RESEARCH DOSSIER
Public Employment Services Coordination with Unemployment Insurance
Abstract. Public employment services (PES) and unemployment insurance (UI) form a consequential institutional interface in labor-market protection. Unemployment insurance compensates insured earnings interruptions under legally prescribed conditions; public employment services provide labor-market information, vacancy matching, guidance, placement, training referrals, and—in many jurisdictions—verification of continuing job-search obligations. Their coordination determines whether a person who loses employment can preserve income, obtain timely assistance, acquire skills, and make a sustainable transition to productive work. This monograph examines that coordination through international labor standards, administrative law, comparative institutional practice, global coverage data, and causal evaluation literature. Its central analytical finding is that coordination quality depends on the alignment of legal entitlement, service access, individualized intervention, procedural safeguards, and measured labor-market outcomes. Organizational integration can reduce fragmentation, but its value depends on accurate eligibility decisions, meaningful employment assistance, and accountable administration. The evidence distinguishes faster exits from benefit rolls from actual transitions into stable employment; it further identifies displacement effects, heterogeneous treatment impacts, and substantial limitations in international comparability. Country cases from the United States, Germany, the United Kingdom, and Australia illustrate distinct legal and administrative arrangements. The synthesis proposes an evidence-based research agenda for assessing the complete path from job loss through benefit receipt to verified employment and earnings outcomes. It advances an original comparative interpretation without asserting a new formal theory or a universal institutional design.
Keywords: Unemployment insurance; public employment services; social insurance; employment activation; job placement; labor-market intermediation; unemployment benefits; conditionality; reemployment; training; benefit administration; service integration; causal evaluation; social protection; procedural rights; comparative labor policy.
Introduction
Unemployment protection connects two domains of public policy that operate on different clocks. An insured worker needs income replacement as soon as a qualifying employment interruption occurs; an employment service seeks a suitable future job, a process that depends on vacancies, qualifications, mobility, care responsibilities, health, and demand conditions. Coordination concerns the full trajectory and the institutional decisions that determine its course. The unit of analysis is therefore a person's sequence of claims, services, obligations, and employment states, nested within a particular legal system and labor market.
The topic has considerable international importance. The International Labour Organization's 2024–2026 social protection assessment reports that 16.7 percent of unemployed people worldwide receive unemployment cash benefits, with major differences among income groups. That statistic measures effective receipt against an unemployment denominator and should not be substituted for the proportion of workers legally insured, the proportion who register with a PES, or the share who receive intensive employment services. The IDB, OECD, and WAPES 2026 global employment-services study documents institutional variation across 78 public employment services in 74 countries in its publication overview; the report's preface describes 78 services in 75 countries. The discrepancy in country counts across publisher descriptions is preserved here rather than harmonized by assumption. Comparative conclusions in this dossier concern documented organizational patterns, not a population-weighted global estimate of coordination quality.
The object of inquiry is the coordination of placement, training, employment counseling, continuing eligibility verification, and statutory unemployment benefits. Related questions concerning qualifying contributions, replacement rates, benefit duration, reserves, sanctions appeals, short-time work, and platform-worker coverage are recognized as adjacent independent Research Dossiers. They enter this analysis only when necessary to explain the employment-service interface.
The method combines textual reading of binding instruments and national rules, administrative service descriptions, official statistics, and peer-reviewed or institutionally authenticated empirical studies. Each jurisdictional proposition is attached to a dated source. Causal evidence is interpreted within its studied population, treatment and comparison, outcome definition, and time horizon. Explicit distinctions are maintained between formal entitlements and effective access; an administrative exit and employment; program participation and program impact; budget savings and social welfare; and cross-sectional association and experimental causal inference.
Conceptual and Historical Foundations
The institutional relationship
Public employment services are publicly mandated institutions that assist jobseekers and employers in bringing labor supply and demand together. Their tools may include vacancies, referrals, interviews, skills assessment, individualized counseling, retraining, employer engagement, and labor-market intelligence. Unemployment insurance is a legally structured income-maintenance arrangement under which an eligible person receives benefits during defined unemployment contingencies. The precise financing, qualifying period, benefit formula, duration, eligibility restrictions, and administering authority are national legal choices. The two institutions may share premises and data while remaining legally distinct.
The ILO Employment Service Convention No. 88, adopted in 1948, requires a free public employment service in countries for which it is in force. Article 1 defines a public employment service's essential duty in relation to labor-market organization; Articles 2 and 3 envisage a nationally directed network and territorially accessible offices. This standard establishes the service infrastructure within which unemployment protection may be administered and activation offered. Ratification and national implementation determine its treaty obligations in any specific state.
The ILO Social Security Convention No. 102, adopted in 1952, addresses minimum unemployment benefit protection in Part IV. Article 20 connects the protected contingency to a capable and available person's inability to obtain suitable employment, and Article 21 defines alternative coverage thresholds. Convention No. 102 establishes minimum standards for accepted branches in ratifying countries. It supplies a benchmark for statutory entitlement and adequacy without imposing a uniform global benefit scheme.
The ILO Employment Promotion Convention No. 168, adopted in 1988, makes the active-passive relationship especially explicit. Article 7 associates employment promotion with social security, employment services, training, and vocational guidance. Articles 20 and 21 condition benefit refusal or reduction in specified cases, including unjustified failure to use placement or training facilities and refusal of suitable employment, while describing the factors relevant to suitability. These are treaty standards for ratifying states that accept their obligations, not claims that every national claimant is governed by identical sanctions.
The trajectory of these instruments reflects the historical development of two linked public functions: insuring income loss and facilitating labor allocation. Early employment offices organized information and matching; contributory unemployment insurance developed systems of verified insured claims; postwar social-security instruments made protection and service provision international normative concerns; modern activation policy introduced systematic links between receipt, search behavior, counseling, and training. The ILO's employment services account situates PES within active labor-market policy and notes collaboration with public, private, and nonprofit organizations. The resulting contemporary institution is a network of tasks and legal authorities, even where administration takes place through a single portal.
