An important part of every projection exercise is the development of plausible assumptions for each of the components of population growth. We based the assumptions on the analysis of past evolution in component trends specific to the Aboriginal populations, supplemented by the most recent work done to project the components of growth for the general Canadian population ¹. For our projections, we elaborated five assumptions of future fertility, one on mortality and two for migration. When combined, we would have ten feasible scenarios of growth for the Aboriginal populations. We selected five scenarios for discussion of the possible future evolution of the size and composition of the Aboriginal populations.

The following sections describe the assumptions for each of the components of growth and the five selected scenarios.

Assumptions

Geographic location - reserves/First Nations communities

In addition to the modifications of the census data done to create the starting point for our projections, which are described in Part 1 of this report, there was one more change related to classification of the North American Indians by place of residence. This change resulted in an expansion of the definition of “reserve” used in this project.

In cooperation with Indian and Northern Affairs Canada, the census geography designates specific units corresponding to census subdivisions as "reserves". There were 1,098 such units in the 2001 Census. For our projections, we added to this number the 41 subdivisions, which are Northern collectives chosen by INAC as quasi-reserves (First Nations communities) because they are associated with a First Nation or an Indian band. The expansion affected the size of the reserve population in Saskatchewan, the Yukon Territory and especially in the Northwest Territories, where 25 of the First Nations communities are found. For the Northwest Territories' Aboriginal population, the consideration of 25 collectives as reserves resulted in over 70% of the 2001 population living on reserves, up from only 2% if quasi-reserves were not included.

The size of the base populations for the three Aboriginal groups by place of residence is given in text table 2.1 .

Text table 2.1
Adjusted Aboriginal population by group and place of residence, Canada, 2001

Fertility assumptions

Over the last 50 years, total fertility for the Aboriginal people has experienced significant change. The trend shows a decline but the level and tempo of this decline varied over time. Prior to the 1970s, the decline was very slow but over the past three decades Aboriginal fertility has dropped by half, from over five children per woman at the end of the 1960s, to below three children near the end of the 1990s. Consistent with the overall Canadian tempo, the fertility level of the Aboriginal Canadian population slowed during the 1980s and early 1990s, but has decreased even more in recent years.

The fertility pattern of the Aboriginal population is not homogeneous. The Inuit, for example, had the highest fertility over the years. More recently, however, the Inuit have experienced a rather sharp decline in their fertility level, from over four children per woman from 1986 to 1991, to slightly above three children from 1996 to 2001. In contrast, the Métis, who are predominantly urban-based, had the lowest fertility level among the Aboriginal peoples in recent years. By the 1996 to 2001 period, the fertility level of this group dropped close to the replacement level of just above two children per woman. The fertility level for the North American Indians fell between that of the Inuit and the Métis, and has remained slightly lower than three children per woman in the past two decades.

Aboriginal fertility varied enormously across provinces and territories. In the1996 to 2001 period, the highest level was observed in Nunavut (close to four children per woman), followed by that in Saskatchewan, the Yukon Territory, the Northwest Territories and Manitoba (slightly more than three children per woman). The lowest levels, just above two children per woman, were characteristic of Ontario and British Columbia. The Aboriginal population has experienced fertility declines in all regions, but the largest drop (around 0.4 child per woman) between the beginning and the end of the 1990s was observed in the high-fertility regions of the Yukon Territory, the Northwest Territories and Saskatchewan.

Fertility assumptions for the Aboriginal groups relates the total fertility rate of each Aboriginal group to the projected total fertility rate for the general Canadian population in the latest population projections of Canada, provinces and territories from 2000 to 2026. The projected total fertility rates by Aboriginal group, high- and low-fertility region, and place of residence were generated based on the assumption that the distance between the Aboriginal and total Canadian level of fertility would not change over the projected period ². The justification for this assumption came from analysis of the past evolution of the fertility of these groups, which showed no evidence of convergence.

Text table 2.2
Total fertility rates by Aboriginal group and assumption, Canada, 2017

For the project, we developed five different future patterns of fertility (text table 2.2 ). First, the fertility level was assumed to remain constant throughout the projection period, at the level estimated for 1996 to 2001: for the North American Indians it was equal to 2.86, for the Métis to 2.17, and for the Inuit to 3.37. While this assumption seemed highly unlikely, we used it as an upper boundary for fertility. Second, the pace of the moderate decline in fertility assumption presumed that the Aboriginal populations would follow the declining trend of the overall Canadian fertility which is assumed to reach 1.35 children per woman by 2017. The decline in fertility is between 0.2 and 0.4 child per woman among the three Aboriginal groups during the projection period. Third, we assumed a slow, gradual decline in fertility over the projection period. The pace of this decline was inferred from the average of the assumption of a constant level of Aboriginal fertility and the assumption of a moderate decline in fertility. Under this assumption, the total fertility rate of the North American Indians declines to 2.71 children per woman by 2017. The decline in fertility among the Métis is only 0.11 children per woman (2.06 in 2017). The total fertility rate of the Inuit is assumed to drop to 3.19 in 2017. For the fourth pattern, the fertility level of each Aboriginal group was projected to decline rapidly and converge to reach 1.8 children per woman by 2026. The impact of this assumption on projected births of the Inuit population would be quite substantial since it assumed a drop of about 1 child per woman, to the level of 2.36 in 2017. By 2017, the total fertility rate of the North American Indians is assumed to reach 2.18 children per woman and 1.93 for the Métis ³. This assumption is proposed as an analytical scenario to observe the results of a rapid decline in fertility and convergence in fertility level among the Aboriginal populations.

