RESEARCH DOSSIER
Platform Worker Income Protection and Unemployment Coverage
Abstract. Digital labor platforms reorganize the allocation, pricing, observation, and termination of paid work, while unemployment protection continues to depend on legal employment status, the accumulation of insured earnings, and nationally defined unemployment contingencies. This monograph investigates how the discontinuous earnings of online and location-based platform workers interact with contributory unemployment insurance, cessation-of-business assistance, general social protection floors, and the administration of portable entitlements. Its comparative method combines primary international instruments, European legislation, national administrative guidance, official surveys, and critically appraised scholarly studies. The central finding is that three independent conditions govern effective protection: inclusion in a legally eligible population, recognition of earnings and contributions over a qualifying period, and proof of a compensable interruption under the relevant benefit rules. Reclassification can improve the first condition and facilitate the second without automatically resolving the third, especially where workers remain active on another platform or experience lost assignments rather than total unemployment. Evidence from the Eurostat 2022 pilot, Eurofound's 2026 online-platform survey, Mexico's administrative registration experience, and national benefit systems shows why a claimed increase in registered platform workers must be distinguished from payment of unemployment benefits. The study specifies the analytical differences among account deactivation, involuntary activity cessation, partial unemployment, income volatility, and temporary incapacity; examines contribution bases and person-centered portability; compares statutory approaches; and identifies auditable measurements and research designs. It contributes a source-grounded comparative synthesis and practical research implications without proposing a new authorial theory or forecasting unobserved policy impacts.
Keywords: Platform work; gig economy; digital labor platforms; unemployment insurance; income protection; social security; dependent self-employment; employment classification; contribution portability; multi-platform work; earnings volatility; deactivation; social protection floors; insurance finance; causal inference.
Introduction: Research Question, Scope, and Method
A platform worker can lose the opportunity to earn through a digital intermediary while continuing to be formally self-employed, maintaining an account on another application, or having no recognized unemployment event under national law. The insured contingency, its legal bearer, and the administrative earnings record are therefore separate determinants of income security. The ILO's 2021 platform-work study documented the diversity of online and location-based arrangements and the role of platform governance in shaping earnings and employment conditions. The joint ILO–ISSA–OECD technical analysis placed social insurance financing and national institutional design at the center of policy comparison (ILO, ISSA, and OECD, 2023).
This dossier asks how statutory status, income reporting, contribution records, and portability determine whether a worker whose platform-mediated earnings stop or decline can obtain timely, adequate, financially sustainable support. The research deliberately treats income protection as a broader set of arrangements than unemployment insurance: it encompasses tax-financed assistance, occupational and sickness protections when relevant to loss of earnings, emergency benefits, and transitional self-employment support. These instruments have different insured events, financing structures, eligibility tests, and benefit purposes. A social assistance payment following loss of orders must be classified as assistance; a sickness benefit following injury remains a sickness or injury benefit even when it stabilizes income.
The method is a structured, source-critical comparison, not a quantitative meta-analysis. Primary materials include ILO conventions and recommendations, the EU Platform Work Directive, an EU Council Recommendation, national statutes and official benefit guidance, and administrative releases. Statistical evidence includes Eurostat's 2022 experimental survey and a 2026 Eurofound–European Labour Authority survey. Scholarly studies are used for explanatory mechanisms and institutional experience, with their sampling and identification limits preserved. Legal positions are dated to November 17, 2026. Instruments adopted but not in force, legislative proposals, eligibility rules, administrative registrations, benefit receipt, and estimated causal effects are identified separately.
The denominator problem is pivotal. The Eurostat experimental release reports that 3.0% of persons aged 15–64 in the participating 16 EU and one EFTA country performed at least an hour of digital-platform work in the preceding year, while 62.4% of platform workers active for at least one hour in the preceding month reported no unemployment protection. These are pilot-survey estimates for a specified geographical and activity definition, not universal global prevalence or proof of claims eligibility. The Eurostat methodological note explains the voluntary country participation and observation framework (Eurostat, 2024). The distinction between having any platform income over twelve months and relying on that income as a principal livelihood changes every policy denominator in the analysis.
Conceptual and Historical Foundations
The economic object and units of observation
Digital labor platforms organize paid tasks using software interfaces and, to varying degrees, algorithmic allocation, monitoring, pricing, performance evaluation, and account access. The EU Platform Work Directive defines its legal scope through the organization of remunerated work and automated monitoring or decision-making systems; its definition expressly distinguishes a person performing platform work from a platform worker who has an employment relationship for the directive's purposes (European Parliament and Council, 2024, arts. 1–2). A descriptive researcher may use the broader phrase platform worker for all individuals earning through platforms, but legal attribution must respect the instrument's narrower terminology.
Work may be online, such as translation, programming, design, or microtasks, or location-based, such as delivery, transportation, care, and household services. A worker can perform tasks for several platforms, switch among employee and self-employed roles, and simultaneously hold conventional employment. The same person can thus be counted several times by platform accounts, once by a person-level labor force survey, and perhaps once or more in contribution accounts depending on system architecture. Employment relationships, gross transactions, billable tasks, taxable income, insured earnings, and recognized unemployment spells are not interchangeable units.
A complete employment-income observation identifies the individual, platform, jurisdiction, period, task payments, fees, reimbursable expenses, net income, statutory status, tax and contribution declarations, benefit coverage, and other employment. A digital account can remain registered after work has ceased; another can be suspended temporarily without extinguishing a worker's broader livelihood. Account counts systematically inflate person counts when workers multi-home or reopen accounts; worker counts can themselves include occasional sellers and renters if broad survey definitions are adopted.
Why the unemployment contingency is difficult to define
Classical unemployment insurance is organized around a legally recognizable interruption of employment, prior earnings or contributions, availability for suitable work, and active job search. The ILO Minimum Standards Convention No. 102 frames unemployment under Part IV as suspension of earnings caused by inability to obtain suitable employment for a protected person who is capable of and available for work. The ILO Convention No. 168 refines the full-unemployment contingency and calls for efforts to extend protection to partial unemployment, temporary reductions in hours, and eligible part-time workers seeking full-time work (ILO, 1952; ILO, 1988). Their operative obligations depend on ratification and, for Convention No. 102, the parts accepted by each ratifying state.
Platform earnings risk has several distinct sources. Demand can fall although an account remains active. An algorithm can reduce visibility and assignments. A rating or fraud system can suspend access. A platform can withdraw from a territory. A person can stop offering services voluntarily, experience illness, or migrate from one app to another. Each event has a different relationship to voluntariness, labor-market attachment, verifiability, and residual earning capacity. Treating every earnings decline as insured unemployment would conflate business revenue risk, partial unemployment, employment termination, health contingencies, and general poverty; treating only formal contract termination as relevant would leave some materially dependent workers outside protection.
