Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Short-Time Work Compensation and Employment Retention

Abstract. Short-time work compensation is an income-protection instrument that finances some of the earnings lost when an employee's hours are temporarily reduced and the employment relationship continues. It can insure the household against an abrupt interruption, preserve valuable worker–firm matches, limit recessionary unemployment, and sustain aggregate demand. Its public value varies with the temporary or persistent character of the shock, the viability of the employer, the incidence of subsidies, and the credibility of the counterfactual against which jobs saved are estimated. This Research Dossier offers a comprehensive comparative synthesis of international labor standards, statutory design in Germany, France and the United States, EU crisis financing, OECD evidence, economic mechanisms and causal research in Europe, the United Kingdom and Australia. It distinguishes legal entitlement, take-up, attributable employment retention, fiscal expenditure, financing, deadweight, error, and social value. It establishes a coherent interpretation of early crisis effectiveness and later reallocation risks, while preserving the methodological differences among published estimates. The study supplies an original integrated analysis of the available evidence and a precise agenda for further evaluation without advancing a new authorial doctrine or causal estimate.

Keywords: short-time work; short-time compensation; employment retention; work sharing; furlough; partial unemployment; social insurance; unemployment benefits; Kurzarbeit; activité partielle; labor hoarding; deadweight; fiscal incidence; policy evaluation; job mobility.

Introduction

The employment relationship organizes pay, human capital, firm-specific knowledge, social insurance, training and future production. A sharp fall in output can interrupt the profitable use of labor without eliminating the economic value of this relationship once conditions recover. The firm then faces a choice between financing idle hours, shortening work schedules, and separating workers. Workers face an associated choice between reduced pay with a continuing contract and unemployment, frequently with considerable search costs. Short-time compensation modifies both decisions through public support for earnings foregone during a temporary reduction in hours. Its design determines which workers are protected, which firms have reason to preserve them, and which fiscal institution finances the interruption.

This economic mechanism has a distinct legal foundation. The ILO Employment Protection Convention of 1988 differentiates full unemployment from partial unemployment and temporary suspension of earnings where the relationship continues. The distinction is crucial: unemployment after dismissal, furlough without work, work sharing with reduced hours, and generalized payroll subsidies all protect income but apply to different statuses and may influence hiring and reallocation differently. Comparative claims become unreliable if an estimate for workers supported by a wage grant is silently used to characterize an insurance benefit for hours not worked.

Two major shocks have furnished much of the empirical record. Hijzen and Venn (2011) analyzed national short-time arrangements during the 2008–09 recession. During the pandemic, governments dramatically expanded standing programs and introduced new ones. The OECD's pandemic employment assessment documents approximately 60 million supported jobs, roughly 20% of employment in the relevant OECD population at the spring 2020 peak. A supported job is a counted participant; a saved job is a causal claim about a counterfactual separation that would have occurred without intervention. These quantities must remain separate throughout analysis.

The research questions are threefold. Which legal contingencies, payroll records and administrative rules determine access to short-time compensation? Under what circumstances can subsidies preserve productive employment and when do windfalls or labor-market lock-in dominate? How should public expenditures, fiscal offsets, debt financing, distribution and longer-run welfare be measured? The investigation uses primary statutes, official agency guidance and audit sources, comparative institutional research, and published original empirical studies. It identifies dated legal rules, beneficiary denominators, treatment and outcome definitions, the evidentiary strength of each finding, and areas where further data would change a conclusion.

A monograph on international social protection must distinguish statutory provision from actual protection. The ILO's 2024–26 protection review reports large gaps in effective unemployment benefit receipt worldwide. That statistic is not a global short-time compensation coverage estimate. Yet it identifies a central institutional issue: workers outside formal employment and contribution records can face the largest earnings shocks while remaining least likely to qualify for payroll-based retention schemes. Every empirical conclusion in the present work is thus bounded by the underlying population, jurisdiction, period and policy environment.

Conceptual and Historical Foundations

Definition and the protected economic relationship

Short-time work compensation is a benefit or publicly financed wage replacement related to a temporary reduction in scheduled or normal hours, generally with the worker's contract intact. It pays for some earnings that would otherwise disappear. A firm's retained employee may continue to work a portion of regular hours; some programs permit full temporary suspension or furlough. The defining feature for this analysis is the preservation of contractual attachment during an interruption in normal paid work, accompanied by income support linked to the interruption.

The economic object is a worker–firm match. Matches accumulate firm-specific skills, routines, organizational trust, client relationships and complementarities among workers. A layoff can destroy these assets even if demand recovers several months later. The employer may be unable to finance idle labor because current revenue has fallen, credit is scarce, or continued production is prohibited. The employee may prefer the security of the original relationship to a costly search in a labor market that has temporarily stopped hiring. Compensation for the hours lost distributes the immediate cost between the household, firm and public financing system.

The neighboring policy instruments require distinct definitions. Work sharing reduces everyone's hours rather than laying off some workers and can finance part of lost wages through unemployment insurance. Furlough may require no labor during a temporary period, but maintain the contract. Wage subsidies sometimes compensate employers for payroll regardless of unworked hours. Ordinary unemployment insurance covers a worker following a qualifying job loss, and social assistance addresses household needs under its own criteria. The OECD policy typology of 2020 and Eurofound's European comparative review document how pandemic interventions bridged and blurred these categories.