In all four above patterns of fertility over the projected period, the underlying assumption is perfect continuity of Aboriginality, meaning that newborn children have the Aboriginal identity of their mother. Our analysis based on the 1996 and 2001 Census data using the continuity index ⁴ demonstrated that this might not be the case. We developed a special, fifth assumption which incorporated into the future fertility pattern of each Aboriginal group its net gains due to “exits and enters” of children identified differently than their mother. Under this assumption, the “modified” total fertility rate would stay constant over the projection period and equal to 3.12 for the North American Indians, 2.48 for the Métis, and 3.47 for the Inuit. The projected 2017 values of fertility rate for the three Aboriginal groups are presented in text table 2.2 .

Finally, we maintained, over the projected period, a disparity of fertility level by place of residence among the three groups specific to high- and low-fertility regions as evidenced by the 2001 estimates.

Mortality assumption

We estimated the 2001 mortality levels for the three Aboriginal groups assuming the relationship between socio-economic conditions and health, and using the estimates of the levels for Registered Indians as derived from the INAC Indian Registry. In continuation with past trends, the projection’s baseline mortality for all three groups was higher than that for the total Canadian population. Text table 2.3 shows that the 2001 life expectancy at birth for the Métis (71.9 years for men and 77.7 years for women), was closest to that of the total Canadian population (77.0 years for men and 82.2 years for women). For the North American Indians, it was estimated to be 71.1 years for men and 76.7 years for women. The lowest life expectancy among the three groups was estimated for the Inuit: 62.6 years for men and 71.7 years for women. The gaps between the Aboriginal groups and the general Canadian population varied from 5 years to 14 years in 2001.

Text table 2.3
Life expectancy at birth by Aboriginal group and sex, Canada, 2001 and 2017

We developed just one assumption for the future trend in Aboriginal mortality for our projection model. Based on past experience, it seems unlikely that the trend in mortality would undergo significant changes over the relatively short projection period of 16 years, making the choice of just one scenario a reasonable decision. We expected declines in Aboriginal mortality if their living conditions and access to quality health care improved. We assumed life expectancy at birth to increase across all Aboriginal populations, but it should remain below the level for the total Canadian population. Although gains in life expectancy would bring it closer to that of the general population, we projected that the pace would slow due to the large proportion of the Aboriginal populations continuing to reside in rural and remote areas with less access to health care. We used the value of life expectancy at birth in 2017 from the medium assumption developed for the most recent projections of general provincial and territorial populations, ⁵ and the distance between the 2001 levels of the Aboriginal groups and the general population to derive life expectancy for the Aboriginal populations in 2017. Text table 2.3  presents the estimated values of life expectancy at birth for the three Aboriginal groups for 2001 and the projected data for 2017.

We assumed that the place of residence would remain a factor of diversification for life expectancy of the Aboriginal peoples, with somewhat higher levels for those living in CMA areas than those living outside. Also, the estimated current regional differences between the relatively lower levels of mortality for the Atlantic provinces, Quebec, Ontario, Manitoba and the Northwest Territories, and higher levels for Saskatchewan, Alberta, British Columbia, the Yukon Territory and Nunavut were assumed to decrease gradually over the projection period.

Migration assumptions

According to the 2001 Census, just over 162,000 Aboriginal people aged older than 4 years reported changing their residential address outside the community during the five-year period between 1996 and 2001. The changes of the address involved migrations between all four types of residence: reserves, large cities, smaller cities and rural areas. These migrants represented about 19% of the Aboriginal population of the same age. The propensity to migrate between communities for the total Canadian population was comparable at 16%. The Aboriginal and general Canadian populations were also similar with respect to the proportion of migrants who moved within the same region (province or territory) vs. those who changed the region of their residence. For the Aboriginal population, 77% migrants stayed in the same region (intraregional migrations) while for all Canadian migrants 80% did so. Although the majority of Aboriginal migrants identified with the North American Indian, their migration relative to the size of their population was of the same magnitude as that for the Métis. The North American Indian migration rate was just below 17% while the Métis migration rate was over 18%. The Inuit migration rate was much lower, at 12%, as was their overall number of migrants - just over 5,000 during the 5-year period prior to the 2001 Census.

To account for the differences in the type of migration flows and their propensity to move, we approached the projection of inter- and intraregional migration differently. For the former, we assumed that over the projection period, the balance between those who left and those who arrived in the region would be a constant annual number based on the trend observed in the census data. For the latter, we estimated the annual migration rates of in-migrants and out-migrants for each place of residence: CMA areas, urban but non-CMA areas, rural areas and reserves. Each province and territory (region) has a set of intraregional rates. These rates were held constant for the 16 years of projections. The annual projected numbers of intraregional migrants, however, varied from year to year because the rates were applied to changing-over-time populations. We also considered the differences among the three Aboriginal groups, by projecting both types of migration separately for the North American Indians and the Métis using observed trends, and by assuming that internal migration had no significant impact on the future size and geographical distribution of the Inuit.