Under a standard contribution-based program, benefit eligibility often requires more than an insured relationship on the last working day. It can depend on a minimum contribution period, historical earnings thresholds, the involuntary character of job loss, availability and search, a waiting period, and a time-limited claim. The appropriate scientific sequence is therefore: legal covered status; insured base; credited months or hours; qualifying event; adjudication; timely payment; net replacement adequacy. Each stage can exclude people included at a prior stage.
From employment-based social security to platform-specific instruments
The Social Protection Floors Recommendation No. 202 advocates nationally defined basic income security for working-age persons unable to earn sufficient income, including through unemployment, alongside essential health care and other life-cycle guarantees (ILO, 2012). It provides a foundation for combining contributory and noncontributory instruments, subject to national implementation. The EU Council Recommendation of 2019 expressly distinguishes formal, effective, adequate, and transparent coverage and recommends that accrued entitlements be preserved, accumulated, or transferred across employment and self-employment statuses (Council of the European Union, 2019, points 8–14).
International labor standard-setting acquired a new platform-specific instrument on June 12, 2026, when the ILO adopted the Decent Work in the Platform Economy Convention No. 193. Its social-security provision requires ratifying members to take measures assuring digital platform workers access to social-security protection on terms no less favorable than those of workers with the same employment-status classification; it also addresses status determination, platform intermediaries, and enforcement (ILO, 2026, arts. 12, 23–24). The ILO instrument profile listed zero ratifications and no entry-into-force date as of November 12, 2026. Thus adoption constitutes an international normative development, not an already binding treaty obligation for every jurisdiction. The Convention does not itself create a national unemployment-insurance account or universal monetary entitlement.
The historical trajectory is one of increasing recognition of employment forms that fluctuate across employers, clients, and jurisdictions. This trajectory supports comparing three institutional pathways—appropriate reclassification into employee schemes, extension of contributory schemes to genuinely self-employed workers, and income floors independent of continuous insured employment. They can coexist within a national system and need not impose identical rules across very different labor-market arrangements.
Income security as a sequence of distinct quantities
Legal access means that legislation opens the relevant branch to a category of worker. Enrollment means that the individual is registered. Contribution coverage means that earnings or qualifying periods are credited. Effective protection requires usable entitlements when a specified contingency occurs. Take-up is the fraction of eligible people actually applying for or receiving benefits, depending on measurement. Adequacy concerns net income replacement, duration, and household needs. These quantities cannot be collapsed into a single statement that a program 'covers' a certain number of workers. The ILO's World Social Protection Report, Chapter 3 discusses the missing middle that may be outside both contribution-based and tax-financed schemes (ILO, 2024).
Employment Classification and the Legal Boundary of Coverage
Status as a jurisdiction-specific legal determination
Platform companies may describe workers as contractors; workers may view themselves as independent; tax, employment, and social-security law can nevertheless use different and sometimes overlapping categories. Control over prices, dispatching, customer relationships, disciplinary systems, freedom to substitute another person, and entrepreneurial discretion are common legal facts, but their legal weight depends on the governing test. A multinational platform's common software does not produce a common multinational employment status.
The United Kingdom's Uber BV v Aslam judgment upheld the conclusion that the claimants were 'workers' for relevant employment protections and addressed the period during which drivers were logged in, in the authorized territory, and ready and willing to accept trips (UK Supreme Court, 2021). Its result concerns statutory worker rights including minimum wage and paid leave. UK worker status is distinct from full employee status; the judgment must not be cited as automatically establishing entitlement to contributory unemployment benefit. The New Style JSA eligibility rules require employee work and Class 1 National Insurance contributions in the relevant period; Class 2 contributions from self-employment alone generally do not qualify, subject to specified exceptions (UK Department for Work and Pensions, accessed 2026). The pairing of these two sources illustrates how a labor-rights judgment and an unemployment-benefit test can answer different questions about the same person's income security.
Spain's Law 12/2021 on digital delivery establishes a presumption of employment for remunerated delivery activities when the firm exercises organization, direction, and control, directly or through algorithmic management, and adds information rights concerning algorithms (Spain, 2021). Its reach is defined by the covered delivery activity and governing facts; it is not a global rule that all online freelancers are employees. Labor classification can shift the channel through which contributions are collected and unemployment rights accrued. Yet entitlements remain conditional on social-security law, qualifying history, and an adjudicated insured event.
The 2024 EU directive adopts a fact-based approach and an effective rebuttable presumption when direction and control are indicated under national law, with the platform bearing the burden of rebuttal under Article 5. Its December 2, 2026 transposition deadline had not yet arrived on the November 17 research cutoff. National implementation measures and actual employment findings therefore require country-level and date-specific verification. The directive's algorithmic-management protections also apply in certain respects to people performing platform work without an employment relationship. These protections may facilitate evidence about assignment and deactivation decisions, but they do not themselves replace contribution requirements in unemployment programs.
Dependency, autonomy, and mixed livelihoods
Economic dependency refers to reliance on platform-derived income, whereas legal subordination concerns facts relevant to an employment relationship. A person can be economically dependent on a single platform while exercising substantial scheduling autonomy. Another can work as a platform employee for a few hours while drawing most income from conventional self-employment. Neither dependence nor flexibility alone resolves status. The research task is to examine the contract, the work performed, the allocation of operational risk, and the specific social-insurance statute.
A worker's dependence also changes over the observation window. A full-time delivery driver can become a part-time driver after taking a conventional job, while a skilled online contractor can replace a principal client with several short projects. One-month and twelve-month dependence classifications may diverge. In household surveys, the most informative measures are the share of total individual earnings derived from platform work, the regularity of those earnings, the number of simultaneous income sources, and the duration of dependency.
Administration must follow judicially determined rights
Reclassification is meaningful for unemployment coverage only if payroll records, platform data, contribution assessments, and agency decisions are reconciled. A court can determine employee status retrospectively while the unemployment authority retains incomplete contribution history for the same period. Employer insolvency, delayed payment, cross-border platform operation, or use of subcontractors can impede implementation. The responsibility to declare workers, collect contributions, and correct historical periods has to be assigned to legally responsible actors, with review and correction rights for individuals.
A statutory presumption can reduce the evidentiary burden of proving employee status. It cannot by itself establish that a person experienced a compensable unemployment spell, nor can it convert platform-reported gross transactions into verified insured wages without an applicable contribution-base rule. The law of coverage and the administration of coverage must therefore be investigated together.
