Population Projections 2024

Published July, 5th. 2024 - (updated on 08. juli 2024)


All calculations are based on tables in Statbank Greenland

The future population

Main results

The population projection shows the results of model calculations that, based on Greenland’s population sex and age structure as of January 1, 2024, as well as historical experiences, describe possible development scenarios towards 2050.

The number of people living and born in Greenland, according to the Main Alternative, will decrease by one-fifth towards 2050.

As with the other alternatives, only the part of the population living and born in Greenland is projected. The part of the population born outside Greenland is assumed to remain constant with the same sex and age distribution as in 2024 in all future years.

Table 1. Population Account
selected years
2019 2024 2029 2034 2039 2044 2049
Start of year 55.992 56.699 55.030 53.001 50.777 48.546 46.390
Born in Greenlnad 50.254 49.957 48.288 46.259 44.035 41.804 39.648
Liveborn 849 729 645 574 542 521 502
Deaths 517 501 522 539 552 543 519
Emigrated 1.364 1.269 1.228 1.188 1.102 1.033 966
Immigrated 933 731 745 735 676 634 581
Born outside Greenland 5.738 6.742 6.742 6.742 6.742 6.742 6.742
End of year 55.893 56.389 54.670 52.583 50.341 48.125 45.988

See table i Statbank

The large cohorts from the 1960s will, in the coming years, reach ages with higher mortality, which is, however, expected to be lower than the mortality experienced by previous generations at the same age. The average age of the population will increase from 35.5 in 2024 to 39.5 in 2050.

While about 800 children were born annually 10 years ago, it is expected that in 10 years, with the current low fertility rate, fewer than 600 children will be born annually.

Due to the opposite trends for children and the elderly, the demographic dependency ratio will not be significantly higher or lower than what has been observed in the past 40 years.

See table in Statbank Greenland

See table in Statbank Greenland


In Figure 2, the age distribution in the four projections from 2000 to 2050 is compared. The differences are most noticeable for children, youth, and the elderly. The animation is a bit fast, but all numbers can be retrieved from the Greenland StatBank Greenland, where five detailed tables are available:

1. Preface

Population projections are the results of model calculations that, based on Greenland’s population’s sex and age structure as of January 1, 2024, describe possible development scenarios towards 2050.

Expectations regarding the size and structure of the future population are crucial for planning puposes. It is the ambition of Statistics Greenland to provide credible forecasts of future developments based on realistic assumptions. Therefore, one of the calculations is referred to as the main alternative, while the other alternatives are intended to enable assessments of the assumptions underlying the main alternative.

There is generally a great deal of uncertainty associated with purely demographic projections of small populations.

2. Methods

2.1 The Theoretical Framework

The theoretical framework used is known in academic literature as the ‘cohort component method’ and is based on table data. Although the data source for population statistics is an administrative personal register (CPR), which has recorded the population’s addresses for the past 50 years, this information is converted into statistical tables in a consolidated population account. In this account, information is summarized by birth cohort and calendar time, distributed by sex and birthplace group.

This allows for a simple, transparent, and reproducible model. In the consolidated table, population status is combined with an account of events. To compile these into a table, a variable ‘Trekant (Lexis)’ is created, containing information on whether events occurred before or after the birthday. With this variable, age at the event can be calculated based on birth cohort and calendar time.

For example, age at the end of a calendar year is equal to the calendar year minus the birth year (2024-1959=65 years). Age at an event during the calendar year is then adjusted depending on whether the event occurs before (65-1=64 years) or after the birthday (65 years).

Approximately 90% of the population is born in Greenland, and the development for this group over time can be described using normal demographic methods. This also applies to those born outside Greenland, but due to the size of this group, projections are associated with much greater uncertainty regarding the theoretical model and random calendar year effects.

It is observed that those born outside Greenland stay in the country for only a few years, and the sex and age distribution of this group is remarkably stable over time. The size, sex, and age distribution are determined by labor market needs rather than the demographic framework.

