Section 1: Method and assumptions

View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Archived

This page has been archived on the Web.

Method

The population projections for Canada, provinces and territories are produced by means of the components method, which has been used for a number of decades both by Statistics Canada and by other provincial, territorial, national and international statistics agencies to project particular characteristics of a population—primarily age, sex and place of residence. Several assumptions are formulated for each component of growth—fertility, mortality, immigration and emigration. Once applied in the form of numbers, rates and quotients to the base population, these assumptions generate births, deaths, immigrants and emigrants. The population of the following year is then obtained by simple demographic accounting, with births and immigrants being added to the base population while deaths and emigrants are subtracted.

The projection model used by Statistics Canada takes account of the interaction between demographic phenomena. For this purpose, the model first constructs a transition matrix that combines the various rates associated with events (births, deaths, immigration, emigration, internal migration). This transition matrix is then applied to the population at the start of the period to obtain the population at the end of the period. This stage also serves to calculate the mid-year population for each year, a figure that is needed in order to calculate events observed during a given year.

Lastly, the population of each province and territory is projected individually, so as to take regional distinctiveness into account. These projected populations can then be summed to obtain a consistent projection of the population of Canada as a whole.

Base population

The base population for these projections is obtained from the preliminary postcensal estimates of the population on July 1^st, 2009 as calculated by Statistics Canada's Demography Division. These estimates by age (from 0 to 99 years and 100 years or over), sex and province/territory of residence are based on the 2006 Census and are adjusted to take account of net undercoverage by age, sex and province/territory. On July 1^st, 2009, the Canadian population was estimated at 33,739,859. Table 1.1 shows the population figures for the various provinces and territories. These estimates are also available at no charge on the Statistics Canada website.

Fertility

Because fertility has a major effect on the size and the age structure of a population, three assumptions are usually formulated for this component. This edition of the population projections is no exception, and these assumptions are based on a careful examination of historical and recent trends at the provincial, territorial, national or international level. The assumptions cover both the future level of fertility and its tempo (mean age at childbearing).

Assumptions

At the national level, the medium assumption used for these projections is a total fertility rate of 1.70 children per woman, which is attained in the first year projected and is held constant over the entire projection period. This is an upward revision of the most recent population projections published in 2005, which suggested a level of 1.51 children per woman.

According to this medium assumption, the mean age at childbearing would continue to rise, going from 29.75 years in 2007 to 30.17 years in 2009, an increase of about 0.4 years. It would then be held constant over the entire projection period.

The low and high assumptions were set at 1.50 and 1.90 children per woman at the national level. These levels are gradually reached by 2013/2014 and are then held constant until the end of the projection period (Chart 1.1). The mean age at childbearing also increases according to these two assumptions, by 0.6 years and 0.2 years according to the low and high assumptions respectively.

Table 1.2 shows, for each province/territory, observed total fertility rates and those projected based on three assumptions derived from those developed at the national level. To build these assumptions, factors of variation were calculated at the national level and then applied to fertility rates by age for each province and territory in order to reach target total fertility rates and mean age at childbearing. In general, mirroring the situation that existed in 2007, Newfoundland and Labrador would have the lowest fertility under all three assumptions, and Nunavut would have the highest.

Rationale for assumptions

Several factors were considered when formulating the fertility assumptions for Canada. First, Canadian fertility has exhibited an upward movement since 2004, with the total fertility rate (TFR) going from 1.53 children per woman to 1.66 in 2007, the last available year for which these projections were produced. The level reached in 2007 was the highest since 1995, indicating that this was indeed a significant increase, observed over a period of just three years. Furthermore, birth data for 2008 released by some Canadian provinces show a continuation of the upward trend, suggesting that the TFR for Canada in 2008 will be even higher than in 2007. For example, in 2008, the TFR already exceeded 1.70 children per woman in Quebec (1.74) and 1.50 in British Columbia, the highest rates since 1976 and 1997 respectively. It was therefore appropriate to take these recent trends into account and to readjust the medium fertility assumption upward in comparison to the previous projection exercise.

The level of the completed fertility rate of Canadian women born between 1960 and 1975 was also considered in developing the medium assumption. The completed fertility rate of a cohort of women, such as those born in 1960, is the average number of children that those women will have actually had during their reproductive life. This indicator is useful for developing fertility assumptions, since it is free from short-term or cyclical variations in fertility. While the TFR in Canada has fallen to levels close to 1.50 children per woman, no cohort of Canadian women thus far has reached a completed fertility rate of less than 1.76 children per woman, suggesting that a medium assumption close to this level is reasonable for mid to long-term projections over a 25-year period.

