Chapter 8
Subprovincial population estimates

Go to text

• 8.1 Definition and relevant concepts
• 8.2 Postcensal population estimates of subprovincial regions
  • 8.2.1 Postcensal population estimates for CMAs and CDs
  • 8.2.2 Postcensal population estimates for economic regions (ERs)
  • 8.2.3 Subprovincial postcensal population estimates by age and sex
  • 8.2.4 Estimate levels
  • 8.2.5 Base population and components of population growth
• 8.3 Intercensal population estimates for subprovincial regions
• 8.4 Population estimates for census subdivisions
  • 8.4.1 Base population of census subdivisions
  • 8.4.2 Postcensal population estimates for census subdivisions

Beginning of text

In addition to estimates at the national and provincial and territorial levels, population estimates are produced for subprovincial regions. Annual population estimates by age and sex are released for census divisions (CDs), census metropolitan areas (CMAs) and economic regions (ERs). Custom requests for population estimates at other geographical levels (e.g., census subdivisions (CSDs), census agglomerations (CAs) and health regions) are also available. Annual population estimates for the subprovincial regions are based on the most recent Standard Geographical Classification (SGC) from 2001 on.

This chapter discusses the estimation methods used to produce postcensal and intercensal population estimates for CDs, CMAs and ERs. The methods used to produce population estimates at the census subdivision level are also presented at the end of this chapter.

8.1 Definition and relevant concepts

A census division (CD) is the general term for provincially legislated areas (such as county, municipalité régionale de comté and regional district) or their equivalents. Census divisions are intermediate geographic areas between the province or territory level and the municipality (i.e., census subdivision).^{Note 1} Census divisions have been established in provincial law to facilitate regional planning, as well as the provision of services that can be more effectively delivered on a scale larger than a municipality. In Newfoundland and Labrador, Manitoba, Saskatchewan, Alberta, Yukon, Northwest Territories and Nunavut, provincial or territorial law does not provide for these administrative geographic areas. Therefore, Statistics Canada, in cooperation with these provinces and territories, has created equivalent areas called census divisions for the purpose of disseminating statistical data. In Yukon, the census division is equivalent to the entire territory.

A census metropolitan area (CMA) is formed by one or more adjacent municipalities centred on a population centre (known as the core). A CMA must have a total population of at least 100,000 of which 50,000 or more must live in the core. To be included in the CMA, other adjacent municipalities must have a high degree of integration with the core, as measured by commuting flows derived from previous census place of work data.

Once an area becomes a CMA, it is retained as a CMA even if its total population declines below 100,000 or the population of its core falls below 50,000. Small population centres with a population count of less than 10,000 are called fringe. All areas inside the CMA that are not population centres are rural areas.

An economic region (ER) is a grouping of complete CDs (with one exception in Ontario) created as a standard geographic unit for analysis of regional economic activity. Within the province of Quebec, economic regions (régions administratives) are designated by law. In all other provinces and territories, economic regions are created by agreement between Statistics Canada and the province or territory concerned. Prince Edward Island and the three territories each consist of one ER. In Ontario, there is one exception where the ER boundary does not respect census division boundaries: the census division of Halton is split between the ER of Hamilton–Niagara Peninsula and the ER of Toronto.

8.2 Postcensal population estimates of subprovincial regions

8.2.1 Postcensal population estimates for CMAs and CDs

The component method is used to produce population estimates for CMAs and CDs. To ensure concordance between the subprovincial and provincial and territorial population estimates by age and sex, two-way raking is used.

