Chapter 3: Projection of fertility

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This page has been archived on the Web.

By Patrice Dion and Nora Bohnert

Skip to text

• Introduction
• Fertility trends
• Survey results
• Methodology
• Fertility of non-permanent residents
• Assumptions
• References
• Notes

Text begins

Introduction

Fertility has a major impact on the size and the age structure of a population. Assumptions are based on an examination of historical and recent trends in fertility at provincial, territorial, national and international levels, results from the Opinion Survey on Future Demographic Trends and evidence from the scientific literature. Trends in fertility are examined using many different indicators from the perspectives of age, parity, period and cohort. Three assumptions (low, medium and high) are formulated.

Fertility trends

Following a sharp decline between the baby-boom period (1946 to 1965) and the early 1980s, Canada’s period total fertility rate (PTFR) has fluctuated between 1.5 and 1.7 children per woman for more than 30 years. This relative stability has been paired with a continued increase in the ages at which most Canadian women are having their children (Figure 3.1).^Note 1 After experiencing the lowest fertility rate ever recorded in Canada in 2000 and 2002 (1.51 children per woman), the period 2003 to 2008 saw the PTFR increase steadily to reach 1.68 children per woman in 2008. At that time, this was taken as evidence that the postponement of fertility may have been approaching an end and that the PTFR was on its way to returning to levels closer to cohort fertility.^Note 2 However, since 2008, the PTFR has declined each year, reaching 1.61 children per woman in 2011.

Description for figure 3.1

Beginning in the 1970s, the age-specific fertility rates (ASFR) of women under the age of 30 began a declining trend which resulted in two milestones in the new millennium: in 2005, for the first time, the fertility rates of women aged 30 to 34 surpassed those of women aged 25 to 29, while in 2011, for the first time, the fertility rates of women aged 35 to 39 surpassed those of women aged 20 to 24 (Figure 3.2). Since 2007, the combined fertility rates of women aged 30 to 39 have been higher than those of women aged 20 to 29. 

Description for figure 3.2

As seen in Figure 3.3, the recently observed downturn in the PTFR was due to the fact that over the period 2008 to 2011, younger women’s fertility rates steadily diminished while older women’s fertility rates were stable. In the preceding period of 2002 to 2007, younger women’s fertility rates temporarily halted their diminishing trend while older women’s fertility rates steadily increased. Thus, while it appeared during the 2002 to 2007 period that the continued postponement of fertility among young women may have been coming to an end, the more recently observed trends suggest that this is not the case. Instead, the postponement of younger women’s fertility is continuing to occur at a stronger level than the ‘recuperation’ of older women (Frejka 2010).

Description for figure 3.3

The delayment of fertility among younger women can also be observed by examining the cumulated fertility experienced to date by successive cohorts of women, displayed in Figure 3.4. Generally, more recent cohorts of women have experienced lower levels of fertility in their early childbearing years compared to older cohorts. Though the cohorts of the 1970s have shown some signs of ‘recovery’ (for example, the 1970 cohort has met, and may eventually surpass, the fertility rates of the 1965 cohort), the more recent cohorts of women born in the 1980s have so far demonstrated relatively lower levels of fertility.

Description for figure 3.4

In an international context, Canada’s fertility has generally followed the year-to-year trends experienced in other industrialized countries, and yet, Canada is also quite unique in terms of its level of fertility. Among industrialized OECD countries, there are two general groupings in terms of fertility. As seen in Figure 3.5, Anglo-Saxon and Nordic countries such as the United States, the United Kingdom and Sweden have experienced relatively high fertility rates close to or above the theoretical replacement level of 2.1 children per woman. In contrast, countries such as Germany, Spain, Italy and Japan have experienced what can be considered low fertility rates over the last decade, all below 1.5 children per woman. Canada’s PTFR tends to lie between these two groups, being close to the OECD average.