Integration as a governance question
The central design problem is how to coordinate acts that have different legal consequences. Registering as unemployed may initiate a benefit claim, trigger labor-market assistance, or do both. A counselor's recommendation may be supportive advice, a condition of continued benefit payment, or a referral that initiates separate administrative proceedings. A failure to attend an appointment may signal a genuine refusal, deficient notification, disability, caregiving conflict, technology failure, or scheduling error. Governance must define who records the event, who establishes the facts, who decides its legal relevance, and how the claimant can contest the result.
Organizational arrangements include a single agency administering both assistance and benefit decisions; separate insurance and employment offices connected by data interfaces; devolved service provision with central statutory benefits; and hybrid public-private placement systems. The 2026 IDB–OECD–WAPES comparative service profiles support the proposition that responsibilities differ across jurisdictions. A shared office is not equivalent to shared legal competence, and a digital handoff is not evidence that services were received.
The European Union's PES Network amendment extends the European network's mandate through 2027 and emphasizes cooperation, inclusion, innovation, and performance comparison. Network coordination occurs across national services; benefit entitlement remains grounded in the applicable national and EU legal arrangements. This distinction matters particularly when comparative policy reports are mistaken for directly enforceable universal claimant rules.
An analytical account of coordination
A complete coordination episode can be observed as a sequence: qualifying job loss; registration; legal eligibility assessment; benefit award and disbursement; employment profiling; agreed service plan; receipt of placement or training; review of work-search conditions; work entry; and follow-up on employment quality. These events may overlap, recur, or terminate at different points. A researcher needs event dates, institutional identifiers, statutory grounds, and observable subsequent outcomes.
Three relationships are analytically decisive. First, benefit access affects the claimant's capacity to search and undertake training by providing resources during unemployment, subject to adequacy and timeliness. Second, service quality affects labor-market information and the ability to obtain suitable employment. Third, administrative monitoring can change the cost of noncompliance and the recording of benefit exits. The third relationship may produce shorter claims even when a durable job has not been obtained. Consequently, the coordination outcome is multidimensional and has to be measured by more than an insurer's benefit-duration series.
The European Commission's labor-market policy database makes a useful statistical distinction: services encompass PES activities; measures comprise training and other activation programs; support covers income maintenance and related financial support. These categories permit accounting analysis without conflating their expenditure, participant counts, or policy functions.
Service Eligibility
Legal entitlement, registration, and actual assistance
Eligibility has several meanings that must be recorded separately. A claimant may qualify for a cash benefit because the insured contingency and contribution rules are satisfied. The same individual may register with the employment service for vacancy referrals without satisfying the insurance qualifying conditions. Other individuals can receive job-search assistance while employed or outside the labor force; conversely, some benefit recipients may receive little intensive support. Service coverage and benefit coverage therefore require distinct denominators.
Formally, let I represent an identified population exposed to unemployment risk, B the set meeting statutory benefit eligibility, R the set registered with the PES, P the set receiving an actual service, and J the set entering verified employment during a specified follow-up interval. The quantities |B|/|I|, |R|/|I|, |P|/|R|, and |J|/|P| answer different descriptive questions. They must never be interpreted as causal effects of coordination. If |B| is based on registered claimants but |I| derives from a household survey, definitions, reference periods, and duplicate episodes must be reconciled first.
International benefit coverage demonstrates the stakes of denominator choice. The ILO global unemployment-protection findings indicate that insured or legally covered workers constitute a different statistical population from unemployed people actually receiving benefits. Informality, exclusions, qualifying criteria, and incomplete uptake can separate legal rules from payment outcomes. The report's 16.7 percent recipient estimate consequently cannot measure the availability of job matching or training.
Service eligibility can depend on registration, residency, work authorization, employability assessment, age, disability status, previous employment, service-capacity priorities, and enrollment in an assistance scheme. Those conditions arise from different statutes and administrative programs. The Eurostat labor-market definitions guide defines labor-force survey unemployment using joblessness, active search, and availability, which differ from the benefit-law definitions or administrative registration rules of particular countries. A comparison must label its unemployment population as surveyed, registered, insured, or receiving benefits.
Registration and timing: the German example
Germany illustrates a legally differentiated sequence. The Federal Employment Agency's jobseeker-registration guidance distinguishes notifying the agency that employment will end from the subsequent act of registering unemployment. Early notification is designed to begin placement before the loss of work. Its unemployment-registration rules explain that unemployment registration is a condition for Arbeitslosengeld and does not arise automatically from earlier jobseeker registration. The agency's public guidance identifies section 141 of Social Code Book III as the legal basis.
The distinction illustrates a continuity problem. Employment services can begin before the insured contingency, while payment depends on formal procedures after that contingency. An effective coordination process therefore sends timely reminders and records whether registration actually occurred, while preserving the claimant's ability to correct identifiable errors under applicable law. Administrative convenience alone cannot change statutory effective dates.
Section 35 of Germany's Social Code Book III describes the Federal Employment Agency's placement offer to jobseekers, apprenticeship seekers, and employers, emphasizing intensified support where integration is expected to be difficult. Section 37 establishes potential analysis and integration planning, including assessment of qualifications and barriers and documentation of reciprocal placement and job-search commitments. Together they show a statutory connection between the individual service process and the administration of unemployment-related obligations. Translation into American English here is analytical; the German statutory text controls the exact legal meaning.
Entry points outside contributory insurance
The United States presents a distinct federal-state architecture. Unemployment insurance is a required partner in the workforce system, and the Department of Labor's RESEA program description explains a structured meeting joining continuing UI eligibility assessment with individual reemployment planning, labor-market information, and access to American Job Center services. Statutory 42 U.S.C. § 506 identifies reemployment outcomes, integrity, workforce-system integration, and referral to partner programs among the goals of RESEA. States retain operational variation within federal requirements. A statutory program covering selected UI claimants cannot be treated as universal service entitlement for all unemployed persons.
The United Kingdom's Jobseeker's Allowance guidance describes a claimant commitment agreed with a work coach, regular appointments, and evidence of job-search activity. This is program-specific claimant guidance; it is not interchangeable with every feature of Universal Credit or with the arrangements of devolved services. It demonstrates how an individualized commitment operates at the point where advice and payment conditions meet.
Australia's JobSeeker mutual-obligation guidance describes job plans, provider appointments, search reporting, and exceptions or modifications for particular circumstances. JobSeeker Payment is a means-tested public income-support arrangement rather than a standard earnings-related contributory unemployment-insurance policy. Its coordination features may be compared with insurance-based activation, but its financing and benefit eligibility must be classified correctly.