The following formula is used to produce the total CMA and CD population estimates:

For each subprovincial region:

Equation 8.1:

$\begin{array}{l}{\text{P}}_{\left(\text{t+1}\right)}{\text{= P}}_{\left(\text{t}\right)}{\text{ + B}}_{\left(\text{t, t+1}\right)}-{\text{ D}}_{\left(\text{t, t+1}\right)}{\text{ + I}}_{\left(\text{t, t+1}\right)}-\left({\text{E}}_{\left(\text{t, t+1}\right)}{\text{ + ΔTE}}_{\left(\text{t, t+1}\right)}\right){\text{ + RE}}_{\left(\text{t, t+1}\right)}\\ {\text{ + ΔNPR}}_{\left(\text{t, t+1}\right)}{\text{ + ΔNinter}}_{\left(\text{t, t+1}\right)}{\text{ + ΔNinfra}}_{\left(\text{t, t+1}\right)}\end{array}$

where, for each subprovincial region

Subprovincial or intraprovincial migration (migration within one same province or territory and between subprovincial regions) is a necessary additional component to estimate migration at the subprovincial level.

8.2.2 Postcensal population estimates for economic regions (ERs)

A different method is used to produce population estimates for economic regions (ERs). In this case the census division's (CD) aggregate method is used. First, the ERs are defined in terms of CDs using the most recent Standard Geographical Classification (SGC) specifications. When the geographic delineation of the CDs and ERs are the same, no adjustment is required; the population estimates for the CDs that make up the ER are simply added together.

However, when the geographic delineation of the CD does not match that of the ER, i.e., when a CD is in more than one ER, distribution of the CD's demographic components are allocated on the basis of its demographic weight in each ER in question. The proportions are referred to as conversion factors. They are calculated using the most recent census counts.

Thus, demographic components (births, deaths and migration) initially measured at the CD level can be allocated to each ER. Using the census division's aggregate method by the ERs' geographic delineation, the population and demographic components of ERs can be estimated.

However, the census division's aggregate method cannot be used to estimate the number of intraprovincial in-migrants and out-migrants, since it overestimates those figures. In-migrants to a given CD from another CD in the same ER should not be counted since the migration occurred within the ER's boundaries. These are false in-migrants. The same is true for out-migrants from one CD to another CD in the same ER: they are false out-migrants. However, the net intraprovincial migration calculated with the CD aggregate method is correct because the false in-migrants and out-migrants cancel each other out. As a result, only the net intraprovincial migration of ERs can be estimated accurately using the CD aggregate method. This is why the estimates for intraprovincial in-migrants and out-migrants are not available at the ER level.

8.2.3 Subprovincial postcensal population estimates by age and sex

The component method is used to produce postcensal population estimates by age and sex for CMAs and CDs. The method is applied to each age-sex cohort in the base population. Chapter 9 describes the application of the cohort component method in detail.

The component method formulas for estimating the population of CMAs and CDs by age and sex are as follows:

For age 0:

Equation 8.2:

${\text{P}}_{\left(\text{t+1}\right)}^{\text{0}}{\text{ = B}}_{\left(\text{t, t+1}\right)}-{\text{ D}}_{\left(\text{t, t+1}\right)}^{\text{-1}}{\text{ + I}}_{\left(\text{t, t+1}\right)}^{\text{-1}}-\left({\text{E}}_{\left(\text{t, t+1}\right)}^{\text{-1}}{\text{ + ΔTE}}_{\left(\text{t, t+1}\right)}^{\text{-1}}\right){\text{ + RE}}_{\left(\text{t, t+1}\right)}^{\text{-1}}{\text{ + NPR}}_{\left(\text{t, t+1}\right)}^{\text{0}}{\text{ + ΔNinter}}_{\left(\text{t, t+1}\right)}^{\text{-1}}{\text{ + ΔNinfra}}_{\left(\text{t, t+1}\right)}^{\text{-1}}$

For ages 1 to 89:

Equation 8.3:

${\text{P}}_{\left(\text{t+1}\right)}^{\text{a+1}}{\text{ = P}}_{\left(\text{t}\right)}^{\text{a}}-{\text{ D}}_{\left(\text{t, t+1}\right)}^{\text{a}}{\text{ + I}}_{\left(\text{t, t+1}\right)}^{\text{a}}-\left({\text{E}}_{\left(\text{t, t+1}\right)}^{\text{a}}{\text{ + ΔTE}}_{\left(\text{t, t+1}\right)}^{\text{a}}\right){\text{ + RE}}_{\left(\text{t, t+1}\right)}^{\text{a}}-{\text{ NPR}}_{\left(\text{t}\right)}^{\text{a}}{\text{ + NPR}}_{\left(\text{t, t+1}\right)}^{\text{a+1}}{\text{ + ΔNinter}}_{\left(\text{t, t+1}\right)}^{\text{a}}{\text{ + ΔNinfra}}_{\left(\text{t, t+1}\right)}^{\text{a}}$

