Admitting Labour Migrants

5. How are numerical limits set and applied?

Abstract

This chapter analyses numerical limits – targets, caps and quotas – that govern the volume of temporary labour migration admissions. It offers an overview of where and how numerical limits are applied, the criteria and processes used to set them, and the design choices and trade‑offs governments face in calibrating them.

What are numerical limits and why are they used?

A numerical limit is a policy instrument that establishes a fixed volume of admissions for labour migrants, either within a specific pathway or across the migration system. While eligibility criteria define who qualifies for admission, numerical limits govern how many foreign workers may enter during a given period. These mechanisms vary in scope, ranging from system-wide caps that regulate overall migration inflows to targeted, pathway-specific limits focussed on particular economic sectors, occupations, or skill levels.

Based on a review of data on 41 OECD and EU countries, this chapter offers a general overview of where and how numerical limits are applied, the criteria and processes used to set them, and the design choices and trade‑offs governments face in calibrating them.

Employer-level limits, which restrict the proportion of foreign workers an individual employer may hire rather than setting a system-level ceiling, represent a distinct variant of numerical limits. While discussed in detail in Chapter 2 regarding their role in employer eligibility, they are included in this chapter in describing the definition of quotas.

Numerical limits are a direct tool for managing labour migration. By fixing a defined volume of admissions, they give authorities a concrete reference point for planning resources and processing capacity, and a clear basis for accepting or rejecting applications once the threshold is reached. Unlike labour market tests, which require case‑by-case assessment, or shortage occupation lists, which demand periodic labour market analysis, a numerical limit can be introduced quickly and applied consistently across a large number of cases. This directness is their primary appeal as they are straightforward to administer and communicate to employers, foreign workers, and the public. However, they are not necessarily the most effective instrument to ensure the matching of labour supply to demand. The clarity of numerical limits can increase the political salience of labour migration. Once a number is publicly stated, it becomes a visible benchmark against which government performance can be measured. If set in primary legislation, adjusting it requires political consensus that may be difficult to build even when labour market conditions warrant it. If the limit is set administratively and revised frequently, it may lose credibility as a meaningful constraint. A limit that is difficult to change may become a focal point for political debate; one that is easy to change risks becoming less meaningful as a constraint.

These properties give rise to a set of inherent trade‑offs. Numerical limits are relatively easy to introduce but difficult to calibrate. A numerical limit is only as good as the forecasts on which it is based, and labour market projections are often imperfect, particularly in rapidly changing sectors. If set too low, limits can create bottlenecks and divert recruitment toward irregular channels. If set too high, they may provide no real constraint. The clarity of a numerical limit can also make systems rigid, often prompting governments to introduce exceptions and carve‑outs to recover flexibility, which in turn may undermine the predictability the limit was meant to provide. The exceptions may become more important than the limit itself in terms of governing admission.

Within the EU, numerical limits on labour migration remain an exclusively national competence. Article 79(5) TFEU explicitly reserves to Member States the right to determine the volumes of admission of third-country nationals seeking work. This means that even where EU-level instruments such as the EU Blue Card set common eligibility criteria, Member States retain full authority to cap or freeze admissions once national limits are exhausted.

Which countries use numerical limits and how are they applied?

Of the 41 OECD and EU countries examined in this study, 26 apply some form of numerical limit to at least one temporary labour migration pathway, covering roughly one-fifth of the pathways analysed. Fourteen countries do not apply any numerical limit, of which 11 are EU Member States. Additionally, Norway is a special case, as it establishes a number of permits that may be granted through a simplified procedure without an individual labour market test.1 Table 5.1 summarises the numerical limits identified across the countries reviewed.

