by them as being inherently associated with specific nationalities of women. It is these essentialised stereotypes, serving as shorthand to convey migrant women’s suitability for tasks, that agencies ‘sell’ to their clients rather than women’s skills or expertise. In effect, agencies promote an “externally defined identity that is purchased by the employer” (Tyner, 1999: 198), an idealised migrant subject.

Agencies’ artificial distinction between candidates also serves to create the illusion of choice for their clients (ILO, 2015a). For instance, agencies may argue that a particular nationality of candidate they have available is more subservient or more passive or docile than another (Cheng, 2003; Lan, 2006). This allows agencies to claim to their clients that one candidate is of better ‘quality’ than another or more likely to be compliant than another, although such characterisations are often deployed for whichever candidate the agency is marketing at that time (Pratt, 1999). In this way, agencies inflate and deflate women’s ‘value’ which impacts on the wages that women are paid (Bizri, 2014; Pratt, 1999). Consequently, women of a particular nationality may face their demands for higher wages being displaced by agencies’ devaluing discourse (Pratt, 1997). This is an example of how agencies’ actions impact on the labour market as a whole rather than only on individual workers (Peck & Theodore, 2001).

Agencies have a powerful financial incentive for complying with employers’ pre-existing prejudices: to ignore them would mean losing income as they, in effect, ‘sell’ (the labour of) domestic worker candidates (Bakan & Stasiulis, 1995). This highlights that the process of recruitment of paid domestic workers is influenced by gendered and racialised divisions of labour, including employer hiring prejudices, within national as well as transnational labour markets (Lan, 2006; Truong, 1996). States and employers are actively implicated in these processes as well as agencies and migrant women (Lan, 2006; Rodriguez, 2010). The activities of placement agencies are influenced by multiple actors within transnational migration markets (Jones, 2014; Shire, 2020).

However, agencies do not only structure expectations of workers and clients, they target specific groups of workers for particular jobs. They seek to generate employer demand for racialised and gendered workers as they profit from discrimination (Peck & Theodore, 2001). As yet, this issue is left unexplored by the migrant domestic work literature. This paper, therefore, also draws on a wider economic geography and sociology literature that theorises the role of temporary staffing agencies—a particular form of placement agency—in targeting women, migrants and people of colour, for low-end jobs (Jones, 2014; Peck & Theodore, 2001; Vosko, 2000).

In low-wage labour markets, workers’ willingness to work in poorly paid jobs with exploitative conditions often matters to employers (and recruiters) more than their skills, qualifications, or job experience (Burawoy, 1976; Waldinger & Lichter, 2003). Their willingness may result from a lack of alternative, better, job opportunities (Peck & Theodore, 2001), a lack of prior experience (Vosko, 2000) and/or their migrant status (Jones, 2014). These factors mean certain groups of workers have relatively less leverage to negotiate better wages and conditions. Agencies are expressly aware of these factors and seek to exploit them to generate business and, therefore, turn a profit (Jones, 2014). This is illustrated by the following three examples.

First, Leah Vosko (2000) forensically analyses how agencies in mid-century Canada deliberately targeted married female homemakers to supply to employers in the then newly expanding white-collar clerical sector. Rather than being driven by pre-existing employer demand for a female workforce, it was agencies which sought out this new group of workers. Agencies embarked on marketing their newly created workforce using the ‘Kelly Girl’ imagery, pitching women to employers as cheaper, more flexible and productive than the existing primary workforce (men), as well as more attractive. In other words, agencies artificially created a market demand for a (married) female workforce. This changed the shape of the national labour market in Canada.

Second, in the U.S. in the 1990s, Peck and Theodore (2001) documented how agencies in Chicago, deliberately targeted unemployed men of specific ethnicities, primarily men of south American and Mexican origin for low-paid day-labour jobs on construction sites. Agencies opened recruiting offices in neighbourhoods, where these communities tended to live, advertising jobs that did not require any prior experience, qualifications nor any formal recruitment process. They also supplied vans to transport the men to construction sites elsewhere in the city. Agencies sought out these previously unemployed workers, because they, like the married female homemakers in Canada, were viewed as cheap, flexible and productive. Third, Jones (2014) found that after the 2004 expansion of the European Union and consequent introduction of free movement to the UK for nationals of the new member states, British agencies opened recruiting offices in working class neighbourhoods in poorer regions of Poland. Agencies installed road-side billboards to advertise opportunities to work in low-paid jobs in agriculture, horticulture and in factories in the UK. Although agencies had access to un- and under-employed workers locally, they opted to recruit from Poland in the expectation that workers from there, many of whom lacked English language skills and knowledge of how to find alternative, better paid, jobs in the UK, would be cheap to hire and more ‘productive’ than UK nationals.