Description for figure 3.5

Between 2005 and 2008, most OECD countries (including Canada) experienced an increasing trend in their fertility rates, a phenomenon that has been attributed by some scholars to more favourable economic conditions and/or to an end in some countries to continued postponement of fertility (see for instance Goldstein et al. (2009), OECD (2011) and Bongaarts and Sobokta (2012)). Since 2008, however, most OECD countries have seen a stabilization or slight decline in their fertility rates. Canada experienced one of the steepest declines in its fertility rate between 2008 and 2010 (-3.0%), following Australia (-3.6%), Spain (-5.5%) and the United States (-7.3%).

Looking at European countries, Lutz et al. (2006) find evidence of demographic, sociological and economic factors that could work together toward lower birth rates in self-reinforcing processes, constituting what they call the “low-fertility trap”. The demographic mechanism is simply the effect of the age structure of the population on the number of births. A sociological mechanism could also be at play in low-fertility countries where young generations, influenced by their environment, develop lower family size ideals. Finally, a third mechanism relates to the relative income hypothesis developed by Richard A. Easterlin (1980), who argued that “Marriage, childbearing and many other aspects of family formation and growth depend crucially on how the ‘typical’ young couple assesses its ‘relative income’, that is, the prospects for achieving the economic lifestyle to which they aspire”.^Note 3 Historically, children tend to experience higher standards of living than their parents did at the same age, in part due to the fact that they generally share parental wealth with fewer siblings, and thus develop higher aspirations. However, they must also bear with the consequences of social security reforms put in place to mitigate the effects of population aging, which tend to have a negative effect on their income.^Note 4 Thus, an increasing gap between the material aspirations of young adults and their relative income may have a depressing effect on cohort fertility levels as well as the timing of births through postponement of childbearing.^Note 5 Martel and Bélanger (1999) found evidence of this phenomenon in Canada for the period 1975 to 1997, linking the interaction between declines in the relative income of young males and changes in female wages to decreases in the fertility rates of women aged 20 to 29.

A study by Goldstein et al. (2003) finds that indeed ideal family sizes could be on the decline in some German-speaking European countries such as Austria, where the PTFR has fallen well below replacement since the end of the baby boom. The most appealing explanation for the authors is that this change is a consequence of the history of low-fertility; in these countries, young generations have “witnessed below replacement fertility for their entire lives”. The authors note that although fertility intentions (in terms of completed family size) rarely actualize in low-fertility countries, this trend could mark a new stage that is indicative of what is to come in other low-fertility countries. However, Edmonston et al. (2010) find no evidence that ideal family size is lowering in Canada to date.^Note 6

At the subnational level, PTFRs vary considerably between Canada’s provinces and territories. In recent decades, the Atlantic provinces have had among the lowest fertility rates in the country, though in 2011, British Columbia registered the lowest rate of 1.42 children per woman (Figure 3.6). Ontario and Quebec had fertility rates closest to the Canadian average (partly a result of their large populations), while the Prairie provinces and the territories were considerably higher than the Canadian average. The highest fertility rate in 2011 and the only to fall above replacement level was that of Nunavut (2.97 children per woman). With the exception of Nunavut, the PTFRs experienced by the provinces and territories have generally been converging over the last 100 years.

Description for figure 3.6

There is also evidence of divergence in terms of age-specific fertility patterns among the provinces and territories in recent years. While at the national level, there were more births to women in their 30s than women in their 20s in 2011, this was only the case for three provinces (Ontario, Alberta and British Columbia) and one territory (Yukon) (Figure 3.7). Among the other provinces and territories, the majority of births were to women aged less than 30. The proportion of all births that were to mothers aged less than 30 was highest in Nunavut (75.1%), New Brunswick (60.9%) and Saskatchewan (60.1%) and was lowest in British Columbia (42.0%). The distribution of births by age of mother in Nunavut demonstrates a uniquely high prevalence of younger mothers, with close to one in five births (18.3%) to women aged less than 20 in 2011—more than three times the average proportion among the other provinces and territories (5.6%).