Therefore, population projections focus on those born in Greenland. To calculate the total population, the base year’s size, sex, and age distribution are kept constant for those born outside Greenland. In recent years, the need for labor has increasingly been met from outside the Kingdom, so birthplace in the population account is divided into Greenland, within the Kingdom, and outside the Kingdom.

All projections are fully calculated and explained by the model. Various expectations for mortality and fertility can be incorporated, while migrations are determined based on a selected time period.

2.2 Model

Both births, deaths, and emigrations can be determined from rates calculated as events from an at-risk population. It is not possible to calculate immigration using this method as the at-risk population is unknown. But for those born in Greenland, the case is different. Here, re-immigrations are derived from historical emigrations.

The population account is divided into birthplace groups, sex, and 1-year age classes, and supplemented with a few satellite tables, all data to determine the projection model parameters is available.

Projection Model

Population size, base year, and historical data

Out:
Deaths, calculated from assumptions about expected mortality rates throughout the projection period.

Emigrations, calculated from emigration rates, where the basis of experience is determined as the average over several years.

In:
Live births, calculated from assumptions about expected fertility rates throughout the projection period, for all women. It is assumed that women born outside Greenland have the same fertility as those born in Greenland.

sex ratio, calculated as the average of all live births since 1973.

Immigrations, calculated from the calculated number of emigrations and assumptions about re-immigrations.

2.3 The Population Account

At a general level, the population account shows the development in population size from one year to the next and explains the reasons for the development through the demographic components. Errors in data are aggregated into a correction item.

The consolidated population account with satellite tables can be found in the StatBank Greenland.

Tabel 2.3 Befolkningsregnskab
Personer født i Grønland
2019 2020 2021 2022 2023 2024
Population, start of year 50.254 50.192 50.367 50.390 50.161 49.957
Birth 849 835 761 746 716 .
Immigration 933 790 635 686 795 .
Emigration 1.364 957 914 1.214 1.248 .
Internal in migration 7.337 6.679 6.560 6.375 6.489 .
Internal out migration 7.337 6.679 6.560 6.375 6.489 .
Death 517 496 498 492 505 .
Correction 37 3 39 45 38 .
Population, end of year 50.192 50.367 50.390 50.161 49.957 .

See table in Statbank Greenland

The consolidated population account with satellite tables at a detailed level constitutes all data for the projection model. A birth cohort can be divided by sex and birthplace group, and its development can be tracked over time, both at the national and regional levels.

  • ‘Area’ is an administratively defined area included in a projection region.
  • ‘sex’ is women and men as well as ‘total’. The default is ‘total’.
  • ‘Event type’ includes status at the beginning and end of the period, as well as demographic event components, including a correction item.
  • ‘Lexis triangle’ is an auxiliary variable used to calculate age at the event based on the period and birth cohort. The calculation requires information on whether the event occurred before or after the birthday.

Example, Events by age, time, and birth cohort for a selected group

2.4 Calculation of Mortality

Here you can find a detailed presentation of the method for calculating mortality, in Danish and thus mortality/life tables based on grouped data. The data basis is the consolidated population account.

3. Data Basis

3.1 Sex, Age, and Birthplace Group

There is a significant difference in the sex and age distribution based on birthplace. Those born outside of Greenland typically move to the country for work and return a few years after immigration. When a person emigrates, they are replaced by a demographically similar person, just a few years younger.

This can be seen in Figure 3.1, which is divided into three birthplace groups to observe differences in sex and age distribution among the three groups over time. Without this division, the development cannot be explained.

3.2 Registration Status

The Central Person Register was introduced in Greenland on January 1, 1973, and everyone who has lived in Greenland (or Denmark) after 1973 is registered here. In Figure 3.2, the registered population born in Greenland is shown as of February 15, 2024. The registered population is divided by birth cohort. Due to death and emigration, only about half of the 1973 birth cohort still resides in Greenland.