In all three examples, agencies were shown to have made strategic business decisions about which groups of workers they would target and channel to low-end jobs. They used workers’ migrant status, sex, and ethnicity as markers of potentially cheap and flexible workforces which they could market to employers. In other words, agencies go beyond simply passively meeting pre-existing employer prejudices but actively seek to generate demand for specific groups of workers (Peck & Theodore, 2001). This explains the drive to recruit migrant workers from poorer countries: agencies’ ability to find and supply flexible workers is a core component of the service they sell to clients (Jones, 2014). To be competitive, agencies must continually seek groups of workers who can be supplied to clients as cheaper and more flexible than others (Peck & Theodore, 2001). This makes structural discrimination a core, but largely hitherto under-recognised, part of their business model in transnational domestic worker labour markets. This article applies a discrimination lens to the end-to-end recruitment process of migrant domestic workers situated within international norms on non-discrimination.

The principle of non-discrimination and international recruitment initiatives

Non-discrimination is a foundational principle of all international human and labour rights, from the Universal Declaration on Human Rights (1948), the International Covenant on Economic, Social and Cultural Rights (1976) and ILO’s Fundamental Principles & Rights at Work (1998). The ILO's principal instrument concerning discrimination in respect of employment and occupation, C111 Discrimination (Employment and Occupation), 1958 refers explicitly to the following seven prohibited grounds of discrimination: race, colour, sex, religion, political opinion, national extraction and social origin (Article 1). The Convention provides that each ratifying state may include in its legal proscriptions such other grounds of discrimination as it considers appropriate, after consultation with employers' and workers' organizations (Article 1). However, this is not always applied to distinctions, exclusions, restrictions, or preferences made by a State party between citizens and migrants. For instance, the UN Convention on the Elimination of All Forms of Discrimination against Women (1979) (CEDAW) applies to migrant women (scope affirmed through General Recommendation 26 on Women Migrant Workers, 2008). Yet, the UN International Convention on the Elimination of All Forms of Racial Discrimination (CERD) (1965) does not (necessarily) apply to migrants. For governments, the ability to legally discriminate both between citizens and non-citizens in terms of rights regarding entry to the territory, residence and employment underpins state sovereignty (Ruhs, 2013). This principle was reaffirmed within the text of the Global Compact on Migration (Assembly, 2018).

For this reason and as countries generally have special regulations governing the employment of foreign nationals, discrimination against migrant workers tends to be dealt with in the three international Conventions on migration and their accompanying Recommendations. The UN Convention on the Protection of the Rights of all Migrant Workers and Members of their Families (1990) places non-discrimination at its core. Provisions on remuneration, hours of work, holiday pay, health and safety, freedom of association and social security, apply “to all migrant workers and members of their families without distinction of any kind such as sex, race, colour, language, religion or conviction, political or other opinion, national, ethnic or social origin, nationality, age, economic position, property, marital status, birth or other status” (Article 1.1). The ILO’s two migration Conventions and accompanying Recommendations—C97 and R086 Migration for Employment (Revised) (1949) and C143 and R151 Migrant Workers (Supplementary Provisions) (1975)—likewise apply the standard of non-discrimination between migrants and nationals in respect of remuneration for work of equal value, access to trade unions and social security (Article 16). C143 allows for restrictions on certain aspects of the employment of foreign nationals (particularly employment in posts connected with the interests of the State), but they limit to two years, as a rule, any other restrictions on freedom of choice of employment. C143 also specifies the minimum equality of rights which must be enjoyed even by migrant workers who are in an irregular situation.

ILO C189, Domestic Workers (2011) is often hailed as a successful outcome of a decade’s long advocacy campaign conducted by civil society organisations and trade unions, as well as a good example of developing a decent work agenda for migrants (Rosewarne, 2013). C189 extended labour standards associated with formal labour markets, to an informal and private form of work, often conducted by migrant women. Non-discrimination is prominent within it: the Preamble notes that: “domestic work continues to be undervalued and invisible and is mainly carried out by women and girls, many of whom are migrants or members of disadvantaged communities and who are particularly vulnerable to discrimination in respect of conditions of employment and of work, and to other abuses of human rights.” (Italic author's own) Its provisions thus require signatory states to eliminate discrimination in respect of employment and occupation, including in relation to social security. The convention did not, however, establish the standard of equal rights between citizens and non-citizens in this regard. While it did provide that remuneration should be established without discrimination based on sex (Art. 11), it did not extend this to discrimination based on national origin, race, or ethnicity.