Description for figure 3.7

In addition to socioeconomic and cultural differences, variations in the PTFR and age-specific rates among the provinces and territories could be in part due to distinctions in public policies that could have an impact on fertility and family size. As with most OECD countries, Canada does not have any explicit policy with regards to fertility, as these issues are generally considered to be part of the private sphere (OECD 2011). However, policies within Canada and other countries have been developed with reference to reducing barriers and costs to having children. Beaujot et al. (2013) note that in Canada, these measures are mostly focused on families with low income. Evaluating the effect that specific policies may have on fertility is often very difficult (Gauthier 2008). Generally, it has been found that while some family benefits may reduce the costs of children, their effects on fertility itself are quite limited, heterogeneous, and often relate more to the timing of births rather than the quantum (OECD 2011; Gauthier 2007; Gauthier 2008; Thévenon and Gauthier 2010). Nonetheless, these timing effects have been found to have an impact on the total fertility rate in some cases.^Note 7

In the case of Canada, there is some evidence of positive but limited impacts of policies on fertility. For example, Morency and Laplante (2010) find small positive impacts of financial aid and parental leave on the first birth of working couples, though the effects vary by the couple’s income and other factors such as job security and homeownership (see also Laplante et al. 2010). The province of Quebec has been the focus of several studies in terms of the impact of various policies on fertility, mainly the province’s introduction of a ‘baby bonus’ cash transfer in the early 1990s and the subsequent subsidized child care and parental leave programs which are more generous and less restrictive compared to the federal programs (Milligan 2005; Beaujot et al. 2013). While the more recent programs seem to have some positive impacts on the labour force participation of mothers (Lefebvre et al. 2011), it is still too early to evaluate their impact on the completed fertility of mothers (Lapierre-Adamcyk 2010).

Survey results

Respondents to the Opinion Survey on Future Demographic Trends gave their views regarding future levels of both period (PTFR) and cohort (CTFR) fertility in Canada. In terms of PTFR, respondents generally anticipated a slight increase. Specifically, the median responses of the most probable estimate of the PTFR were 1.65 children per woman for 2018 and 1.67 children per woman for 2038 (Figure 3.8).

Description for figure 3.8

Somewhat in contrast, respondents anticipated a slight decline in cohort fertility rates in the future. Compared to the most recently completed fertility rate of 1.81 children per woman for the 1962 cohort, the median survey responses to the most probable estimate of the completed fertility of the 1980 and 1990 cohorts were 1.75 and 1.78 children per woman, respectively (Figure 3.9).

Description for figure 3.9

In supporting their estimates, survey respondents mentioned trends that could alternatively suggest a small increase or a small decrease in fertility in the future. Many respondents anticipated that sociocultural trends such as delayed union formation, union instability, diversification of family types, and the increasing educational attainment and labour force participation of women will persist and cause further declines in fertility in the future. The trend of increasing mean age of childbearing suggested to some respondents that fertility levels will decline in the future, due simply to the biological limits of fecundity as, if women increasingly delay childbearing, they could increasingly face difficulties in achieving their desired number of children. For others, it was expected that the mean age of childbearing will cease to continue to increase (again, for biological reasons), and eventually, PTFRs will rise as the ‘timing effects’, which have contributed to lower PTFRs in recent years, lessen. Other arguments supporting an increase in fertility levels in the future included the potential impact of higher fertility among recent immigrants, as well as the fact that as the western provinces grow proportionally in size, their higher fertility levels could have more influence on Canada’s overall fertility rate.

Methodology

As Preston et al. (2001) state, the total fertility rate is the “single most important indicator of fertility”. They define it as “the average number of children a woman would bear if she survived thorough the end of the reproductive age span and experienced at each age a particular set of age-specific fertility rates” (Ibid, page 95). These age-specific fertility rates can be observed during specific periods in order to obtain the period total fertility rate (PTFR), or, over the course of the reproductive life of a cohort of women, in which case we obtain a cohort total fertility rate (CTFR), also known as cohort completed fertility.