Insurance, stabilization and matching mechanisms

The household insurance effect arises because workers would otherwise bear a large immediate earnings loss during hours reductions. The employment effect arises when preserving the relationship avoids a costly separation and future vacancy-filling. The macroeconomic effect arises as maintained incomes support demand in a period of negative aggregate shock. Each effect has an independent counterfactual: the worker's alternative transfer and earnings, the firm's separation decision and subsequent match formation, and aggregate spending in the absence of support. A program can provide consumption insurance even if it saves relatively few jobs, and it can preserve firm matches even where aggregate demand is temporarily constrained by public restrictions.

Financing also changes incentives. A fully subsidized lost hour may be attractive to firms that would have retained workers without support. An employer contribution can reduce that windfall but create liquidity difficulties for otherwise viable enterprises. A worker benefit that replaces too little pay can fail to protect living standards, while an extremely high replacement rate sustained during recovery can reduce the private incentive to return to ordinary hours. These effects depend on benefit formulas, wage caps, funding sources, duration, administrative verification and the intensity of the shock.

The policy's insurance logic is strongest when the negative event is temporary, external and difficult for individual employers or workers to control. Its matching logic is strongest when long-run joint productivity remains positive. The Giupponi and Landais (2023) examine this economic interaction using Italian institutional variation and employment data. Their focus on welfare and labor hoarding demonstrates why causal retention must be considered jointly with lost output, transfers and the alternatives workers would otherwise pursue. The present study treats that literature as a source of mechanisms and bounded estimates rather than deriving a universal numerical rule.

Historical development and changed crisis objectives

European short-time schemes existed before the pandemic and were repeatedly adapted to economic crises. The OECD Great Recession study assessed the extent to which short-time working mitigated the collapse in labor demand during 2008–09. Evaluating the schemes became difficult because their adoption and generosity were correlated with the same shocks that caused unemployment. Germany's Kurzarbeit played an especially important role in this debate, but its experience was embedded in national bargaining arrangements, social insurance administration and the recession's distinctive industrial composition.

The pandemic created an additional challenge: governments directly halted or restricted many businesses while the productive potential of affected employees had not necessarily disappeared. Subsidized retention could therefore avoid separations driven by a public-health emergency and permit prompt restart. The OECD 2021 chapter on retention schemes documents the unprecedented simultaneous expansion of short-time arrangements, furlough systems and wage support. Programs sometimes prioritized speed and broad access over fine-grained screening, appropriately reflecting the administrative cost of delay at the acute stage, while opening substantial later questions about windfalls and improper payments.

A temporary shock can become structural when consumers permanently change behavior, production shifts geographically, energy inputs become durably more expensive or technology makes existing tasks obsolete. The OECD analysis of job mobility emphasizes that movement toward more productive jobs contributes to long-term growth. The resulting challenge is dynamic: early retention can prevent inefficient displacement, while late retention may postpone desirable mobility. Scientific analysis has to identify when the marginal costs and benefits change, not simply classify an entire scheme as successful or unsuccessful from a single year of payroll counts.

The institutional scope of comparative evidence

A national scheme is shaped by its payroll records, collective bargaining, employment contract rules, contribution history and revenue capacity. Legal conditions that appear alike can result in very different effective coverage because establishments differ in liquidity, administrative competence and workers' documented status. International reports can identify patterns, but they may aggregate unique participants, monthly recipients, jobs, full-time equivalents and hours. An apparent difference in national take-up or cost can reflect the denominator instead of a policy effect.

The ILO social protection coverage report is relevant to these exclusion questions, though unemployment benefit receipt and retention-scheme access are different variables. The historical lesson is that employment-linked transfers are effective only for relationships that government agencies can legally recognize, finance and verify. Effective access requires analysis of the worker, firm, payroll and claims process in addition to the nominal benefit statute.

Short-time Eligibility

International standards and national benefit definitions

The ILO Convention No. 168, Article 10 establishes partial unemployment and temporary suspension of earnings with continuing employment as international protection contingencies. Its provisions operate in the context of national obligations, ratification and implementation, rather than issuing an identical payment schedule worldwide. Legally relevant distinctions include employee versus self-employed status, a genuine reduction in work, the reason the hours were lost, minimum loss thresholds, the duration of entitlement, and whether the employer maintains insurance contributions.

Eligibility analysis needs separate measurement of legal population, administrative applicants, approved claims, payments and ultimate validated entitlement. A worker can satisfy statutory requirements while never receiving a payment because of inadequate employer action or documentation. A claim can initially be paid but later recouped following audit. Statistical receipt does not establish that a worker would have been dismissed without support. For every country, the observation window and the unit of counting must be documented.

Germany: Kurzarbeitergeld and the loss-of-work test

In Germany, SGB III §95 identifies requirements for ordinary short-time compensation, including a substantial loss of working hours and earnings, establishment eligibility, personal eligibility and notification. SGB III §96 requires a qualifying economic cause or unavoidable event, a temporary and unavoidable loss, and ordinarily that at least one-third of employees at an establishment experience an earnings loss exceeding 10% during the claim month. Emergency regulatory exceptions from earlier periods must be matched to their relevant effective dates rather than substituted for ordinary requirements.