The ILO did not comprehensively address the activities of recruitment and placement agencies in relation to employment until the adoption of C181, Private Employment Agencies, in 1997 (Vosko, 2000). Reflecting the wording of C111, C181 specifies that: “[i]n order to promote equality of opportunity and treatment in access to employment and to particular occupations, a Member shall ensure that private employment agencies treat workers without discrimination on the basis of race, colour, sex, religion, political opinion, national extraction, social origin, or any other form of discrimination covered by national law and practice, such as age or disability.” (Article 5.1). Non-binding accompanying Recommendation R188 further requires that “Private employment agencies should be prohibited, or by other means prevented, from drawing up and publishing vacancy notices or offers of employment in ways that directly or indirectly result in discrimination on grounds such as race, colour, sex, age, religion, political opinion, national extraction, social origin, ethnic origin, disability, marital or family status, sexual orientation or membership of a worker’s organization.” (Art. 9). Discrimination based on national extraction (rather than national origin) means distinctions on the basis of place of birth, ancestry or foreign origin and is not necessarily related to citizenship. In other words, the sole international Convention and accompanying Recommendation that explicitly addresses recruitment clearly provides for non-discrimination, but not specifically in relation to distinctions between nationals and migrants.

Cross-border recruitment is barely mentioned in C181’s provisions, as its primary focus was on the casualised agency employment which had become endemic in north American and European labour markets (Vosko, 2000). Consequently, during the subsequent decade, as ‘abusive recruitment’ emerged onto international policy agendas, the ILO came to view this as a substantive gap in the global governance of recruitment and migration (ILO, 2015b). In response, in 2014, the ILO Director-General Guy Ryder unveiled the global Fair Recruitment Initiative as a key pillar of the ILO’s Fair Migration Agenda. The Initiative had three objectives: (1) to prevent human trafficking; (2) to protect the rights of workers from abusive and fraudulent practices during the recruitment process; and (3) to reduce the costs of labour migration and enhance development outcomes. The ‘global migrant recruitment market’ was, the ILO claimed, ‘anarchic and in need of proper coordinated regulation to ensure fair and ethical hiring of workers from all countries’ (ILO, 2014).

In 2016, the ILO published the General Principles and Operational Guidelines (GPOG) for Fair Recruitment (ILO, 2016b), later embedded within the GCM (Assembly, 2018). The GPOG operationalised, what had come to be referred to as, ‘fair recruitment’ in international labour standards and human rights, with a specific reference to the ILO’s Declaration on Fundamental Principles & Rights at Work (1998). Discrimination is addressed in GPOG General Principle 1: “Governments have an obligation to respect, protect and fulfil internationally recognized human rights, including fundamental principles and rights at work, and other relevant international labour standards, in the recruitment process. This includes respect for, and protection of, the right to freedom of association and collective bargaining, and prevention and elimination of forced labour, child labour and discrimination in respect of employment and occupation”. However, the non-discrimination clauses as worded in C181 (Article 5) and R188 (Article 9) were missing as were references to the forms which discrimination might take as outlined in C111: namely, sex, national extraction, social origins, ethnicity, race and religion. In addition, while the GPOG does reference the migrant Conventions, C97 and C143, it does not specifically reflect the non-discrimination articles within them, instead adding a rather vague sub-clause to General Principle 1: “This obligation applies with respect to all workers recruited into, within or from their territory and/or jurisdiction.” This article contends that the standard of non-discrimination (as set out in C97 and C143) is an essential component to decent work for migrants and as such should be applied more strongly to private sector recruitment. The following section outlines the research design and context to the research.

Research design and context

For the empirical case-study sections, this paper draws on material gathered via a total of 156 qualitative, in-depth, semi-structured interviews undertaken by an Arabic and Bangla-speaking research team. This formed part of a multi-country research project to analyse recruitment firms’ strategies and the resulting impact on workers’ employment conditions. The research included 59 interviews with owners or senior managers of licensed recruitment agencies in Amman, Jordan (18), Beirut, Lebanon (21) and Dhaka, Bangladesh (20). Interviews lasted between 40 min and 3 h. Interviewees were sampled from among the legally licensed agencies in each country (Jordan N = 100, Lebanon N = 700, Bangladesh N = 28). The sample were all small businesses, employing between two and ten members of staff. As is common in this sector, they, therefore, relied on an extensive network of recruitment agencies in the capital cities of other countries, including Dhaka, to source women willing to live and work in private households (ILO, 2015a).

A semi-structured interview guide facilitated data-collection about agency histories, recruitment processes, business partnerships, finances, relationships with clients and with government departments. Thirty supplementary interviews with informal brokers (dalals) in Bangladesh were also conducted, 30 with returned women migrants 37 with government officials, UN and NGO representatives in the three countries to elaborate the context to the recruitment. The article draws on analysis of all the empirical data. However, as the article’s focus is the structural discrimination which occurs in the countries of destination (the ‘demand’ side), quotes are primarily derived from interviews with agencies in Jordan and Lebanon.