The legal position at the 2026 cutoff
The European Commission's 2023 implementation assessment documented heterogeneous progress following the nonbinding 2019 Council Recommendation. The more recent EU directive lays down binding obligations that must be transposed nationally by the specified deadline, while Convention No. 193 requires ratification and entry into force to bind a member as a treaty. Spain has a nationally operative delivery-worker presumption. These are different legal stages. Treating them as simultaneous universal insurance coverage would produce both a false legal statement and a false quantitative estimate.
Contributions, Earnings Bases, and Financing
From transactions to insurable earnings
A platform observes transactions, a worker receives payouts, a tax authority recognizes income under national tax rules, and an insurance institution credits a statutory contribution base. These values can differ markedly. Let gross customer payments to worker i in period t across platforms be G(i,t); platform fees and transaction adjustments be F(i,t); necessary work expenses be X(i,t); and other taxable adjustments be A(i,t). A simple descriptive net-work-income measure is:
Y(i,t) = G(i,t) − F(i,t) − X(i,t) + A(i,t).
This is an accounting illustration, not a universal legal tax base or a new theoretical model. Each jurisdiction determines which costs are deductible, when income is recognized, how mixed personal and business expenses are treated, and whether a wage-like payment rather than business profit is the insured base. An insurance contribution could be C(i,t) = r × B(i,t) for a legally defined assessable base B, subject to contribution floors, ceilings, employer shares, government subsidies, and qualifying thresholds. The chosen basis materially affects low-earning, high-expense drivers and couriers. Treating the full customer fare as the worker's net income would overstate ability to contribute and possible replacement benefits.
Payout timing also matters. Platforms may remit several settlements within one month, apply refunds in later months, or charge fees that obscure the correspondence between a task and recognized income. For multi-platform workers, administrative rules require period alignment, unique identifiers, reconciliation of duplicate transactions, and adjustment procedures. A platform's settlement record is valuable evidence but is not itself proof that earnings were credited for unemployment insurance.
Contribution coverage of intermittent workers
Conventional unemployment insurance may employ payroll contributions and minimum work, earnings, or insurance periods. A highly intermittent platform worker can remain formally insured and still fail a qualifying-period test. Consider an illustrative threshold of six credited months in twelve. A worker earning substantial income in four months and none in eight could contribute more in total currency than another worker with six low-earning months but fail that hypothetical month-based test. Changing the definition of a credited period changes eligibility independent of contribution yield; this example is deliberately hypothetical and asserts no country's precise rule.
The Council's access-to-protection Recommendation urges that qualifying and benefit rules not obstruct access simply because of a labor-market status and that self-employed contribution and entitlement calculations account for fluctuating income (Council of the European Union, 2019, points 9, 12–14). Its recommendation for portability recognizes that a person should not necessarily lose accumulated protection by moving between an employee job, self-employment, and another industry. Legislative implementation, rather than the recommendation itself, determines the individual entitlement.
Thresholds can be defined by time, earnings, or a combination. Earnings-based systems accommodate nonstandard hours but are sensitive to variable prices, seasonality, contribution ceilings, and reporting errors. Time-based systems recognize attachment but may penalize microtasks and irregular bookings. A mixed criterion may be necessary to reconcile proof of attachment with contribution adequacy; the effect has to be assessed with real earnings histories rather than inferred from general claims about flexibility.
Mandatory, voluntary, and subsidized participation
Mandatory insurance pools workers exposed to varying risks, supports predictable contribution collection, and helps contain the adverse selection that can arise if participation starts only when loss of work becomes likely. Voluntary schemes can accommodate heterogeneous self-employed preferences and enterprise risk, but enrollment can favor workers anticipating higher claims. Public subsidies may extend access where low earnings make proportional contributions unaffordable. These design trade-offs follow established insurance economics; they are not a claim that one arrangement is optimal across all jurisdictions.
Germany's voluntary unemployment insurance rules offer a concrete example. Under stated conditions, people beginning self-employment can apply within three months; prior insured employment or a qualifying income-replacement entitlement is generally required, and the contribution is flat rather than indexed to business turnover (Federal Employment Agency, accessed 2026). This design supports continuity for qualifying transitions from employment but does not grant every long-standing independent platform worker automatic access. The German federal service guide stated a monthly self-employed premium of EUR 102.83 for 2026, with half-rate treatment in the founding and following year; the amount is a dated program parameter, not an international benchmark (Federal Republic of Germany, 2026).
Contribution subsidies can be evaluated by net public cost per additional effectively protected worker rather than simply by total registrations. If enrollment expands but waiting periods and qualification rules prevent claims, the policy may raise contribution receipts without appreciably improving unemployment protection. Conversely, eligibility expansion can improve income security with a longer fiscal maturation period. Both effects require evidence on claims incidence, duration, benefit amounts, administration, and changes in worker behavior.
Actuarial adequacy and common shocks
The expected benefit expenditure of a group over an accounting year depends on the distribution of qualifying interruptions, average paid duration, the insured base, replacement formula, benefit ceilings, and take-up. For exposition only, a simplified product E[benefit outgo] ≈ N × p × d × b × q uses N exposed persons, p probability of a qualifying spell, d expected paid periods conditional on such a spell, b average benefit per period, and q a simplified claims realization factor. In practice, these variables are correlated, eligibility evolves with time, and multi-spell models or simulations are needed; gross financial requirements also include administration and reserves. The expression is standard expected-cost decomposition rather than an original actuarial doctrine.
Risk pooling is constrained by correlated shocks. A citywide mobility restriction, a platform market exit, or a severe recession can interrupt work for many contributors simultaneously. Claim incidence and contribution revenue may move adversely at the same time; contribution rates calibrated to ordinary individual unemployment will not necessarily finance systemic platform disruption. Premium and reserve design must specify fiscal backstops, cyclical sensitivity, reinsurance where appropriate, and the state role in extraordinary events. Public unemployment funds may act as automatic stabilizers, while private income-loss insurance can face exclusions, waiting periods, limited policy duration, and expensive monitoring of voluntary withdrawal.
The role of the public income floor
The ILO's 2024–2026 unemployment-protection assessment reports that only 16.7% of unemployed persons globally receive unemployment cash benefits and identifies platform and self-employed workers among groups with gaps (ILO, 2024). This is a global unemployed-person coverage measure, not the percentage of platform workers insured. Tax-financed income guarantees can protect individuals who fall outside contributory eligibility and help stabilize households after earnings shocks, but their targeting, adequacy, work incentives, funding, and responsiveness require separate assessment.
Means tests often operate at household level, whereas platform earnings are recorded at person or account level. A partner's earnings, savings rules, assets, and timing of income recognition may determine access. A person might be economically dependent on platform work yet fail a household means test; another might receive assistance even though prior platform work never generated contributions. Scientific accounting must keep the two mechanisms separate to avoid falsely attributing income support to unemployment insurance.