The Federal Employment Agency's guidance describes an ordinary replacement of 60% of net earnings lost, or 67% under the relevant child condition. It also describes a 24-month temporary maximum for 2025–26, rather than the standard 12-month maximum; the duration applied to any claim depends on its historical starting point and governing rules. The March 2026 ministry explanation situates the instrument within employment preservation during temporary reductions. Legal analysis thus needs the month, applicable exception, employee earnings base and employer filing chronology.

Consider a worker who normally earns a net monthly wage W and whose actual net wage for reduced hours is w. The earnings shortfall is W − w. Under a stylized German ordinary calculation, compensation approximates 0.60(W − w), with a 0.67 factor where the child condition is met, subject to statutory computation rules. This illustration is a representation of the replacement logic, not an individual benefit calculation or an assertion that all net-wage differences are administratively calculated by a simple subtraction. It highlights why comparing percentages across jurisdictions without specifying the wage base is misleading.

The employment relationship remains a central condition. Temporary contracts, seasonal arrangements and employment status may affect eligibility. Periodic reduction, hours worked, payroll taxes, employer compensation and training are different reporting variables. A German employer's total public subsidy can exceed the worker's immediate cash compensation when associated employer costs receive additional relief during exceptional periods. Fiscal analysis must specify those components.

France: activité partielle and gross-wage compensation

France's ordinary activité partielle compensates a worker placed on reduced hours by an employer under legally recognized circumstances. The dated Code du travail R5122-18 specifies an hourly indemnity equivalent to 60% of the relevant gross hourly wage reference within applicable ceilings. The official French worker guidance describes the categories of employment covered and the employer's payment role. The worker indemnity and the public reimbursement rate to the employer are distinct amounts; special arrangements may use separate parameters.

France's gross-reference formula therefore cannot be compared arithmetically with Germany's 60% of lost net remuneration without reconstructing taxes, caps, reference hours and actual earnings. A nominally identical replacement percentage can produce different net compensation. When support is paid through payroll and reimbursed later, the delay can create a liquidity requirement for the firm. This requirement influences access for small enterprises and may generate selection by employer financial strength even where employees are legally covered.

French research also demonstrates why eligibility is a necessary but insufficient causal indicator. Cahuc, Kramarz, and Nevoux (2021) show heterogeneous effects across firms with different shock exposure. A legally eligible establishment with little risk of layoffs can obtain a lawful subsidy that creates a fiscal windfall. A severely exposed establishment can preserve otherwise valuable jobs because the subsidy closes a temporary financing gap. The legal category alone cannot classify either outcome.

United States: federal-state work-sharing insurance

The American short-time compensation instrument is generally administered through state unemployment insurance programs under a federal statutory definition. The U.S. Code §3306(v) describes employer participation under a voluntary short-time plan, a workweek reduction of at least 10% and at most 60%, payment proportional to the unemployment compensation otherwise payable to the worker, and maintenance of certain benefits subject to statutory conditions. States retain important implementation choices, and program participation and eligibility can differ.

The federal proportional logic can be expressed as STC = rU, where r is the eligible fraction of hours reduced and U is the worker's otherwise payable unemployment benefit for full qualifying unemployment. The person also earns pay for hours actually worked. This differs from a direct percentage of a net wage shortfall. For example, a 20% reduction in hours generally implies payment of approximately 20% of the relevant unemployment insurance entitlement, rather than a 20% wage replacement rate. State-specific computation and caps govern actual payments.

The option to spread reduced hours among employees instead of selecting a smaller number for layoff can preserve employer skill pools and distribute income risk. It also means the same number of covered employee relationships may conceal substantial reductions in aggregate hours and labor income. A valid causal outcome should include hours and earnings as well as the simple existence of a contract. U.S. evaluation additionally requires sensitivity to differences across states in program adoption and employer awareness.

European Union financing: SURE and national schemes

The European Union's SURE Regulation of 2020 established temporary financial assistance to member states, primarily to help finance national short-time arrangements and similar protective measures. It did not institute one pan-European employee indemnity with uniform eligibility. The European Commission's final SURE implementation report records €98.4 billion disbursed in loans to 19 member states. Loan disbursement, approved national spending and cash benefits to individual workers must not be summed as independent fiscal burdens.

This distinction is central to international public finance. A national government can pay an employer reimbursement financed in part by borrowing from a supranational facility. The worker receives one compensated loss, the member state incurs a public expenditure and a debt obligation, and the Union holds a loan asset and corresponding financing liabilities. The same financing chain appears in several public accounts. Proper consolidation avoids double counting while retaining interest, repayment and risk details.

Eligibility gaps and administrative observability

Formal payroll and contribution records facilitate fast eligibility verification. Workers with irregular hours, informal employment, dependent self-employment, new contracts or multiple simultaneous jobs may face complex or incomplete entitlement. Employer-managed applications can also place the worker's protection behind the firm's choice to file. The Eurofound's 2024 comparative account describes changes in European access and conditionality during the pandemic, while the ILO's global evidence on protection gaps identifies wider problems in employment-linked coverage.

The scientific denominator should therefore be defined as workers within the relevant statutory scope and affected by a qualifying reduction, not all workers in a national labor force unless a broad employment-population rate is explicitly intended. A take-up rate using the employed population can be informative about overall reach, but it is not a fraction of those legally entitled. It is likewise unsuitable as a direct measure of the policy's causal contribution to employment continuity.

