Kafala sponsorship systems

A family of systems joins employment to legal presence

Kafala is a comparative label, not the title of one regional statute. An ILO framework describes the Gulf Cooperation Council states, Jordan, and Lebanon as using different combinations of laws, administrative rules, social practices, and exceptions. It identifies the family resemblance through five questions: whether a worker is tied to the employer for entry, permit renewal, termination, transfer, and exit. The same report dates the modern Gulf arrangements to the 1960s and 1970s while limiting its country comparison to public information available in January 2017.¹

That variation matters. A country can remove an exit permit but retain an employer-linked residence process; permit job changes for workers covered by a general labor law but exclude domestic workers; or state a right that remains difficult to use in practice. The ILO's later comparative summary describes Qatar as allowing workers, including domestic workers, to change employers after notice, while describing Saudi mobility rules as covering workers under the Labor Law and excluding domestic workers.² “Kafala abolished” is therefore too coarse a conclusion unless it identifies the worker category, control point, legal date, and operating practice.

The central delegation is public. The state authorizes entry and residence but lets a named sponsor, usually the employer, control or initiate decisions that affect those permissions. The employer also directs the work and may be the party accused of violating the contract. A dispute over wages, hours, safety, or resignation can consequently change the worker's immigration position as well as the employment relationship.¹

Temporary migration does not logically require that fusion. A state can admit a worker for employment, set a reasonable notice or job-search period, and administer status without giving one employer a veto over transfer or departure. The institutional question is narrower than whether temporary labor migration continues: which actor should control each decision when the worker and employer disagree?

Recruitment cost becomes leverage when exit threatens legal status

Recruitment connects origin-country brokers, destination-country agencies, employers, ministries, and workers who rarely hold equal information. Employers may pay substantial placement costs; workers may also pay agents or borrow before departure. If a promised job or wage changes, returning home can leave the worker with debt, while an employer may view early departure as the loss of an investment. Sponsor control converts that commercial conflict into leverage over mobility.¹

Lebanon's employer study shows how intermediaries become operational governors. Among 1,200 employers of live-in migrant domestic workers in four urban areas, 87.4 percent reported recruiting through an agency. Employers relied on agencies for information and mediation, even though the recruiter earned revenue from keeping the employer as a client. Two in five reported directing the worker's first three months of salary to the agency rather than the worker. The study's 29 interviews found employers describing passport retention as security for recruitment spending.³

The same employer survey reported that 22.5 percent always or sometimes locked the worker inside and 94.3 percent retained the worker's passport. Those are employer reports from a probability cluster sample, not allegations sampled only among complainants. They still describe selected urban regions, live-in work, and self-reported behavior; social desirability could make some violations less visible rather than turn the percentages into national prevalence estimates.³

An “absconding” process deepens the leverage. The ILO framework defines the term as an administrative offense attached to leaving a sponsor without permission and records that loss of status can expose even a worker escaping abuse to arrest, detention, and deportation. It also describes false reports, employer failure to renew documents, and work for a different employer as paths into irregular status.¹ A complaint mechanism cannot supply an equal remedy if the worker becomes deportable before the underlying dispute is heard.

Lebanon makes the workplace, residence, and private home coincide

Lebanese domestic workers were excluded from general labor protections in the 2019 legal and interview record, and their residence was tied to the employment relationship. Amnesty International's 32 interviews with women migrant domestic workers documented repeated reports of excessive hours, missing rest, withheld pay, restricted movement and communication, inadequate food and housing, violence, and denied health care. Amnesty classified eight interviewed cases as forced labor and four as trafficking under its legal analysis.⁴ Those classifications establish what the organization found in those cases; they do not estimate how often each abuse occurred among all workers in Lebanon.

A larger ILO survey gives a different but converging view. It interviewed 1,541 women migrant domestic workers in Beirut and Mount Lebanon in 2014. Half said papers were held against their wishes, reported work averaged 10.5 hours a day, only 36 percent reported a weekly day off, around one third reported sometimes being locked inside, and more than 60 percent said they had no one to turn to in an employment conflict.⁵ Because recruitment occurred at churches, markets, embassies, nongovernmental organizations, call centers, and money-transfer outlets, the survey could not reach workers never allowed out of the household; its authors warn that conditions may therefore appear better than the inaccessible population's conditions.⁵

Private-home inspection is only part of the problem. The employer may control the worker's food, sleep, telephone, documents, and door while also supplying the facts that immigration authorities use. The worker may have no colleague who observes the workday and no independent address from which to pursue a claim. Formal notice and complaint rights consequently depend on communication, safe housing, document possession, lawful status during a dispute, and an institution that can hear evidence without returning the worker to the accused employer.

Workers have also built institutions under those constraints. Anti-Racism Movement and the Ethiopian migrant-domestic-worker-led group Mesewat conducted participant-driven research with nine active organizers. The report describes a membership-based group of more than 200 people providing financial, medical, emotional, social, and case support while facing time, language, legal- recognition, and deportation constraints.⁶ Its participants are not a proxy for every nationality or worker in Lebanon. Their record is valuable for a different reason: workers defined organizing needs and proposed action rather than appearing only as respondents to a state, employer, or advocacy institution.

