"Hayk's Arrow" Armenian Strategic Center
Official statistics portray a population still growing in aggregate but with low recent fertility and substantial differences by age, place and sex.
Official statistics portray a population still growing in aggregate but with low recent fertility and substantial differences by age, place and sex. The strategic issue for Armenia is how these cohorts translate into skills and economic capacity, while keeping administrative estimates, census coverage and permanent-migration definitions distinct.
The detail below uses WDI/UN demographic estimates and ILO modeled labour series, not national census counts or a merged population total. Common-year comparisons preserve the distinction between labour-force participation, employment and age composition.
| Indicator | Value and reference period |
|---|---|
| Female labour-force participation, age 15+ILO modeled female labour force—employed plus unemployed—divided by female population aged 15+. Participation is not employment. | 60.8 % of female population aged 15+2025Estimate[116][125] |
| Male labour-force participation, age 15+ILO modeled male labour force—employed plus unemployed—divided by male population aged 15+; the same age threshold as the female series. | 66.0 % of male population aged 15+2025Estimate[117][125] |
| Female employment-to-population ratio, age 15+ILO modeled employed women divided by all women aged 15+; excludes unemployed jobseekers from the numerator. | 57.0 % of female population aged 15+2025Estimate[118][125] |
| Male employment-to-population ratio, age 15+ILO modeled employed men divided by all men aged 15+; it is not the employment share of the working-age 15–64 population. | 62.9 % of male population aged 15+2025Estimate[119][125] |
| Urban population shareUN-smoothed population series using each country’s national urban definition; urban boundaries and classifications are not identical. | 58.6 % of total population2025Estimate[120] |
| Total age dependency ratioPopulation under 15 plus population 65+, divided by population 15–64. Age groups do not identify actual earners, carers or financial dependence. | 43.6 persons per 100 people aged 15–642025Estimate[121] |
| Older-age dependency ratioPopulation aged 65+ per 100 people aged 15–64, using the WDI/UN population series; not a pensioner-to-worker ratio. | 13.2 persons per 100 people aged 15–642025Estimate[122] |
| Child-age dependency ratioPopulation under 15 per 100 people aged 15–64. It describes age composition, not the cost or quality of childcare and schooling. | 30.4 persons per 100 people aged 15–642025Estimate[123] |
| Life expectancy at birthPeriod life expectancy: expected years if that year’s mortality patterns persisted; not the observed lifetime of a birth cohort. | 74.6 years2024Estimate[124] |
Azerbaijan’s modeled participation gap is 5.3 percentage points in 2025, smaller than the gaps in the same international series for Armenia and Georgia. It is not evidence that Armenian residents face the same labour conditions or that the smaller gap means better pay, formal jobs or equal care burdens. Armenia can use the comparison to ask which participation constraints its own evidence identifies, while keeping job quality and individual preference separate. The employment-to-population rows help distinguish participation from actual employment, but still do not identify working hours, earnings or occupational shortages. [116][117][118][119]
The international age structure puts a larger child-related than older-age component into Azerbaijan’s dependency ratio. For Armenia, that is a difference in the composition of potential service demand, not a forecast of another country’s future power or available labour. Actual school, care and health demand depends on residence, household arrangements and provider capacity. The national official Azerbaijani population and birth series already shown elsewhere retain their own dates and definitions; they should not be spliced into this UN denominator. Armenia’s essential-service staffing choices must still be justified by its own vacancies, turnover and voluntary recruitment evidence. [123][122][121]
Labour rows are ILO modeled estimates for age 15+, whereas age-dependency ratios use ages 15–64 as the denominator. UN population estimates and national census-rebased series may differ; they must not be spliced. Urban definitions remain national. Life expectancy uses the common 2024 observation, while the other rows use 2025; it is not a forecast of an individual’s lifespan. ILO models combine reported and imputed observations and may retain an older employment definition where changing standards would create a break. The exact model-to-survey reconciliation has not been established here; treat labour comparisons as qualified, not a ranking of national labour-market performance. [116][121][120][124][125]
Azerbaijan's large working-age population and expanded vocational pipeline do not establish either an available workforce for Armenia or a model Armenia can copy.