For age group 90 and over:

Equation 8.4:

$\begin{array}{l}{\text{P}}_{\left(\text{t+1}\right)}^{\text{90+}}{\text{ = P}}_{\left(\text{t}\right)}^{\text{89+}}{\text{ - D}}_{\left(\text{t, t+1}\right)}^{\text{89+}}{\text{ + I}}_{\left(\text{t, t+1}\right)}^{\text{89+}}\text{ - }\left({\text{E}}_{\left(\text{t, t+1}\right)}^{\text{89+}}{\text{ + ΔTE}}_{\left(\text{t, t+1}\right)}^{\text{89+}}\right){\text{ + RE}}_{\left(\text{t, t+1}\right)}^{\text{89+}}\\ {\text{ - NPR}}_{\left(\text{t}\right)}^{\text{89+}}{\text{ + NPR}}_{\left(\text{t+1}\right)}^{\text{90+}}{\text{ + ΔNinter}}_{\left(\text{t, t+1}\right)}^{\text{89+}}{\text{ + ΔNinfra}}_{\left(\text{t, t+1}\right)}^{\text{89+}}\end{array}$

where, for each subprovincial region

To ensure concordance between the subprovincial estimates and the provincial and territorial estimates by age and sex, two-way raking is used.

Special treatment for preliminary postcensal estimates for Quebec and British Columbia

A different method is used to calculate preliminary postcensal population estimates for census divisions (CDs) and census metropolitan areas (CMAs) in Quebec. The population estimates by age and sex produced by the Institut de la statistique du Québec (ISQ) are used. These population estimates are based on data from the Fichier d'inscription des personnes assurées (FIPA), the insured persons register, from the Régie de l'assurance-maladie du Québec (RAMQ).

For British Columbia, preliminary postcensal estimates at the CMA and CD levels are calculated by applying the total population growth rates provided by BC Stats, British Columbia's statistical agency, to the previous year's estimates produced by the Demography Division. The total preliminary postcensal estimates are then distributed by age and sex using the Demography Division's component method. The British Columbia population estimates used to calculate the rates are produced using a regression model based on data from residential Hydro services and Ministry of Health Client Registry data as symptomatic indicators.

To ensure concordance between the subprovincial estimates and the provincial totals by age and sex, two-way raking is used.

8.2.4 Estimate levels

For subprovincial regions in Quebec and British Columbia, the specific methods described in the previous section are used only for preliminary postcensal estimates. For updated and final postcensal estimates, the component method is used.

For the subprovincial regions in other provinces and territories, the difference between preliminary and final postcensal population estimates lies in the timeliness of the components. When all the components are preliminary, the population estimate is deemed preliminary postcensal (PP). When the components are all final, the estimate is deemed final postcensal (PD). Any other combination of levels is considered updated postcensal (PR).

8.2.5 Base population and components of population growth

Base population

A full description of the methodology used to calculate the postcensal base population is described in Chapter 2. In the Demographic Estimates Program, the base populations for subprovincial regions are derived from the five-year censuses between 2001 and 2011. Population counts at the provincial, territorial and subprovincial levels are subject to the same adjustment procedures outlined in Chapter 2, unless otherwise noted. To estimate census net undercoverage (CNU) at the subprovincial level, provincial and territorial CNU rates by age and sex are applied to census subdivisions (CSDs), which are aggregated to create the base population of higher subprovincial levels (census metropolitan areas (CMAs) and census divisions (CDs) in the province).