Lebanon's 2020 revised standard contract attempted to add unilateral termination and other protections, but Human Rights Watch reports that recruitment agencies successfully challenged it and that the administrative court treated harm to agency interests as a basis for blocking the contract.⁷ An ILO care-economy assessment conducted during 2023 still described migrant domestic workers as subject to kafala and outside labor-law coverage, found limited legal and policy progress, and warned that weak protection and oversight can also reduce care quality.⁸ The two records are consistent on continued structural dependency, although neither substitutes for the underlying court file or a complete 2024 legal code.

Qatar separates formal reform from implementation evidence

Qatar supplies a useful test because the government makes specific, auditable claims. Its communications office says the country abolished the no-objection certificate for changing employers, removed exit-permit requirements for most workers, extended the exit change to domestic workers subject to notice, and introduced a minimum wage covering domestic work. The same page describes wage monitoring, complaint channels, dispute committees, and a worker-support fund.⁹ These are the government's account of its laws and institutions, not independent proof that every eligible worker used them without delay or retaliation.

The ILO Evaluation Office's independent evaluation confirms meaningful procedural change and a persistent implementation gap. It reports 669,198 job- change applications approved and 364,053 rejected from September 2020 through October 2023. Those are applications, not a count of unique workers or proof that every approved transfer completed without wage loss. The evaluators also found employers still requesting no-objection certificates, incomplete worker awareness, limited progress for domestic workers, delays in dispute resolution, and gaps in impact data.¹⁰

The evaluation is unusually candid about its own limits. Its mixed-method study included 87 informants across government, employers, worker representatives, unions, nongovernmental organizations, academia, and international bodies, but the team did not interview ordinary workers directly. It evaluated a Qatar- funded ILO technical-cooperation project, used government and project data, and reported stakeholder concerns about data reliability and transparency.¹⁰ Its institutional access and independent Evaluation Office oversight make it a strong program record; they do not turn it into a representative worker survey or a national causal estimate.

Amnesty's earlier investigation preserves direct accounts that the evaluation could not collect. Of 105 women who were currently or recently employed as live-in domestic workers and contacted during 2019–2020, 90 reported regularly working more than 14 hours, 89 reported no weekly day off, and 87 reported passport confiscation.¹¹ Amnesty explicitly says the access-driven sample was not statistically representative. The research also straddles the 2020 mobility changes, so it demonstrates serious implementation failures around that reform period but cannot measure their later trajectory or estimate current prevalence.

The evidence supports both a change and a warning. Qatar replaced important formal employer permissions with state procedures, and a large volume of workers sought job changes. Employer resistance, absconding retaliation, domestic-worker isolation, and remedy delays nevertheless remained operating problems. Government statements, program evaluation, and worker interviews disagree most sharply when a legal entitlement is treated as a completed social outcome; their different methods and dates explain part, but not all, of that gap.⁹¹⁰¹¹

A mobility event reveals where authority still sits

The cleanest operating test follows a worker attempting to leave one employer: who holds the passport; whether wages and status remain valid; whether the employer can cancel documents or file an absconding report; whether the worker has safe housing and information; how long a transfer or complaint takes; and whether a final decision can be appealed. An enacted statute is one input to that sequence, not the result.

The ILO's comparative recommendations separate the functions directly. Entry, residence, and work permission would not name one controlling employer; workers could renew their own status, terminate with notice, change employers without consent or status loss, and leave without sponsor approval. Labor law and the contract would govern the employment dispute, while public immigration institutions would govern presence in the country.² Employers would retain ordinary powers to direct work and enforce lawful notice and contract terms, but not the power to convert resignation into irregular migration.

Measures should distinguish inputs, process, and outcomes. Laws, guidance, staffing, and digital platforms are inputs. Applications, processing times, complaints, inspections, appeals, wage recovery, and status continuity are process measures. Safe exits, completed transfers, reduced recruitment debt, fewer retaliatory reports, durable worker voice, and better job matching are outcomes. Counts of approved applications establish administrative activity; they do not answer the other questions without denominators, unique-person counts, completion data, worker reports, and disaggregation by sector, gender, nationality, wage, and living arrangement.¹⁰

The organizational profile follows the chain of control

State bureaucracy supplies immigration and labor authority, while market capital supplies demand, recruitment finance, and employer leverage; those are the two authority sources. Decisions occur in central-executive ministries, in rule-bound administrative hierarchies, and at the frontline-local level of a firm or household. Ownership is not applicable to the regime as a whole; partnership-network describes the operational chain among agencies, employers, ministries, and origin-country actors rather than an equity claim.¹³