Azerbaijan's large working-age population and expanded vocational pipeline do not establish either an available workforce for Armenia or a model Armenia can copy. Its ageing and unequal preschool access instead show how aggregate capacity can coexist with service constraints. The distinctive Armenian choice is whether to strengthen continuity in an underserved essential service or finance another broad training intake. Begin with a documented Armenian vacancy or turnover problem, not a demographic comparison. Retaining qualified staff through workable conditions can preserve residents' access, but simply attracting a worker from another underserved locality may redistribute the shortage rather than solve it. [20][90][114]
Give regional-staffing preparation priority where a verified shortage threatens an essential service. This is a conditional choice: the Azerbaijani evidence cannot identify an Armenian locality or profession in need. Use Armenian vacancy, qualification, workload and turnover evidence to decide. Compared with another enrolment target, retention asks whether existing trained people can deliver more reliable service without coercion. The first 90 days concern diagnosis and agreed conditions, not a promise to fill posts. [20][28][48][90]
Proposed horizons run from authorisation. Later stages depend on the conditions below; these are analytical choices, not adopted deadlines or funding commitments.
In the proposed first 90 days, select one verified Armenian essential-service staffing gap using existing records and voluntary staff and user interviews. Separate an unfilled budgeted post, unaffordable vacancy, excessive workload and missing qualification. Reconcile only the resident and age definitions needed to establish demand; name the service decision before gathering more information. [20][28][48]
During the proposed three-to-twelve-month phase, test a voluntary retention package addressing the documented cause: supervision, predictable hours, travel, training or care access where relevant. Compare service continuity and worker well-being with baseline conditions. Check effects on neighbouring providers so a successful recruitment is not counted as a net gain when another essential service loses capacity. [28][48][83][95]
Beyond twelve months, retain only affordable arrangements with verified service gains. Where the constraint is a genuine skills pipeline, add employer-linked paid transitions with transferable qualifications and proper licensing. Update staffing plans as local demand changes; keep participation voluntary and do not impose residence obligations or population targets. [90][1][8][20]
The binding constraint is recurrent qualified service capacity, not a room, a national headcount or a one-time recruitment payment. Working conditions, safe staffing and an operating budget must remain viable after the trial. Armenian authorities need lawful employment and eligibility rules and a provider able to maintain the service. Aggregate Azerbaijani ratios cannot justify Armenian ethnic selection, coercive mobility restrictions or relaxed professional standards. [20][28][48][95]
Put care-access first when qualified people and vacancies already exist but care hours prevent willing residents from working. Put paid-transitions first when a willing employer identifies a genuine qualification-to-job gap instead. If current provision is adequate, no new programme is justified by regional comparison alone. The choice must follow the Armenian service bottleneck and residents' preferences, not a desire to match Azerbaijan's enrolment or population scale. [83][90][7][8]
Azerbaijan's growing total population coexists with smaller child cohorts and more older residents. For Armenia, this changes the meaning of demographic comparison: a larger population does not mechanically translate into a proportionately larger employed workforce, fiscal surplus or permanent strategic advantage. Care demands, health, qualifications and participation intervene. The implication is to judge Armenia's own capacity through productive and healthy working lives, rather than adopt a numerical race in births or migration. Azerbaijan's pattern can inform regional comparisons of labour availability and future service demand, but it is not evidence of hostile intent. Policies should improve the welfare and opportunities of residents, with individual rights independent of any external demographic benchmark. [19][20][30][1]
The larger flow of Azerbaijani school graduates and the expansion of vocational enrolment create potential productive capacity, but qualifications become economic strength only when employers use them well. Armenia's relevant exposure is competition for customers, investment and skilled work, not the mere existence of young people across the border. An Armenian response can be concrete and domestic: connect training to paid work, remove avoidable barriers to participation and retain people through credible opportunities. Copying enrolment targets would be a weak response if graduates do not find suitable jobs. Regional education comparisons are useful when they identify a specific skill or transition problem and a way to improve Armenian outcomes, rather than rank populations as inherently stronger or weaker. [79][81][90][7][8]
Azerbaijan's permanent-migration register captures a narrow category and cannot describe every worker abroad, temporary resident or border crossing. Its household and health evidence likewise concerns different periods and populations. For Armenia, careless comparison can misdirect staffing, housing and integration policy, especially when official totals are treated as ready-made measures of available labour. Family care obligations and geographic access to services influence whether a resident can work even without moving. The useful lesson is to link Armenian workforce policy to actual residence, skills, care needs and employer demand, while preserving privacy. This allows mobility and participation to be addressed as practical service issues rather than as a competition to accumulate population numbers. [9][12][66][111]