Benefit Portability and Comparative Institutional Outcomes
What portability must preserve
Portability has a financial meaning, an administrative meaning, and a legal meaning. Financial portability preserves the value of contributions or credited periods when a person changes employer, platform, contractual status, or locality. Administrative portability means the acquiring institution can recognize prior periods without forcing the claimant to reconstruct years of earnings from disconnected accounts. Legal portability establishes the circumstances under which a recognized record can count toward a benefit under another program. A transfer of data alone accomplishes none of these without compatible statutory rules. The EU Council Recommendation on access specifically calls for transparency and transferability of entitlements across jobs and statuses, subject to national systems (Council of the European Union, 2019).
Consider a worker with four months of insured payroll employment, three months of platform earnings classified as independent business income, and five months of mixed assignments through several platforms. There is no universal twelve-month unemployment-contribution record arising simply because twelve calendar months elapsed. Portability requires first knowing which portions are insurable under the applicable scheme, whether concurrent periods count once, how minimum earnings enter contribution crediting, and which laws determine aggregation. A record system that preserves all receipts can still produce zero additional qualifying months if the underlying statute excludes self-employment. Conversely, a statutory aggregation rule can be ineffective without accurate cross-employer reporting.
Person-centered records should retain worker identifiers, platform or payer identifiers, work periods, reported remuneration, assessment bases, contributions assessed, contributions actually paid, corrections, and appeals. A key design question is whether the worker loses an accrued entitlement because a platform fails to remit its obligation. The answer is jurisdiction-specific and calls for a distinction between a worker's recognized insured period and public collection of an employer debt. The insurer, tax administration, labor inspectorate, and platform operator may hold different pieces of evidence; an auditable record must expose how conflicts are resolved.
The European Union: status harmonization and social-security limits
The Platform Work Directive (EU) 2024/2831 addresses determination of employment status, algorithmic management, and transparency, with national transposition due December 2, 2026 (European Parliament and Council, 2024, art. 29). Its legal-presumption mechanism must be read carefully: Article 5(3) states that the presumption does not apply to proceedings concerning tax, criminal, or social-security matters unless a Member State elects to apply it under national law. Thus the presence of a presumption in labor proceedings does not automatically establish unemployment-insurance contributions or benefit eligibility in a social-security proceeding. National implementing legislation, judicial classifications, and national insurance law remain decisive. On November 17, 2026 the transposition deadline was still ahead; no claim of complete uniform implementation is justified.
The 2019 Council Recommendation includes workers and self-employed persons within its reform vision and treats effective coverage, adequacy, transparency, and transferability as connected objectives. As a Recommendation, it is a policy instrument rather than a directly enforceable uniform unemployment benefit. The European Commission's implementation report issued in 2023 documents varying national measures and persistent gaps, establishing a comparative baseline rather than a conclusion about each state's law in late 2026 (European Commission, 2023).
Spain: employee presumption and self-employed cessation protection
Spain's Law 12/2021 on platform delivery adopted a presumption of employment for delivery and distribution activities organized through digital-platform algorithmic management, subject to its operative statutory terms (Spain, 2021). For those properly classified as employees, the ordinary social-security and unemployment framework becomes relevant; the presumption is neither a universal rule for every online freelancer nor a guarantee of benefit payment after an earnings decline. Its practical reach depends on enforcement, factual work arrangements, and compliance with registration and contribution duties. Vieira's qualitative study of Spain's reform emphasizes the gap between formal rulemaking and observed industry implementation; qualitative observations do not by themselves measure a national causal increase in unemployment-benefit take-up (Vieira, 2025).
Separate from employee unemployment insurance, Spain maintains a self-employed cessation-of-activity protection regime administered through social-security arrangements. The Spanish employment service's guide to cessation of activity identifies the distinct self-employed contingency and routes for information and claims (SEPE, accessed 2026). It should not be merged statistically with the unemployment entitlement of payroll workers: conditions, contribution arrangements, proof of cessation or reduction, and competent administrators differ. Where partial cessation of independent activity is recognized by the applicable provisions, it should be distinguished from the loss of a single platform account while the insured business continues to earn elsewhere.
France: narrow transitional benefits for qualifying independent workers
France's allocation des travailleurs indépendants (ATI) is a targeted benefit for qualifying self-employed workers facing specified business-cessation circumstances. The French public employment service's ATI conditions describe eligibility linked to prior independent activity, business circumstances, revenue and resources, and registration as a job seeker (France Travail, accessed 2026). The detailed France Travail benefit guidance states that ATI is paid for a limited period, generally 182 calendar days, and explains its applicable statutory amount and conditions (France Travail, accessed 2026). This instrument supplies an example of limited independent-worker protection; it is not equivalent to open-ended insurance for ordinary weekly platform-order fluctuations.
The distributional concern is that a worker can have substantial dependence on one platform yet lack the legally specified business-cessation event or documented prior period. A conditional benefit may protect some business failures while leaving abrupt algorithmic deactivation and modest but persistent earnings losses unaddressed. Empirical evaluation should measure initial applications, eligibility adjudication, payments, denial reasons, and the overlap with general assistance. Using registrations or aggregate self-employed population alone would obscure the program's effective reach.
Germany: elective continuity from insured employment
Germany's self-employed unemployment insurance option offers a bridge for eligible people who enter self-employment after an insurance-covered work history and meet timely application conditions (Federal Employment Agency, accessed 2026). The federal service specification provides dated 2026 payment information and eligibility explanation. Voluntary entry can preserve a route to benefits for some career transitions, but its entry window and previous insurance conditions exclude many workers whose platform activity has long been their only work. The separate distinction between prior salaried contribution history and subsequent independent income is essential.
A flat contribution design simplifies collection but creates different effective contribution rates across income levels. A fixed monthly charge constitutes a larger fraction of low net platform earnings than of high net earnings, while benefit rules may depend on prior standardized or earned assessments rather than contemporaneous task revenue. Research should estimate take-up by net-income quantile, rather than extrapolate from the posted monthly charge to universal affordability.
United Kingdom: an intermediate worker category and contribution class
In Uber BV v Aslam, the UK Supreme Court upheld the finding that the drivers at issue had statutory worker rights for the claims before it, including minimum-wage and paid-leave implications (UK Supreme Court, 2021). British 'worker' status is a legal category distinct from the standard 'employee' category, and labor-rights judgments must not be interpreted as automatic enrollment in every contributory unemployment benefit. The official New Style Jobseeker's Allowance eligibility guidance ties entitlement to specified National Insurance contribution conditions, principally Class 1 contributions (UK Government, accessed 2026). A self-employed platform worker's payments under other National Insurance classes do not, solely by virtue of the payments, establish eligibility for this benefit.