Table 5.1. Numerical limits across OECD countries

Country	Pathway	Type	Limit category	Oversubscription mechanism (how excess demand is addressed)
Australia	Skilled Employer Sponsored Regional (subclass 494)	Target	Programme‑Specific	Not exceeded – planning instrument
	Skilled Work Regional (subclass 491)	Target	Programme‑Specific	Not exceeded – planning instrument
Austria	Seasonal (Tourism, Agriculture, and Forestry)	Cap	Sector-Specific	Flexible – up to 50% above limit in tourism peak months; 30% in agriculture
Bulgaria	Single Work and Residence Permit	Quota	Employer-Level Limit	Hard stop, but ministerial exception possible
Canada	TFW Program – Low-wage stream	Cap	Employer-Level Limit	Sector exemptions apply
	International Mobility Program	Target	General System-Wide	Flexible
	International Experience Canada	Cap	Programme‑Specific	Not exceeded – planning instrument
	TFW Program	Target	General System-Wide	Flexible
Chile	All permits	Quota	Employer-Level Limit	Exemptions for irreplaceable specialists
Croatia	All permits	Quota	Employer-Level Limit	No further applications accepted
Czechia	Employee Card (incl. Economic migration programmes)	Cap	Sub-Category / Stream	Priority queue; non-priority applications accepted without processing priority
Denmark	Positive List (skilled work – social and healthcare)	Cap	Occupation-Specific	Occupation deleted from list once limit reached
Estonia	Temporary Residence Permit for employment (incl. EU Blue Card)	Cap	General System-Wide	No further applications accepted
Germany	Employment of certain nationals (Western Balkans Regulation)	Cap	Programme‑Specific	No further applications accepted
	Seasonal employment	Cap	Programme‑Specific	No further applications accepted
Greece	Dependent employment	Cap	Programme‑Specific	No further applications accepted
	Seasonal Workers	Cap	Programme‑Specific	No further applications accepted
Hungary	Guest Worker Permit	Cap	Programme‑Specific	No further applications accepted
Ireland	General Employment Permit (specific occupations)	Cap/Quota	Occupation-Specific and employer-level	No further applications accepted
Israel	General Work Visa (B/1)	Cap	Sector- or Occupation-Specific	No further applications accepted
Italy	Seasonal work (Decreto Flussi)	Cap	Programme‑Specific / Sub-Category / Stream	System closes on click day
	Subordinate work residence permit (Decreto Flussi)	Cap	Programme‑Specific / Sub-Category / Stream	System closes on click day
Japan	Special Skilled Worker I (SSW1)	Cap	Sector- and Occupation-Specific	No further applications accepted
	Technical Intern Training Program (TITP)	Quota	Sector- and Occupation-Specific and employer-level	No further applications accepted
Korea	E‑7 (Specialty workers)	Cap/Quota	Sector- or Occupation-Specific and Employer-Level	No further applications accepted
	E‑8 (Seasonal)	Cap	Sector- or Occupation-Specific Limit	First-come, first-served among local governments
	E‑9 (Non-professional employment)	Cap/Quota	Sector- or Occupation-Specific/ Sub-Category / Stream and Employer-Level	No further applications accepted
	H‑2	Cap	Programme‑Specific Limit / Sub-Category / Stream Limit and Employer-level	Lottery
Lithuania	Temporary Residence Permit	Cap	General System-Wide	Salary threshold override above quota
Mexico	All work permits	Quota	Employer-Level	Exceptions if no Mexican specialists available
New Zealand	China Skilled Workers Visa	Cap	Sector- or Occupation-Specific Limit / Sub-Category / Stream	First-come, first-served
	Recognised Seasonal Employer (RSE)	Cap	Programme‑Specific	Employer allocation system
Romania	All work permits (incl. EU Blue Card, Permanent Workers, Seasonal)	Cap	General System-Wide	No further applications accepted
Spain	Seasonal Worker (GECCO)	Cap	Sector- or Occupation-Specific	Review and reallocation during year
Switzerland	B Permit	Cap	Programme‑Specific	No further applications accepted
	L Permit (short-term employment)	Cap	Programme‑Specific	No further applications accepted
Türkiye	Temporary work permit	Quota	Employer-Level	Exceptions for FDI, high-tech, strategic sectors
United Kingdom	Seasonal Worker Visa	Cap	Sector- or Occupation-Specific Limit / Sub-Category / Stream	No further applications accepted
United States	E‑3 Specialty Occupation Professionals (Australia)	Cap	Programme‑Specific Limit / Sub-Category / Stream	No further applications accepted
	H‑1B Specialty Occupations	Cap/Quota	Sector- or Occupation-Specific Limit / Sub-Category / Stream	Lottery; from 2026, ranked lottery
	H‑2B Non-Agricultural Workers	Cap	Programme‑Specific	Lottery

Source: OECD Secretariat.

Table 5.1 shows that programme‑specific and sector-specific limits are the most common forms of numerical limit across OECD countries, while system-wide limits are the least common, used only by Estonia, Lithuania, and Romania. Highly skilled pathways are less frequently subject to numerical limits, with Switzerland and the United States as notable exceptions where programme‑level limits apply across the skill spectrum. Employer-level limits are concentrated among non-European OECD Members with just three EU countries utilising them (Bulgaria, Croatia and Ireland).

In terms of the use of numerical limits on pathways structured by the EU legal migration framework, six Member States apply numerical limits to permits issued under the Seasonal Workers Directive – Austria, Germany, Greece, Italy, Romania, and Spain. The EU Blue Card is generally not subject to national numerical limits. In Romania, the rules in place until August 2026 subject it to the national limit on work authorisation; pathways introduced in August 2026 separate the highly qualified into a stream outside the annual limit.