In all three countries, the recruitment and employment of migrant domestic workers was deeply gendered and racialised as well as primarily facilitated by placement agencies. In Jordan and Lebanon, by 2013, employers disproportionately hired migrant women from South and Southeast Asia as live-in domestics, and especially from Bangladesh, via agencies (Frantz, 2013; ILO, 2015a; Pande, 2013). In Jordan, employers were legally required to hire a migrant domestic worker through an agency. In Lebanon, while not a legal requirement, most employers continued to recruit via agencies for practical and cost reasons (ILO, 2015a). This form of employment was also, for many households, a practical and economic necessity given the lack of public health and social care systems (Raghuram, 2012). In effect, the recruitment of migrant domestic workers was encouraged by the state to subsidise the lack of publicly available health and social care.

As elsewhere in the region, most migrant domestic workers ‘lived-in’ on 2-year employment contracts, their immigration status ‘tied’ to their employers (the ‘sponsor’ or ‘kafeel’). Women were paid often extremely low wages, subjected to long—often 24/7 on-call employment—working hours, and multiple and sometimes invasive forms of exploitation (Amnesty, 2019; Frantz, 2013; Pande, 2013). Migrant domestic workers were omitted from coverage of national Labour Codes in either country, hence were not covered by employment protections extended to citizens, including the right to non-discrimination. This included a ban on participation in trade unions. Over the 2000s, global civil society as well as within Lebanon and Jordan grew increasingly vocal about the role of placement agencies in women’s abuse. These campaigns, however, focused on the need for greater protection rather than the inherent discrimination within the recruitment business. In particular, campaigns and international policy coalesced around the recruitment fees paid by women and recouped through their salaries or loan repayments at high interest rates which meant that women faced working for several months for no income, unable to remit monies home (ILO, 2015a).

In Bangladesh, overseas domestic work provided a vital economic lifeline for women who found it challenging to find economic opportunities at home (Danneker, 2009). At the time of the research, Lebanon and Jordan were especially important destination countries for Bangladeshi women (ILO, 2015a). It was, however, prohibited by Bangladesh until 2003, when it was partially lifted for women aged between 25 and 45 albeit with some highly gendered conditions. Women were required to submit a permission letter from a male guardian to the authorities to be allowed to migrate and had to attend (and pay for) a lengthy residential training course in domestic work (Siddiqui, 2011). By 2012, almost 40,000 women migrated for overseas domestic work annually, accounting for approximately five percent of the total official (regular) migration. However, the unofficial total, including irregular migration, was likely to be far higher (ILO, 2015a).

After 2003, the Bangladesh government actively sought to negotiate Memorandums of Understanding (MoU) with the governments of the main destination countries with the aim of protecting its overseas nationals, especially women (ILO, 2015a). These also served to ‘open’ destinations for Bangladeshi women to migrate to. By 2015, Bangladesh had signed 12 agreements on labour migration, including with Jordan (although not with Lebanon). To be consistent with Bangladesh law, the MoU with Jordan stipulated that only Bangladeshi women aged between 25 and 45 could be recruited by employers who were required to bear the full cost of women’s migration, including air fares and visa fees. To be consistent with a recently adopted Jordanian law, the agreement also required employers to provide women with suitable accommodation, food and to cover their medical needs via a life insurance policy. Moreover, employers were required to open a bank account for women and to deposit their salary monthly. However, unsurprisingly, the agreement lacked any enforcement measure, and its provisions were widely flouted (ILO, 2015a). The following empirical case study is utilised to illustrate how placement agencies in Jordan and Lebanon, with the assistance of those in Bangladesh, engage in discriminatory mechanisms to turn a profit.

Discriminatory marketing of Bangladeshi women

On the streets of Lebanon and Amman images on agency door-signs, on billboards and in newspapers, depict brown or black women who are often in maid’s outfits. As the clients, also mainly female, entered their offices recruiters questioned them about their needs. How many people live at home? Are any of them children or elderly? Has the client just married? Does she have a job outside the home? Does the client already know which nationality of domestic worker she wants? Using this information to highlight candidates who might be of interest, agents presented their clients with a portfolio of bio-data forms containing details about candidates’ nationality, name, age, level of education, marital status as well as an accompanying photograph. Serving as a Curriculum Vitae some forms also include some limited information about candidates’ experience in household tasks, such as cleaning, sewing, shopping, care of children and elders. Specific skills and experience for the job, however, according to recruiters, featured little in these conversations. These were not important to their clients, as one put it: “I provide someone that they can mould and adapt as they like.” (Office Manager for 18 years, Amman). For this reason, women’s education history or foreign language ability was not regarded as important: “if she is educated a bit then this means she can comprehend what is being asked of her.” (Agency Owner, 10 years, Beirut). This was regarded as sufficient.