Theoretical and practical considerations

Most agencies frame their projection assumptions in terms of PTFRs, mainly because projection inputs take the form of age-specific fertility rates on a yearly basis, and because the CTFR can be calculated only for cohorts of women who have already reached the end of their reproductive years. However, period measures are affected by changes in the timing (tempo) of fertility of successive cohorts, and thus, they can be misleading indicators of actual cohort fertility (quantum). For instance, if in a given year, delayed fertility leads to a decrease in the PTFR, it does not necessarily imply a decrease in cohort fertility if those women ultimately recuperate those births at older ages. For Sobotka (2003), the postponement observed over recent decades in low-fertility countries has rendered the PTFR an inadequate indicator of the quantum of fertility.

Since its variations reflect solely changes in the number of children that cohorts of women have, the CTFR is, in contrast to period measures, much more stable, and is generally more appropriate for use in projections.^Note 8 As Li and Wu (2003, page 303) state, “Demographers generally agree that cohort fertility measures are better than period measures at reflecting how well a society is replacing itself”.

The challenge then, as van Imhoff (2001, page 24) explains, is how “to arrive to statements about family formation processes from a cohort perspective from data that are essentially collected on an annual basis, i.e., from a period perspective”. Some adjustment procedures to remove the ‘tempo effects‘ inherent to the PTFR have been proposed in the literature, such as that proposed by Bongaarts and Feeney (1998). However, the evidence of the validity of these tempo-adjusted measures as estimators of cohort fertility are at best mixed (Ní Bhrolcháin 2011). While some have found that the Bongaarts-Feeney adjustment is generally robust (Zeng and Land 2000), and that deviations from main assumptions “will introduce only minor errors in estimates of the quantum and tempo effects components…” (Bongaarts and Feeney 2000, page 563), the capacity of this measure to isolate pure quantum effects has also been much criticized (van Imhoff and Keilman 2000; Kim and Schoen 2000; Kohler and Philipov 2001).^Note 9 These considerations convinced Ní Bhrolcháin (2011) to advocate for an explicit forecast of cohort fertility, as a more transparent and versatile way to estimate the ultimate mean family size of cohorts not having reached the end of their reproductive years.

Several projections based on CTFR have been conducted for Canada in the literature (see for instance Li and Wu 2003; Schmertmann et al. 2012; Myrskylä et al. 2013). Recently, Myrskylä et al. (2013) projected CTFRs by extrapolating age-specific fertility rates five years into the future based on trends observed over the past five years using a variation of the Lee-Carter method (Lee and Carter 1992) which is generally used to project period mortality.^Note 10 Using birth data for Canada up to 2007, they projected CTFRs of 1.84 children per woman for both cohorts of women born in 1975 and 1979, which implies an eventual end to the long-term declining trend of cohort fertility. They obtain similar results in other countries: in fact, their projections show a leveling off or an increase of cohort fertility in most countries that experienced low period fertility over the last few decades. The authors obtained better results in posteriori comparisons than other methods such as using unlimited linear extrapolation or simple ‘freeze rate’ methods.

However, as was mentioned earlier, the year 2007 marked the end of a period of fertility increase which began in 2003; in the subsequent years of 2008 to 2011, period fertility decreased. Applying the same model to the most recent data available, considerably different results are obtained. Figure 3.10 shows the result of three projections. In the first two projections, age-specific rates are extrapolated for five years before freezing them, using two different reference periods: firstly, the period 2002 to 2007 is used, similarly to Myrskylä et al. (2013); secondly, the most recent data is used, that is, the 2006 to 2011 period. It can be seen that the results differ substantially depending on the reference period. In the early years of the projection, the results are indistinguishable because the women of these cohorts have already passed through most of their reproductive years; thus, only the later years (where fertility rates are relatively low) are extrapolated. However, the trends for later cohorts differ, and cohort fertility ends up declining further over the longer term when using the most recent reference period. A third projection was made using a 10-year reference period, in which the rates were extrapolated for 10 years before being held constant: the resulting CTFRs also show a declining trend in the long run, but less pronounced than in the projection based on the 2006 to 2011 period.