Three types of numerical limits

Numerical limits can be broadly categorised into three main types: targets, caps, and quotas. Each serves a distinct purpose and is implemented with varying levels of stringency, depending on how migration planning is embedded in a country’s broader economic and demographic strategies.

Targets set an expected or desired volume of admissions without imposing a hard ceiling. Unlike caps, they function as flexible benchmarks, allowing governments to adjust processing priorities in response to shifting economic conditions without necessarily stopping admissions once a threshold is reached. They are commonly seen in settlement countries such as Australia and Canada, where administrative planning levels are set annually through whole‑of-government processes and linked to broader demographic and labour market objectives. Targets are more prevalent in permanent residence streams (not covered in this report) than in temporary labour migration. While this flexibility enables a swift policy response to changing conditions, it can also create challenges in calibration when admissions fluctuate significantly against projected goals. A variant of the target is the trigger limit, used by Norway. A defined number of permits for skilled workers and 3 500 for seasonal workers does not stop admissions when reached but instead makes a labour market test mandatory for further hiring. This design preserves a degree of labour market protection while avoiding the rigidity of a hard ceiling.

Caps establish a fixed ceiling on admissions that cannot be exceeded once reached. Unlike targets, which allow for adjustment, a cap functions as a binding constraint: once the limit is filled, no further applications are approved unless an explicit exemption applies. Caps are applied across a range of pathway types but are particularly common in seasonal and lower-skilled programmes, where governments seek an enforceable boundary on programme size and a clear mechanism for managing employer demand. Examples of this model are present in Austria, Germany, Greece, Hungary, Italy, Korea, New Zealand, Norway, Romania, Spain, the United Kingdom and the United States. They can operate at different levels of granularity: Korea’s E‑8 seasonal worker visa applies national ceilings that are further divided by local government and sector, allowing for fine‑grained volume control while still maintaining an overall limit. Germany’s Western Balkans Regulation illustrates a different dynamic: the annual cap of 50 000 was doubled from its original level in response to sustained employer demand, yet has continued to be fully subscribed, suggesting that the limit constrains rather than reflects actual recruitment needs. The tighter a cap is designed, the more likely it is to generate oversubscription, requiring additional mechanisms to manage excess demand.

Caps may be disaggregated. The maximum number of admissions may be broken down into caps by sector, region, nationality, or occupation, determining in advance not just how many workers may be admitted but where those places will be directed. Israel illustrates this mechanism through its sector-specific quotas for construction and agriculture, which are further divided through bilateral agreements with specific sending countries, linking admission volumes to both industry needs and diplomatic priorities. Italy’s Flow Decrees partition its annual national ceiling into reserved sub-streams by nationality and by seasonal versus non-seasonal work. Disaggregation offers precision but also rigidity: once places are assigned to a particular sector or nationality group, unused slots may be difficult to reallocate even when demand elsewhere is unmet.

Quotas are allocations of shares; rather than setting limits, they set a fixed percentage of the total cap or target, rather than fixed numbers. In practice, quotas are rarely used, since fixed numerical caps by category are set. In this chapter the term quota is used only when referring to a percentage limit. The countries using quotas on the share of foreign workers in the employer’s workforce as the main form of limit are Mexico (10%), Chile (15%, for firms with more than 25 employees), and Türkiye (a ratio of one foreigner to five Turkish employees).

How limits are applied to admission: from general to specific

The impact of a numerical limit depends not only on its type but on the level at which it is applied. Table 5.1 shows that OECD countries apply numerical limits across several distinct operational categories: general, or disaggregated by sector, occupation, etc.

General system‑wide limits

Only a small number of countries analysed in this study use single national ceilings that apply across all or most migration pathways. Currently, only Canada, Estonia, Lithuania, and Romania apply general ceilings. These ceilings are often tied to demographic markers, such as a fixed percentage of the permanent population (as in Canada, Estonia and Lithuania), or are defined annually based on variables such as economic growth, projected labour shortages, or the number of applications received in the previous year (as in Canada and Romania). System‑wide caps represent the bluntest form of numerical control: while straightforward to administer, they do not target particular labour‑market needs and therefore offer limited capacity to address differentiated or sector‑specific shortages.

In Canada, in 2024, the government committed to reduce the share of temporary migrants to less than 5% of the population by the end of 2027, equivalent to a 43% decrease relative to the previous year. This was done partly to ease pressures on housing, infrastructure, and services, and to ensure sustainability of the immigration system. Annual arrival targets were established in the 2025-2027 Immigration Levels Plan for both the International Mobility Program and the Temporary Foreign Worker Program. Targets were revised downward again within the 2026-2028 Levels Plan, including a target of 60 000 TFWP arrivals for 2026, declining to 50 000 in subsequent years. Canada is working toward the broader objective of reducing the volume of non-permanent residents to less than 5% by the end of 2027, while aligning immigration measures to where they deliver the greatest impact: filling key labour and skill gaps, strengthening the economy, and supporting communities across the country.