Finally, the medium assumption was established in light of the recent evolution of fertility in numerous industrialized countries. As in Canada, fertility in many countries has also risen in recent years, suggesting that the increase in Canada is not an isolated phenomenon. For example, fertility in the United Kingdom went from approximately 1.65 children per woman in 2001 to 1.90 children per woman in 2007 (Chart 1.2). It is interesting to note that Canada falls between two quite distinct groups. A first group, consisting of the United States, Australia and the United Kingdom (all Anglo-Saxon countries like Canada), France and Sweden, usually exhibits higher fertility. By contrast, Italy, Japan, Spain, Germany and Switzerland usually exhibit lower fertility. In light of these considerations, the median assumption was developed to preserve Canada's intermediate position while taking account of the upward movement observed in almost all these countries.

The medium assumption also has the mean age at childbearing rise quickly from 29.75 years to 30.17 years. This is an increase of approximately 0.4 years, which extends a long upward trend seen in recent decades. Indeed, in 2005 the fertility rate of women aged 30 to 34 years exceeded that of women aged 25 to 29 years, showing that Canadian women are having their children at increasingly older ages. This assumption seems reasonable in light of the uninterrupted rise thus far in female labour force participation rates and the even greater upward shift in the age pattern of childbearing in some countries such as Australia.

The low assumption was set at 1.50 children per woman in view of how Canadian fertility has evolved in recent decades. This level roughly corresponds to the lowest level ever recorded in Canada (1.51 children per woman in 2000 and 2002). It also corresponds to the lowest fertility levels recorded in 2007 in Canadian provinces or territories (Newfoundland and Labrador, Nova Scotia, New Brunswick and British Columbia). This level is also slightly higher than the fertility recently recorded within the group of "low fertility" industrialized countries consisting of Germany, Japan, Italy, Switzerland and Spain. A final point is that some phenomena historically associated with low fertility are still on the rise: common-law unions, the participation of women in the labour force and their average level of schooling. Possibly these phenomena will continue to exert downward pressure on Canadian fertility in the future.

The low assumption is associated with a greater upward shift in the age pattern of fertility, with the mean age at childbearing rising 0.6 years to 30.33 years. This level has already been recorded in some industrialized countries and therefore appears reasonable in light of past trends for this indicator.

The high assumption was set at 1.90 children per woman in light of the international context: this level is approximately the average level of fertility recently recorded in "high fertility" industrialized countries such as the United States, France, Australia, the United Kingdom and Sweden. This level is also slightly lower than the average fertility recorded in 2007 for the provinces where the highest fertility levels were recorded (Manitoba, Saskatchewan and Alberta). Lastly, it should be noted that this level has not been observed in Canada since 1972, and it is still below the replacement level (approximately 2.10 children per woman).

According to the high assumption, the mean age at childbearing would rise from 29.75 years to 29.96 years. The recent example of a number of countries, especially Australia, shows that a slight upward shift in the age pattern of fertility is not incompatible with a significant rise in the fertility level. Since the mean age at childbearing has not declined in Canada since 1975, the high assumption adopted provides for that age to rise in the coming years, although the rise is more modest than under the low and medium assumptions.

Mortality

As with fertility, three assumptions were formulated regarding future Canadian mortality, a demographic component that can have a major effect on the size of Canada's old and very old populations. In the context of demographic aging, these groups are of particular interest.

The three assumptions formulated suggest a continuation of the secular decline of Canadian mortality, and hence a life expectancy at birth that would continue to rise in the coming years. The trends since 1921 are clear; they show that mortality has greatly declined over the past 90 years in Canada. Between 1921 and 2006, life expectancy grew by nearly 25 years for females and 22 years for males. However, the rate of increase in life expectancy has varied from one period to another; overall, the increase has been less rapid in recent decades than earlier in the twentieth century, especially for Canadian women.

The method used to project mortality is the same as was used in the previous edition of the population projections. It draws on a parametric model developed by Li and Lee (2005). That model is used to project mortality rates by age and sex in a consistent manner for all provinces, which largely reflects the fact that Canadian standards ensure a high degree of similarity in the health insurance plans of the provinces and territories.

The parameters of the Li and Lee model are derived from the change in mortality rates recorded over a given period, in this case the 1981 to 2006 interval. This period was chosen not only because it captures the recent pace of the decline in Canadian mortality, but also because it has the advantage of reflecting the narrowing of the gap in life expectancy between males and females.