Prior to generating the demographic estimates for each component, two-way raking is used to ensure concordance between the subprovincial and the provincial or territorial totals by age and sex.

Two-way raking involves simultaneously adjusting the differences by assuming consistency between the following:

1. the sum of the population of subprovincial regions by province or territory and the total provincial and territorial population, and
2. the sum of the population of subprovincial regions by age and sex and the provincial and territorial distribution by age and sex.

Since CD and CMA boundaries do not remain stable over time, component data are adjusted to respect the boundaries of the Standard Geographical Classification (SGC), as defined in the most recent census. This ensures a stable geographical universe for the whole reference period.

Births and deaths

The numbers of births and deaths at the census division (CD) and census metropolitan area (CMA) levels are taken directly from the vital statistics database of Statistics Canada's Health Statistics Division. For CMAs, births and deaths have been calculated using this data source since 2007/2008.^{Note 2}

Births and deaths estimates are categorized as final when they are directly taken from the Health Statistics Division's vital statistics. They are then adjusted to the provincial and territorial totals using two-way raking to ensure their concordance.

When no births or deaths data are available, preliminary provincial or territorial estimates are broken down, using the most recent known subprovincial distribution derived from Health Statistics Division's vital statistics, to produce estimates by region. In that case, the births and deaths estimates are categorized as preliminary. They are then adjusted to the provincial and territorial totals using two-way raking to ensure their consistency.

Immigration

Since we do not use subprovincial immigration data from Citizenship and Immigration Canada (CIC), the most recent known subprovincial distribution derived from the T1FF^{Note 3} is used to produce immigrants estimates by subprovincial region. Because the data are available only by broad age groups (0-17, 18-24, 25-44, 45-64, 65+), they are broken down by age and sex based on the distribution from the most recent census or NHS (starting in 2011). The distribution stems from the NHS mobility question on place of residence one year ago. Since 2011/2012, NHS distributions have been modelled to minimize the impact of outliers found in some subprovincial regions, mostly for smaller geographies. To ensure their consistency, subprovincial estimates are then adjusted to the provincial and territorial totals using two-way raking.

The difference between preliminary and final estimates lies in the timeliness of the sources used to estimate this component. Since the subprovincial estimates of immigrants are adjusted to provincial and territorial estimates, the level of subprovincial estimates will be the same. Immigration estimates are preliminary the first year and final the following year.

Net non-permanent residents (NPRs)

At the subprovincial level, there are no reliable administrative data sources available to directly estimate net NPRs. To compensate for this lack of data, the provincial and territorial NPR estimates by age and sex are broken down by subprovincial region based on the distribution from the most recent census or NHS (starting in 2011). Since 2011/2012, NHS distributions have been modelled to minimize the impact of outliers found in some subprovincial regions, mostly for smaller geographies. To ensure their consistency, subprovincial estimates are then adjusted to the provincial and territorial totals using two-way raking.

For the 2005/2006 and 2010/2011 years, the net NPRs are calculated using two different distributions—the 2001 and 2006 censuses for the year 2005/2006, and the 2006 Census as well as the 2011 NHS for the year 2010/2011. This approach assumes that the two distributions are similar. If the two distributions vary by the regional breakdown of NPRs, the net NPRs for 2005/2006 and 2010/2011 will absorb all the changes attributable to the difference between the two distributions that were used. For this reason, the net NPRs for 2005/2006 and 2010/2011 should not be compared with the rest of the historical series.

Since the subprovincial estimates of the net number of NPRs are adjusted to provincial and territorial estimates, the level of the subprovincial estimates will be the same. NPR estimates are preliminary the first year and updated the following year. They become final two to three years after the reference year, when all other components are also final.

Emigration

As with immigrants, the number of emigrants at the subprovincial level is derived from the T1FF. Because the estimates are available only by broad age groups (0-17, 18-24, 25-44, 45-64, 65+), they are broken down by age and sex based on the provincial or territorial distribution. They are then adjusted to the provincial and territorial totals using two-way raking to ensure their consistency.