Disaggregated limits

The general system limits can be divided by programme, sector, occupation, region or locality, employer type, nationality and other criteria. These are often used together. Italy, for example, has a labour pathway which includes cap-subject and cap-exempt categories, with multiple disaggregation.

Programme‑specific limits apply a numerical ceiling to an individual temporary labour migration pathway rather than to the system as a whole. This is one of the most commonly used approaches across the countries reviewed. Because they apply to only one channel, programme‑specific limits allow governments to shape the scale of admissions within a specific pathway, adjust volumes annually, or pilot new schemes in a controlled way without altering the broader migration framework. Hungary’s Guest Worker Permit, for example, combines an annual ceiling with restrictions to designated occupations, employer categories, and partner countries. Switzerland applies programme‑level caps to its L and B permits, with annual volumes of 4 000 and 4 500 respectively, set by the Federal Council based on labour market and economic conditions.

Examples of programme‑specific caps vary widely across countries. Germany’s Western Balkans Regulation provides a capped entry channel for lower‑ and medium‑skilled workers from a defined region (certain nationalities). Hungary’s Guest Worker Permit for temporary recruitment similarly combines an annual ceiling with restrictions to designated occupations, employer categories, and partner countries. Australia applies programme‑level caps in selected parts of its temporary system, such as certain Labour Agreements or regional employer‑sponsored pilots, where nomination ceilings help align intakes with identified labour market needs. Across these cases, programme‑specific caps enable targeted temporary migration while maintaining clear and predictable parameters for how each pathway operates.

Sector-specific limits set ceilings for admissions within a defined industry or broad economic sector. They are most commonly used in seasonal and non-professional work in sectors like agriculture, fisheries, hospitality, and food processing, where labour demand is high-volume, time‑bound, and sensitive to production cycles. Austria’s seasonal worker cap incorporates flexibility: the annual ceiling can be exceeded by up to 50% in tourism peak months and 30% in agriculture, provided the annual average is maintained. Korea’s E‑9 programme applies sector-specific ceilings that are further divided by country of origin, linking volume controls to bilateral recruitment agreements. Spain’s GECCO seasonal programme sets limits driven by provincial-level submissions of crop-specific labour needs, which are reviewed and reallocated during the year in response to changing demand.

Occupation- and sector specific limits set ceilings for defined job categories or industries. In practice, sector‑wide caps are most often used in lower‑skilled or seasonal work, such as agriculture, fisheries, hospitality, or food processing, where labour demand is high‑volume, time‑bound, and sensitive to production cycles. By setting ceilings for particular industries, authorities can try to match inflows to sector‑level workforce needs, regulatory oversight capacity, and cyclical or structural patterns of demand. Occupation‑based caps, although operationally similar, tend to appear in mid‑skilled segments, such as care work, building maintenance, or specific technical fields, where shortages are narrower and defined by particular skill requirements rather than entire industries.

Occupation lists are discussed in more detail in Chapter 4; these lists may also be associated with caps. Denmark’s Positive List applies a cap of 1 000 permits to social and healthcare occupations, with the occupation removed from the list once the cap is reached. Ireland’s General Employment Permit applies annual caps to specific occupations identified through ministerial regulation.

The methodology underpinning calculation of sector and occupation limits varies considerably across countries and is not always publicly disclosed, which can limit transparency for employers and applicants planning international recruitment (OECD, 2024[1]).

Japan’s SSW1 system applies numerical ceilings to both sectors and linked occupational groups, from nursing care to building cleaning, food service, and industrial machinery, based on labour market assessments. Switzerland incorporates sector‑based allocation considerations into the annual distribution of L‑ and B‑permits, directing limited permit volumes toward industries with persistent recruitment pressures, such as hospitality, construction, and manufacturing. In general, sector‑ and occupation‑specific caps enable governments to direct temporary migration into the labour market segments where it is most needed: whether broad industries reliant on seasonal labour or specific occupations facing structural mid‑skilled shortages.

Sub‑category or stream limits apply within a broader migration programme, restricting a specific subset of applicants rather than the entire visa category. These limits typically govern nationality-based allocations, bilateral sub-streams, or programme segments, enabling policymakers to adjust the composition rather than just the total volume of admissions. Italy’s Flow Decree divides its annual ceiling across sectors, nationality groups and across seasonal versus non-seasonal sub-streams, reflecting bilateral mobility partnerships and differing labour demands. Czechia applies limits by individual diplomatic mission based on processing capacity, with applications accepted outside the priority cap without processing priority. Only employers approved within the Economic Migration Programmes – endorsed by the Ministry of Trade and relevant sector organisation – are able to access visa issuance slots. The United States has a number of capped pathways related to commitments in trade agreements. For example, its annual ceilings under the H‑1B1 programme for workers from Chile and Singapore; bilateral trade agreements imposed the creation of distinct visa categories to ensure that applicants would not be excluded from an oversubscribed H‑1B pathway.