The mortality of the territories was projected using the method based on recent differences in relation to the national average, since the populations of these regions are too small to obtain reliable parameters for the Li-Lee model. Unlike in the previous edition, each territory was treated individually this time.

The low and high assumptions were based on the 1% confidence intervals of the ARIMA model used to project a parameter of the equation used in the Li-Lee method. The use of these intervals yields a reasonable range for the future course of life expectancy for the provinces and territories.

Assumptions

Under the medium assumption, Canadian females' life expectancy at birth would increase from 82.9 years in 2006 to 87.3 years in 2036, a gain of 4.4 years in 30 years (Table 1.3). By comparison, the gain was 5.1 years between 1976 and 2006. For males, life expectancy would rise from 78.2 years in 2006 to 84.0 years in 2036, an increase of 5.8 years. During the period 1976 to 2006, the gain was larger, namely 7.8 years. The more rapid increase in male life expectancy means that the mortality gap between the two sexes would continue to narrow in the coming years, going from 4.7 years in 2006 to 3.3 years in 2036.

From one province to another, differences in life expectancy would be fairly small in 2036, mirroring the situation in 2006 and reflecting the relative homogeneity of mortality in Canada. Among the provinces, Newfoundland and Labrador would have the lowest life expectancies in 2036, with 81.7 years for males and 85.5 years for females, and British Columbia would have the highest, at 84.5 years for males and 87.8 years for females.

The low assumption would have life expectancy rise only 3.1 years for females and 4.1 years for males by 2036; at that point, life expectancies would be 86.0 and 82.3 years respectively. On the other hand, the high assumption would have life expectancy rise more rapidly, by 5.5 years for females and 7.2 years for males, comparable to the rate recorded during the last 25 years. Under this assumption, life expectancy in 2036 would reach 88.4 and 85.4 years for females and males respectively.

All the assumptions suggest that the existing gap between male and female life expectancies will continue to narrow.

Rationale for assumptions

The future course of mortality has been much debated within the scientific community. Some people may believe that the extent to which life expectancy can continue to increase is unknown since mortality is already quite low at numerous ages. If knowledge, medicine and technology continue to advance, it is likely that future gains will have to be made against chronic illnesses such as cancer, which has been the leading cause of death in Canada for several years, or against diseases that mainly affect the very old population. Others think that the growing prevalence of various conditions among the young, such as obesity and the diseases associated with it, including diabetes and hypertension, could considerably slow the increase in life expectancy in the coming decades. Some also think that life expectancy at the population level is already approaching a limit, given existing health care systems. In this context, future gains would require major public investment, which cannot be taken for granted. In contrast, many scientists remain optimistic, stressing that life expectancy continues to increase despite the high level that it has already reached in some countries. Since the limits of human life are well beyond 100 years, these researchers stress that there is still considerable room for the life expectancy of a population to increase.

Whatever the case, Canadians, both male and female, have one of the world's highest life expectancies at birth. A comparison of life expectancies recorded during the 2007 to 2009 period shows that Canadian females were behind only French, Spanish, Swiss, Australian and Japanese females, with the latter at the forefront with a life expectancy of 86 years. Canadian males, for their part, were behind only Japanese, Swedish, Swiss, Australian and Icelandic males, with the latter leading with a life expectancy of 80 years. The assumptions developed take this situation into account; they propose life expectancies in 2036 that have never been observed to date, which would help maintain Canada among the countries with the highest life expectancies for both males and females.

International immigration

For some years now, Canadian population growth has been based more on migratory increase (i.e., excess of immigrants over emigrants) than on natural increase (excess of births over deaths). Considering that this situation is expected to continue over the coming decades because of low fertility and an aging population, the assumptions concerning this component take on great importance in this projection exercise.

Both the number of immigrants admitted annually to Canada and the immigration rate have sometimes varied considerably from year to year. For example, the immigration rate went from nearly 9.0 per thousand in 1993 to less than 6.0 per thousand in 1998, five years later. Even larger variations had been recorded in earlier decades. To take account of the sometimes fluctuating nature of immigration, three assumptions on future levels of immigration to Canada were formulated, as is usually the case in Statistics Canada projections.