Employment Retention

Causal estimands: recipients and jobs saved

A paid claim identifies participation. A job saved is an employment relationship that would have ended in the absence of the relevant policy, over a specified observation window, but survived because of it. The causal problem is to estimate an unobserved outcome for supported firms and workers: whether the relationship would have continued without support. It requires more than observing workers after receipt or comparing subsidized employees with all others.

For an aligned period, define N as eligible supported relationships and A as the number for which support causally prevents separation. The proportion A/N describes attributable retention among recipients if estimation assumptions and units are defensible; (N − A)/N is the complement, sometimes interpreted as direct deadweight. This simple accounting equality is not an original theory and does not capture indirect effects on competitors, job creation, worker welfare, inflation or economy-wide demand. It is useful for preventing confusion between program scale and program impact.

Self-selection creates two opposite distortions. Firms with the largest negative shocks are most likely to apply, so their workers might show poorer outcomes even when the benefit strongly improves their prospects. Alternatively, financially sophisticated or administratively capable firms may be disproportionately likely to apply and also to retain workers for unrelated reasons. Evaluations must account for pretreatment shocks, firm dynamics, worker composition, policy variation and possible changes in who enters the sample.

Great Recession findings and Germany

The Hijzen–Venn international comparison investigated national schemes during the 2008–09 downturn, an important historical benchmark because a contraction in manufacturing and global demand affected countries very differently. It documented the plausible stabilizing role of short-time work while confronting the limitations of aggregate comparisons. Where industrial structure, labor protection and policy generosity jointly determine participation and separations, a cross-country correlation cannot cleanly identify the jobs attributable to compensation.

Balleer and coauthors (2016) analyzed Germany using a structural business-cycle model and distinguished the role of automatic rule-based short-time arrangements from discretionary policy variation. In their setting, rule-based operation contributed to employment stabilization in a way not mirrored by discretionary variation. The result is conditional on their model, time period and specified shock processes; it does not mean discretionary emergency extensions are necessarily ineffective under completely different circumstances.

Germany's institutional experience illustrates two requirements of credibility. First, a preexisting administrative and legal framework can deliver payments quickly when businesses face an identifiable downturn. Second, automatic thresholds must still be adapted when an exceptional shock changes the proportion of establishments affected. A rule that is efficient during an ordinary recession can exclude businesses facing a general shutdown, whereas an indiscriminate prolonged exception can later create unnecessary subsidy.

Firm heterogeneity in French research

The French heterogeneous-effects analysis examines firms that vary in shock exposure and finds substantial job preservation among severely adversely affected employers, alongside windfall gains when the employment relationship was much less threatened. This contrast explains why a high number of recipients can coexist with large direct benefits and nontrivial deadweight. The relevant feature is the difference between expected separations with and without intervention, rather than simply the absolute number of claims.

Firm finances matter. A temporary revenue shock may force a heavily indebted but otherwise productive firm to separate employees if it lacks financing for wage bills. Another firm with ample liquidity can preserve the same workforce without public aid. A subsidy can therefore be valuable to preserve efficient matches where liquidity binds, yet largely redistributive to owners where it does not. Conditions such as verified revenue reductions, employer contributions or time-limited thresholds can improve targeting, though they also introduce administrative error and barriers for the most constrained firms.

The original literature should be read at the level of its identified establishments, sample construction and shock definitions. Its estimates do not constitute a universal numerical coefficient for other economies or later periods. Nevertheless, the heterogeneous mechanism provides a strong analytical basis for studying why average program effects change as a recession passes.

Italian labor hoarding and welfare

Giupponi and Landais's Italian study uses social-security records and institutional variation to analyze subsidized labor hoarding, employment and welfare. The term labor hoarding refers to a firm's decision to keep more employees attached than current production immediately requires. A subsidy can make that decision privately feasible and socially desirable when labor demand is expected to resume, and socially costly where productivity has permanently shifted.

The Italian evidence highlights the need to evaluate the preserved match against alternative jobs. A separated worker may face prolonged unemployment and skill decay, or may enter a higher-productivity employer. The comparison depends on economic conditions, worker skills, mobility, retraining and the local vacancy market. Evaluations with short horizons may find highly favorable retention effects but miss future lock-in; long-run studies can similarly overlook the avoided harm of initial job loss.

Pandemic expansion and measured effects

The emergency expanded participation at an unprecedented rate. The OECD's October 2020 assessment reported roughly 50 million jobs receiving support by May 2020, and the 2021 Employment Outlook chapter reported a peak around 60 million jobs across its coverage. These two figures refer to different observational snapshots and potentially different reporting sets; they are compatible as descriptions of scale. Neither measures how many separations the programs prevented.

The United Kingdom introduced a broad furlough arrangement through the Coronavirus Job Retention Scheme. The government's final CJRS evaluation records around 11.7 million employments supported across the scheme, about £70 billion of payments, and an estimated four million jobs directly protected under the evaluation's counterfactual. Furloughed workers initially performed no work for the employer; later flexible arrangements allowed partial work. The outcome must be interpreted as an evaluation-based estimate rather than as a simple administrative count or a directly observed avoided separation.