Allocation rules and rights
Service prioritization frequently uses risk assessment, previous employment, skills, duration, and barriers. Risk-based triage can provide intensive attention where the expected marginal gain is greatest, while also creating incentives to select clients easiest to place. A credible eligibility record documents the assessment variables, human review, service offered, service actually delivered, and the reasons any alternative was selected.
Conditionality requires legal precision. Under ILO Convention No. 168, refusal of suitable employment or unjustified failure to use specified facilities can have benefit consequences within the convention's prescribed framework; suitability assessments expressly consider individual and labor-market circumstances. Sound administration distinguishes a genuine refusal from barriers to attendance, inaccurate notification, and impossibility of compliance. A workable coordination record therefore contains accessible appointment notices, reasons for exemptions, evidence relied upon, the deciding authority, and routes for administrative or judicial challenge under national law.
The OECD's 2025 activation-requirements comparison examines reporting, monitoring, suitable-work rules, and sanctions across unemployment and minimum-income schemes. Its composite strictness indicator is valuable for describing formally prescribed obligations; by itself it does not establish whether obligations were fairly implemented or whether a specific claimant obtained sustainable employment.
Eligibility coordination is successful when timely and lawful benefit decisions coexist with an intelligible service offer, actual access to relevant assistance, and reliable handling of obligations. This defines the legal and administrative base on which the analysis of interventions can proceed.
Activation Interventions
What employment activation actually delivers
Activation denotes interventions intended to promote entry or reentry into employment among people receiving unemployment-related support or otherwise seeking work. It encompasses employment counseling, job-search assistance, vacancy matching, skills diagnosis, formal training, vocational rehabilitation, wage subsidies, work experience, and sometimes requirements to participate or demonstrate search effort. These instruments vary in intensity, cost, timing, participant selection, and the employment margin they address.
The European Commission's labor-market intervention taxonomy distinguishes services, measures, and support; that classification is particularly useful for attributing costs. A payment to replace lost earnings is an income-support expenditure, while counselor time, referrals, and publicly financed training create separate service or measure expenditures. Even when one institution pays for all three, effectiveness requires measuring them separately before assessing interactions.
An operational activation episode starts with diagnosis. Employment histories, skills, preferred occupations, transport access, disability accommodations, caregiving constraints, language proficiency, and employer demand determine plausible pathways. A service provider may then issue vacancy referrals, arrange interviews, refer a claimant to training, or assess participation in a subsidized placement. Administrative data commonly record the action taken, while evidence of a completed course, an accepted job, sustained earnings, or continued social insurance contribution may require linkage to independent sources.
The ILO's employment-transition guidance treats employment services and active labor-market programs as institutions for adjustment, inclusion, and resilience. This supports examining training and placement alongside income support. It does not establish a causal benefit for every counseling meeting or compulsory training assignment. The quality of each intervention depends on the target population, counterfactual employment opportunities, and how the program is implemented.
Counseling, information, and monitoring
Counseling can reduce search frictions when jobseekers possess incomplete information about vacancies or lack effective job-search strategies. It can also identify skills that are poorly signaled to employers. Its impact can be modest when vacancies are scarce, when services add little information, or when monitoring changes reported behavior without improving matching.
The Dutch controlled experiment examined by van den Berg and van der Klaauw (2006) provides an instructive distinction. Their counseling and monitoring experiment reports no evidence that the combined interventions altered the exit rate to employment in the study setting, while monitoring shifted activity from informal to formal search channels. The result cautions against equating more visible or more easily documented search effort with greater employment probability. It is a specific experimental result from a historical Dutch program, not a general conclusion that counseling never helps.
Meyer (1995) reviewed earlier American unemployment-insurance experiments in Lessons from the U.S. Unemployment Insurance Experiments. The article addresses reemployment bonuses and job-search programs, including differences between service interventions and eligibility checks. Its evidentiary value is historical and methodological: well-defined interventions can be evaluated by randomized assignment, and a faster benefit exit must be distinguished from an improved labor-market outcome. The present dossier relies on the publisher's bibliographic record and the independently available abstract of the related NBER analysis of UI experiments, without claiming replication of every experiment reported in the original review.
Administrative monitoring has at least three mechanisms. It can induce additional search effort, improve the information available to the claims authority, and change the risk of losing benefits for noncompliance. These mechanisms may reinforce or offset one another. Increasing appointment frequency could help some people through counseling while creating burdens for others whose work schedules, travel requirements, disability accommodations, or childcare responsibilities are poorly recognized. The estimated treatment effect is therefore inseparable from how the support and compliance components were combined.
A properly specified evaluation should identify whether its treatment is an actual employment service, heightened monitoring, a scheduled interview, a credible sanction risk, or a bundle. If the bundle affects the outcome, the researcher should avoid attributing its complete impact to the counseling component alone. This distinction is crucial to the U.S. RESEA case, whose objectives combine placement assistance with benefit integrity.
Training as an investment in employability
Training differs from immediate placement in its time structure. A person attending a course can temporarily reduce job-search intensity while acquiring qualifications expected to improve future employment and earnings. Evaluation may therefore show low or negative short-run effects during participation and more positive effects after program completion. That temporal pattern becomes invisible when program performance is judged only by quick exits from cash benefits.
Card, Kluve, and Weber (2018), in their meta-analysis of active labor programs, summarize estimates from more than 200 studies and report average effects close to zero in the short run, with more positive results two to three years after completion. They also identify heterogeneity by program type and participant characteristics, with comparatively larger average gains for human-capital interventions and certain groups. These are study-level tendencies, not a transferable percentage effect for an unexamined national program.
The interaction with unemployment insurance is immediate. During approved training, national law determines whether beneficiaries remain eligible, face modified availability tests, receive additional allowances, or transition to another support category. The same course may be classified as a voluntary service for a nonbeneficiary and a mandatory activity for an insured claimant. The monetary value of the intervention to the individual includes both training resources and the income consequences of enrollment. Researchers should trace these components separately.