Assumptions

The immigration assumptions were formulated for two periods. Firstly, the low, medium and high assumptions were, for the first three years of projections, directly drawn from the 2009 Immigration Plan as formulated by Citizenship and Immigration Canada (CIC). This range of immigrants, set out by CIC in its immigration plan for 2009, is considered the best estimate of the level of Canadian immigration in the short term. The Immigration Plan called for admitting between 240,000 (low assumption) and 265,000 (high assumption) immigrants to Canada in 2010. The medium assumption was set at the middle value of the range, namely 252,500 immigrants annually.

After the first three years of projection, the assumptions are based on immigration rates rather than numbers. Thus, the numbers of immigrants admitted to Canada evolve along with the projected population growth. The use of rates serves to prevent these rates from mechanically decreasing with every future population increase, an effect generated systematically when the assumptions are based on immigrant numbers held constant over the projection period. Finally, unlike in the previously published projections, where the numbers of immigrants reached in 2031 were held constant until 2056, the present assumptions use immigration rates until the end of the projection period for Canada, in 2061.

Under the medium assumption, the immigration rate would be 7.5 per thousand in Canada during the entire projection period, except, of course, for the first three years when the figures used are drawn from the Immigration Plan (Chart 1.3). That level is the average annual immigration rate for the period from 1991 to 2008.

The low assumption instead suggests that the immigration rate would be 6.0 per thousand between 2012/2013 and 2060/2061 in Canada. This level roughly corresponds to the lowest immigration rate recorded since 1991. Finally, the high assumption sets the immigration rate at 9.0 per thousand over the entire projection period, a level that instead corresponds to the highest rate recorded since 1991.

The provincial distribution of Canadian immigration has changed in recent years. For example, a larger proportion of immigrants are settling in Manitoba and Alberta. Conversely, the proportion of immigrants who choose Ontario has been declining for several years, going from 60% to approximately 50%. To take these important changes into account, the distribution of immigrants admitted to Canada in the various provinces and territories was based on the average provincial distribution of immigrants during the period 2005/2006 to 2007/2008. This distribution is held constant over the entire projection period.

The same period was also used to distribute immigrants by age and sex. In regions that receive large contingents of immigrants, namely in Ontario, Quebec, British Columbia and Alberta, the distribution recorded for the province was used. In the case of Saskatchewan and Manitoba, the average distribution for these two provinces was used. Finally, the average national age and sex distribution was used in all other regions.

Table 1.4 shows the actual (2008/2009) and projected (2035/2036) immigrant numbers for Canada, the provinces and the territories according to the three assumptions used. Under the medium assumption, Canada would receive nearly 334,000 immigrants in 2035/2036, nearly half of whom (166,300) would settle in Ontario.

Rationale for assumptions

Immigration is a difficult component to project in the long run. In the short run, immigration levels are determined by the department responsible. Therefore, in addition to looking at recent immigration levels, it is appropriate to take other factors into consideration when formulating assumptions.

The high assumption sets the immigration rate at 9.0 per thousand, a rate that could reflect an increase aiming to meet Canada's future labour force needs. It is to be expected that many workers—most belonging to the baby boom generation, born between 1946 and 1965—will leave the labour force in the coming years to join the ranks of the elderly. Immigration could serve to bolster the labour force, which will need among other things to meet the needs of a rapidly growing elderly population.

The low assumption, which sets the immigration rate at 6.0 per thousand, is also plausible considering several factors. First, the economic integration of immigrants has seemed in recent years to be more difficult than it was previously. Second, the rapid economic development of some countries that sources of many recent immigrants to Canada—China and India in particular—could reduce the pool of potential candidates for immigration. Increased competition from other industrialized countries for skilled immigrants could also affect Canada's ability to recruit the immigrants that it needs.

Emigration

Statistics Canada's population estimates and projections programs identify three components of emigration: emigrants, returning emigrants and the net number of persons temporarily abroad.

Emigrants

Emigrants are Canadian citizens or landed immigrants who have left Canada to settle permanently in another country. The number of emigrants has shown little change, averaging around 45,700 persons since the early 1990s. In light of this, a single assumption was formulated, taking the average emigration rates by age, sex and province/territory recorded during the period 1991/1992 to 2007/2008 and holding them constant throughout the entire projection period. Thus formulated, the assumption has the number of emigrants increasing from 48,800 in 2009/2010 to roughly 59,200 in 2035/2036 according to the medium-growth scenario (M1).

Returning emigrants

Returning emigrants are Canadian citizens or landed immigrants who emigrated from Canada and returned to settle there. Their number—a modest one—is estimated at approximately 23,500 annually since the start of the 2000s. The vast majority of returning emigrants are concentrated in four provinces: Ontario, Quebec, Alberta and British Columbia.