Australia's JobKeeper took a different path with a broad wage subsidy and employment-attachment conditions. Bishop and Day (2020) estimated at least 700,000 employment relationships retained in the first four months of their study under eligibility-threshold and extrapolation assumptions. Bradshaw, Deutscher, and Vass (2023) later estimated 300,000–700,000 additional jobs preserved through a distinct discontinuity approach. Both findings depend on eligibility thresholds and populations and should remain separate estimates rather than being arithmetically averaged. Their disagreement in bounds and timing is an analytic reason to inspect identification, not evidence of a contradiction.

Spain's ERTE arrangements combined temporary furloughs and reductions of hours with strong administrative adaptation. The OECD evaluation of Spain's ERTE describes participation approaching one-quarter of the workforce in April 2020. The assessed gains in job continuity were linked to the acute pandemic context and specific eligibility changes. A quarter of the workforce receiving support is a measure of reach; the causal number preserved is determined by the policy evaluation's counterfactual model.

The OECD's February 2026 Belgium assessment provides an informative recent estimate. In that report, retention measures are estimated to have averted a 12.9% fall in employment at the peak of the pandemic; for every 100 subsidized workers, 55 jobs were estimated to be preserved, implying a modeled 45% deadweight share. This is a study-specific causal interpretation, not a count of fraudulent claims or a share directly audited. The report also identifies weak effects for teleworkable occupations and potentially adverse employment effects of continuing broad protection well after restrictions were lifted.

Outcomes beyond continued contracts

Avoided layoffs can reduce wage scarring, unemployment insurance claims, and the loss of firm-specific knowledge. Such effects can persist after support ends. Yet the number of contracts preserved on an observation date alone does not measure hours supplied, productivity, earnings, job quality or employee mobility. A worker with a continuing contract at almost zero hours and inadequate income may have experienced limited insurance benefit despite appearing retained in administrative records.

The OECD's worker-mobility analysis and European Commission's 2025 synthesis draw attention to productivity and transitions following the immediate crisis. Lock-in can arise when an otherwise mobile worker remains at a weak employer because subsidized attachment is temporarily more attractive than entry into a growing firm. It can also arise indirectly if public protection of incumbents discourages otherwise viable firms from hiring displaced or new entrants.

The empirical outcome set should include person-months employed, separations, earnings, hours, business survival, return to normal schedules, movement to new firms, unemployment spells, vacancies and value added. These variables are related but not interchangeable. Strong early preservation can coexist with a later mobility penalty, and measured fiscal cost can coexist with substantial nonfiscal social gains.

Fiscal Performance

Budget arithmetic and correct classification

The immediate fiscal cost of short-time compensation is the cash paid by a public authority or the accrued liability it incurs. This amount can include a benefit replacing lost earnings, reimbursement to employers, associated payroll contribution relief, financed training, administration and compliance spending. The costs depend on the claimant population, lost hours, benefit base, caps, program duration and wage distribution. Public accounts must additionally identify whether the same amounts appear as spending, fund transfers, advances or borrowings.

A general accounting identity for a defined time interval is: gross program outlays = cash compensation plus employer reimbursements and other subsidies plus administrative costs. A counterfactual net fiscal impact additionally incorporates avoided alternative unemployment payments, tax and social-contribution receipts resulting from greater employment, other welfare spending, financing and compliance costs. These are accounting distinctions whose numerical values must be estimated, not new causal laws. A proper estimate specifies whether it is cash-basis, accrual-basis or a discounted present value.

The European Commission's SURE final report illustrates a common double-counting risk. The €98.4 billion recorded is loan financing for 19 countries, not a separate €98.4 billion cash benefit paid on top of national disbursements. The 2020 SURE Regulation sets out the lending mechanism. A country might spend national resources financed partly by EU loans, with repayment obligations affecting future budgets. Fiscal consolidation must record the benefit expenditure once and the corresponding financing separately.

Social and exchequer valuations

A transfer from taxpayers to a supported household is a cost to the public budget but also income to a recipient. A welfare analysis asks whether preserving productive matches, maintaining consumption, avoiding social damage and limiting unemployment generates benefits that justify real resources and distributional effects. It must also consider deadweight, employer profits, distortions and the fiscal cost of raising funds. An exchequer analysis narrows the perspective to government cash flows, obligations and receipts.

The UK CJRS value-for-money evaluation offers an explicit comparison. Its central evaluation puts the program's net social present value at approximately £50 billion while estimating a negative £25 billion net exchequer present value. The estimated four million directly protected jobs and about £70 billion total payments sit within these larger modeled accounts. The positive social and negative fiscal results are compatible because the perspectives include different benefits and costs. These are the British government's modeled estimates, not independent recomputations within this dossier.

Program duration affects both measurements. A transfer that preserves a job for one month may be more or less efficient than a transfer that prevents a long unemployment spell, depending on the counterfactual. Dividing an entire multiyear program's nominal payments by jobs estimated saved at a single peak date generates a distorted 'cost per job saved' ratio. Units should be standardized as employee-months, full-time-equivalent job-years or another defined comparable measure, with all numerator expenses matched to the same period and population.

Australian gross payments and public accountability

The Australian National Audit Office's audit records roughly AUD 88.8 billion in JobKeeper payments and addresses administrative implementation. The Treasury's independent JobKeeper evaluation assesses overall performance, distribution and policy operation. The audit figure is a gross payment measure; neither an audited dollar nor a supported employment relationship proves that one additional job was maintained. Comparison with the 2023 Australian causal employment study requires matching exposure windows and allowing for uncertainty in extrapolating local eligibility effects.