Description for figure 3.10

Description of method

The previous considerations show that in a context of volatility of fertility rates, it is difficult to extrapolate future levels of cohort fertility without making somewhat arbitrary—but heavily significant—choices about the reference period. The problem is made worse when attempting to project the provinces and territories separately: not only do the various regions experience differing trends over the same period in some cases, but in some regions, volatile year-to-year trends can be the result of very small population sizes. That said, the Lee-Carter variant method used by Myrskylä et al. (2013) holds many advantages. It is simple, transparent, and can be adapted to create different assumptions. Moreover, it consistently translates changes in overall levels of fertility to plausible changes in the ASFRs using trends observed in the past. For these reasons, the method was used to calculate the age-specific fertility rates serving as inputs in the projection in conjunction with targets established in terms of PTFR at the Canada level.^Note 11

Briefly, the method consists of a variation of the Lee-Carter model:

$$\begin{array}{cccc}{f}_{x,t}=a_x+b_x{K}_t& & & (3.1)\end{array}$$

In this equation, $f_{x,t}$ is the age-specific fertility rate at age x and time t; $a_x$  is the age-specific fertility rate of the most recent period (the baseline rate); $b_x$ is a vector of parameters measuring the changes related to each specific age over time, estimated as the average annual change in age-specific fertility rates during the reference period; and $K_t$ is the time component which is projected. For a reference period of 10 years, the $K_t$ parameters are estimated using a regression of $\left(f_{x,t}-\stackrel{\wedge }{a_x}\right)$ on $K_t{\stackrel{\wedge }{b}}_x$ for years t-9 to t. For a more detailed description of the method, the reader should refer to Myrskylä et al. (2013).

The method is adapted to match the ‘hybrid bottom-up’ approach generally employed in the projections, which requires the production of separate projections for the individual provinces and territories while observing a main set of assumptions for Canada. Three separate assumptions are proposed: low, medium and high fertility. In a first step, a specific reference period is chosen for each assumption, reflecting the general desired trend in terms of PTFR and CTFR at the national level. The ASFRs and resulting PTFRs are projected at the Canada level using the selected reference period. The period 2001 to 2011 was selected for the medium assumption because its extrapolation produces a ‘moderate’ evolution of a slight increase in period fertility and slight decrease in cohort fertility rates at the Canada level, an evolution which was supported by the ‘most probable’ estimates provided by respondents to the Opinion Survey on Future Demographic Trends. The selected reference period for the low assumption is 2008 to 2011, a period during which the PTFR for Canada as a whole declined steadily. In contrast, the reference period for the high assumption is 2002 to 2008, a period of fairly steady increases in the PTFR at the Canada level. In the low and high assumptions, the selected PTFR and CTFR targets at the national level are reached not only through the selection of the reference period but also by adjusting the weight given to the rate of change over the reference period, the $K_t$ (time trend) factor. In all assumptions, the ASFRs are extrapolated from 2012 up to 2021, after which they remain constant. However, the values extrapolated for 2012 are not utilized for the projection of births, which commences in 2013.

The production of a Canada-level extrapolation is only done as an intermediary step in order to build the provincial and territorial rates. Thus, as a second step, the method is repeated for each province and territory, using the same reference periods as the Canada level when possible.^Note 12 PTFR targets for each province and territory are set to match, in proportion, the projected change in the PTFR for Canada in the previous step. For instance, if the Canada-level PTFR was projected to decrease by 12% between 2011 and 2021 under the low-fertility assumption, the desired target of PTFR in 2021 for each province and territory would be 12% lower than its observed level in 2011. This target is reached by obtaining the appropriate time factor, $K_t$ , through iterations. In using PTFR targets, the same variation between the low and high assumptions is obtained in every province and territory, thus providing an acceptable range of projected fertility outcomes. It also ensures that the projection of all provinces and territories will provide the PTFR target assumed at the Canada level.^Note 13 As with the projection at the national level, ASFRs are extrapolated from 2012 up to 2021,^Note 14 after which they remain constant, with the exception of the province of Quebec. In the case of Quebec, observed ASFRs for 2012 and 2013 became available at the time of assumption-building; thus, the extrapolation for Quebec begins in 2014 from observed 2013 ASFR values.^Note 15