As in the preceding edition of the projections, the assumption formulated for returning emigrants is based on the relationship that exists between them and emigration, since the persons likely to return to Canada to settle are those who emigrated at some point in their lives. The projected number of returning emigrants is obtained by applying a returning migration rate to the number of emigrants generated by the projection model. This rate is set at 44.5%, the average for the period 1991/1992 to 2007/2008, the same period as for the assumption on emigration.

Persons temporarily abroad

Persons temporarily abroad are Canadian citizens or landed immigrants who are living abroad temporarily and no longer have a usual place of residence in Canada. The data available on this component yield an annual balance, rather than a number. A balance is the result of two flows, in this case consisting of persons leaving Canada (temporarily) and those returning.

Since the early 2000s, the annual net number of persons temporarily abroad has remained stable, averaging slightly more than 20,000, and is mostly attributable to four provinces: Ontario, Quebec, Alberta and British Columbia.

The assumption adopted in the present projections consists in taking the net annual average figure recorded during the period 2005/2006 to 2007/2008, namely 21,180 persons, and holding it constant until 2061. This level is very close to the average net figure for the period 1991/1992 to 2007/2008, reflecting the fact that this component has undergone few sizable variations over the past two decades.

Once again, the projections use a distribution by age, sex and province/territory derived from the average recorded for the period 2005/2006 to 2007/2008.

Non-permanent residents

Non-permanent residents include the following persons: persons residing in Canada who claim refugee status, persons who hold a work permit, a ministerial permit or a student visa, as well as members of their family living with them in Canada. The net annual number of non-permanent residents increased considerably in recent years, going from roughly 9,000 persons in 2004/2005 to more than 70,000 in 2008/2009. The increase was especially sizable for Alberta, British Columbia and Saskatchewan. Also, a new CIC class of immigrants, the Canadian Experience Class (CEC), allows non-permanent residents of Canada who meet certain conditions to file an application for permanent residence. This class could make the status of non-permanent residents in Canada more attractive and thus lead, in the coming years, to the maintenance of a high net annual figure. This is of particular interest in a context where many Canadians will soon be leaving the labour force.

The assumption adopted for non-permanent residents was therefore changed from the the previously published projections to reflect the recent changes. The assumption adopted suggests that the net annual figure would show a linear decline from approximately 51,000 in 2007/2008 to 0 in 2018/2019, during a period when departures from the labour force will quite probably be numerous. After 2018, the net number of non-permanent residents would be maintained at zero for the rest of the projection period. Thus, under this assumption, the stock of non-permanent residents in the Canadian population would rise from 500,000 in 2008 to just over 750,000 in 2018.

Again to reflect the recent changes, the distribution by age, sex and province/territory corresponds to the distribution recorded during the period 2005/2006 to 2007/2008. This distribution would be held constant over the entire projection period.

Interprovincial migration

Of all the components of population growth, interprovincial migration assumptions often generate the largest variations as to the projected size and growth of the population of the different provinces and territories. The strong volatility of interprovincial migration makes projection difficult; from year to year, variations in interprovincial net migration can be quite sizable and can, for example, shift from positive to negative. Also, because it is often associated with structural or economic factors (Coulombe, 2006; Bernard et al., 2009), interprovincial migration adds considerable uncertainty to projections at the provincial/territorial level.

For all these reasons, and in order to propose a reasonable range of possible future situations, four separate assumptions are provided for this component. Also, in developing these assumptions, a major constraint must be taken into account: the net interprovincial migration figures for all the provinces and territories must sum to zero.

Assumptions

The method used to project interprovincial migration is the same as in the preceding edition of the projections: first, out-migration rates by age and sex for each province/territory are calculated on the basis of reference periods selected by means of an analysis of past trends. Second, origin-destination matrixes are used to allocate interprovincial out-migrants among Canada's other provinces and territories. These origin-destination matrixes are calculated using the same reference periods as for the out-migration rates in order to keep the model consistent.

The four assumptions, each reflecting a past reference period, are as follows: "historical" trends (1981 to 2008), 1988 to 1996 trends, 2001 to 2006 trends and 2006 to 2008 trends.