The State Statistical Committee reports 10.2624 million residents at the beginning of 2026, including 5.5742 million in urban areas and 4.6882 million in rural areas. These are national administrative estimates, not an independently verified headcount. The library retains separate tables for population change, average annual population, density and settlement size so that different denominators are not silently combined. [22][23][18]
The national population-change series rises from 9.593 million at the beginning of 2015 to 10.2624 million at the beginning of 2026. Aggregate growth therefore coexists with the lower fertility described below. For Armenia, this combination points to demographic momentum: today’s total can keep rising even as future entering cohorts become smaller. District and settlement distributions show where that momentum is concentrated, with implications for labour availability, urban services and the location of future economic demand. [19][14][17]
The official total fertility series records 1.4 children per woman in 2025, with 1.2 in urban areas and 1.6 in rural areas. Its historical series also records 1.4 in 2024. Fertility rates describe a period pattern, not the completed family size of every woman. Birth totals, age-specific fertility and birth order provide necessary additional context for interpreting the aggregate. [30][35][36]
The registered-birth series falls from 166,210 in 2015 to 95,875 in 2025, while natural increase falls from 111,513 to 36,041. This is a substantial slowing of renewal even though births still exceed deaths in the official series. For Armenian planners, smaller successive birth cohorts can reshape the regional balance of school demand and later labour-market entrants. Maternal-age and birth-order tables help distinguish delayed family formation from changes in the number of potential parents. [25][39][36]
The mean age at first childbirth rose from 23.3 years in 2015 to 24.5 in 2025; the mean for all births increased from 25.1 to 26.5. Delayed births can temporarily depress a period fertility rate, but the simultaneous contraction of young children shows a real reduction in current school-entry cohorts. The urban all-birth mean of 27.7 years versus 25.5 in rural areas also implies different local timetables for family formation and care demand. [33][39][35]
UNFPA’s November 2023 factsheet reports 112 boys per 100 girls born in 2022, after 116 in 2021, in its analysis of son preference and gender-biased sex selection. This is a historical birth-ratio indicator, separate from the total fertility rate. It is retained with its older reference year rather than presented as the sex ratio of 2025 or 2026 births. [115]
The official life-expectancy table reports 77.1 years at birth in 2025, including 74.9 for men and 79.3 for women. This period measure is not a guarantee of an individual lifespan. Separate cause-of-death, infant, maternal and working-age mortality tables allow the national average to be examined alongside age and sex differences rather than treated as a complete health assessment. [31][48][46]
The same official life-expectancy series records 75.6 years in 2015, 72.7 in 2020 when rounded, and 77.1 in 2025. It therefore shows recovery after a sharp interruption, rather than a smooth uninterrupted improvement. For Armenia, the strategic implication is that healthy working life and care needs matter alongside population size. Cause-of-death and working-age mortality tables give the more specific evidence needed to assess which health pressures could affect participation and productivity. [31][48][58]
WHO’s 2025 profile models 2023 current tobacco use at 19.0% among people aged 15+, with 37.8% for men and 0.1% for women. These are age-standardised estimates, distinct from raw survey prevalence. The pronounced sex difference adds a specific risk-factor dimension to the mortality tables; it does not establish an individual cause of death or present-day prevalence. [113]
The international-migration series records 3,400 arrivals for permanent residence and 2,000 departures in 2025, a net inflow of 1,400. Its definition is crucial: permanent-residence registration is not a count of all border crossings, seasonal work or temporary migration. The country-of-origin and age-sex tables add composition but do not remove this underlying coverage limitation. [9][10][12]
The registered permanent-migration balance increased from plus 700 in 2024 to plus 1,400 in 2025, remaining small relative to the national population and annual natural increase. This suggests that the movement captured by this specific register is not the main numerical driver of aggregate growth. For Armenia, the country and age composition is more useful than the net figure alone when examining labour links, household connections and possible exposure to changes in neighbouring countries’ employment conditions. [9][10][12]
The household-size table is based on censuses, with its latest displayed census in 2019; it records an average of 4.3 members per household. It must not be relabelled as a 2025 household survey. Annual marriage and divorce registers answer a different question: legal events recorded in a year, rather than the full number or lived circumstances of families. [66][62][60]