Means-tested support and its earnings adjustments require separate examination. For example, a person working sporadically through applications may receive an income-tested public benefit under household rules while never receiving contributory unemployment insurance. An impact study must label the payment correctly and not interpret the size of the Universal Credit recipient population as the coverage of unemployment insurance.
Canada and the United States: special benefits and state employment tests
Canada's Employment Insurance self-employed framework admits participating self-employed persons to specified special benefits rather than ordinary regular unemployment benefits as a direct consequence of opting in (Employment and Social Development Canada, accessed 2026). The official self-employed eligibility page sets the participation and qualifying requirements. Sickness, caregiving, and maternity benefits can stabilize earnings after distinct life events, but they are not compensation for losing platform orders while able and available to work. That categorical distinction is especially important in cross-country datasets built from aggregated 'social insurance' expenditure.
The United States primarily relies on state unemployment-insurance systems with federal parameters, and entitlement generally turns on covered employment, earnings, separation circumstances, availability, and state law. The government's independent-contractor eligibility explanation explains that generally self-employment is not covered as conventional unemployment-insurance employment (United States Department of Labor, accessed 2026). The federal benefits portal directs applicants to states and makes the administrative decentralization explicit (USAGov, accessed 2026). A particular worker's classification, state test, and any temporary special public program must be checked by jurisdiction and date. Pandemic-era expansions are historical interventions and cannot be presented as permanent ordinary coverage in November 2026.
Mexico: registration, contribution thresholds, and the insured branch
Mexico's IMSS digital-platform worker program illustrates a distinct route to social-security registration through platform-mediated employment and a pilot implementation beginning in July 2025 (Mexican Social Security Institute, 2025–2026). The IMSS January 2026 continuation announcement records the administrative continuation and clarifies operational arrangements (IMSS, 2026). Official enrollment releases distinguish workers registered during the pilot from persons whose net earnings meet applicable thresholds for full branch coverage. That distinction makes the ratio of accounts or registrations to actually protected people a core evaluation variable.
Most importantly, the broad IMSS package of social-insurance branches is not proof of a conventional unemployment-cash-insurance entitlement. The covered branches include legally specified sickness and maternity, employment risks, invalidity and life, retirement and old age, and childcare/social benefits. The question studied in this dossier concerns compensable unemployment; it must not be answered affirmatively by citing five social-security branches without examining an independent unemployment instrument. Mexico therefore provides evidence for administrative inclusion and payroll-like reporting, with the unemployment-benefit question remaining a separate legal inquiry.
Cross-border platform work and portable protection
A platform may be incorporated abroad while work is performed locally; an online contractor can accept clients in several countries; a worker can reside in one jurisdiction and perform tasks in another. The country of company incorporation, place of task performance, habitual residence, applicable employment law, tax residency, and social-security affiliation can diverge. A platform-facing digital interface does not determine the applicable social-security legislation by itself. Bilateral or regional coordination arrangements may aggregate qualifying periods or prevent double contributions under specific conditions, but recognition of another country's earnings and export of an unemployment benefit always require examination of the governing instrument.
Cross-border research should begin with a fact pattern and trace each rule separately. For a two-country year it should identify the competent insurer, legal employment status in each period, earnings credited to each account, treatment of overlapping work, whether benefits are exportable, and the appropriate appeal path. Population-wide statements such as 'portable platform insurance across Europe' would be overinclusive without country-level coordination detail and the specific type of benefit. The research purpose is to establish which rights are actually portable, to whom, and through which lawful data transfer.
Evidence Verification, Measurement, and Sensitivity
A measurement architecture with explicit denominators
Coverage statistics in this field are unusually vulnerable to numerator–denominator mismatch. At least seven quantities should be separately observed: (1) persons who did any platform work within a reference interval; (2) persons for whom it supplied a material share of earnings; (3) legally covered persons under a specified unemployment program; (4) insured contributors with credited qualifying periods; (5) persons who experienced an administratively recognized qualifying interruption; (6) claimants who applied; and (7) claimants paid a benefit. Each ratio answers a different question. The ratio of covered persons to all ever-active platform accounts can be distorted by duplicate accounts, whereas the ratio of benefit recipients to all insured platform workers measures realized claims under observed risks and administrative behavior, not legal eligibility alone.
For a legal coverage rate, the relevant denominator is the precisely defined working population in the jurisdiction and period and the numerator is people legally within the named scheme. For effective protection, the numerator must correspond to demonstrable access or payments and must be paired with a denominator that experienced the relevant contingency where applicable. A percentage of all unemployed receiving any unemployment cash benefit, as reported in the ILO's World Social Protection Report chapter, is different from a percentage of platform workers who could qualify if deactivated (ILO, 2024).
The Eurostat 2022 pilot release is valuable because it observes people rather than administrative accounts and includes self-reported protection. Its participating-country and reference-year restrictions are essential, as is the difference between the preceding-year prevalence figure and the last-month-active denominator for the 62.4% uninsured estimate. Eurostat's methodological documentation should accompany any use of the estimate. It cannot determine a claimant's eligibility under the detailed law of a particular jurisdiction, validate contributions paid, or establish the effect of a new directive that postdates the survey.
The difference between survey, administrative, and causal evidence
A survey measuring perceived protection can identify awareness gaps and economic dependence but may misclassify statutory rights. An agency extract can prove that contributions were assessed or benefits transferred, while omitting unregistered workers and revealing little about their counterfactual income. A court judgment establishes rights in a litigated factual and legal setting and may not capture routine compliance elsewhere. A legislative text establishes formal entitlement when in force but cannot by itself measure take-up or timely payment. Peer-reviewed case research can explain mechanisms and implementation conflict without offering population-level effect sizes.
The Eurofound–ELA 2026 online-work investigation is especially useful for employer–worker control, working conditions, and informational asymmetries among surveyed online platform workers (Eurofound and ELA, 2026). Its recruited respondents and country selection are not the same sampling frame as a probability-representative EU labor-force survey. The findings are therefore evidence of mechanisms and experiences within a studied group, not a calibrated estimate of the EU-wide coverage share. Research methods must declare the specific sample frame, response method, and platform category before comparing published percentages.
The Mexican IMSS case illustrates a converse limitation. An official registration portal documents enrollment operations, but a registration does not disclose which unemployed people will be paid a cash benefit, nor does a high number of registered records automatically imply a corresponding number of continuously protected people. Analytic linkage of unique individuals, reporting periods, net earnings and applicable contribution thresholds is required.