Training allocation also creates a selection problem. Persons referred to lengthy courses may differ in their baseline job prospects from those offered immediate placement. A naive comparison of employment after six months can reward the program that selects job-ready individuals and penalize a training service that targets more complex barriers. Valid comparisons require randomization, defensible quasi-experimental identification, or sufficient covariate and longitudinal information to explain the selection process. Longer-horizon outcomes should include completed credentials, employment stability, wages, hours, and recurrent unemployment.
Profiling and individualized service intensity
Employment agencies may estimate the probability of prolonged unemployment to prioritize intensive support. The OECD's AI and employment-services assessment documents increasing use of artificial intelligence for matching, profiling, and client services, alongside issues of explainability, bias, governance, and evaluation. Profiling is a decision-support technique only insofar as the resulting allocation rules are accountable to the actual objective of effective assistance.
A risk score can measure predicted difficulty of obtaining work but does not itself measure treatment benefit. Consider two applicants, one with a very low baseline reemployment probability and a small expected response to the available service, the other with a moderate baseline probability and a large expected response. A triage rule based solely on predicted unemployment duration may give priority to the first, while a rule maximizing service impact may prioritize the second. Equity, legal entitlement, geographic access, and distributional priorities can warrant still different allocations. No single statistical score settles these normative and institutional questions.
Digital service systems can improve reach through remote appointments, vacancy alerts, and integrated identity verification. Their outcomes depend on accessibility, connectivity, language, data quality, procedural alternatives, and whether clients actually complete the service. A successful online log-in is an administrative event, not proof of counseling or employability improvement.
The EU Artificial Intelligence Act identifies certain uses of AI in access to essential public assistance and social benefits as high-risk, subject to the law's specified scope and application dates. A benefit determination has different legal implications from a voluntary career recommendation. For comparative research, it is necessary to document the precise automated function, human oversight, appeal channel, and legal basis instead of treating all PES algorithms as interchangeable.
Institutional case: United States RESEA
RESEA is a particularly explicit integration mechanism. Under 42 U.S.C. § 506, the program is organized to improve reemployment outcomes, reduce weeks of benefit receipt through employment, enhance integrity, and connect UI claimants with the workforce system. The statute establishes requirements for evidence-rated interventions and evaluation of interventions without the designated ratings. The current codified text incorporates amendments through 2024 and should be used in preference to superseded summaries when describing contemporary statutory requirements.
The Department of Labor's RESEA implementation guidance describes individualized assessment, career information, reemployment plans, enrollment in Wagner–Peyser employment services, and coordination with other American Job Center programs. Program design and take-up differ by state and selected claimant population; the federal description is not a state-by-state evaluation.
The Department cites a Nevada intervention associated with shorter average UI duration, lower average benefits paid, and better employment outcomes in the examined setting. Its reported average difference of 1.82 weeks and $536 less paid per participant, together with the stated savings ratio, should be attributed to that referenced intervention and its methodology. They must not be presented as universal RESEA effects or as proof that benefits withdrawn without employment improve welfare. A defensible fiscal evaluation would additionally identify program costs, wage gains, tax effects, possible claim reentry, comparison-group outcomes, and whether accelerated benefit exits came from actual work.
The distinctive significance of RESEA lies in joining an eligibility review with an identifiable service intervention. That design makes it possible to observe coordinated contacts, but it also complicates causal attribution between information, counseling, monitoring, and benefit enforcement. Evaluation should describe the treatment as delivered and, where feasible, vary components independently.
Work Transitions
From administrative status to economic outcomes
The main outcome of employment coordination is a transition in labor-market activity that is economically and legally meaningful. A person may stop receiving UI because they obtain a job, exhaust entitlement, fail a procedural requirement, move to another benefit, leave the jurisdiction, cease registering, or experience a data error. These are different endpoints. The observed administrative exit rate cannot be used as a direct estimator of the job-finding rate without reason-specific exit codes and independent employment evidence.
For analysis, define an individual's states as insured unemployment with payment, registered unemployment without payment, participation in training, employment, other nonemployment, and unknown status. A transition from paid benefits to employment can be counted only if a job is observed within a prespecified window and meets an explicit employment definition. Follow-up should distinguish any earnings from stable contracted work, minimum hours, job duration, wage level, and recurrent unemployment. The chosen classification reflects a research purpose and should remain consistent throughout each comparison.
The OECD–European Commission activation-requirements review demonstrates the diversity of formal search and compliance requirements. More demanding reporting rules can increase observed exit from an administrative benefit register without producing a proportionate improvement in employment. That possibility requires explicit testing rather than inference from regulation alone.
Evaluation horizons and the job-match quality problem
Short-horizon outcomes measure the speed with which someone obtains any work; medium-horizon outcomes detect job retention, hours, and early earnings; longer-horizon outcomes reveal skill acquisition, wage progression, recurrent unemployment, and benefit dependence. A program can dominate at one horizon and perform poorly at another. Training can create a temporary participation effect, while aggressive rapid placement can lead to repeated low-quality job spells. Thus, duration is one dimension of performance rather than a complete definition.
Card, Kluve, and Weber's evaluation meta-analysis supplies an empirical reason to use several horizons: average postprogram effects change over time and differ across intervention types. Evaluation designs should report the observation window measured from assignment and, if appropriate, completion. Starting the clock at training completion while ignoring time spent in training introduces a survivorship and accounting problem. Employment measured 24 months after assignment and employment measured 24 months after completion answer different questions.
The relevant population includes repeat claimants. A first job may end quickly, and the worker can reenter a benefit system with altered eligibility. A longitudinal measure of weeks employed, cumulative earnings, and recurrent claims captures this dynamic more reliably than a binary employment flag at a single date.
Displacement and general-equilibrium effects
A successful individual placement need not represent a net increase in employment in a local labor market. Firms may hire a participant instead of a similar nonparticipant, and scarce vacancies may be redistributed among competing jobseekers. Crépon, Duflo, Gurgand, Rathelot, and Zamora (2013) directly tested this concern in France through a clustered randomized placement experiment. Their design varied treatment saturation across 235 labor markets and then randomized eligible jobseekers within those markets. It found initial gains for treated youth, but also evidence that some gains displaced other eligible jobseekers, with small overall net benefits in the studied setting.