However, these statements also reflected that agents knew little about their candidates, usually having never met them. Their sales pitch to clients instead focused on candidates’ nationality which enabled them to construct a façade of matching suitability between employee and employer:

“It depends on the conditions of the family. For example, if the client is almost 70, and his father is 100, I will recommend Bangladeshi of 43 [years old]. She knows how to cook, bake, and has probably worked in Jordan before, so knows a bit of Arabic…. Bangladeshis are especially suited to managing the household.” (Agency Owner, 15 years, Beirut)

For their sales pitch, agents also drew on what they could discern from candidates’ photographs: their age and looks, which they associated to nationality. Agents explained to their clients that if the woman appears too thin or indeed too large, that she would be unlikely to be fit enough to carry out her tasks. They also debated whether their clients would prefer a nationality of woman who would ‘look good’ in their service, which could, according to interviewees, add to the reputation, position and status of the family household, or whether their clients would feel threatened by an attractive employee: “Pretty/not pretty is also a factor. Some women [employers] feel intimidated by pretty domestic workers; but others want them to look nice.” (Owner, Mount Lebanon, Beirut). What agents (and their clients) defined as “pretty” was related to skin colour; this also was related by them to women’s nationality, a sign of employers’ pre-existing prejudice.

“I tell them Bangladeshi are fair anyway, but Sri Lankas come very dark and they prefer a fairer complexion because the kids may get scared. In this case, Bangladeshi is better.” (Agency Owner, Amman)

Agents emphasised their Bangladeshi candidates’ obedience, docility and trustworthiness to their clients, characteristics nevertheless viewed as globally universal requirements for live-in domestic workers (Bakan & Stasiulis, 1995; Constable, 1997; Cheng, 2003; Lan, 2006). For instance, one interviewee explained to his clients that women from Bangladesh tend to have “long patience and stay at home” (Recruiter, Hazmieh, Beirut) by which he signalled that Bangladeshi women were obedient and would not seek to leave the employer’s house. Another interviewee advised his clients that Bangladeshi women “are good and calm” (Recruiter, Mount Lebanon) and would be, therefore, content to take orders from Madam (the client) without complaint. These ‘characteristics’ were often juxtaposed with the supposed negative ones which interviewees associated with other nationalities:

“With Bangladeshi domestic workers, it is the women that are decision-makers, unlike Sri Lankan or Filipino women. It is the women there that rule in the household and even get their way with agencies. She has authority.” (Office Manager, Hazmieh, Beirut)

In this instance, ‘authority’ was presented as a positive trait. Which ‘characteristics’ they associated with particular nationalities depended not just on what ‘type’ of woman this client wanted, but also served as a strategic mechanism to persuade clients to hire their own recruits rather than those of their competitors (Bakan and Stasiulus, 1995; Constable, 1997). Agents used the information they had gleaned at the start of the conversation to promote the nationalities of candidates they had available at that time (Pratt, 1999).

In Jordan and Lebanon, the selection process took place entirely without candidates’ involvement. Unlike in other countries in Asia (e.g., see Constable, 1997 on recruitment in Hong Kong) agencies did not offer their clients the possibility to conduct interviews with potential candidates over Skype or telephone. In making hiring decisions, clients were, therefore, wholly reliant on the rhetoric deployed by their chosen agent, who positioned themselves as ‘experts’ in specific nationalities to justify their value to clients (Jones, 2014). This was because the success of their business was contingent on persuading clients that their actions added value beyond simply placing job adverts (Constable, 1997): “I’ve been recruiting from Bangladesh for 3 years so have a good sense of what the quality of the women is like.” (Recruiter, Amman).

In summary, agents encouraged their clients to think that they should select a ‘type’ of domestic worker not on skills or experience but according to her national origins (Lan, 2006). This matters not just because it is an example of prejudicial hiring—although it does, and it is. It also matters, because recruiters’ discriminatory marketing of domestic work candidates impacted on how Bangladeshi women were ‘priced’ in the labour market by agencies and by employers. How agencies describe their candidates to clients has consequences for workers. As Geraldine Pratt (1997), writes, if women are constructed as other than employees and their jobs as not quite jobs, their lesser wage levels and conditions are legitimated. In Jordan and Lebanon, agencies categorized Bangladeshi women at the bottom of the fee and wages hierarchy as the following section shows.

Pricing Bangladeshi women

At the time of this research, interviewees in Jordan and Lebanon charged their clients fees of between US$1,300 and US$5,600 to recruit migrant women. Prices were differentiated according to candidates’ nationalities (Table 1) with the highest fees set for recruiting women from the Philippines and the lowest for women from Bangladesh. This table demonstrates the central role which agencies played in constructing a hierarchy of financial values placed on migrant women (see also Bizri, 2014).