The difference between preliminary and final estimates lies in the timeliness of the sources used to estimate this component. Since the subprovincial estimates of emigrants are adjusted to provincial and territorial estimates, the level of the subprovincial estimates will be the same.

Net temporary emigration

At the subprovincial level, provincial and territorial net temporary emigration estimates by age and sex are broken down based on the subprovincial distribution of emigrants. They are then adjusted to the provincial and territorial totals using two-way raking to ensure their consistency.

The difference between preliminary and final estimates lies in the timeliness of the net temporary emigration estimates.

Returning emigrants

As with immigrants and emigrants, the number of returning emigrants at the subprovincial level is derived from the T1FF. Because the estimates are available only by broad age groups (0-17, 18-24, 25-44, 45-64, 65+), they are broken down by age and sex based on the provincial or territorial distribution. They are then adjusted to the provincial and territorial totals using two-way raking to ensure their consistency.

The difference between preliminary and final estimates lies in the timeliness of the sources used to estimate this component. Since the subprovincial estimates of returning emigrants are adjusted to provincial and territorial estimates, the level of the subprovincial estimates will be the same.

Interprovincial migration

Interprovincial migration by broad age groups and sex for subprovincial regions is derived from the T1FF for each subprovincial region. The estimates by broad age groups and sex are broken down by age based on distributions stemming from the most recent census or NHS (starting in 2011) mobility question on place of residence one year ago. Since 2011/2012, NHS distributions have been modelled to minimize the impact of outliers found in some subprovincial regions, mostly for smaller geographies. Subprovincial estimates are then adjusted to the provincial and territorial totals using two-way raking to ensure their consistency.

The difference between preliminary and final estimates lies in the timeliness of the sources used to estimate this component. Since the subprovincial estimates of interprovincial migrants are adjusted to provincial and territorial estimates, the level of the subprovincial estimates will be the same.

Intraprovincial migration

As with interprovincial migration, the components of intraprovincial migration by broad age groups and sex are derived from the T1FF for each subprovincial region. The estimates by broad age groups and sex are broken down by age based on distributions stemming from the most recent census or NHS (starting in 2011) mobility question on place of residence one year ago. Since 2011/2012, NHS distributions have been modelled to minimize the impact of outliers found in some subprovincial regions, mostly for smaller geographies.

These sources are used for both preliminary and final estimates.

The difference between preliminary and final estimates lies in the availability of the T1FF data used to estimate this component.

Since there are no reliable data sources for preliminary intraprovincial migration estimates, the data for the most recent year, for which final estimates are available, are used. The assumption that intraprovincial migratory behaviours for the current year are similar to those for the previous year for which final estimates are available is adopted.

8.3 Intercensal population estimates for subprovincial regions

Intercensal estimates for census divisions (CDs), census metropolitan areas (CMAs) and economic regions (ERs) are produced much in the same manner as intercensal estimates at the provincial and territorial level (see Chapter 1 for information on the methods). There are three main steps in the production of intercensal estimates:

• the correspondence of the geographic boundaries between the two censuses
• calculation of the error of closure
• linear distribution of the error of closure (residual deviation).

To ensure geographical concordance, the base populations and components of population growth must be adjusted according to geographical boundaries at the time of the most recent census. For areas whose geographical boundaries changed between the two censuses (as measured by the SGC), historical conversion factors are used based on population transfers at the census subdivision level during the most recent intercensal period. In general, corrections to CDs, CMAs and ERs are minor.

Error of closure is defined as the difference between postcensal population estimates on census day and the population enumerated in that census adjusted for census net undercoverage (CNU). The error of closure is spread evenly over the intercensal period, based on the number of days in each month. Intercensal estimates by age and sex are adjusted the same way (i.e., by distributing the error of closure evenly across the age and sex cohorts). As with postcensal estimates, the intercensal subprovincial estimates by age and sex are adjusted to provincial and territorial estimates using two-way raking to ensure their consistency.