The inferential consequence is substantial. A program may improve the treated claimant's job probability while offering a smaller aggregate labor-market gain. The difference is particularly material in weak-demand labor markets, where the available vacancies are not expanding. Evaluations should therefore distinguish individual treatment effects from effects on untreated workers, participating firms, and total employment. Saturation designs, regional comparisons, employer data, and credible structural or quasi-experimental approaches can help characterize these channels, though each introduces additional assumptions.
Displacement effects do not erase a participant's private gains or distributional relevance. Helping a long-term unemployed or disadvantaged worker compete for existing vacancies may have equity value even when net employment is unchanged. The analyst should state the chosen evaluative objective—beneficiary welfare, distributional redistribution, aggregate jobs, insurer expenditure, or public finances—rather than treating one measure as a proxy for all of them.
International institutional contrasts
Germany's social insurance–employment agency interface gives a distinctive role to early jobseeker registration, eligibility registration, placement duties, and individualized integration planning. The Federal Employment Agency's registration sequence reveals how statutory timing and advance placement can be coordinated across the transition from employment into unemployment. Observing the timing of the three steps is essential to evaluating delays and interruptions.
The United States RESEA arrangement provides an integration entry point for selected unemployment claimants in a federal-state insurance structure, documented by the Department of Labor program record. The responsible state agencies and workforce partners may have distinct responsibilities. Treatment and nonparticipant populations must be specified before cross-state estimates are compared.
The United Kingdom's Jobseeker's Allowance claimant process connects a work coach's agreement, appointment cycle, and evidence of search with ongoing benefit responsibilities. The statutory and practical context of Jobseeker's Allowance needs to be distinguished from the much broader Universal Credit system when constructing a country case. Australia offers a different interface in which JobSeeker mutual obligations link means-tested payment administration to employment-service providers and personalized plans. These schemes illustrate functional coordination across different underlying types of income maintenance.
International classification requires more than naming a country as an integrated or separated model. Even where the same organization administers services and transfers, payment authority, provider contracting, local discretion, quality assurance, and appeals may be legally separate. A comparative case matrix should identify each function and its controlling institution.
Economic and fiscal consequences
The fiscal consequences of coordination include benefit payments, administrative costs, counselor time, training and placement expenditure, taxes on subsequent earnings, employer recruitment costs, and expenditure on other public programs. A reduction in UI outlays can arise from improved employment or from curtailed access to benefits; the two pathways have different welfare effects. An employment service that increases stable earnings may generate broader gains that are omitted from a benefit-fund account.
A simple budget identity clarifies the distinction. For a fixed cohort and horizon, net program fiscal effect equals avoided UI payments plus additional fiscal receipts plus reductions in other public expenditure, minus service and training costs and any replacement support. Each term needs a counterfactual estimate. It is an accounting identity, not an original model or an estimate of causal impact. Because benefit and tax records may be held by different agencies, coordination of research data is itself a prerequisite for sound assessment.
Where formal insurance has limited reach, a job-placement intervention can benefit noninsured workers without producing any change in UI payments. Conversely, a program targeted only at benefit recipients may miss workers in informal employment or those without sufficient contribution histories. The ILO's unemployment-protection coverage evidence establishes the importance of keeping insured and noninsured populations analytically separate.
The transition analysis therefore favors a joint appraisal of verified employment, earnings, job durability, household income, administrative burden, effective coverage, and public costs. No single administrative performance indicator answers all of these questions.
Evidence Verification and Sensitivity
Source hierarchy and legal dating
The source hierarchy in this inquiry gives priority to the enacted text of legal instruments for normative propositions; to official agency instructions for operational practice; to statistical documentation for measurement definitions; and to studies with identifiable designs for claims about effects. Publisher metadata and abstracts confirm the identity and reported result of a study but do not permit independent reproduction of its estimates. When evidence is limited to an abstract or program description, the associated claim is confined to that level of access.
Treaty obligations require a ratification and implementation inquiry. ILO Conventions Nos. 88, 102, and 168 are authoritative texts, but membership, branch acceptance, reservations or declarations, legal effective dates, and national implementing legislation determine their practical legal reach. Domestic benefit provisions similarly change over time. U.S. 42 U.S.C. § 506 is read from a U.S. House compilation indicating legislative currency through October 2026; it reflects the 2024 amendment, but program implementation within particular states remains a separate evidentiary issue. Germany's Social Code Book III, British claimant guidance, and Australian mutual-obligation instructions are treated as sources of jurisdiction-specific administrative rules at the access date.
A legal rule cannot be inferred from an organization's secondary policy summary where the controlling law is available. Conversely, a statutory provision does not establish the empirical frequency or quality of actual administration. These categories are recorded in the evidence ledger as normative, administrative, statistical, or causal sources.
Measurement and denominator sensitivity
The most frequent statistical failure is denominator substitution. Registered unemployment, household-survey unemployment, benefit eligibility, current benefit receipt, and service-program enrollment are overlapping but different sets. Even within insured unemployment, counts of people, claims, and benefit weeks are not equivalent. A person can have multiple spells, and one spell can span a fiscal boundary. The European Commission labor-market policy statistics methodology records stocks, flows, expenditures, and intervention categories separately. These fields should be retained rather than combined into an undefined activation total.
The ILO's reported 16.7 percent unemployment-benefit coverage ratio is a population indicator, not an institutional effectiveness score. It cannot support country-level inferences about PES service quality without harmonized participation and claimant records. Comparing a national program's reemployment rate among voluntarily attending claimants with another country's employment rate among all registered unemployed would create severe composition bias.
Income and fiscal figures require further standardization. Dollars and euros must be associated with price year, nominal or real measurement, exchange convention, purchasing-power adjustment if used, and fiscal or calendar accounting. A per-participant saving should state whether it includes only UI cash payments, program administration, other transfers, and tax revenue. The Nevada figures reproduced in the U.S. Department of Labor summary are historical treatment-specific claims; this dossier does not update them for inflation or infer a contemporary statewide budget forecast.
Identification and selection
Let Y(1) denote the employment outcome under an assigned service intervention and Y(0) under a defined counterfactual. Their difference is a causal contrast conceptually, but only one potential outcome is observed for each participant. Randomized assignment estimates an average contrast subject to implementation, attrition, spillovers, and statistical uncertainty. Nonrandom assignment requires identifying assumptions that make observed differences informative about the counterfactual.