Employer-level limits

Employer‑level limits restrict the proportion or number of foreign workers an individual employer may hire. In most cases, they are quotas, defining the maximum share of foreign employees, or temporary foreign worker employees, an employer is allowed to have. Unlike system‑wide or programme‑specific ceilings, these caps influence how foreign labour is distributed across firms and sectors rather than how many workers enter overall. Their main function is to prevent concentration of migrant workers within particular companies, ensure that employers continue to invest in recruiting and retaining local staff, and reduce vulnerabilities associated with heavy reliance on foreign labour. As a result, employer‑level caps are often used in areas where there is strong and persistent employer demand for lower or mid‑skilled roles, and where regulators aim to balance flexibility for employers with safeguards for the labour market as a whole. Canada imposes firm-level caps (10%‑20% at a given worksite, depending on sector). In Türkiye, at least five Turkish citizens must be employed for each foreign employee at the workplace, although there are relaxations and exemptions, with reductions in the five‑to‑one ratio. Bulgaria limits third-country nationals to 20% of the average workforce, or 35% for small and medium enterprises. Ireland’s 50:50 rule prevents the granting of an employment permit where more than half of the employer’s workforce are non-EEA nationals, with ministerial waivers available in specific circumstances. Croatia applies ratios of 20% or 10% depending on whether the occupation is on the shortage occupation list. Outside Europe, Korea caps foreign workers at 20% of insured Korean staff across its E‑7, E‑9, and H‑2 pathways.

Which criteria are considered to set numerical limits?

The methods used to set numerical limits vary considerably across countries and are not always publicly disclosed. Across the systems reviewed, limit-setting combines labour market indicators with institutional capacity, social partner input, and in some cases international commitments. The resulting number does not always reflect a precise technical assessment – in several cases it reflects political considerations, processing constraints, or bilateral negotiations as much as labour market evidence.

Labour market need is the most commonly cited basis for setting annual ceilings, though the quality and transparency of the underlying methodology varies widely. Where methodologies are published, countries draw on vacancy data, seasonal production cycles, and projected shortages. Spain bases seasonal admissions on provincial submissions of crop-specific labour requirements. Japan’s SSW1 programme is informed by sectoral forecasts produced by line ministries, though the transparency of these assessments has been noted as an area for improvement. Slovak Republic draws on vacancy data from manufacturing and the automotive industry. In agriculture, tourism, fisheries, and food processing, ceilings are often linked to anticipated production peaks. These examples illustrate that labour market evidence is a common starting point, but the rigour and openness of the analytical process differ considerably across systems.

Consultation with social partners is a common feature of limit-setting in many countries, though the degree to which social partner positions are actually reflected in the final number varies. Austria follows a tripartite model where seasonal ceilings are determined through dialogue involving the Ministry of Labour, the public employment service, and social partners. Ireland uses stakeholder input through its Employment Permits Review to determine which roles should be subject to numerical controls. Finland draws on tripartite consultation when considering whether quantitative restrictions are appropriate for particular occupations. In these cases, consultation serves both an informational function – incorporating sectoral knowledge – and a legitimising one, building consensus around politically sensitive decisions. However, consultation does not always translate into acceptance of social partner positions, and final decisions typically remain with government.2

Administrative and oversight capacity also shapes limit levels, though the relationship can run in both directions. Institutional capacity constrains how many applications can be processed, effectively setting a practical ceiling on admissions regardless of labour market demand. A cap can determine how many staff and resources are allocated, meaning that administrative capacity is shaped by the limit rather than the reverse. Czechia allocates Employee Card caps across diplomatic missions based on processing capacity. Greece adjusts migration planning numbers according to the capacity of regional offices to assess employer requests. Türkiye uses administrative capacity considerations in setting employer‑level hiring limits tied to the Ministry of Labour’s monitoring ability. These examples highlight that numerical limits often reflect the state’s capacity to process and enforce as much as its assessment of labour market need.