Registered marriages declined from 68,773 in 2015 to 48,546 in 2025, while divorces rose from 12,764 to 20,675. These event series, combined with the census household structure, indicate changing family-formation patterns rather than a direct count of stable or unstable families. For Armenia, they provide context for housing, care and labour mobility. The strategic issue is how household obligations and resources influence participation, not a moral judgement derived from a marriage-to-divorce ratio. [62][66][64]
MICS 2023 estimates that 15.0% of women aged 20–24 first married or entered a union before 18: 8.4% in urban areas and 25.4% in rural areas. These retrospective survey histories include cohabiting unions, unlike registered marriages alone. They describe the sampled cohort’s earlier experience, not the number or rate of marriages newly occurring in 2023. [111]
The youth collection distinguishes the population aged 14–29 from labour-force tables generally using ages 15–29. These groups are not interchangeable. School enrolment, qualifications, employment and unemployment cover different stages of transition into adult work. Several educational-attainment tables remain census-based, while labour tables include 2024–2025 observations; both the age definition and reference year must travel with any comparison. [2][2][1]
The 2025 youth labour table records 1.1845 million economically active people aged 15–29, including 1.0855 million employed and 99,000 unemployed. That is a capacity measure with an employment structure, not simply a count of young residents. For Armenia, comparing qualification and sex-specific participation can reveal where Azerbaijan’s demographic potential is realised or constrained. Training abroad and domestic educational attainment add information about the future skills pipeline, without guaranteeing that qualifications translate into productive jobs. [1][7][2]
External context: the UNICEF-supported MICS surveyed households in May–August 2023 and published its findings in September 2024, providing a distinct survey basis for conditions affecting women and children. UNFPA’s historical implementation profile uses older demographic baselines and projections. These documents broaden the national-register picture, but neither should replace 2025 observations; differences in period, sample population and estimation method must remain explicit. [111][112]
The age pyramid is already changing beneath aggregate growth. At the beginning of 2026 the official series counts 539,100 children aged 0–4, compared with 851,400 in 2015. Meanwhile the population aged 65 and above rises from 570,700 to 1,086,400, calculated by adding the published 65–69 and 70-plus groups. These are observed administrative age groups, not a projection. Their coexistence means that the demand for primary schooling and the demand for elder care can move in opposite directions within the same growing country. [20][28]
A broad 15–64 working-age band still contains about 7.166 million people in the 2026 table, calculated from ten five-year groups. That potential is filtered by employment, health, unpaid care and skill matching. It is not an army size, a count of taxpayers or an employed workforce. For Armenia the useful comparison is how many people produce income and services per dependent resident. A neighbour can retain a large labour pool while simultaneously facing rising retirement and care costs, reducing the fraction of public resources available for other purposes. [20][48][7]
The education pipeline shows why births and current labour entry should not be confused. Complete-secondary graduates increased from 88,600 in 2015 to 111,400 in 2025, even while births declined. Larger earlier cohorts are still progressing through school. At the beginning of 2026 vocational institutions reported 40,820 students and 17,544 graduates, against 35,262 and 14,764 a year earlier. These are concrete training volumes; whether they raise industrial capacity depends on occupation, teaching quality and employer demand. Preschool provision also matters because accessible care influences whether parents can remain in paid work. [79][81][83][90]
Care obligations connect population structure to household purchasing power. The 2025 mortality register attributes 33,555 deaths to circulatory diseases and 10,102 to neoplasms, identifying a substantial chronic-disease burden without measuring disability. The child-service series reports 3,274 paediatricians at the beginning of 2026, while 471,796 children received monthly social benefits averaging 209.12 manats. Service counts and benefit recipients describe different populations. Together they show that human-capital policy requires continuing health and household support, rather than a one-time expansion of classrooms or construction. [55][95][100]
UNESCO’s 2026 case study cites the 2023 MICS showing early-childhood programme attendance among five-year-olds at 70% in the richest household-wealth quintile and 55% in the poorest. The 15 percentage-point gap survives national expansion. These are survey participation measures by household wealth, not a count of available preschool places or the current year’s administrative enrolment rate. [114][111]
Smaller child cohorts alongside more older residents can raise care costs and slow future labour-force renewal, even while total population still grows. [20][30]
Education enrolment and a large working-age population do not remove employment and health constraints; unrealised skills can leave demographic scale below its economic potential. [1][7][95]
Expanding vocational training can support industrial and service employment if qualifications match real vacancies; the larger current graduate flow provides a near-term skills resource. [90][81]
Accessible childcare and targeted family support can make existing human capital more usable by reducing constraints on parents’ paid work. [83][100]
These scenarios explore how documented drivers could affect Armenia. They are conditional analytical paths, not predictions or assigned probabilities.