It should be noted that the method could not be used to achieve specific (pre-defined) CTFR targets. Indeed, for cohorts of women who have already entered their fertile years, the eventual CTFR reached will depend in part on the ASFRs already observed and in part on the projected ASFRs. For instance, the completed CTFR level of the 1980 cohort will be determined in part by ASFRs that have already been observed, that is, when these women were aged 10 to 31, and in part by future rates from 2012 to 2035, as these women move through the ages 32 to 54. Thus, imposing a target CTFR for the 1980 cohort would imply no assumptions about the ASFRs under age 32, although these ASFRs will nevertheless impact the projected number of births, and the projected CTFR of subsequent cohorts of women.

In fact, although the CTFR targets were considered in the selection of reference periods for projection at Canada level in the first step for consistency, clearly, Myrskylä et al.’s methodology, intended principally for the projection of cohort fertility rates, is used here to project targets in terms of period fertility measures. It nonetheless holds many useful features in regards to the projection of cohort fertility rates. Its main strength is its ability to model variations in ASFRs in a sound manner independently for each province and territory. Since past variations in ASFRs observed in the individual provinces and territories are often weakly correlated, it is indeed preferable not to impose a single future path of ASFR evolution common to all of them. Using each province and territory’s own past trends to project their future trends should lead to more plausible variations in ASFRs and number of births.^Note 16 As a result, the eventual CTFR levels to be attained in each province and territory are not determined a priori in conjunction with the desired Canada levels; rather, they evolve independently from the national level. More precisely, since each province and territory has a distinct structure of $b_x$ (composition of changes by age) which is applied at different intensities, and given that cohort fertility is affected by $b_x$ structures, the effect on CTFR at the Canada level is unknown prior to the projection. Ultimately, the CTFR reached in a province or a territory is only known for cohorts of women who enter their fertility years in or after 2021; in these cases, the CTFR will equal the PTFR which is held constant as of 2021.

Fertility of non-permanent residents

In past editions, non-permanent residents (NPRs) were always assumed to have fertility behaviours identical to Canadian permanent residents (PRs). However, it is doubtful that this is the case, given that their residence in Canada is temporary. While it is possible to provide distinct fertility parameters for NPRs, since they are projected separately from the PR population,^Note 17 specific information about NPR fertility is not available through Canada’s Vital Statistics system, from which births data are used in the calculation of fertility rates.

This edition incorporates, for the first time, a distinct series of fertility parameters for the population of NPRs in Canada. To do so, fertility rates were calculated using the 2011 National Household Survey (NHS) in conjunction with the own-children method. Specifically, the own-children method consists of identifiying infants living in families and linking them to the women likely to be their mothers in the NHS, thus allowing for the calculation of an estimation of ASFRs.^Note 18 While holding limitations, this approach is judged to result in a net improvement in capturing the fertility behaviour of NPRs as compared to assuming that this group holds fertility rates that are identical to the PR population.^Note 19

Due to the small numbers involved in some areas (in terms of both births to NPR women and the population of NPR women), it was only viable to calculate a base-year ASFR schedule for NPR women at the Canada level. Specific low, medium and high fertility assumptions for NPR women were created through evolving the base-year NPR ASFRs to follow the main fertility assumptions for the total population in each province and territory.