Hierarchy coordinates permits and enforcement, markets match labor and employers, standards set contracts and labor protections, and rule-and-ritual captures recurring sponsorship and renewal procedures. Knowledge moves top-down through official rules, bidirectionally through applications and complaints, and peer-to-peer through worker networks. Financial measures include fees and wages; operational measures include permits, transfers, and complaints; behavioral measures include employer compliance; informal measures include trust, reputation, and worker-to-worker knowledge.³⁶¹⁰

Market feedback can expose poor job matching, while doctrinal revision changes laws and administrative interpretations. Adaptation can therefore take the form of central reconfiguration, local iteration in agency or employer practice, and slow institutional change in enforcement and social norms. Customers, workers, the state and public, and communities are plausible beneficiary groups, but the profile does not imply that all benefited. Financial extraction, suppressed voice, externalized harm, capture by employers or intermediaries, and fragility when status depends on one employment relationship are the principal failure risks.²⁶¹⁰¹²

Six themes define the case

Authority, legitimacy, and acceptance, delegation, decentralization, and responsibility, structure, hierarchy, and scale, cooperation, incentives, and organizational equilibrium, work design, productivity, and automation, and governance, stewardship, and accountability receive the highest emphasis. The regime delegates public authority into a scaled public-private chain, uses status dependency to stabilize cooperation, shapes who may leave a job, and raises the accountability problem of an employer holding both contractual and immigration leverage.

Coordination, decision-making, strategy, and culture receive medium emphasis: the case turns on multi-actor handoffs, bounded choices under debt and legal risk, reform responses to labor-market pressure, and informal norms about passports, mobility, gender, race, and voice. Purpose, measurement, knowledge, and learning receive lower emphasis because they illuminate particular parts of the mechanism without organizing the whole account. Innovation, executive attention, and organizational ignorance score zero: entrepreneurship, a single executive sensing system, and ignorance as a separately demonstrated strategy are not developed enough in the evidence to define the case.¹²

Related paths are comparisons and interpretive lenses

SEWA cooperatives are a governance contrast: member workers pool authority to bargain and build services, while sponsorship places public-status leverage with the buyer of labor. The contrast does not claim that one institution influenced the other.

Lebanon's confessional power-sharing system is jurisdictional context. It concerns how Lebanese public authority is allocated across confessional institutions; it is not presented as the sole cause of migrant domestic-worker sponsorship.

Fujimori's reproductive-health program is a bounded coercive-delegation analogue. Both cases ask what happens when public goals are implemented through frontline relationships in which a vulnerable person may be unable to refuse without penalty. The programs, jurisdictions, harms, and histories are otherwise distinct, and no influence claim is intended.

The six defining idea links above are organizing lenses. Authority locates the source and acceptance of sponsor power; delegation asks where responsibility moved; structure traces the chain; cooperation examines incentives that keep it stable; work design centers exit and bargaining; and governance asks who can review, remedy, and change the arrangement. These classifications do not assert historical or intellectual lineage.¹²

Effects and evidence boundaries

Evidence is strongest for formal control points, Lebanese domestic-work practice, worker and employer accounts in bounded samples, Qatar's reform procedures, and the distance between legal change and implementation. It is not strong enough to assign one prevalence rate to the region, one net social effect, or one post-reform trajectory. Country, sector, worker category, year, and method must travel with every result.²⁵³¹⁰

Affected-worker evidence is present but uneven. The Lebanon survey reached a large number of workers yet excluded those confined inside; Amnesty's Lebanon and Qatar samples preserve detailed accounts without representing national populations; the Mesewat project gives participants unusual control over knowledge but concerns a small group of Ethiopian organizers; and the Qatar evaluation did not interview ordinary workers directly.^5461110 There is no comparable community-controlled source across multiple countries, sectors, nationalities, genders, wage levels, and legal statuses.

Membership, nonhuman life, ecosystems, and future generations remain research-needed domains. The regime has no member constituency, and the public evidence used here does not trace ecological effects or long-run outcomes for workers' children, origin communities, destination care systems, or labor institutions. Absence from these materials is not evidence of no effect.

Paths into deeper study

• Build country-and-sector matrices for entry, renewal, resignation, transfer, exit, absconding, domestic-work coverage, document possession, complaint, appeal, and status during a dispute, each tied to the law and effective date.
• Follow representative cohorts from recruitment through arrival, first job, attempted transfer, complaint, and departure; separate applications from unique people and completed, retaliation-free outcomes.
• Compare worker, employer, agency, court, ministry, embassy, and worker-led records for the same disputes rather than inferring a whole process from one participant's account.
• Measure recruitment payments and debt at origin and destination, identify who ultimately bears each cost, and test whether mobility reform changes wages, turnover, productivity, and job matching.
• Fund multilingual, worker-controlled research that can safely include people confined in homes, workers in irregular status, men and women, and sectors beyond domestic work and large firms.
• Add adjudicated case files and current administrative rules to test false absconding, passport recovery, status cancellation, wage recovery, detention, deportation, and agency accountability.
• Add environmental and longitudinal research before assigning ecosystem or intergenerational effects.