Conditional scenario, not a forecast: if the larger graduate flow and vocational expansion find productive jobs, Azerbaijan’s economic capacity could increase without faster population growth. Armenia would face stronger service and industrial competition, alongside possible complementary business demand. [90][1][81]
Conditional scenario, not a forecast: if smaller entering cohorts coincide with rising elder-care needs, household and fiscal resources may shift towards care. Armenia’s regional comparisons would then need productivity and participation measures rather than treating Azerbaijan’s total population advantage as a fixed economic multiplier. [20][30][55]
These are analytical recommendations based on the evidence, not adopted government decisions or commitments by the named institutions. Timelines and measures below are proposed implementation tests.
Prepare an Armenian cohort-and-service baseline linking schools, care demand and labour entry under clearly stated residence and age definitions.
Prepare an Armenian cohort-and-service baseline linking schools, care demand and labour entry under clearly stated residence and age definitions. Start by reconciling existing Armenian administrative and survey sources; use Azerbaijani data only as an attributed comparator, with census and migration definitions kept separate.
Pilot paid work placements tied to specific vacancies identified by Armenian employers, with written learning objectives and supervised tasks.
Pilot paid work placements tied to specific vacancies identified by Armenian employers, with written learning objectives and supervised tasks. Start with existing vocational providers and willing firms; admit eligible residents through transparent criteria and provide portable evidence of skills rather than training attendance alone.
Identify Armenian communities where care schedules prevent willing residents from accepting existing jobs, then pilot compatible childcare or elder-care hours.
Identify Armenian communities where care schedules prevent willing residents from accepting existing jobs, then pilot compatible childcare or elder-care hours. Start with voluntary household and employer interviews and use existing service providers where safe, rather than construct facilities before demand is known.
Select Armenian communities with documented shortages in essential services and build voluntary retention packages around working conditions, training, predictable schedules and practical transport.
Select Armenian communities with documented shortages in essential services and build voluntary retention packages around working conditions, training, predictable schedules and practical transport. Begin with vacancy and turnover records, then interview current staff and service users before choosing financial or non-financial measures.
The count includes distinct documents whose content was retrieved and checked for relevance to this profile. Translations, duplicate links and generic homepages are excluded. A document may support more than one topic when its content is relevant to each. Historical documents retain their original dates; they do not establish that a project or policy remains unchanged today.
Sources labelled “Excerpt checked” were reviewed only in the retained relevant passage. They can support the cited claim but are excluded from the full-document total.
Filters use the source date recorded in this library. Indicator reference periods are shown separately.
Two annual observations from the official birth series.
These are annual registered births, not fertility rates. Intermediate years are not plotted; no continuous path is implied.
| Observation | Value · Births | Reference period | Status | Sources |
|---|---|---|---|---|
| 2015 | 166,210 | 2015 | Reported result | |
| 2025 | 95,875 | 2025 | Reported result |
The detail below uses WDI/UN demographic estimates and ILO modeled labour series, not national census counts or a merged population total. Common-year comparisons preserve the distinction between labour-force participation, employment and age composition.
On smaller screens, scroll the table sideways to compare all three countries.