From Denmark’s Statistics series ‘Statistical Yearbook’, the annual birth numbers before 1973 are known. Some births, from cohorts born before 1973 are not registered in Greenland’s statistical population registers because they died or emigrated before 1973. The top red curve shows the annual number of live births since 1918, so the top white area accounts for the number of people who have never been registered in the CPR, as they died (a few emigrated) before the establishment of the CPR or have not lived in Greenland after 1973.

Source: See table in Statbank Greenland https://bank.stat.gl/BEESTFSGRL

Some of the foreign-born people currently living in Greenland were born to Greenland-born parents while they were, for example, studying and residing in Denmark. Conversely, some of those born in Greenland and now living outside Greenland were born to foreign-born parents while they resided in Greenland.

Thus, almost all Greenland-born individuals with non-Greenland-born parents over the age of 10 live in Denmark. Of those who have one Greenland-born parent, about half still reside in Greenland.

Source: See table in Statbank Greenland https://bank.stat.gl/BEESTFSGRL

3.3 Fertility

3.2.1 Calendar Year Fertility

Up until 1970, Greenlandic women had many children. The calculated calendar year fertility showed that each woman could expect to give birth to more than 6 children on average if they gave birth according to the calendar year experiences. From the beginning of the 1970s, fertility halved after access to modern contraceptive methods was introduced. Since Law No. 232 on Termination of Pregnancy came into effect on July 1, 1975, more than half of all pregnancies have ended in induced abortion.

Calendar year fertility remained at around 2.5 children per woman until 2010, after which it declined to about 2.1 children per woman. In recent years, fertility has further decreased, and in 2023, the lowest annual birth rate was recorded for the entire period. The total fertility rate was estimated at 1.8 children per woman in 2023.

https://bank.stat.gl/BEEBBSF1 

On average, women giving birth have become older over the past 20-25 years. Figure 3.3.1b shows the average age of first to fourth-time mothers born in Greenland.

See table in StatBank Greenland

3.3.2 Cohort Fertility

The calendar year fertility shown above indicates how many children a woman would have if she gave birth according to observed fertility patterns within a given short period. This measure does not reflect how many children she actually wants to have or how many she has had.

Cohort fertility shows how many children women from specific birth cohorts have had on average. To calculate this, women are followed through all their fertile ages, from 15 to 49 years.

In 2023, it has been 50 years since the Central Population Register (CPR) was introduced in Greenland, allowing analysis of Greenlandic fertility for 20 complete birth cohorts (1960-1980). From Figure 3.3.2a, it can be seen that birth cohorts from 1970, 1975, and 1980 had around 2.5 children on average, while women from the 1960 and 1965 cohorts had around 2.1 children.

Birth cohorts after 1990 have only passed through a part of their fertile ages, but as shown in Figure 3.3.2b, they are likely to have fewer children than women from the 1960 and 1965 cohorts.

3.3.3 Regional Fertility

In Figure 3.3.3, regional differences in age-specific fertility rates (calculated in 5-year periods) since 2000 are shown. Fertility differences are most pronounced up to the age of 28. It is noted that almost no women give birth after the age of 43.

The decline is particularly notable in settlements with fewer than 3000 inhabitants, which historically had significantly higher fertility rates compared to larger educational cities.

Regional Age-Specific Fertility

3.4 Mortality

Over the past 10 years, life expectancy (remaining life expectancy for infants) has remained largely unchanged. Figure 3.4 illustrates the development over continuous 2- and 5-year calculation periods.

Figure 3.4b shows the development in life expectancy at selected ages. It is noted that the decline in infant mortality until approximately 2015 has significantly contributed to the longer life expectancy, which has increased by 5.7 years for women and 6.8 years for men from 2003 to 2023.

Infant mortality is now so low that future improvements in life expectancy must occur through reductions in non-natural causes of death (accidents and suicide) and lifestyle factors (tobacco and exercise). Using broad strokes, Table 3.4 shows very small changes in life expectancy for the elderly. A 67-year-old can now expect to live 1 year longer than 20 years ago, the same as 10 years ago.