Caps are reviewed – sometimes as part of an annual exercise, as in Korea, the United Kingdom, Czechia, Germany, Italy and Hungary. This can lead to changes, and even countries with annual review sometimes adjust the caps during the period of coverage. The United Kingdom provides an example of rapid increase in caps in response to immediate labour demand. It introduced its Seasonal Agricultural Worker programme to address changes in labour supply following its exit from the European Union, reviving a scheme that had previously been obviated by free movement from the EU. Initially launched with a cap of 2 500 in 2019 to test the pilot. The cap was increased to 10 000 mid-2020 to address labour shortages, which were partly COVID-related. When free movement ended in 2021, the cap was raised to 30 000 and further increased in 2023 to allow 45 000 horticultural workers, an additional 2 000 in poultry and 10 000 to assign. The numbers have since fallen slightly in line with a political commitment to reduce reliance on migrant labour. Korea, likewise, has released additional numbers during the year.

Many countries review previous years’ utilisation to recalibrate annual ceilings, though this is not universal practice. Lithuania revises its low and medium-skilled employment cap by examining uptake levels. Hungary incorporates utilisation patterns from earlier schemes into the design of its guest worker pathways. Romania adjusts its national work permit cap based on the volume of applications received the previous year. Such feedback mechanisms can help maintain alignment between limits and actual demand over time. Where they are absent, limits may remain fixed even when conditions have changed significantly.

The United States has certain caps fixed in legislation (H‑1B, H‑2B) and cannot change them without congressional action, limiting the ability to adjust to changing economic conditions. The H‑1B cap is related to commitments made by the United States under WTO GATS Mode 4, so restrictions in the number of H‑1B visas below the cap of 65 000 may have implications for trade agreements.

International agreements may include commitments to offer labour migration opportunities to nationals of the partner country. For example, Italy has, in the past, published a quota reserved for nationals of designated partner countries – who co‑operate on migration management – within its flow decrees. Japan has a labour migration channel with a fixed annual commitment to admit a certain number of candidate nursing and care workers from several origin countries.

Overall, these criteria show that limit-setting is rarely a purely technical exercise. It combines labour market evidence with administrative, political, and diplomatic considerations, and the weight given to each varies considerably across countries and pathway types. In several systems, the process is not fully transparent, making it difficult to assess whether the resulting number reflects actual labour market conditions or other factors.

Oversubscription, under-utilisation

Administration of caps means checking applications against remaining allocations; this can be done at different points in the process. In Italy, the provincial-level labour offices receive allocations and cease issuing approvals once the specific allocations are exhausted. In Germany, the cap for the Western Balkan pathway is monitored by the Federal Employment Agency (BA), which ceases to issue new approvals once the number is reached. Although numerical limits are designed as fixed ceilings, in practice they frequently diverge from actual admission levels in both directions. On the one hand, demand may exceed available slots, generating oversubscription and requiring mechanisms to allocate scarce places among competing applicants. On the other hand, authorised places may go unfilled due to administrative bottlenecks or misalignment between the limit and actual recruitment patterns.

Countries have developed different mechanisms to manage this. For example, in the United States H‑1B programme, demand consistently surpasses the annual allocation of 65 000 for the general pool and 20 000 for advanced-degree holders. Since 2024, USCIS has implemented a weighted selection model in which employers indicate the highest wage level that the offered salary meets or exceeds, with higher wage levels accruing more entries in the lottery pool and thereby increasing the probability of selection for higher-paid roles. Beneficiary-centric controls introduced in the same period prevent multiple registrations for the same individual. Korea’s H‑2 programme similarly uses a lottery when annual demand exceeds the national cap. Spain’s seasonal agricultural programme manages oversubscription through waiting lists that carry unmet employer requests into subsequent rounds, while Romania reallocates unused allocations within the cap across sectors mid-year. The United States H‑2B programme allows for authorisation of supplemental releases when persistent demand exhausts the fixed 66 000 annual ceiling. Japan adjusted the initial sectoral targets set under the SSW1 system when utilisation diverged significantly from pre‑pandemic projections: two sectors – food processing and manufacturing – substantially exceeded their initial targets without triggering a suspension of admissions, though the revised targets set for 2025 are being applied more strictly, with some sectors such as restaurants having already reached their ceiling with no new admissions permitted. In Switzerland, annual caps are allocated to cantons, but the State Secretariat for Migration (SEM) holds a federal reserve, which it can release to cantons which request additional work permits.

Targets can be adjusted rapidly to signal policy control, they may prove difficult to implement across diverse categories of temporary migrants when underlying employer demand remains high. In Canada, the introduction of limits to the International Mobility Program and the Temporary Foreign Worker Program in 2025 required sharper cuts in the second half of the year than in the first in order to reach the target (Jedwab, 2025[2]). In 2025, new worker arrivals fell below the established target by approximately 100 000.

Under-utilisation presents a different challenge. Authorised places may go unfilled when administrative bottlenecks slow the procedural chain between approval and actual admission.