An important fiscal classification separates lawful deadweight from improper payments. Deadweight arises when public assistance reaches eligible workers and firms whose employment would have persisted without the support. Program error arises from incorrect applications of eligibility or calculation rules; fraud involves deliberate misrepresentation. The British government's fraud and error assessment treats compliance losses as a separately evaluated component, while its overall social cost-benefit work also considers deadweight. An estimate of 45% deadweight in Belgium is accordingly not equivalent to a statement that 45% of claims were fraudulent.

Benefit rate, employer contribution and liquidity

The government's optimal contribution depends on the employment and income effects of each additional unit of support. If employers bear none of the cost of retaining underused labor, subsidies can finance continuing contracts that were never at risk. If employers bear too much, heavily constrained but productive firms can be pushed into layoffs. An employer co-payment can be a screening mechanism only to the extent that access to credit and the expected profitability of the retained match permit meaningful participation.

The French evidence on windfall effects suggests stronger net benefits under severe shocks, while Belgium's post-peak experience highlights diminishing returns to continuing expanded terms. The European Commission's policy conclusions recommends close attention to withdrawal, firm incentives and worker transitions. These studies support a dynamic evaluation of eligibility, co-financing and exit rather than the mechanical application of one generosity level to every phase of an economic disruption.

Payment timing changes effective coverage. Where the employer must advance compensation before reimbursement, a liquidity-constrained enterprise may be unable to comply even when the worker is eligible. Direct worker payments can reduce employer cash pressure but require distinct beneficiary verification. In either case, delayed disbursement can weaken the mechanism that is supposed to prevent urgent layoffs. The relevant administrative statistic is time from loss of work to reliable payment, alongside the eventual amount and accuracy of that payment.

Fiscal exposure and exit

During a systemic shock, claims rise just as wages, profits and tax receipts decline. Contribution-financed unemployment insurance reserves can therefore suffer simultaneous revenue and expenditure shocks. Tax-financed benefits face the broader sovereign budget constraint; borrowed finance smooths immediate obligations but introduces interest and repayment risk. The EU's SURE implementation record demonstrates one way of distributing crisis financing across levels of public administration while leaving national eligibility rules in place.

Withdrawal becomes economically important as production resumes. The OECD 2022 review of scheme adaptation describes average scheme participation declining from around 20% at the early-pandemic peak to 1.3% of employment at the end of 2021. These numbers describe changing reported use rather than attributing the entire reduction causally to any one program design. As ordinary hours recover, benefit expenses should respond to actual losses and later claims should be assessed against remaining economic need.

A credible actuarial or fiscal stress test states the number of eligible workers, participating proportion, hours fraction, wages, benefit formula, duration, return-to-work rate, compliance rates, employer contribution and interest assumptions. Correlation among affected firms must be included because a recession is a systemwide event rather than an independent set of small shocks. The necessary sensitivity analysis can be reported without inventing a numerical forecast or pretending to observe unprovided national microdata.

Evidence Verification and Sensitivity

Evidence classes and source dating

This monograph relies on a hierarchy of evidence. Statutes establish legal contingencies and entitlement requirements within a jurisdiction and time. Administrative guides describe implemented procedures and may change after the effective date of the underlying law. Audits establish the accounting and compliance evidence they actually test. Peer-reviewed or working-paper studies can estimate causal effects within their designs and assumptions. Comparative institutional assessments establish patterns only if definitions and reporting periods have been harmonized.

The German §95 entitlement provision and §96 work-loss conditions, the dated French compensation rule and the 2024 federal work-sharing definition were treated as dated legal sources. The German agency's current instructions describe a limited temporary 2025–26 duration exception. The research does not retroactively apply that rule to all German claims in 2020, nor does it infer French indemnity from a German lost-net-wage base.

The source verification profile in EVIDENCE.json distinguishes primary documents, accessible official reports, publisher summaries and original academic abstracts. The level of access matters. A publisher abstract can establish a study's stated question, methods and broad result, but may not permit verification of all robustness checks, tables or subgroup estimates. The present synthesis limits those claims accordingly and does not assert independent replication of paywalled analyses.

Quantitative comparability

The observable entities include workers, employments, claims, establishments, employers, benefit months and hours. A worker can have multiple jobs; a participant may receive assistance across several reporting periods. An establishment is not always equivalent to the enterprise that owns it. A claim filed does not always result in an accepted paid benefit. The CJRS cumulative evaluation describes both supported employments and an estimated counterfactual jobs-saved quantity, which should not occupy the same column of a causal outcomes table without a distinction.

The comparative empirical numbers below preserve their original observation units and source attribution.

Source and period Reported quantity Scientific meaning
OECD, spring 2020 Approximately 60 million supported jobs Participating jobs, not attributable jobs saved
United Kingdom, 2020–21 11.7 million supported employments; about 4 million directly protected jobs Cumulative administrative coverage and separate estimated causal effect
Australia, early pandemic 300,000–700,000 preserved jobs in a 2023 study Extrapolated causal employment effect from eligibility thresholds
Spain, April 2020 Nearly one-quarter of the workforce on ERTE Temporary support coverage, not the causal effect
Belgium, pandemic peak Averted 12.9% employment loss; 55 saved jobs per 100 supported OECD model-dependent causal assessment
EU SURE, through December 2022 €98.4 billion of loans to 19 member states Government financing rather than additional benefit payments

The table's observations come from the OECD 2021 comparison, British official evaluation, Australian Treasury study, OECD Spanish ERTE report, OECD Belgian analysis and European SURE implementation report. They should not be combined into a pooled international saved-job total. The sampling, time horizons, estimands, policy definitions and financing types are fundamentally different.