For a coordination intervention, selection can occur at several stages: becoming insured, filing a claim, registering with the PES, being selected for a program, attending an appointment, accepting a vacancy, and remaining observable in administrative data. Conditioning on a later stage can distort the estimate of an earlier policy. Treatment effect estimates should identify whether they concern intention to treat, actual participation, or a particular selected subgroup.
Interference among claimants is another central challenge. The French saturation experiment demonstrates that untreated individuals' outcomes can change when other jobseekers receive assistance. A design assuming every jobseeker's treatment has no effect on anyone else may exaggerate aggregate gains in a fixed-vacancy environment. Randomization at labor-market level, saturation variation, or supplementary equilibrium analysis is valuable when the objective is net employment rather than only participant outcomes.
Heterogeneous impacts call for careful interpretation. Age, gender, disability, migration status, skill level, unemployment duration, and local demand may interact with the same intervention. The Card–Kluve–Weber evidence synthesis reports systematic variation by participant group and horizon. A pooled average with no subgroup composition data may conceal policy-relevant distributional differences.
Procedural error and digital exclusion
Claims and services depend on administrative information that can be wrong. Incorrect employer contribution records, identity mismatches, missed notifications, duplicate appointments, poorly explained conditions, inaccessible digital interfaces, and delayed appeals may affect benefit continuity without reflecting job-search behavior. Reliable administrative data should distinguish a suspected eligibility issue, an adjudicated decision, an appeal, a reversed decision, and a payment correction.
Employment-service automation can increase routing speed, but it creates model and interface risks. The OECD 2024 analysis of AI in PES identifies transparency, explainability, stakeholder involvement, and evaluation as governance priorities. A case study should record a model's target variable, date of training data, observed error rates, distribution of errors, opportunities for human override, and pathways to contest a consequential decision. This is particularly important when service prioritization and benefit compliance share a data infrastructure.
An apparent increase in program efficiency may merely shift cost onto claimants who must travel, find childcare, provide documentation, or negotiate inaccessible digital systems. Those costs belong in a social evaluation even when they are absent from the insurer's administrative budget. They should be measured rather than assumed, using suitable survey and administrative methods.
Sensitivity tests for a serious evaluation
At minimum, analysis should vary the definition of employment; the horizon from assignment; the treatment inclusion criterion; the handling of benefit exhaustion; the treatment of missing wage records; the interpretation of nonparticipation; the duration of postemployment follow-up; and the geographic labor-market boundary used to assess spillovers. Results should show absolute values and denominators, relative and percentage-point changes where appropriate, and uncertainty intervals.
An additional sensitivity exercise separates the components of coordinated service. One comparison might contrast counseling plus eligibility assessment against ordinary service; another could examine enhanced service without additional monitoring, if the legal and practical intervention supports such comparison. Without independent variation, it is inappropriate to claim that one component caused the total program effect.
Current evidence also has an institutional coverage limitation. Most readily documented detailed conditionality systems come from advanced economies with contributory insurance or robust income-support administration. The ILO's comparative unemployment report establishes much weaker effective protection in lower-income settings. Extrapolating formal insured-claimant results to informal workers, people with no benefit entitlement, or conflict-affected labor markets would exceed the support of the cited studies.
Finally, reproducibility requires a traceable bibliography and source-to-claim ledger. This manuscript's companion EVIDENCE.json distinguishes inspected legal text, official program descriptions, full institutional summaries, and article abstracts. It records where a claim is an interpretation of several verified facts rather than a numerical finding from an original dataset. No person-level administrative microdata, national compliance audit, causal reestimation, external peer review, or original survey has been performed for this dossier.
Scientific Synthesis and Research Agenda
A comparative reading of institutional coordination
The evidence supports an account of PES–UI coordination as an ordered set of legal and administrative relationships. Benefit entitlement must be established from authoritative rules; a service must be accessible and actually delivered; activation must be suited to the person's circumstances and labor-market opportunities; monitoring must distinguish compliance from administrative error; and subsequent employment must be measured independently of benefit status. Coordination is strongest when information can move reliably across these stages while each decision retains a clear accountable authority.
This interpretation arises from the combined reading of international labor standards, country-level arrangements, and evaluation evidence. ILO Convention No. 88 defines the public service function; Convention No. 102 establishes unemployment protection benchmarks; and Convention No. 168 expressly relates income support to active employment promotion. These instruments support examining service and protection together while preserving national legal specificity.
Comparative empirical evidence provides a complementary lesson. The 2026 global PES report shows institutional diversity, including uneven involvement in unemployment income support. The 2018 program meta-analysis shows that employment effects differ by time horizon and participant group; the French placement experiment makes displacement observable; and the Dutch monitoring experiment shows that recorded search channels may change without an observed improvement in transition rates. Taken together, these sources justify multi-outcome evaluation and reject a simple identification of administrative activity with economic success.
Conditions for effective coordination
The first condition is clarity of entitlement. A benefit claimant must be able to identify the qualifying rule, submitting authority, relevant time limit, payment decision, and applicable avenues of review. A service user must likewise understand registration and referral requirements, including which activities carry benefit consequences. Clear notice and accessible case handling reduce avoidable gaps between substantive rights and effective receipt.
The second condition is a coherent service pathway. A timely referral should lead to an appointment, an individualized needs assessment, a feasible intervention, and observable follow-up. The pathway accommodates different labor-market situations: rapid vacancy referral for job-ready claimants, sectoral retraining for workers displaced by technological change, language or accessibility assistance where necessary, and employer-side engagement in weak-demand regions. The available evidence supports these as types of intervention to evaluate, not a fixed universal sequence.
The third condition is separation of legally consequential decisions from service-provider incentives. Contracting a provider to maximize rapid placements can lead to attention being concentrated on easy-to-place clients; sanctions triggered by automated nonattendance flags can misclassify claimants with valid reasons. Transparent decision rules, documented human review, reasoned notifications, and correction mechanisms help prevent incentives to manipulate apparent performance.
The fourth condition is evaluation of actual outcomes. The statutory U.S. RESEA program provides an example of explicitly connecting reemployment, evidence-based interventions, and program integrity. Its framework also illustrates why independent evaluation of employment and earnings is necessary when the same process assesses eligibility. Quality of employment, benefit continuity, displacement, administrative burden, and net fiscal consequences should be reported alongside weeks of UI receipt.