Shock definitions and competing explanations
A fall in task income may reflect a platform's unilateral access restriction, technological reallocation, seasonal demand, consumer switching, worker decisions, planned vacation, sickness, relocation, changes in prices, or movement toward another job. A policy evaluation must separate an exogenous or involuntary earnings interruption from endogenous reductions in supplied labor and from adverse changes in work capacity. A worker who loses one platform while increasing hours elsewhere may have suffered platform-specific loss but not total unemployment. A worker logged in without available orders may be available for work despite having no earnings and may struggle to document that fact under traditional certification rules.
Account deactivation offers an apparently observable event but is a noisy proxy for income loss. Some suspensions follow alleged misconduct or identity concerns, some are immediately reversed, and many workers have several accounts. Conversely, a platform may leave an account active while reducing task visibility or payouts sharply. Sensitivity analyses should vary the allowed deactivation duration, the minimum pre-event earnings share, a multi-platform earnings offset, and the treatment of appeals and reversals. The chosen claim trigger has fiscal effects and changes incentives for both platform operators and workers.
Threshold sensitivity and illustrative accounting
Let a worker's legally assessable earnings in each of twelve months be B(t), and let T be a hypothetical minimum monthly threshold for crediting a covered month. The count of credited months is Q(T) = sum of indicators that B(t) reaches T, subject to national law. For an illustrative sequence of B(t) equal to 0, 0, 200, 250, 300, 450, 500, 700, 0, 100, 650, and 750 monetary units, Q(300) equals six, while Q(500) equals four. The same annual income therefore generates different qualifying histories when the legal threshold changes. This is an arithmetic sensitivity example, not observed national data or a statutory benchmark.
An annual net-income floor can create different outcomes for a worker with low monthly income but sustained attachment, while an annual earnings test may recognize the worker whom a monthly test excludes. The research should vary deductible work costs, treatment of earnings from several platforms, minimum qualifying periods, benefit waiting days, and the generosity of partial-unemployment adjustments. Results must distinguish a model's assumptions from a program's actual eligibility statute. A finding that 'portability improves coverage' is incomplete without specifying whether the improvement concerns record continuity, eligible person-months, successful claims, or actual net income after transfers.
Administrative errors, legal appeals, and unequal observability
Platform data can be highly granular while entitlement data are coarse, delayed, or inaccessible to workers. Missing tax identification numbers, duplicated profiles, mismatched dates, undocumented commission charges, and disagreements over employee status can prevent a formally eligible claimant from proving past earnings. An insurance regime should be evaluated on correction time, error rates, information rights, appeal accessibility, and recovery of unpaid employer contributions, not only on contribution receipts. The EU Platform Work Directive introduces transparency and safeguards around automated monitoring and decision-making; these rights can matter to evidence production, even though the directive's legal presumption is not automatically operative in social-security litigation (European Parliament and Council, 2024).
Workers with little digital access, migrants, irregular work histories, carers with fragmented careers, and people experiencing disability may be less able to navigate multi-agency claims. A register constructed from platform compliance records will often omit the very workers who are difficult to classify or have the weakest evidence. Equity assessments should therefore report both inclusion of historically unregistered groups and the remaining attrition through eligibility, application, adjudication, and payment. An aggregate expansion rate without group breakdowns may conceal increased disparities.
Research design for credible policy evaluation
Useful quasi-experimental comparisons could exploit a dated legal reform affecting a defined platform category, comparing pre- and post-reform trajectories with plausibly comparable groups. A difference-in-differences design would need to investigate pre-trends, simultaneous labor-law or tax changes, anticipatory reclassification, enforcement timing, platform exits and entry, and changes in reporting incentives. A discontinuity at an earnings or tenure threshold could identify local behavioral responses if manipulation and sorting around the threshold were credibly excluded. Matched administrative records could compare spells of benefit receipt after deactivation with earnings before and after loss of platform access, retaining workers with multiple income sources.
The principal outcomes should be measured benefits per qualifying shock, claim approval and appeals, replacement of net income, duration until reemployment or return to platform activity, net public expenditure, and continuity of insurance records. Outcomes such as well-being and poverty require appropriately sourced household data and a counterfactual; administrative benefit totals cannot establish them alone. Voluntary insurance take-up is particularly sensitive to selection: people who enroll may be systematically different in expected risk, financial literacy, and prior employment history. Rigorous inference should identify which differences are observed and which remain unmeasured.
Evidence horizon and the interpretive boundary
This monograph has a November 17, 2026 research cutoff. The 2024 EU Platform Work Directive's December 2, 2026 deadline is a future legal milestone; it must not be reported as a completed EU-wide reform. The ILO Convention No. 193 profile reports adoption in June 2026 and zero ratifications as of the November 2026 instrument view; treaty entry into force has not occurred. The 2022 Eurostat estimates are historical measurements and cannot be silently updated to the research date. Official benefit websites are operational snapshots that may change as national law and administrative interpretations develop. Country-specific figures in this study have their own observation dates and must not be treated as a pooled, harmonized international panel.
Scientific Synthesis and Research Agenda
Source-grounded synthesis
The comparative evidence establishes that protecting income in platform work requires the legal scope of a scheme, recognition of an insurance record, and verification of a covered contingency to function together. These are analytically separable stages in the ordinary administration of social insurance. Their separation explains why some widely discussed reforms produce consequential labor rights without automatically delivering unemployment cash protection. Spain's delivery-worker presumption changes employment classification for a defined category; the EU directive's presumption has an explicit social-security-proceeding limitation; Germany supports voluntary insurance continuity for eligible transitions; France provides targeted assistance after defined business cessation; Canada offers self-employed contributors particular special benefits but not ordinary EI regular unemployment; and Mexico improves registration under social-security branches without establishing a general unemployment-cash benefit for all registered platform workers.
This comparative mapping identifies three recurrent implementation problems. First, the worker's economic loss is often partially hidden by the legal binary of employed versus unemployed; ongoing activity on another application can coexist with significant lost income. Second, earnings traceability does not ensure contribution recognition because net income, statutory bases, and qualifying-period rules differ. Third, multiple programs that stabilize household income have distinct contingencies and should not be collapsed into a single 'unemployment coverage' statistic. These findings are an interpretive synthesis of documented institutions and should be tested with worker-level linked data rather than elevated into an uncommissioned new doctrine.
The ILO's technical policy analysis and social protection assessment support extending adequate, sustainable protection to workers across employment forms while maintaining attention to funding and administrative reach (ILO, ISSA, and OECD, 2023; ILO, 2024). In operational terms, the evidence favors transparent recognition of the individual worker's contribution history, clear legal definitions of covered loss, effective rights to dispute missing records, and the complementary role of public protection floors. The allocation of financing among worker, platform, client, and state remains a national choice subject to distributive and fiscal analysis.