Identification strategies and interpretive constraints

A difference-in-differences design requires a credible counterfactual trajectory for untreated workers or firms. Eligibility cutoff designs identify local effects near the threshold under continuity assumptions and protection against manipulation. Instrumental variables require relevance and credible exclusion restrictions. Structural simulations use behavioral relationships and calibration rather than identifying every effect from directly observed policy assignment. A descriptive international panel can be valuable for assessing broad patterns while remaining vulnerable to unobserved country differences.

The 2020 RBA JobKeeper paper and 2023 Treasury JobKeeper paper use threshold-related variation but differ in sample, timing, estimand and generalization. The French establishment analysis distinguishes stronger shocks from windfall situations. The Italian quasi-experimental study investigates labor hoarding and welfare, while the German business-cycle model evaluates a specified macroeconomic mechanism. These distinct designs can jointly improve conceptual understanding while producing estimates that remain conditional on their contexts.

The temporary-shock premise itself is unobserved when a program begins. Policymakers often cannot identify which enterprises will recover until the shock has passed. Subsequent survival can be affected by the program, causing bias if analysis conditions only on firms that remain open. Administrative thresholds can induce behavioral changes, including manipulation of hours or reported payroll. Economic recovery coinciding with subsidy termination can conceal benefits that occurred earlier; persistent low-productivity matching can conceal costs that arise later.

Sensitivity domains

The most consequential sensitivity choices concern the share of work interruptions that would end without dismissal, the time required for normal production to resume, employers' liquidity constraints, workers' alternative job opportunities, the treatment of employer contributions, and the net benefit that workers would obtain if unemployed. If a large share of recipients would never have been separated, the fiscal cost per additional saved job rises. If an early layoff causes unusually long unemployment and the firm later restarts, the value of preserving the match rises.

Distributional conclusions are sensitive to the omission of workers outside formal payrolls, gender and caregiving constraints, firm size, disability, migration status and local sector composition. Fiscal results depend on whether a model counts foregone taxes, avoided unemployment support, social discounting and public financing costs. The ILO's international coverage evidence helps place program inclusiveness in a broader context, while the Eurofound comparative study documents changes in access across pandemic schemes.

An explicit source-to-claim audit must avoid three particularly damaging errors: treating a count of participants as an estimated treatment effect, equating a loan receipt with a final cost of the underlying program, and equating modeled deadweight with audited fraud. The UK employment and welfare estimates, European SURE report and Belgian deadweight assessment provide exact examples in which these distinctions change interpretation. Every contextual source hyperlink in this manuscript is mapped to a bibliographic record and evidence entry.

Access and research limitations

The study draws on available official law, government reports, institutional chapters and research-publisher abstracts. No confidential employer–employee panel, proprietary econometric code, state-by-state U.S. administrative microdata, original survey or unpublished ministerial decision file was analyzed. Regression discontinuities and national welfare estimates were interpreted from their publication descriptions, not independently re-estimated. The research provides a comparative scholarly synthesis and methodological analysis, rather than a new numerical causal result or an exhaustive world legal inventory.

Source-dating matters particularly for recent German and French statutory provisions and for policy programs that were temporary by construction. The authoritative law and administrative page should be revisited whenever a later edition is compared, without rewriting the historical rules applicable to previously observed outcomes. This manuscript's research cutoff is November 17, 2026. Its claims concern the documents and date ranges stated, and its assessment does not purport to predict unobserved future reforms.

Scientific Synthesis and Research Agenda

The conditions associated with valuable retention

The strongest substantive interpretation supported by the reviewed evidence is that public compensation produces its highest additional job-preservation value when disruption is sufficiently severe to threaten employment, the shock is temporary, and the affected match is expected to be economically productive after recovery. The French severe-shock evidence supports heterogeneous effects by exposure; Australian threshold-based estimates and Spanish ERTE evaluation identify preservation during acute pandemic restrictions; and the British government's final evaluation attributes substantial avoided job loss to its furlough program. Exact effects depend on each study's sampling and identification choices.

This pattern also explains lawful windfalls. A firm that would retain its workers regardless of subsidy may legally qualify and collect compensation, generating income support or an employer transfer with limited additional preservation. The same program can be valuable to liquidity-constrained firms facing temporary prohibitions on ordinary activity. During an early, wide-reaching shock, broad rapid access may be justified by the expected cost of delay. During recovery, continued broad access requires stronger evidence of current employment risk and firm viability.

The Belgium's 2026 OECD assessment adds a second temporal dimension. Large early benefits can coexist with weakening results or negative longer-run employment effects where expanded provisions persist after the acute crisis. The European Commission's 2025 synthesis and OECD's job-mobility analysis indicate that productivity and labor reallocation deserve explicit weight in the recovery phase. The relevant policy choice is not between preserving all existing jobs indefinitely and permitting every initial separation; it is an evidence-guided transition between insuring temporary attachment and supporting productive worker mobility.