Table 3.4 Life Expectancy at Selected Ages
Individuals born in Greenland
Expected Life Expectancy Conditional Expected Life Expectancy
Men Women Men Women
2003 2013 2023 2003 2013 2023 2003 2013 2023 2003 2013 2023
0 61 66 68 67 72 73 61 66 68 67 72 73
30 38 40 41 41 43 45 68 70 71 71 73 75
40 29 31 32 32 34 35 69 71 72 72 74 75
50 21 23 23 23 25 26 71 73 73 73 75 76
60 13 15 15 15 17 18 73 75 75 75 77 78
67 9 10 10 11 12 12 76 77 77 78 79 79
80 4 4 4 5 5 5 84 84 84 85 85 85

The development in population survival is illustrated in Figure 3.4c, where survival for an ‘artificial’ generation of 1000 women and men is shown since 1986. Each calendar year is calculated based on experiences in the previous 5-year period.

Although the previous decline in life expectancy in recent years has stopped, there is no reason to believe that mortality will not continue its previous decline in the future. Birth cohorts born after 1950 have lower mortality rates than those born before. It is reasonable to assume that the lower mortality observed for birth cohorts after 1950 will also apply as these cohorts reach older ages.

Figure 3.4d shows survival for selected generations. It is evident that survival is higher for younger generations compared to older generations.

Changes in sex- and age-specific mortality rates are calculated and shown in Figure 3.4e for the past 10 years. The change is expressed in percentage as the difference in the number of survivors in the age group compared to a birth cohort that is 20 years older.

For projections, it is assumed that the development observed over the past 20 years will continue in the future. This means that the percentage changes for the period 1998/2003 to 2018/23 evenly distributed by sex and age over 20 years correspond to future changes. The percentages are smoothed using LOESS.

3.4.2 Uncertainty

The small population size significantly affects quotient calculations due to random annual variations. Figure 3.4.2 illustrates the smoothing effect by calculating mortality rates over respective 2-year (2022/23) and 5-year periods (2019/23). Mortality generally increases with age, except for the first year of life.

Even when calculated over a 5-year average, Figure 3.4.2b shows significant random annual variations. In model calculations, these variations are eliminated by assuming that the observations come from a sample drawn from a much larger population, where the ‘true’ mortality can be determined. This is achieved here using a parameter-free method developed by Anastasia Kostaki (2000) and incorporated in the R package MortalityLaws. The method estimates mortality rates in 1-year age groups based on mortality rates defined by birth cohort and period (b-groups) over continuous 5-year periods.

In Figure 3.4.2b, this is shown for the 5-year period 2019/23.

3.5 Migrations

Over time, Greenland has almost consistently lost a portion of its population to the outside world. Throughout the period 1993 - 2023, there were a total of 37514 emigrations and only 26188 immigrations. On average for the entire period, 365 individuals born in Greenland net-emigrated annually.

3.5.1 Emigrations

Approximately 10% of the annual number of live births, or about 75 children, are born to women born outside of Greenland. Children born to parents both born outside of Greenland typically emigrate before starting school. Adjustments for this have not been made in the calculations. Greenlandic society statistics have a long tradition of classification by place of birth, which is why this division is chosen here. Greenland Statistics also categorizes the population by ‘extended place of birth’, where information on parents’ and grandparents’ birthplaces is used for delineation. This statistical concept must be fully implemented in StatBank Greenland’s table collection before it can be applied.

Unlike many other countries where projections focus on expectations for future (net) immigration, Greenland’s population projections focus on emigrations. Since the population base is known, emigration ratios, etc., can be calculated using standard demographic methods.

Emigration Rates

Age-specific emigration rates are calculated based on population accounting, and detailed emigration rates can be seen in the Statistical Bank table BEDOUTREG.