Italy provides the most glaring example of how difficult it is to set numerical limits, how difficult it is to make them perform and how oversubscription can lead to underutilisation. The main labour migration pathway to Italy since 1998 has been through the annual Flow Decree, which establishes a cap. The cap is subdivided between seasonal work and non-seasonal work, into multiple categories (by nationality, by employment type, for persons who trained in authorised origin-country programmes, for specific sectors and occupations, etc.), as well as for changes of status of persons already in Italy on other grounds. The cap is further subdivided into regional and provincial allocations, so that crossing some of the categories with nationalities or sectors may lead to very small numbers of available slots for persons who must meet multiple criteria (e.g. nationality and sector). The Flow Decree, often published late in the calendar year, can only be renewed using the previous year’s allocations if there is no general planning directive. From 2016 to 2020, the cap was unchanged, and kept at zero for new non-seasonal arrivals through general employment channels. Since 2021, the cap has increased progressively, and moved to a three‑year planning cycle from 2023. The annual intake for 2025 was set at 165 000, of which 71 450 were for non-seasonal employment categories.

Applications open on a specific date, called “Click Day” as applicants wait to click and send their file as soon as the application window opens. Applications for non-seasonal employment have consistently exceeded the cap – often by hundreds of thousands.

Nonetheless, the consistent demand, and the recent increases in the allocation have not corresponded to actual entries. A substantial share of Flow Decree authorisations do not result in completed admissions, due to delays across approval (nulla osta) issuance, visa processing, and contract finalisation (Coresi, 2026[3]). For example, of the 2023 allocation of 82 700 places, there were hundreds of thousands of applications. 74 100 were approved, but only 60 000 invitations to apply for a visa, of which 39 000 were still waiting for visa appointments at the end of January 2024; 17 400 workers entered Italy and received a residence permit – and these included other categories of admission. Further, the share of places in the Decree which lead to actual entry and residence permits has been shrinking in recent years as visa waits grow longer. In December 2023, about 600 000 requests were filed and 127 000 assigned under the cap. More than a year later, 16 000 of these requests had been approved, and 9 500 residence permits issued.

Romania also illustrates this gap, although its experience is more recent and partly due to cap design. In 2023, approximately one in three approved applications led to an actual entry, a consequence of several compounding factors. Not all employers who receive an authorisation proceed to make a job offer; long consular delays in key origin countries can extend the time between authorisation and arrival to eight or nine months; applicants may lack the financial means to travel to a consulate or find more attractive employment in the meantime; and visa refusal rates remain high even where work permits have been approved. Knowing this, some employers file more applications than they genuinely need, inflating apparent demand. Authorisations rather than visa issuances are counted against the cap, meaning that even where a visa is never issued or a worker never arrives, availability within the cap is reduced. Romania’s cap had for many years been underutilised and served primarily as a signalling mechanism – communicating to employers that international recruitment was possible within a governed framework – rather than as a binding constraint on flows. It began to be exhausted only from 2022 onwards, when the cap of 100 000 was reached on 13 December of that year. Even then, additional authorisations were granted for workers already admitted in previous years, illustrating how formally binding ceilings can be subject to administrative flexibility in practice (OECD, 2025[4]).

When ceilings are reached, recruitment may also shift toward parallel channels that operate outside the cap framework. Italy provides a clear illustration: while the Flow Decree sets annual admission limits for large segments of non-EU labour migration, other routes – including EU Blue Cards, intra-company transfers, and high-skilled permits – operate independently of cap constraints. In Denmark, a high salary allows the applicant to avoid caps in the Positive List pathway. In the United States, the H‑1B channel is capped, and otherwise eligible candidates may try for other uncapped categories such as the O‑1 visa, if they qualify. This means that the formal ceiling does not govern total admissions but shapes their composition, channelling some flows through capped routes while others remain uncapped.

Denmark’s Positive List institutionalises the shift from the capped to the uncapped categories: rather than exempting categories from the cap, it removes the occupation from the Positive List entirely once the cap of 1 000 is reached. Further applications must qualify for one of the other pathways to Denmark, such as the salary-level pathway, for the remainder of the cap period. Such measures increase responsiveness but also mean that caps function as partial constraints that influence the distribution of admissions rather than defining their overall scale.

Exemptions may curtail the real impact of numerical limits

Almost all of the limits presented in this chapter include exemptions and exceptions which remove specific categories of applicants from numerical control. The range of exemptions is wide, and can reflect individual characteristics (nationality, education level), job characteristics (highly-qualified, strategic), and employer (innovative firms may have specific exemptions).