1. Historical trends assumption (1981 to 2008)

The purpose of this assumption is to reflect broad long-term trends in internal migration in Canada. Accordingly, the reference period for this assumption is the longest of the four assumptions developed for these projections, and it was set to extend from 1981 to 2008. In particular, the analysis covering such a period identifies movements that are usually favourable to Western Canada, especially Alberta and British Columbia, and to Ontario, and less favourable to Eastern Canada, including Newfoundland and Labrador, Nova Scotia, New Brunswick and Quebec. This scenario generates the most favourable net interprovincial migration figures of any of the four assumptions for Ontario and Manitoba and the least favourable for Newfoundland and Labrador, the Northwest Territories and Yukon (Table 1.5).

1. 1988 to 1996 trends assumption

This assumption is based on the trends recorded during the period 1988 to 1996, which were favourable to British Columbia, with that province registering an average annual net interprovincial migration of almost 34,000 persons. This is the highest figure in all four assumptions for that province. Because of the continuing expansion of the emerging economies in Asia and the reduced presence of the manufacturing sector—then in crisis—in the province's economy, British Columbia has, in relative terms, been spared the economic difficulties experienced elsewhere in Canada. This assumption also generates the most favourable net migration for Prince Edward Island, Nova Scotia, New Brunswick, the Northwest Territories and Nunavut.

Conversely, this period generates the most unfavourable net interprovincial migration of all four assumptions for the three Prairie provinces: Manitoba, Saskatchewan and Alberta.

1. 2001 to 2006 trends assumption

The expansion of Alberta's economy, in particular the petroleum industry, had a not inconsiderable effect on internal migration movements in Canada. Indeed, Alberta has succeeded in attracting large numbers of workers for the past fifteen years, and the 2001 to 2006 assumption is designed to reflect this phenomenon. Obviously, this is the assumption that generates the most favourable figures of all four hypotheses for Alberta, but also for Quebec, which experienced strong economic growth during these years. Conversely, this scenario is the least favourable to British Columbia, whose average annual net interprovincial migration was barely more than 3,000.

1. 2006 to 2008 trends assumption

Recently there have been a number of major changes in interprovincial migration. After 22 straight years of negative annual net migration for Saskatchewan, the trend reversed in 2006/2007, and that province has since had positive and steadily rising net migration. Thus in 2007/2008, Saskatchewan gained more than 4,000 persons in its exchanges with Canada's other provinces and territories.

The situation both in Newfoundland and Labrador and in Manitoba has also considerably changed quite recently. In Newfoundland and Labrador, for example, annual net interprovincial migration went from an average of -4,000 during the period 2004/2005 to 2006/2007 to -528 in 2007/2008. Net migration even became positive in 2008/2009 (preliminary data) after being negative without interruption for 24 years. Finally, Ontario's net migration has greatly deteriorated, plunging to levels not seen since the late 1990s, at nearly -17,000 persons per year on average during the period 2006/2007 to 2007/2008.

These changes are probably related in part to Canada's recent economic situation. Saskatchewan has seen major development in the mining and petroleum sectors that has stimulated its economic growth. Major investments in Newfoundland and Labrador have created a demand for skilled labour. Lastly, problems in the automotive sector in Ontario have contributed to the deterioration of its net interprovincial migration.

The 2006 to 2008 assumption was developed to reflect these major changes. It generates the most favourable, or the least unfavourable, net interprovincial migration figures for Saskatchewan, Newfoundland and Labrador and Yukon and the most negative figures for Ontario, Quebec, New Brunswick, Nova Scotia, Prince Edward Island and Nunavut.

Projected figures

Charts 1.4 to 1.16 show the observed and projected net interprovincial migration figures for each province and territory during the period from 1981/1982 to 2035/2036 according to the different scenarios proposed. ¹  Among other things, these charts show some effects of applying constant out-migration rates to populations whose size and structure may vary over time. The use of rates of out-migration, compared with the use of fixed numbers, has a number of advantages, one being that they take account of both the size of populations at risk of migrating and their evolving age structure. However, because the population at risk of leaving a region does not necessarily evolve at the same rate as populations at risk of coming to that region, there can be a change over time in the net migration rate.

This phenomenon may be explained as follows: if the population of a region grows less rapidly than the populations of other regions, that region will see its net migration improve, since departures will decline and arrivals will increase. The opposite is also true. Thus, provinces with lower growth rates, such as the Atlantic provinces, will see their net interprovincial migration improve as the projection advances in time; conversely, provinces with high growth rates, notably Ontario, will see their net migration deteriorate over time. This effect is due to the fact that the assumptions are based on constant out-migration rates and on invariable origin-destination matrixes. However, the effect is moderate in most regions.

Next