In Figure 3.5.1, emigration rates are shown for selected 5-year periods. Post-secondary school stays in Denmark significantly influence migration for 16-17-year-olds, where up to one-fifth attend boarding school after 10th grade. The age-distributed pattern is surprisingly stable over time, and it is observed that emigration rates for most ages are somewhat lower in the Covid-19 years, 2020 and 2021.



https://bank.stat.gl/BEEOUTREG

3.5.3 Immigration

Immigration cannot be calculated in the same way as emigration because the at-risk population is not currently known with the necessary level of detail, despite almost all emigrations having Denmark as their destination. Instead, information on return migrations is used, which in the projection model determines the expected number of Greenland-born immigrants.

Although the annual net emigration averages individuals and a Greenland-born person can expect to emigrate approximately twice in their lifetime on average, this suggests that the number of return migrations must be nearly as high. Emigration is not a permanent departure from Greenland but an event that will have some duration—months or years—as many will move between Greenland and Denmark many times, for example, due to high school, secondary education, or work, family relationships, or other reasons.

3.5.4 Return Migrations

On average, two-thirds of emigrants return to Greenland within 10 years. This average is too coarse for projection purposes, hence the same measure is now calculated by sex and age.

Nine out of ten individuals who emigrate at the age of 16 return to Greenland within 1 year.



4. Calculations

4.1 Structure and Sequence

Greenland Statistics annually publishes tables with projection results in the StatBank Greenland. These tables include a main scenario as well as several alternative projections.

Each projection alternative is documented by separate scripts containing the programming code that selects data from the StatBank Greenland and saves it along with derived calculations in an Excel file.

The Excel file contains all relevant data and serves as complete input to the projection module, where another Excel file is generated with calculation results. One Excel file is produced for each alternative.

Finally, the table presented in the StatBank Greenland is created using a package (pxmake) developed for Greenland Statistics, which is also available on GitHub. The full documentation for this process can be found in the script that generates this document.

For users executing the code themselves, Pxwin for Windows (from www.Scb.se) can be used to explore the tables.

4.2 Assumptions

All projections intended to be presented in a StatBank Greenland table must share the same dimensions regarding regions, sex, age, birthplace, and time.

Regarding the time dimension, the base year is the latest population census. The projection horizon is not a technical limitation but rather a matter of communication. A longer time horizon introduces greater uncertainty in the calculation results.

Initially, data in the form of a spreadsheet containing all data must be prepared. Therefore, ‘DoBase_main.qmd’ is edited and executed in the programming environment to generate (in Danish)DoBase_main.html along with Base_main.xlsx.





Once the main scenario is finalized, ‘DoBase_main.qmd’ serves as a template for an unlimited number of alternatives, such as ‘DoBase_alt1.qmd’ and ‘DoBase_alt2.qmd’. These are edited and executed in the programming environment, resulting in (in Danish) DoBase_alt1.html, DoBase_alt2.html, and DoBase_alt3.html along with their corresponding xlsx foundational data files.

4.3 Projection Module

Each projection alternative is processed using the projection module, and the results are saved in a cluster of Excel files. The cluster is named with a common prefix.

Here, the specified alternatives are loaded and utilized by the projection module.

5. Tables

Calculation results are saved in a number of Excel files, for example: 2024v1_projection_main.xlsx

2024v1_projection_main.xlsx

2024v1_projection_alt1.xlsx




5.1 for Statbank/Pxwin

From the projection cluster, relevant data is selected and aggregated into tables using the R package pxmake, which translates them into tables to be presented in Statistikbanken.

5.1.1 BEXP24.px

In this table, projection results for 2024 are aggregated. The table itself covers the period from 1999 to 2050.



5.1.2 BEXP24CALC.px

In this table, historical and future calculated events are aggregated in a population account without information on birth years.




5.1.3 Fertility and Mortality in Projections

To facilitate straightforward comparison of expectations for future fertility and mortality, these are aggregated into separate tables.

5.1.3a BEXP24FERT.px





5.1.3b BEXP24MORT.px





5.1.4 BEXP24EMR.px