| Indicator | 🇦🇲 Armenia | 🇦🇿 Azerbaijan | 🇬🇪 Georgia |
|---|---|---|---|
| Female labour-force participation, age 15+% of female population aged 15+Read with stated qualifications | 51.1 % of female population aged 15+2025EstimateSource 1Source 2 | 60.8 % of female population aged 15+2025EstimateSource 1Source 2 | 56.7 % of female population aged 15+2025EstimateSource 1Source 2Source 3 |
| Male labour-force participation, age 15+% of male population aged 15+Read with stated qualifications | 69.5 % of male population aged 15+2025EstimateSource 1Source 2 | 66.0 % of male population aged 15+2025EstimateSource 1Source 2 | 73.5 % of male population aged 15+2025EstimateSource 1Source 2Source 3 |
| Urban population share% of total populationRead with stated qualifications | 65.9 % of total population2025EstimateSource | 58.6 % of total population2025EstimateSource | 61.2 % of total population2025EstimateSource |
| Older-age dependency ratiopersons per 100 people aged 15–64Matched period and definition | 21.7 persons per 100 people aged 15–642025EstimateSource | 13.2 persons per 100 people aged 15–642025EstimateSource | 25.2 persons per 100 people aged 15–642025EstimateSource |
| Child-age dependency ratiopersons per 100 people aged 15–64Matched period and definition | 28.5 persons per 100 people aged 15–642025EstimateSource | 30.4 persons per 100 people aged 15–642025EstimateSource | 32.3 persons per 100 people aged 15–642025EstimateSource |
| Life expectancy at birthyearsMatched period and definition | 78.3 years2024EstimateSource | 74.6 years2024EstimateSource | 74.7 years2024EstimateSource |
ILO modeled female labour force—employed plus unemployed—divided by female population aged 15+. Participation is not employment. Same WDI model series and year, but not interchangeable with national LFS. ILO may retain older definitions to avoid breaks; Georgia changed national standards in 2020. No complete country-specific reconciliation of the model-to-survey gaps is established here.
This is a common-age modeled participation comparison. It can prompt Armenian investigation of barriers, but does not prove that care services alone explain the difference or that every non-participant seeks work.
ILO modeled male labour force—employed plus unemployed—divided by male population aged 15+; the same age threshold as the female series. Same WDI model series and year, but not interchangeable with national LFS. ILO may retain older definitions to avoid breaks; Georgia changed national standards in 2020. No complete country-specific reconciliation of the model-to-survey gaps is established here.
Read together with the female row and employment ratios. Participation includes unemployed jobseekers and is not a count of immediately available skilled labour.
UN-smoothed population series using each country’s national urban definition; urban boundaries and classifications are not identical. Common year and WDI/UN compilation, but each country retains its national urban classification; compare with qualification.
National urban definitions differ, so the percentages do not identify identical settlement types or service access. Armenian programmes need location-specific evidence.
Population aged 65+ per 100 people aged 15–64, using the WDI/UN population series; not a pensioner-to-worker ratio. Same international modeled/estimated series and reference year; it does not replace national survey or census results.
Older-age pressure differs from child-related demand. Armenia should plan accessible health and care using actual local needs; this is not a pensioner-to-employed-worker ratio.
Population under 15 per 100 people aged 15–64. It describes age composition, not the cost or quality of childcare and schooling. Same international modeled/estimated series and reference year; it does not replace national survey or census results.
The indicator separates age composition from fertility targets and school quality. It supports broad planning, not an obligation for households to change family decisions.
Period life expectancy: expected years if that year’s mortality patterns persisted; not the observed lifetime of a birth cohort. Same international modeled/estimated series and reference year; it does not replace national survey or census results.
The common 2024 estimate describes mortality conditions, not healthy life expectancy, personal prognosis or healthcare-system performance alone.
Labour rows are ILO modeled estimates for age 15+, whereas age-dependency ratios use ages 15–64 as the denominator. UN population estimates and national census-rebased series may differ; they must not be spliced. Urban definitions remain national. Life expectancy uses the common 2024 observation, while the other rows use 2025; it is not a forecast of an individual’s lifespan.
All values are rounded to one decimal; original precision and indicator metadata are retained in the research notes. Source counts do not imply independent corroboration.
Several countries or indicators can share one underlying data release. Multiple source links do not mean independent confirmation.
These dated records preserve the earlier research edition. Later source releases and statistical revisions are reflected in the detailed analysis above.
Reference period: Population estimate at beginning of 2025
The official series lists 10,224,900 people on 1 January 2025, with 54.4% in urban places. This is a start-of-year national estimate, not a census conducted that day.
Reference period: 2024 registered births and deaths
Official statistics record 102,310 births and 58,909 deaths in 2024. Their difference is natural increase; it excludes migration.
This is a curated starting collection, not an exhaustive country database or a live feed. Each record identifies its reference period and links to original evidence. “Announced” does not mean funded, ratified or operational; those distinctions are stated in the text. Armenia implications are editorial analysis, not source quotations. Census counts, population estimates and forecasts must not be treated as interchangeable. Source-check dates do not change the underlying data period.