The fifth condition is adaptation to institutional context. In high-income settings with extensive insurance, the dominant interface may be insurance case management and employment-service personalization. Where effective unemployment benefit coverage is low, the immediate priority may be accessible job-placement infrastructure and lawful pathways to expand income protection. The ILO's effective coverage assessment provides the international empirical grounding for this difference.
A practical empirical specification
An evaluable research design would follow a defined cohort from the point of job loss, claim registration, or random service assignment. The chosen entry point must be explicit because each captures a different selection process. The minimum observational fields would include jurisdiction, date, insured status, statutory qualifying basis, registration timestamp, benefit decision, payment weeks and amount, referral and service dates, attendance, exemptions, administrative review, training duration, employer placement, subsequent earnings, job stability, and recurrent claims.
The core outcomes are cumulative employment weeks, earned income, sustained employment at prespecified horizons, actual payment receipt, time from qualifying job loss to first payment, and administrative or participant cost. Secondary outcomes can include completed credentials, occupational mobility, geographic mobility, and transitions among unemployment, employment, informal work, education, and nonparticipation. The observation window should be long enough to capture training effects and recurrent unemployment.
A randomized evaluation of integrated case management could compare an enhanced coordination protocol with current practice among otherwise eligible claimants. Assignment should be audited for fidelity; intention-to-treat estimates should include all those assigned, not only those who completed sessions. Where randomization is infeasible, phased implementation, a credible discontinuity, or a carefully specified difference-in-differences design may supply conditional evidence. The design should also investigate untreated jobseekers and local vacancy conditions to detect displacement.
Cross-country research requires a common data dictionary before pooling outcomes. A statutory unemployment insurance beneficiary in Germany, a selected American RESEA participant, a British JSA claimant, and an Australian means-tested JobSeeker recipient belong to different entitlement and service populations. Data harmonization should retain these differences rather than erase them through a generic label. Standardized outcome definitions can coexist with separate legal and financing classifications.
Institutional implications
For lawmakers, the relevant question is how legal rules ensure benefit continuity and protect rights while making employment assistance operationally effective. For administrators, it is whether data and responsibility flow correctly among the insurance authority, caseworker, service provider, employer, and reviewing body. For evaluators, it is whether the evidence identifies the causal contribution of service coordination rather than simply documenting changes in payments or registrations.
The comparative record supports investments in interoperable records, service-quality measurement, longitudinal employment follow-up, and transparent allocation criteria. It supports controlled experimentation with intervention intensity and training duration where legally permissible and ethically administered. It also supports carefully limited use of profiling systems, independently monitored for different error patterns across claimant groups.
These are analytically derived institutional implications, not a new theory, doctrine, or formally validated design standard. The evidence does not identify a universal optimum for sanction intensity, staffing ratio, office merger, or automated allocation. Such claims would require new cross-system data, explicit welfare weights, implementation-cost evidence, and controlled causal comparison.
Research priorities
The most consequential open empirical question is the marginal effect of integrating claim assessment and actual employment assistance while keeping the intensity of monitoring separately identifiable. A second priority is long-term job durability and wage progression following training versus immediate placement. A third is spillover effects in labor markets with limited vacancy creation. A fourth is access for people whose income losses fall outside contributory unemployment insurance, including informal and intermittently employed workers. A fifth is the reliability and distributional effect of algorithmic profiling and remote service delivery.
These questions can be studied with linked administrative histories, representative surveys, employer data, preregistered outcome definitions, and institutional comparisons that carefully specify national law. Progress depends on recognizing that service coordination has administrative, social-insurance, and labor-market dimensions and that each demands its own valid evidence.
Conclusion
Public employment services and unemployment insurance jointly govern the passage from job loss to possible reemployment. Their coordination is a matter of legal entitlement, administrative execution, service access, and labor-market outcome measurement. International conventions establish standards for employment services and unemployment protection; national law determines eligibility and institutional powers; evaluation research identifies where intervention effects vary, where benefit exits may be misread, and where individual placement gains may displace other workers.
The comparative evidence yields a consistent scientific conclusion: coordinated systems should be evaluated on the complete claimant and jobseeker trajectory, including the accuracy and timing of benefit payments, meaningful service delivery, procedural fairness, verified earnings and sustained employment, distributional effects, and public resource costs. Organizational integration is valuable when it improves those outcomes under the applicable legal framework. Its effectiveness must be demonstrated with source-grounded longitudinal and causal evidence.
The original contribution of this dossier is a structured synthesis of those relationships, supported by verified legal texts, international statistical definitions, official country practices, and empirical evaluation research. Its findings remain bounded by the cited dates, jurisdictions, populations, and observation horizons. The monograph establishes an evidence-based comparative analysis and a concrete research agenda for public employment services coordination with unemployment insurance.