The effect of exemptions is, in some pathways, that the stated ceiling may not correspond to the number of workers actually admitted. Exemptions, in fact, have important effects. In the United States, the annual H‑1B cap of 65 000, plus 20 000 for applicants with Master’s degrees, does not apply to universities, non-profit research institutions, or government research organisations, which may continue recruiting throughout the year regardless of cap exhaustion. In Fiscal Year 2024, 140 000 petitions for initial employment were approved. In Italy, the administrative bottlenecks leading to underuse of the annual admission caps mean that the actual initial job starts in Italy are substantially higher than the actual issuance of residence permits under the annual decree, and therefore are largely for categories not covered by the cap. The increased attractiveness of cap-exempt categories such as training in origin can be an unintended side effect of dysfunctions in the administration of issuance of cap-subject admission.

Key takeaways

Numerical limits are applied in 27 of the 41 countries covered in this review, covering roughly one fifth of the pathways analysed. The most common type of numerical limit is the cap, present in 26 pathways. Targets are the least common, only utilised by Australia.

Their prevalence reflects their primary appeal: they appear administratively simple and easy to communicate. By fixing a defined volume of admissions, they give authorities a concrete reference point and a basis for managing flows. However, the analysis shows that the relationship between a stated ceiling and actual admissions is rarely straightforward. Oversubscription mechanisms, administrative bottlenecks, parallel uncapped channels, and formal exemptions mean that the headline figure often captures only a portion of actual admissions. The attractive simplicity of numerical limits begins to be lost when they must be segmented and disaggregated according to multiple parameters.

The methods used to set limits vary considerably and are not always transparent. Labour market evidence – vacancy data, sectoral forecasts, employer surveys – is the most commonly cited basis, but vacancy data alone may not accurately capture actual shortages, and where methodologies are not publicly disclosed, it is difficult to assess the extent to which the number reflects labour market conditions or other considerations. Consultation with social partners is widespread but does not always translate into acceptance of their positions. Administrative capacity, bilateral commitments, and political considerations all shape the final number alongside labour market evidence.

The clarity of a fixed ceiling can also increase the political salience of labour migration. Once a number is publicly stated, however, it becomes a visible benchmark that may be difficult to adjust, whether upward or downward, without attracting political attention. The United States and Canada illustrate different dimensions of this challenge: in the former, legislative entrenchment has prevented adjustment despite persistent oversubscription; in the latter, administratively determined targets, allow the government to limit admissions in order to manage the size of the temporary worker population. The frequent gap between limits and actual admissions does not necessarily indicate a failure of the limit setting exercise. A limit that is consistently oversubscribed may mean that it is effective in protecting the domestic labour market, or that it is excessively restrictive and creating bottlenecks. The governance of a pathway may include adjusting the limit upwards when it is approached, so that it never plays a constraining role. A limit that is rarely reached may reflect a function primarily as a signal of political control rather than a binding constraint.

Many exemptions exist across many OECD countries, where shortage occupations, highly skilled roles, or certain categories of employers are not counted against numerical limits. The exemptions may be more important than the general rule.

Assessing whether a numerical limit is working therefore requires clarity about what it is intended to achieve, and that objective is not always made explicit in the examples reviewed in this chapter.

References

[3] Coresi, F. (2026), IV rapporto di monitoraggio sugli ingressi per lavoro, Ero Straniero, https://erostraniero.it/wp-content/uploads/2026/03/DossierFlussi2026CES.pdf (accessed on 7 April 2026).

[2] Jedwab, J. (2025), Temporary Migrant Admissions in Canada: What a deeper dive into the numbers tells us about the first half of 2025, https://acsmetropolisca-wpuploads.s3.ca-central-1.amazonaws.com/wp-content/uploads/2025/10/22122708/Immigration-Trends-in-Canada-June-2025-part-2-Temporary-Migrants.pdf (accessed on 20 August 2026).

[4] OECD (2025), OECD Reviews of Labour Market and Social Policies: Romania 2025, OECD Reviews of Labour Market and Social Policies, OECD Publishing, Paris, https://doi.org/10.1787/f0532908-en.

[1] OECD (2024), Recruiting Immigrant Workers: Japan 2024, Recruiting Immigrant Workers, OECD Publishing, Paris, https://doi.org/10.1787/0e5a10e3-en.

Notes

← 1.  The 14 countries without any numerical limit are: Belgium, Colombia, Costa Rica, Finland, France, Iceland, Latvia, Luxembourg, the Netherlands, Poland, Portugal, the Slovak Republic, Slovenia, and Sweden.

← 2.  Court cases may also influence decisions: Colombia had a 15% quota for foreign workers; when it was abolished, the professional order of engineers litigated to maintain it exclusively for engineers. See Concepto Sala de Consulta C.E. 2226 de 2015 Consejo de Estado - Sala de Consulta y Servicio Civil - Gestor Normativo - Función Pública.