Migration trends, contrary to fertility and mortality which show a high degree of inertia, display significant changes in intensity and directions. For that reason, projection of migration is subject to a high degree of uncertainty. For this project, we used two assumptions for migration. Both were developed employing the approach described above but were based on two different data sets: the 2001 and 1996 Censuses. The first assumption was an extrapolation into the future of migration patterns observed among the two Aboriginal groups during the second-half of the 1990s (1996 to 2001). In the second assumption, we projected numbers and geographical distribution of migrants belonging to the two groups suggesting that the Aboriginal peoples might change their residence according to the patterns observed among them in the beginning of the 1990s (1991 to 1996). Although there were some differences in the directions of interregional migrations, it was in the intraregional migration intensity levels where we could observe significant shifts between these two periods.

Chart 2.1
Annual net interregional migrants for the North American Indian population by assumption and province/territory, 1996 and 2001

Chart 2.2
Annual net interregional migrants for the Métis by assumption and province/territory, 1996 and 2001

Alberta and Newfoundland and Labrador were the biggest winners from interregional migration of the North American Indians during the late 1990s - over 300 and almost 200 people annually, respectively (chart 2.1 ). These gains were mostly at the expense of British Columbia, Ontario (over 200 fewer arrivals than departures) and, to a lesser extent, Quebec (a negative balance of about 100 persons). This is a reversal of the trend from the beginning of the 1990s, when migration of this Aboriginal group resulted in net losses for Alberta (almost 500 annually) and Newfoundland and Labrador (less than 100) but gains for British Columbia (over 300). Saskatchewan, from being a relatively important recipient of the North American Indian migrants from other regions, became less so by the end of the 1990s. For the Métis, interregional migrations at the end of the 1990s were favourable for Alberta, which was gaining about 500 persons annually, while British Columbia had a net loss of almost 300 persons per year (chart 2.2 ). Again, this is a reversal of the trend from the beginning of the 1990s.

As we already mentioned, for the North American Indians and the Métis, intraregional migrations are much more important than interregional movements. Also, although intraregional migration trends over the 1990s were fairly stable, the levels between the first- and the second-half of this decade were different. The estimated annual net rates of intraregional migration are shown in text table 2.4 .

Text table 2.4
Annual net intraregional migration rates by Aboriginal group, place of residence and assumption, 1996 and 2001

For the North American Indians, reserves were consistently the “catching areas” for migrants but their importance decreased over time. For the other three types of residence, their loses of the North American Indian population were not of the same importance – rural areas outside reserves were depleting more at the end of 1990s but urban areas in general were losing relatively fewer of their residents than they did at the beginning of the period.

Intraregional migration of the Métis was consistently not in favour of urban areas outside CMAs. On the other hand, over the 1990s, this migration was a factor of the growth of the rural Métis population, but its importance was decreasing. And finally, there was a reversal of the migration trend with destination to metropolitan areas - it seems that the Métis in large cities at the end of the 1990s were gaining population as opposed to losing it at the beginning of the period.

Five projection scenarios

We selected five scenarios to illustrate feasible demographic futures of the Aboriginal population in Canada by 2017. We produced projections showing a high population growth (Scenario A). For these projections, we assumed that the number of births for the three Aboriginal groups would be the result of constant fertility, the same as estimated by us for 2001 using the census data; mortality would be declining; and migration trends would continue their course as observed during the second-half of the 1990s. Two other series of projections were produced by using the same assumptions for mortality and migrations as in Scenario A, but different assumptions about future fertility. For Scenario B, designated as a medium-growth scenario, we assumed a slow decline in fertility over the projected period. The third series of the future size of the Aboriginal populations, Scenario D, was derived by assuming that fertility for all three groups would decline rapidly and the tempo of decline would be dictated by a converging trend to reach 1.8 children per woman by 2026. To demonstrate the possible range in the future geographic concentration of the Aboriginal populations, we produced a fourth series of projections according to Scenario C. This scenario had the same assumption for mortality and fertility as Scenario B, but migrations were estimated by extrapolating into the 2001 to 2017 period the trends observed at the beginning of the 1990s. And finally, the special scenario, Scenario S, projected a demographic profile of the Aboriginal populations into 2017 by assuming that there would not be a perfect continuity in the transfer of Aboriginality from mother to child, and then by extrapolating the 2001 trends of the Aboriginality transfer into the 16 year projection period. In this scenario, assumptions of mortality and migrations were the same as in Scenario A and B.

It is important to remember that the accuracy of the future size, composition, geographic concentration and demographic profile of the Aboriginal populations given by projections depends on the underlying assumptions of the components of the populations' change. These assumptions were developed taking into consideration the results of research on relevant topics done in the past. They are intended to show a plausible variation in the future evolution of the Aboriginal populations but should not be considered as predictions of this future.