Table 1 Recruitment fees charged by nationality
Full size table

Interviewees argued that in part the fee hierarchy reflected the actual costs involved in locating and transporting women from their home country to Lebanon or Jordan. These costs included the cost of flights, visas, permits, as well commissions that were paid to recruiting agencies in Bangladesh (and other countries) to find women and email their bio-data forms for placement agency portfolios. However, like any private sector business, agencies also set their prices according to what they thought they could feasibly charge their clients. This tended to reflect how desirable a particular nationality of candidate was to clients at any one time. Interviewees in both countries explained that they charged more for candidates from the Philippines, because they were highly sought after by the upper classes (see also Pratt, 1999; Guevarra, 2009; Lan, 2006). They, therefore, explicitly looked for clients from affluent neighbourhoods to supply candidates from this country to. These were not, however, the clients to which agencies in either country pitched Bangladeshi women. Instead, agents actively promoted Bangladeshi women to those clients who they perceive to be of lower income status and for whom the cost of recruiting candidates from the Philippines was out of reach.

Interviewees explained that they assessed potential clients’ likely household financial circumstances as they walked through their office door. They judged this through prior knowledge of this client, how they were dressed, which area of the city they lived in, household size and through responses provided to their initial set of questions about their own employment status and household needs. Their assessment of how much money a particular client could afford to spend informed which nationality of candidates they marketed to them. A well-dressed client, wearing jewellery who described a large house might be pitched a candidate from the Philippines. In contrast, Bangladeshi women were pitched to working class clients:

“We advise them on who to recruit depending on their conditions at home and their financial needs which determines what nationality to choose; Bangladeshi domestic workers are cheaper.” (Owner, Amman)

This enabled recruiters in Jordan and Lebanon to profit from the supply of domestic workers to even the most financially constrained clients, of whom interviewees were often dismissive: “They want something cheap. They don’t have money to eat but they want a domestic worker. I can give them a Bangladeshi domestic worker.” (Female owner, Ain Remayni, Beirut) Another asserted that: “if they [employers] have financial constraints, it will have to be a Bangladesh domestic worker.” (Agency Owner, Mount Lebanon, Beirut). Put simply, placement agencies enabled even clients with low incomes to have private home-based help and care as a coping strategy in the lack of state support (Raghuram, 2012). For agencies, this provided an additional market for them to which they could supply Bangladeshi women.

To justify their racialised fee hierarchy to the researchers (Table 1) as well as to their clients, agents argued that Filipino women were ‘high quality’, described as educated, light-skinned and highly trained. On the other hand, Bangladeshi women were spoken about as ‘low quality’, even as agents sought to persuade their clients that they were especially suitable as domestic workers. The following quotes were typical:

“The quality of the Bangladeshi is less than Indonesian and Filipino domestic workers so that only employers that are really in need will employ one.” (Owner, Amman)

“Bangladeshi domestic workers are without a doubt the worst nationality to recruit from... You can think of it as one Sri Lankan domestic worker equals four Bangladeshi domestic workers. Their minds are very primitive.” (Owner, Amman)

Agencies in both Lebanon and Jordan commonly deployed the term ‘primitive’ to describe Bangladeshi women to rationalise their place at the bottom of the labour market hierarchy (see also Cheng, 2003). A blatant but nevertheless normalised comment in this context, the term ‘primitive’ enabled agencies to express their views that women from Bangladesh were uneducated. It was also undoubtedly a racialised and classed comment. Despite this, they also reassured their clients that Bangladeshi women could still do the job:

“I tell them that 90% of the time Bangladeshis get married young and this can be okay. They usually are educated until 7th grade; I tell them this is ok. I tell them if she – the Bangladeshi - is from a rural area, then she is primitive, and won’t know how to use appliances or to clean well, but she probably won’t run away. So that’s good.” (Agency Owner, Amman, Jordan)

After selecting a candidate, agencies required their clients to pay a deposit, with the balance of the fee payable on arrival of the new employee. This could be as little as US$100.

