Sources and calculations

Each numbered reference links to the original publisher. The notes below identify what each dataset measures and show how the calculations can be reproduced. Research checked between 26 September and 26 October 2026.

Data definitions and calculations

Citizenship recipients and processing times

The route table counts certificates issued in 2025, not arrivals or applications. Shares use 29,919 as the denominator. The Department’s “S15 residency” category is the route for spouses and civil partners of Irish citizens; “Granted International Protection” covers refugees and people with subsidiary protection. Children (23.2%) and spouses or civil partners (9.8%) together make up 33.0% of certificates. The source does not link recipients to earlier employment permits.

The January 2026 certificate table gives 29,919 for 2025; a later answer gives 29,300 certificates conferred. The site uses the first series consistently. The 2023 total of 18,265 comes from the original table because a suppressed cell prevents a complete sum of route figures. Applications and grants in the same year are not a matched group and cannot yield an approval rate. Processing figures are medians, not minimum waiting periods. ^[1] ^[2] ^[3]

Employment permits

The sector table includes renewals and all permit types. It counts permits, not unique workers, citizenship recipients or HSE hires. Critical Skills and General permits total 17,168 + 19,478 = 36,646; dividing by 39,398 gives 93.0%.

The official sector extract sums to 39,398, eight more than the annual series total of 39,390. The table and its shares consistently use the sector extract. ^[4] ^[5]

EU residence requirements

The residence chapter counts the total residence needed before an ordinary application. Where a country requires permanent residence for its naturalisation period, the years needed to obtain permanent residence are added. The starting point is Figure 2 of the European Parliament brief (GLOBALCIT data with 2025 updates); each country was rechecked in September 2026.

Fifteen countries were confirmed on an official government page, read directly on 28 September 2026; Ireland is taken from the Act itself:

Austria: 10. Official source
Cyprus: 8 (12 months plus 7 of the previous 10 years). Official source
Denmark: 9. Official source
Estonia: 8 (5 on a permanent basis). Official source
Finland: 8 (5 with language skills). Official source
France: 5. Official source
Germany: 5 (three-year fast track abolished October 2025). Official source
Italy: 10. Official source
Latvia: 5. Official source
Luxembourg: 5. Official source
Netherlands: 5. Official source
Portugal: 10 for non-EU nationals, 7 for EU and Portuguese-speaking countries. Official source
Slovenia: 10. Official source
Spain: 10. Official source
Sweden: 8 (from 6 June 2026). Official source

Countries where permanent residence must come first, and how long it takes:

Czechia: permanent residence after 5 years of continuous residence. Source (official page read directly).
Latvia: permanent residence permit after 5 years with temporary permits; naturalisation needs 5 years with it. Source (official page read directly).
Poland: EU long-term residence after 5 years of legal stay (study time counts half); naturalisation needs 3 years with it. Source (Office for Foreigners; its text did not load for automated reading, so confirmed through search results quoting it. By 26 October 2026 the original page on the long-term resident’s permit redirected to the Office’s home page, which is linked here).
Slovakia: permanent residence after 5 years of uninterrupted legal residence; naturalisation needs 8 years of continuous permanent residence. Source (IOM Migration Information Centre, read directly for the 8-year rule; the 5-year rule confirmed through search results quoting its permanent-residence page).
Bulgaria: permanent residence after 5 years of continuous residence; naturalisation needs 5 years with it (secondary sources; official sites did not state it in readable form).

In Estonia (8 years, of which 5 permanent), Croatia (8), Lithuania (10), Denmark (9) and Austria (10, of which 5 settled), permanent or long-term status is required but falls within the stated period, so nothing is added.

For Belgium (5), Croatia, Greece, Hungary (8), Lithuania, Malta (about 5: 4 of 6 years plus 12 months) and Romania (8), official websites refused automated access or did not state the period; the brief’s figure was confirmed against current secondary sources and no change in force was found. Romania requires 8 years of legal domicile; whether time before long-term residence counts could not be confirmed from an official source, so it is shown at 8 with that caveat. Poland has bills for longer periods before its parliament, and Belgium’s 2026 reform concerned loss of nationality, not the residence period.

Averages are the author’s calculation: 218 ÷ 27 = 8.1 years. The figures are general minimums; many countries count only some kinds of permission or allow shorter periods for language skills, marriage or refugee status, so equal numbers of years are not always equally demanding.

Citizenship by nationality, and permits five years earlier

The nationality table is the Department’s original spreadsheet for PQs 482–483. Yearly totals match the route table, apart from a few suppressed small values (at most 29 certificates a year). India’s share is 6,298 ÷ 29,919 = 21.1% in 2025 and 1,026 ÷ 10,002 = 10.3% in 2016. Certificates rose by 29,919 − 9,766 = 20,153 between 2021 and 2025; India’s rise of 6,298 − 746 = 5,552 is 27.5% of that. ^[6]

The permits-against-certificates chart compares first-time (“new”) employment permits issued in 2015–2020 with certificates in 2021–2025, about the five years of residence the ordinary route requires. It covers the ten non-EU nationalities with the most certificates in 2021–2025; EU/EEA and UK citizens are excluded because they need no permit. Nationality names differ slightly between the two sources (“China” and “China (Including Hong Kong)”); they are matched by hand. The datasets cannot be linked person by person, and certificates include partners and children, so the chart compares nationalities, not individuals. ^[7]

Nurses and midwives

NMBI Table 2 classifies registrants by qualification location, not nationality. Its non-EU group includes the UK. The register covers public and private practice and some practitioners abroad; it is not an HSE staffing count. NMBI’s definitions include assumptions about recent entrants.