A practical sequence for policy assessment
A jurisdiction considering reform should first map current eligibility across employee, intermediate, dependent self-employed, and independent statuses with exact statutory references. It should then measure the economically dependent population using person-level surveys, identify how platform reports become assessed and paid contributions, and audit the pathway from a verifiable earnings shock to a paid claim. This sequence reveals whether the immediate bottleneck lies in classification, missing contributions, the legal definition of unemployment, insufficient benefits, or application barriers. Policymakers need separate instruments for distinct bottlenecks; a single reclassification statute cannot be assumed to solve all of them.
Next, actuarial and fiscal work should define the eligible exposure, collect historical covered-earnings distributions, estimate correlated platform shocks, consider behavioral responses, and model financing across the business cycle. The analysis should compare mandatory extension, voluntary opt-in, income-tested supports, and hybrid public arrangements on consistent measures: additional effectively insured worker-years, qualifying shocks covered, average after-tax benefit replacement, administrative cost, distributional incidence, and solvency under stress. A registry count is informative as an implementation measure but insufficient as a welfare outcome.
Finally, release rules should ensure that the worker can see the income and contribution record on which the institution relies, correct errors, and challenge decisions. Platforms and public authorities should specify which transactions they report, which deductions they apply, and how cross-platform employment histories are reconciled. Meaningful enforcement depends on timely transmission, independent audit, and an accessible procedure for the individual claimant. These are administrative consequences of existing rights-based and social insurance principles, not a new theoretical system.
Testable research questions
The international research agenda turns on observable transitions: Does platform earnings reporting increase recognized insured periods for people with intermittent jobs? Does an employee-status reform increase unemployment-benefit applications and successful payments after qualifying separations? Do earnings-based qualifying rules protect more genuinely dependent workers than month-count rules, and at what additional cost? Does integrating multiple platform earnings into one contribution record improve rights without double counting concurrent periods? How do temporary account deactivations compare with sustained income losses in predicting welfare need? Does a subsidy reduce exclusion or chiefly transfer financing for workers who would have enrolled anyway? Each question requires precise identification of the treated population, the outcome, the period, and the legal contingency.
The Eurostat experimental data make cross-country measurement gaps visible, while the Eurofound–ELA study suggests important questions about algorithmic control and information access. Neither source supplies the linked longitudinal contribution, claim, and earnings records needed to calculate causal effects of reform across jurisdictions (Eurostat, 2024; Eurofound and ELA, 2026). The next evidence priority is the lawful linkage of these variables with strong protections for privacy, correction, and analytic transparency.
Limitations and authorial contribution
The original contribution of this dossier is the comparative reconstruction of the entitlement pathway, the separation of employment-status rules from unemployment-benefit conditions, the insistence on person-level denominator discipline, and the cross-country interpretation of the observed policy examples at a fixed legal cutoff. The arithmetic scenarios are explanatory sensitivity exercises using ordinary actuarial logic. The research does not estimate a universal platform-worker uninsured rate, derive new causal effects, interpret law for an individual claimant, or claim independent fieldwork or original microdata. The cited texts and administrative descriptions are primary records of particular legal and institutional arrangements; qualitative studies and experimental surveys are weighed according to their actual methods. Future research should prioritize direct administrative claimant records and independent evaluation over attractive but unverified platform registration headlines.
Conclusions
Platform work has brought a persistent mismatch into unemployment protection: applications and transactions are observable in real time, while entitlement still depends on the legal character of work, qualifying histories, and the definition of an insured interruption. Source-based comparative analysis demonstrates that effective income protection arises from the intersection of covered status, credited contributions and a compensable event. Program enrollment is relevant, but evidence of a paid qualifying claim is a distinct and later observation.
Employment classification is consequential but cannot stand in for insurance adjudication. Spain's sector-specific employment presumption, the labor-status rights identified in the UK Uber litigation, and the EU's platform-work legislation operate under different legal scopes. The directive expressly limits its presumption in social-security proceedings absent national extension. For a platform worker seeking unemployment cash benefits, the controlling law is the applicable national or coordinated insurance entitlement at the time of the claim.
Contributions require a reliable definition of the earnings base. Gross customer payments, net platform payouts, taxable business profit, and insurable remuneration may diverge; the gap is especially material to workers who finance fuel, vehicles, data access, equipment and platform fees. Contribution records must identify concurrent and sequential platform engagements without double counting. Portability of data improves the ability to establish a claim, while portability of rights requires additional substantive legal recognition.
National systems illustrate distinct design choices. Germany permits qualifying continuity through voluntary unemployment insurance for self-employment; France operates a restricted independent-worker cessation benefit; Canada's self-employed Employment Insurance option provides designated special benefits; the United States continues to distinguish covered employment by state-law tests; and Mexico's IMSS platform integration addresses broader social-insurance registration without establishing universal unemployment cash protection. These regimes cannot be placed on a single ordered scale without specifying the insurance event, covered population, earnings base, benefit formula and financing.
Measured coverage must be distinguished from subjective reports, administrative registrations, eligible contributors and paid claims. The Eurostat pilot's 62.4% without reported unemployment protection is a 2022 participating-country estimate restricted to digital-platform workers active in the preceding month, while the ILO's 16.7% global unemployment-benefit receipt figure concerns all unemployed persons. Neither answers the same question as an individual platform worker's legal entitlement after deactivation. Both nonetheless underscore why policymakers need better source-matched denominators and records.
Effective protection also depends on fair administration. Accurate earnings feeds, contribution histories, correction rights, transparent decisions, accessible appeals and carefully defined treatment of partial interruptions are essential to turning statutory eligibility into actual income replacement. Tax-financed social protection floors remain important where contributory systems exclude workers or cannot respond quickly enough to severe shocks; they should be analyzed under their own statutory and fiscal criteria.
By November 17, 2026 the platform economy had a newly adopted international convention and an approaching EU directive transposition deadline, alongside diverse established national systems. The 2026 ILO platform-work Convention was not yet in force according to the ILO instrument profile. Future reforms should be evaluated against observable improvements in recognized coverage, contributions, benefit payment, adequacy, distribution and fiscal sustainability rather than inferred from legislative adoption alone. The enduring scientific task is to measure how well each institutional pathway protects a worker when paid platform activity actually disappears.
References
Council of the European Union (2019). Council Recommendation of 8 November 2019 on Access to Social Protection for Workers and the Self-Employed (2019/C 387/01). Official Journal of the European Union, C 387, 15 November. Council Recommendation of 8 November 2019 on Access to Social Protection for Workers and the Self-Employed (2019/C 387/01).
Employment and Social Development Canada (2026). Digest of Benefit Entitlement Principles: Chapter 24, Benefits for Self-Employed Persons. Government of Canada, accessed November 17. Digest of Benefit Entitlement Principles: Chapter 24, Benefits for Self-Employed Persons.