Beneficiary design and public administration

Effective implementation must reconcile speed, accurate payroll identification and equitable access. A preexisting social-insurance infrastructure can verify normal hours, earnings and contribution histories quickly, while an emergency wage grant can reach a different population. Neither can be evaluated solely on announced benefit rates. Processing delays, claims reversals, employer financing constraints, worker awareness and application costs can determine whether a statutory entitlement provides effective immediate income protection.

For international comparisons, each country report should display legal eligibility, approved claims, unique recipients, paid person-months, lost hours, payroll replacement, administrative expense and paid recoveries separately. The public's understanding improves when dashboards identify precisely whether a reported number is an expenditure, a borrowing facility, a participant count, an error estimate or a causal employment outcome. In a policy domain where large numbers attract attention, correct classification is itself a substantial scientific service.

The French administrative payment guidance and German agency's worker guidance show that wage bases and payment procedures differ even among established European arrangements. The U.S. statutory work-sharing rule gives a further example in which reduced hours determine a proportional unemployment benefit rather than a percentage of previous net lost wages. Harmonized statistical reporting must reconstruct these denominators and payment channels before making performance claims.

A research agenda for credible comparisons

Future studies should join administrative work-hour histories to employer revenue shocks, financial constraints, worker earnings and mobility, unemployment insurance, and national fiscal accounts. Privacy-preserving linkage can make it possible to identify which subsidized relationships were realistically at risk, which survived only briefly, and which sustained earnings after a return to ordinary hours. Research should document treatment exposure as claim-months and estimated unworked hours, rather than merely whether a firm ever participated.

Useful quasi-experiments include eligibility cutoffs, policy phase-outs, staggered local restrictions, program duration thresholds and unexpected changes in employer contributions. Each opportunity requires careful checks that the threshold did not simultaneously alter other supports or encourage manipulation. An evaluation should state the population to which its local estimate applies. Translating one employment effect into international policy recommendations requires checking external validity, labor institutions and macroeconomic conditions.

The evaluation horizon should extend beyond the subsidized months. Relevant outcomes include separation and reemployment, full-time-equivalent earnings, return to ordinary schedules, firm survival, worker transitions to more productive employers, vacancies, new business entry, tax receipts and welfare support. Such an integrated panel would clarify whether early retention merely shifts layoffs forward or genuinely protects productive capacity. The comparison should also include workers who lacked employment-linked entitlements and could benefit from alternative protection.

Public finance studies should distinguish gross cash claims, national financing, subsidized hours, avoided unemployment benefits, net budget costs and total social welfare. Sensitivity analysis should vary shock persistence, counterfactual separation risk, program participation, employer liquidity, fraud, deadweight and discount rates. A harmonized cost per attributable full-time-equivalent job-year could then be estimated for comparable settings, with credible uncertainty intervals and explicit treatment of employment spillovers.

Original analytical contribution

The authorial contribution of this dossier is the integration of legal entitlement, the worker–firm match, program selection, causal employment estimation and fiscal incidence within a single comparative scholarly account. It establishes conceptual clarity about the quantities being measured and relates published estimates to the institutions that generated them. This synthesis follows established economic and social-insurance research; it does not institute a proprietary theory, principle or model. Its value lies in making the available evidence logically compatible, exposing unsuitable comparisons and identifying questions whose answer would materially change public judgment.

The central conclusion of that synthesis is that the value of short-time compensation is contingent on the timing of a shock and the prospects of the threatened employment relationships. Policies can have large social benefits during acute temporary interruption and still require substantial exit and mobility provisions after recovery. Strong evaluation must preserve those phases rather than average them into one unconditional verdict.

Conclusion

Short-time work compensation belongs to the core set of instruments connecting social insurance and active employment protection. It finances part of the earnings lost when work hours contract and employment continues. The benefit can protect household consumption, preserve productive workplace knowledge and reduce separations that would be inefficient under a temporary interruption. Its legal architecture varies materially among Germany's lost-net-pay based compensation, France's gross-wage based indemnity, U.S. unemployment-insurance work sharing and emergency furlough or wage subsidy schemes.

Empirical evidence demonstrates meaningful employment preservation in selected settings, especially during acute pandemic restrictions and severe firm-level disruptions. It also identifies substantial lawful deadweight, differing distributions of support, and potential obstacles to labor reallocation when expanded schemes persist into recovery. International headline figures cannot be aggregated without reconstructing the period, beneficiary unit, counterfactual, hours base and financing mechanism. The number of jobs receiving payments is not the number of jobs saved. An EU loan is a form of financing rather than an extra worker benefit. A modeled counterfactual deadweight percentage is distinct from an audited fraud rate.

The fiscal and welfare dimensions are complementary. An expenditure may have a substantial net cost to the treasury while preserving social value through earnings, employment continuity and a stronger recovery. Conversely, a low administrative cost per recipient cannot by itself establish net employment gains. The most informative measurement keeps workers, jobs, equivalent hours, funds, loans, welfare and public accounts in their correct categories.

A sound policy response provides accessible support when temporary disturbances threaten viable employment and then adjusts conditions, employer contributions, labor-market services and exit provisions as economic prospects become clearer. The scientific standard for its evaluation is a dated legal account, a transparent observation unit, a credible causal comparison and a reconciled fiscal perspective. This dossier supplies that comparative synthesis on traceable source material, defines the limits of inference and identifies the evidence required for the next stage of research.

References