“I don’t take the full price from the employer right away. I take a down-payment, just anything – could be as little as US$100. Once the employer receives the girl, she has to pay the remaining.” (Office Manager, Beirut)

This racialised hierarchy of fees was also reflected in the wages which agents advised their clients should pay to their new employees; as with the selection of candidates, workers were removed from this negotiation. In Lebanon, an interviewee reeled off to the researcher that women from the Philippines would ‘cost’ US$4,500 to recruit and should be paid US$250 to 300 in wages; Ethiopian domestic workers would cost US$2,300 and should be paid wages of between US$150 and US$200 per month; domestic workers from the Cameroon would cost US$2,000 in recruitment fees and should be paid US$200 a month during the first year. Bangladeshi women, interviewees in both Jordan and Lebanon, advised their clients, should be paid between US$100 and US$150 per month. Agencies’ roles in categorising women at the bottom of the hierarchy, therefore, enabled employers to pay them the least (Pratt, 1999). This emphasises the material consequences for women of agencies’ racialised and gendered marketing narratives. The following section illustrates how agencies deliberately started to recruit from Bangladesh specifically, because this country was identified as a source of cheap(er) female labour.

Discriminatory sourcing of Bangladeshi women

Agencies included in this research only ever recruited female migrants for domestic jobs such as cleaning, caring, shopping, and cooking; this was so taken for granted by interviewees that a question about it prompted surprise. None had ever supplied either a man or a Lebanese or Jordanian citizen for live-in domestic jobs, although male migrants were sometimes supplied for gardening, household security or driving jobs. In Lebanon, interviewees conveyed that in the long-ago past some offices had supplied Lebanese women as part-time, live-out, housekeepers, as well as women from Syria, Egypt and the Kurdish territories, but that this business had long since disappeared. In Jordan, agents interviewed could not recall that Jordanian women had ever been hired as domestics. Agencies had switched to recruiting migrant women or established their business to do so. They specifically sought out migrants willing to work for low pay, in poor often exploitative working conditions and who would not or could not complain about their treatment (Jones, 2014; Peck & Theodore, 2001). As a result, agencies’ main business—and source of profits—was the direct recruitment of women from their home countries. As has been found elsewhere, it was migrant women who were highly profitable commodities to agencies (Lindquist, 2010).

At the time the research was conducted, the agencies in Amman offered their clients women from Bangladesh, Philippines and Sri Lanka, specialising in candidates of these three nationalities. In contrast, the agencies in Beirut, in addition to recruiting primarily from Bangladesh, supplied women from a far wider additional range of countries: Cameroon, Kenya, Ghana, Madagascar, Togo, Burkina Faso and Ethiopia, and Sri Lanka. These countries were not randomly selected by agencies; they were specifically targeted. Interviewees argued that women from all these countries were especially suited to these jobs, being grateful for the little income they received.

“I like to recruit Bangladeshi girls from 25 to 30. They have good health and are mature. They are not challenging. Through experience I have realised this is a good nationality and age group.” (Recruiter, Hazmieh, Beirut)

Agencies in Jordan and Lebanon were not free to recruit women from any country in the world. Where they could recruit was influenced by migration regulations in migrants’ origin countries as well as their own (Jones, 2014). This is a key difference from the activities of the agencies which deliberately targeted female homemakers in Canada (Vosko, 2000), and the agencies in Chicago that sought men of Latin American ethnicities and nationalities (Peck & Theodore, 2001). However, what they did have in common with these agencies is to continue to profit they needed to secure a steady and guaranteed supply of candidates (Peck & Theodore, 2001). Interviewees for this study explained that they needed to secure between 20 and 50 bio-data forms containing the details of migrant women each month to keep their clients’ interest and their business afloat. This was not just because of a high demand for workers, but because they needed to generate interest from clients, to create the illusion of choice (ILO, 2015a): “Customers need to see many applications to choose which one they want.” (Owner, Amman).

Agencies in Jordan began recruiting from Bangladesh in 2012 to 2013, according to interviewees in Amman, because they were facing a shortage in the supply of migrant women from other countries. In 2010, the government of Indonesia had banned the recruitment of women to Jordan. This followed an earlier ban by the Philippines government. Interviewees repeatedly emphasised what a challenge such migration bans were to their business model which required a constant flow of workers: “We were and are very limited in what nationality to advise employers on, because so many bans.” (Agency Owner Amman, Jordan) Bangladesh, with its plentiful supply of workers willing to work overseas due to a lack of alternatives at home and also a Muslim country, was an ideal replacement source country:

“The need for domestic workers per month in Jordan was 2,000 in 2010; 60% were from Indonesia, and 40% were from Sri Lanka and Philippines. When Indonesia banned their workers, this was a great problem…. We needed a fast action, which was Bangladesh. Because of the large supply that they have they were easy.” (Agency Manager, Amman)