Eurofound and European Labour Authority (2026). Online Platform Work in the European Union: Implications for Working Conditions and Access to Information. Publications Office of the European Union, Luxembourg, report EF25005, ISBN 978-92-897-2562-0, DOI 10.2806/0981358. Online Platform Work in the European Union: Implications for Working Conditions and Access to Information.
European Commission (2023). Report on Implementation of the Council Recommendation on Access to Social Protection. Directorate-General for Employment, Social Affairs and Inclusion, January 31. Report on Implementation of the Council Recommendation on Access to Social Protection.
European Parliament and Council of the European Union (2024). Directive (EU) 2024/2831 on Improving Working Conditions in Platform Work. Official Journal of the European Union, November 11; arts. 1–5, 29. Directive (EU) 2024/2831 on Improving Working Conditions in Platform Work.
Eurostat (2024). Digital Platform Employment: Experimental Statistics from the 2022 Labor Force Survey. Eurostat Statistics Explained / Eurostat News, July 18. Digital Platform Employment: Experimental Statistics from the 2022 Labor Force Survey.
Eurostat (2024). Methodological Note: Digital Platform Employment. European Commission, Eurostat. Methodological Note: Digital Platform Employment.
Federal Employment Agency (Germany) (2026). Freiwillige Arbeitslosenversicherung (Voluntary Unemployment Insurance). Bundesagentur für Arbeit, accessed November 17. Freiwillige Arbeitslosenversicherung (Voluntary Unemployment Insurance).
Federal Republic of Germany (2026). Unemployment Insurance on Application: Federal Service Portal. Bundesportal, service 99038007017000, accessed November 17. Unemployment Insurance on Application: Federal Service Portal.
France Travail (2026). Allocation des Travailleurs Indépendants: Les Six Conditions Spécifiques. Public ATI portal, accessed November 17. Allocation des Travailleurs Indépendants: Les Six Conditions Spécifiques.
France Travail (2026). L’Allocation pour les Travailleurs Indépendants (ATI). France Travail benefits guidance, accessed November 17. L’Allocation pour les Travailleurs Indépendants (ATI).
Government of Canada (2026). Employment Insurance Benefits for Self-Employed People: Eligibility. Canada.ca, accessed November 17. Employment Insurance Benefits for Self-Employed People: Eligibility.
Instituto Mexicano del Seguro Social (2025–2026). Seguridad Social para Empresas y Personas Trabajadoras de Plataformas Digitales. IMSS official digital platform program. Seguridad Social para Empresas y Personas Trabajadoras de Plataformas Digitales.
Instituto Mexicano del Seguro Social and Secretaría del Trabajo y Previsión Social (2026). IMSS y STPS Informan sobre la Reforma Laboral en Materia de Personas Trabajadoras de Plataformas Digitales. Official press release No. 002/2026, January 2. IMSS y STPS Informan sobre la Reforma Laboral en Materia de Personas Trabajadoras de Plataformas Digitales.
International Labour Organization (2026). Convention No. 193: NORMLEX Instrument Profile and Ratification Status. NORMLEX, status accessed November 17. Convention No. 193: NORMLEX Instrument Profile and Ratification Status.
International Labour Organization (2026). Decent Work in the Platform Economy Convention, 2026 (No. 193). International Labour Conference, adopted June 12, official NORMLEX convention text. Decent Work in the Platform Economy Convention, 2026 (No. 193).
International Labour Organization (1988). Employment Promotion and Protection against Unemployment Convention, 1988 (No. 168). International Labour Organization, convention text. Employment Promotion and Protection against Unemployment Convention, 1988 (No. 168).
International Labour Organization (2012). Social Protection Floors Recommendation, 2012 (No. 202). International Labour Conference, Recommendation. Social Protection Floors Recommendation, 2012 (No. 202).
International Labour Organization (1952). Social Security (Minimum Standards) Convention, 1952 (No. 102). NORMLEX, Part IV. Social Security (Minimum Standards) Convention, 1952 (No. 102).
International Labour Organization (2021). World Employment and Social Outlook 2021: The Role of Digital Labour Platforms in Transforming the World of Work. International Labour Office, Geneva. World Employment and Social Outlook 2021: The Role of Digital Labour Platforms in Transforming the World of Work.
International Labour Organization (2024). World Social Protection Report 2024–26: Chapter 3, Extending Social Protection to All. International Labour Office, chapter 3. World Social Protection Report 2024–26: Chapter 3, Extending Social Protection to All.
International Labour Organization (2024). World Social Protection Report 2024–26: Chapter 4, Income Security and Social Protection. International Labour Office, chapter 4. World Social Protection Report 2024–26: Chapter 4, Income Security and Social Protection.
International Labour Organization, International Social Security Association, and Organisation for Economic Co-operation and Development (2023). Providing Adequate and Sustainable Social Protection for Workers in the Gig and Platform Economy. G20 Employment Working Group technical paper, February 6. Providing Adequate and Sustainable Social Protection for Workers in the Gig and Platform Economy.
Kingdom of Spain (2021). Ley 12/2021, de 28 de septiembre, por la que se modifica el texto refundido de la Ley del Estatuto de los Trabajadores. Boletín Oficial del Estado, September 29; disposition additional 23. Ley 12/2021, de 28 de septiembre, por la que se modifica el texto refundido de la Ley del Estatuto de los Trabajadores.
Servicio Público de Empleo Estatal (Spain) (2026). Protección por Cese de Actividad de las Personas Trabajadoras Autónomas. SEPE, accessed November 17. Protección por Cese de Actividad de las Personas Trabajadoras Autónomas.
Supreme Court of the United Kingdom (2021). Uber BV and Others v Aslam and Others [2021] UKSC 5. Supreme Court of the United Kingdom, case UKSC/2019/0029. Uber BV and Others v Aslam and Others [2021] UKSC 5.
United Kingdom Government (2026). New Style Jobseeker’s Allowance: Eligibility. GOV.UK administrative guidance, accessed November 17. New Style Jobseeker’s Allowance: Eligibility.
United States Department of Labor (2026). Can I Get Unemployment Benefits If I Am an Independent Contractor or Self-Employed?. Unemployment.gov, federal claimant guidance, accessed November 17. Can I Get Unemployment Benefits If I Am an Independent Contractor or Self-Employed?.
United States Government (2026). Unemployment Benefits. USAGov, accessed November 17. Unemployment Benefits.
Vieira, Tiago (2025). An Incomplete Double Movement: Spain’s Legislative Strategy to Platform Couriers’ Reclassification. International Labour Review 164(4): 1–20. DOI: 10.16995/ilr.23799. An Incomplete Double Movement: Spain’s Legislative Strategy to Platform Couriers’ Reclassification.