8.4 Population estimates for census subdivisions

Census subdivision (CSD) is the general term for municipalities (as determined by provincial and territorial legislation) or areas treated as municipal equivalents for statistical purposes (e.g., Indian reserves, Indian settlements and unorganized territories). They are classified into 54 types according to official designations adopted by provincial, territorial or federal authorities. Two exceptions are subdivision of unorganized (SNO) in Newfoundland and Labrador, and subdivision of county municipality (SC) in Nova Scotia, which are geographic areas created as equivalents for municipalities by Statistics Canada, in cooperation with those provinces, for the purpose of disseminating statistical data.

Because there are no components of sufficient quality at that level, the population of CSDs is not estimated using the cohort component method, as is done with census divisions (CDs), census metropolitan areas (CMAs), provinces and territories. The method used consists of applying the CD growth rate to the base population of its CSDs. Two sets of data are necessary in this method: the base population of CSDs and the annual population estimates for CDs.

8.4.1 Base population of census subdivisions

A base population is the population at the beginning of a period, used as a starting point for the estimation process. At the CSD level, the base population is the population count by age and sex for five-year censuses, adjusted for coverage errors.^{Note 4} The census data are adjusted as follows:

• adjustment to take population reviews into account. Because there are no population reviews by age and sex, calibration by age and sex is carried out to ensure consistency with the total counts adjusted for population reviews.
• adjustment of the population for census net undercoverage (CNU). Given that coverage studies are not designed to produce subprovincial-level estimates of CNU, provincial and territorial rates by age and sex are used.
• addition of independent population estimates for incompletely enumerated Indian reserves by age and sex.

8.4.2 Postcensal population estimates for census subdivisions

CSD populations are estimated in four steps: (1) CD-level population estimates by age and sex are first calculated; (2) CD-level growth rates are calculated by age and sex; (3) these growth rates by age and sex are applied to the corresponding CSD population estimates; (4) CSD-level estimates are adjusted to ensure consistency with the CD-level estimates by age and sex.

Step 1 – Estimating the CD population

The first step consists of estimating the CD-level populations by age and sex using the component method described previously in this chapter.

Step 2 – Calculating CD growth rates by age and sex

The second step involves calculating population growth rates by age and sex for each CD. The formula is as follows:

Equation 8.5:

$\text{GR\_}{}_{\text{j}}\text{C}{\text{D}}_{\left(\text{t, t+1}\right)}^{\text{a}}\text{ = }\frac{\text{Pop\_}{}_{\text{j}}\text{C}{\text{D}}_{\left(\text{t+1}\right)}^{\text{a}}-\text{ Pop\_}{}_{\text{j}}\text{C}{\text{D}}_{\left(\text{t}\right)}^{\text{a}}}{\text{Pop\_}{}_{\text{j}}\text{C}{\text{D}}_{\left(\text{t}\right)}^{\text{a}}}$

where

Step 3 – Estimating postcensal population estimates of CSDs

The third step consists of applying the CD-level growth rates by age and sex to the corresponding CSD base populations. The formula is as follows:

Equation 8.6:

$\text{Pop\_}{}_{{}_{{}_{\text{j}}\text{C}\text{D}}\text{i}}\text{C}{\text{SD}}_{\left(\text{t+1}\right)}^{\text{a}}\text{ = Pop\_}{}_{{}_{{}_{\text{j}}\text{C}\text{D}}\text{i}}\text{C}{\text{SD}}_{\left(\text{t}\right)}^{\text{a}}\text{ + }\left(\text{Pop\_}{}_{{}_{{}_{\text{j}}\text{C}\text{D}}\text{i}}\text{C}{\text{SD}}_{\left(\text{t}\right)}^{\text{a}}\text{ × GR\_}{}_{\text{j}}\text{C}{\text{D}}_{\left(\text{t, t+1}\right)}^{\text{a}}\right)$

where

Step 4 – Calibrating the CSD-level population estimates to ensure consistency with the CD population estimates

Finally, to ensure consistency between the CSD and CD population estimates by age and sex, the CSD population estimates are adjusted by age and sex using two-way raking.

Notes