However, for agencies in Jordan recruiting from Bangladesh only became legally possible in 2012 to 2013 as the MoU between the Bangladesh and Jordan governments was adopted. Interviewees recalled that their agency association had actively lobbied the Jordan government for this MoU precisely, because it would enable their members to access an alternative source of cheap labour. Prior to the signing of the agreement, representatives from the Jordanian agency association had even visited their counterpart—BAIRA (Bangladesh Recruiting Industry Association)—in Dhaka to negotiate the commissions they would pay them to send them Bangladeshi women. Their purpose was to maximise their profits through ensuring both a plentiful supply of women as well as low commissions, which they would pass on to their clients in the form of lower recruitment fees, enabling even working-class clients to ‘purchase’ a domestic worker. At the time of signing the agreement between Jordan and Bangladesh in 2012, the Jordanian Labour Minister was quoted in the press as saying that: “The opening of a new source country should contribute to minimising recruitment costs, which we expect to drop to as low as JD1500 (approx. US$2000 in 2013) after reaching unprecedented figures up to nearly JD4000 (approx. US$5000 in 2013).” (Cited in ILO, 2015a).

On the other hand, agencies in Lebanon began recruiting from Bangladesh a few years earlier, in 2006: “After the war in 2006, domestic workers were banned to come so we resorted to working with Bangladesh. They were cheap.” (Owner, Ramlit El Bayda, Beirut). In this country, in addition to facing labour shortages due to migration bans by origin countries (Philippines, Sri Lanka, Indonesia), agencies also struggled to recruit because of successive conflicts and political instability in the country. As the Bangladesh government lifted their prohibition on women migrating for overseas domestic work in 2003, Bangladesh provided an immediate solution for agencies’ shortage of labour. The final section outlines an agenda for decent work based on non-discrimination.

Rights-based decent work inclusive of non-discrimination

Placement agencies are the primary gatekeepers to employment for millions of migrant workers (Goh et al, 2017; Xiang & Lindquist, 2018). In the past decade, policymakers have increasingly associated recruiters’ practice of charging fees to migrants with debt bondage and other poor employment conditions (UNDOC, 2015; ILO, 2015b). International action has approached abusive agencies as individual ‘unscrupulous’ bad actors, which can be tackled with comprehensive and better targeted regulation, codes of conduct and training. At the same time, increasingly vocal global recruitment industry figures promote their valuable contribution to ‘efficient’ labour markets as well as enabling migrants to migrate safely (WEC, 2020). On this basis, they are commonly welcomed into international dialogues on migration to help devise regulation. Yet, to date, agencies’ responsibility for systemic and discriminatory racialised and gendered employment patterns have been largely ignored by policymakers.

Drawing on 156 interviews in Jordan, Lebanon and Bangladesh, this paper has clearly evidenced two things. First, that agencies remain at the centre of a globally deeply exploitative system of temporary labour migration, with racialised and gendered inequalities at its core (Bakan & Stasiulis, 1995; Parrenas, 2012; Pratt, 1999; Truong, 1996). Second, agencies systematically constructed the value of Bangladeshi women at the bottom of a highly racialised (transnational) labour market hierarchy (Lan, 2006; Bizri, 2014). Agencies do not merely facilitate mobility; they structure the opportunities and the outcomes of migration for millions of workers in Asia. What they sell—temporary, flexible, cheap and marginalised workers—fuels demand for this type of labour from employers while impacting detrimentally on labour organising as women are pitched against each other (Bakan & Stasiulis, 1995; Cheng, 2003; Pratt, 1997). What agencies do impacts, therefore, on migrants and citizens alike as their activities shape norms and expectations across labour markets (Peck & Theodore, 2001).

These findings powerfully emphasise the need to urgently explore how the current system of recruitment of migrant domestic workers can be dismantled. The structural roles of placement agencies are accepted as the norm by international organisations, yet actions aimed merely at better managing the worst excesses of the recruitment industry will only ever marginally improve migrants’ experiences. To substantiate ‘decent work’ for migrants as well as citizens requires more exposing and boldly attending to the systemic issues. C181 (of which the global recruitment industry is supportive) and R188 clearly establish the standards of non-discrimination on the grounds of sex, national extraction and social origin. No recruiter should recruit or advertise vacancies on the basis of these groups. The ILO migrant worker Conventions (C97 and C143) firmly establish the standards of non-discrimination between migrants and citizens in matters of employment. Although not widely ratified they do nevertheless establish clear global norms. Although numerous Conventions allow opt-outs for States in the name of sovereignty, agencies are not state actors; they are private enterprises. The industry should be held to account by international organisations for all the discriminatory recruitment of migrants.

Moreover, as a final point worth making, migration research in the Middle East is increasingly addressing discrimination, human rights violation and migrants’ rights issues from the Islamic ethics perspective (Jureidini and Hassan, 2020). Islamic ethical and legal traditions may offer an alternative international moral and legal paradigm to combat discrimination, exploitation and violations of migrants’ rights. According to scholars from the region, these faith-based norms could create a more powerful normative universe to inspire activism in the region (El Fadl, 2020) and as such should be taken seriously.