Non-EU-educated shares: 42,432 ÷ 86,948 = 48.8% of practising registrants; 39,760 ÷ 79,194 = 50.2% of patient-facing registrants. ^[8]

Employment, education and public finances

The employment comparison uses citizenship, ages 20–64, and 2024 Eurostat figures reproduced in ESRI’s June 2026 review, footnote 19. Education uses citizenship and ages 25–64. Welfare receipt uses birthplace and ages 15–65. These populations should not be treated as the same group or as a profile of citizenship applicants. ^[9] ^[10] ^[11]

The OECD table comes from the same ESRI review (printed pages 20–22), which reports OECD (2021) results for Ireland averaged over 2006–2018. Each row adds another layer of public spending to what is charged to each group: first individual benefits and services, then congestible public goods such as police, transport and water, then pure public goods such as defence and debt interest. The last row is the fiscal ratio for individual items. The period includes the post-2008 deficits and excludes recent surpluses. ^[9]

The fiscal table reproduces the JRC estimate discussed in section 3.1.2 of the ESRI review: annual income tax and social-insurance contributions less benefits per person, for 2015–2019. It excludes a full allocation of public-service costs. ESRI’s broader conclusion draws on several studies with different accounting methods. ^[9] ^[12]

The family-benefit figures come from Table 4.1 of ESRI’s welfare study, using CSO SILC data. The category includes universal Child Benefit. The survey covers private households, not the population in asylum accommodation. ^[11]

Earnings and care

The earnings figures describe employments active for at least 50 weeks in 2024, not household disposable income. (56,356 − 44,816) ÷ 44,816 × 100 = 25.7%. The Scheme has not selected the mean as a threshold. The comparison illustrates the consequences of choosing an income measure. ^[13] ^[14]

Women accounted for 181,592 of 299,128 unpaid carers in Census 2022: 60.7%, rounded to 61%. These are population counts, not citizenship-applicant counts. ^[15]

Test-centre planning estimate

Annual sittings = 41,511 applications × 2 tests × 1.20 = 99,626.4, rounded up. Annual centre capacity = seats × sessions per day × 200 operating days. Divide demand by capacity and round up to obtain the number of centres. Each centre is assumed to have one test room and each session to accommodate one test.

With 20 seats and two daily sessions, this gives 13 centres. Exempting 30% of applicants from both tests reduces that to nine. Removing retakes gives 11. Using only 80% of advertised seats raises the base case to 16.

This is a planning scenario, not an official forecast. The application total includes minors. Exemptions, recognised qualifications, combined tests and existing venues may reduce demand. Speaking assessments, accessible provision and regional coverage affect capacity. A launch backlog is excluded. ^[2]

Legal scope and transition

The General Scheme is a proposal. Regulations will determine the income amount, specified supports and testing details. No particular maternity, caring or disability payment is described here as already disqualifying. A language-test waiver does not itself waive an income condition. ^[14] ^[16]

The offence discussion identifies duties included in the proposed Schedule; an offence must still be established under the applicable law. General waiver powers must also be considered. “No time limit” describes the scheduled condition’s wording, not a claim that no exception could ever apply.

The eight revocations include the two since April 2025. Earlier litigation disrupted the procedure, so a low count cannot by itself establish that wider powers are unnecessary. The 2024 IHREC statement concerns the earlier amendments; the Commission’s October 2026 submission addresses the Scheme itself. The 2026 proposal removes consultation on procedural regulations, not consultation before individual revocation decisions. ^[17] ^[18] ^[19]

Children, residence examples and doctoral study

The January 2022 example assumes uninterrupted qualifying residence. The Scheme protects submitted applications from Head 9’s new residence exclusions but does not provide comprehensive protection for everyone already resident. No claim is made that a further change in three to five years has been announced. ^[14]

The doctoral example assumes four years entirely on Stamp 2 followed by qualifying residence, with all other conditions met. Employed researchers may hold other permissions. Nine or twelve years is time to eligibility, excluding application processing. ^[20]

The child provisions change qualifying parental residence for future births. They do not raise the three-of-four-years test to eight years or remove other birth-citizenship and minor-naturalisation routes. The 6,949 certificates for minors do not measure how many children the proposal would affect. The published route table does not split minors by age or by legal route (section 15B or section 16), so it cannot show how many were 14 or older or how many needed the good-character waiver that Head 8 would remove. ^[14]

Download the evidence

Data tables and calculations

Official source register

Department of Justice · General Scheme, 21 September 2026