References
Brioscú, Ailbhe, Anne Lauringson, Anne Saint-Martin, and Theodora Xenogiani. 2024. A New Dawn for Public Employment Services: Service Delivery in the Age of Artificial Intelligence. OECD Artificial Intelligence Papers. OECD Publishing. DOI: 10.1787/5dc3eb8e-en. https://www.oecd.org/en/publications/a-new-dawn-for-public-employment-services_5dc3eb8e-en.html
Card, David, Jochen Kluve, and Andrea Weber. 2018. What Works? A Meta Analysis of Recent Active Labor Market Program Evaluations. Journal of the European Economic Association 16 (3): 894–931. DOI: 10.1093/jeea/jvx028. https://doi.org/10.1093/jeea/jvx028
Crépon, Bruno, Esther Duflo, Marc Gurgand, Roland Rathelot, and Philippe Zamora. 2013. Do Labor Market Policies Have Displacement Effects? Evidence from a Clustered Randomized Experiment. Quarterly Journal of Economics 128 (2): 531–580. DOI: 10.1093/qje/qjt001. https://doi.org/10.1093/qje/qjt001
European Commission. n.d. Labour Market Policies (LMP) Database: Methodology and Categories of Interventions. Directorate-General for Employment, Social Affairs and Inclusion. Accessed November 17, 2026. https://employment-social-affairs.ec.europa.eu/policies-and-activities/analysis-evaluation-impact-assessment-and-databases/databases-and-indicators/labour-market-policies-lmp-database_en
European Parliament and Council. 2020. Decision (EU) 2020/1782 of 25 November 2020 Amending Decision No 573/2014/EU on Enhanced Cooperation between Public Employment Services. Official Journal of the European Union L 400: 7–10. https://eur-lex.europa.eu/eli/dec/2020/1782/oj/eng
European Parliament and Council. 2024. Regulation (EU) 2024/1689 of 13 June 2024 Laying Down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act). Official Journal of the European Union. https://eur-lex.europa.eu/legal-content/EN-DA/TXT/?uri=CELEX%3A32024R1689
Eurostat. n.d. Labour Market: Basic Definitions. European Statistics Learning Hub, European Commission. Accessed November 17, 2026. https://ec.europa.eu/eurostat/web/european-statistics-learning-hub/statistics-for-beginners/labour-market-definitions
German Federal Employment Agency. n.d. Wie Sie sich arbeitslos melden [How to Register as Unemployed]. Bundesagentur für Arbeit. Accessed November 17, 2026. https://www.arbeitsagentur.de/arbeitslos-arbeit-finden/arbeitslosengeld/ihre-schritte-wenn-sie-arbeitslos-werden/wie-sie-sich-arbeitslos-melden
German Federal Employment Agency. n.d. Wie Sie sich arbeitsuchend melden [How to Register as a Jobseeker]. Bundesagentur für Arbeit. Accessed November 17, 2026. https://www.arbeitsagentur.de/arbeitslos-arbeit-finden/arbeitslosengeld/ihre-schritte-wenn-sie-arbeitslos-werden/wie-sie-sich-arbeitsuchend-melden
Germany. 1997, as amended. Sozialgesetzbuch Drittes Buch (SGB III), § 35, Vermittlungsangebot [Placement Offer]. Official federal legislation portal. Accessed November 17, 2026. https://www.gesetze-im-internet.de/sgb_3/__35.html
Germany. 1997, as amended. Sozialgesetzbuch Drittes Buch (SGB III), § 37, Potenzialanalyse und Eingliederungsvereinbarung [Potential Analysis and Integration Agreement]. Official federal legislation portal. Accessed November 17, 2026. https://www.gesetze-im-internet.de/sgb_3/__37.html
Inter-American Development Bank, Organisation for Economic Co-operation and Development, and World Association of Public Employment Services (IDB/OECD/WAPES). 2026. The World of Public Employment Services 2026: Delivering in a Changing World of Work. Inter-American Development Bank. DOI: 10.1787/4aebd080-en. https://www.oecd.org/en/publications/the-world-of-public-employment-services-2026_4aebd080-en.html
International Labour Organization (ILO). 1948. Employment Service Convention (No. 88). Adopted July 9, 1948. NORMLEX. https://www.ilo.org/dyn/normlex/en/f?p=1000%3A12100%3A0%3A%3ANO%3A%3AP12100_ILO_CODE%3AC088
International Labour Organization (ILO). 1952. Social Security (Minimum Standards) Convention (No. 102). Adopted June 28, 1952. NORMLEX. https://www.ilo.org/dyn/normlex/en/f?p=1000%3A12100%3A0%3A%3ANO%3A%3AP12100_ILO_CODE%3AC102
International Labour Organization (ILO). 1988. Employment Promotion and Protection against Unemployment Convention (No. 168). Adopted June 21, 1988. NORMLEX. https://www.ilo.org/dyn/normlex/en/f?p=1000%3A12100%3A0%3A%3ANO%3A%3AP12100_ILO_CODE%3AC168
International Labour Organization (ILO). 2023. Employment Services and ALMPs to Support Transitions. February 28, 2023. https://www.ilo.org/resource/article/employment-services-and-almps-support-transitions
International Labour Organization (ILO). 2024. World Social Protection Report 2024–26: Universal Social Protection for Climate Action and a Just Transition. Chapter 4, section 4.2.6, Unemployment Protection as an Integrated Approach. Geneva: ILO. https://webapps.ilo.org/static/english/reports/flagship/world_social_protection_report_2024-26/Chapter%2004.html
Meyer, Bruce D. 1995. Lessons from the U.S. Unemployment Insurance Experiments. Journal of Economic Literature 33 (1): 91–131. https://www.jstor.org/stable/2728911
Meyer, Bruce D. 1992. Policy Lessons from the U.S. Unemployment Experiments. NBER Working Paper 4197. National Bureau of Economic Research. DOI: 10.3386/w4197. https://www.nber.org/papers/w4197
Mroczka, Joanna, Daniele Pacifico, and Eddie Chen. 2025. Comparative Assessment of Activation Requirements for Unemployment and Minimum Income Benefit Recipients. OECD Social, Employment and Migration Working Papers. DOI: 10.1787/86ec18af-en. https://www.oecd.org/en/publications/comparative-assessment-of-activation-requirements-for-unemployment-and-minimum-income-benefit-recipients_86ec18af-en.html
Services Australia. n.d. Mutual Obligation Requirements: JobSeeker Payment. Government of Australia. Accessed November 17, 2026. https://www.servicesaustralia.gov.au/mutual-obligation-requirements?context=51411
United Kingdom, Department for Work and Pensions. n.d. Jobseeker's Allowance (JSA): Your JSA Claim. GOV.UK. Accessed November 17, 2026. https://www.gov.uk/jobseekers-allowance/your-jsa-claim
United States Department of Labor, Employment and Training Administration. n.d. Reemployment Services and Eligibility Assessment (RESEA) Grants. Accessed November 17, 2026. https://www.dol.gov/agencies/eta/american-job-centers/RESEA
United States Congress. 1935, as amended through 2024. 42 U.S.C. § 506: Grants to States for Reemployment Services and Eligibility Assessments. Official U.S. Code compilation, laws in effect through October 7, 2026. https://uscode.house.gov/us/usc/119-120/t42/s506/
van den Berg, Gerard J., and Bas van der Klaauw. 2006. Counseling and Monitoring of Unemployed Workers: Theory and Evidence from a Controlled Social Experiment. International Economic Review 47 (3): 895–936. DOI: 10.1111/j.1468-2354.2006.00399.x. https://doi.org/10.1111/j.1468-2354.2006.00399.x