Since separate fertility assumptions were created for the NPR population, it was then necessary to derive fertility assumptions for the remaining PR population. The PR fertility assumptions were set to equal, at each age and for each province and territory, the balance of fertility which remained after NPR fertility was subtracted from the fertility of the total population (i.e., the fertility of NPRs and PRs combined). Like for the NPR population, the fertility assumptions for PRs evolve with time in proportion with the assumptions determined previously for the whole population.

Figure 3.11 displays the base-year (2011) age-specific fertility rates calculated for NPRs, PRs and the total population of Canada as a whole. Results show that, as expected, NPR females exhibit lower fertility than PR females (experiencing an estimated PTFR of 1.08 and 1.64 children per woman, respectively, in 2011). The NPR population also displayed a distinct age structure of fertility, with an older mean age at childbearing (32.4 years) than PRs (30.1 years). As for the differences between the fertility of the total population and PR fertility, those are minimal because of the small weight of NPRs in the total population. The extrapolated trends described earlier for the total population and the changes in ASFRs that they convey are applied to both the PR and NPR fertility schedules, thus maintaining consistency with the assumptions for the total population in terms of both PTFR and CTFR indicators.

Description for figure 3.11

In the following section, assumptions are presented for the total population only; however, fertility inputs into the projection program are separated for PRs and NPRs, which together sum to the total population fertility assumptions.

Assumptions

The analysis of trends in regards to past fertility and the results from the Opinion Survey on Future Demographic Trends lead to the elaboration of three distinct assumptions of low, medium and high fertility. In all assumptions, age-specific fertility rates are frozen 10 years after the beginning of the projection. This implies that the PTFR from that point will remain the same, and that the CTFR will eventually converge to the PTFR level in the long term. It also implies that changes in the age structure of fertility rates stop, which is consistent with the assumption that further postponement is limited due to biological limits to fecundity and to the fact that women can only reduce time between births to a certain extent.

Under the medium assumption, at the Canada level, the period total fertility rate increases slightly from the most recently observed level of 1.61 children per woman in 2011 to 1.67 in 2021, after which it is held constant. Under this assumption, Canada would continue its long-term trend of holding a PTFR below 1.70 children per woman, but levels would be considerably above the lowest observed levels of the early 2000s. Reflecting recent trends, for Canada as a whole, fertility postponement and recuperation would continue to occur at similar levels to one another, resulting in a near-stabilization of the PTFR. Incidentally, a PTFR of 1.67 children per woman matches the median ‘most probable’ long-term estimate provided by opinion survey respondents.

In the high assumption, the PTFR increases from 1.61 children per woman in 2011 to 1.88 in 2021, after which it holds constant. A PTFR value of 1.88 children per woman, while recently observed in Australia and the United States, was last observed for Canada as a whole in 1973. An increase in fertility of this magnitude could occur if, for example, age-specific fertility rates among women in their thirties continue their increasing trend or if certain subpopulations with higher fertility grow in share within the Canadian population.

In the low assumption, the PTFR decreases from 1.61 children per woman in 2011 to 1.53 from 2021 onward. A PTFR of 1.53 children per woman is slightly above the lowest recorded level for Canada (that being 1.51 children per woman in 2000 and 2002) and still falls above levels recently observed in certain ‘low-fertility’ industrialized countries such as Italy, Germany and Spain. Such an evolution could occur if, for example, young women increasingly delay the onset of childbearing to an extent that completed fertility is lower simply due to the biological limits of fecundity; or if, as some experts suggest, various sociocultural trends such as delayed union formation and the increasing educational attainment and labour force participation of women evolve in a manner which promotes lower fertility.

Figure 3.12 shows the projected levels of the PTFR for the low, medium and high assumptions at the Canada level,^Note 20 while Table 3.1 summarizes the projected CTFR and PTFR for all assumptions, for Canada, provinces and territories. The results show that the provinces and territories greatly differ in their projected CTFRs, a result of the fact that in projecting each region on the basis of its own past trends, each region preserves its own (heterogeneous) path.

Description for figure 3.12

The evolution of the mean age at childbearing for each of the low, medium and high assumptions follows the changes observed during the selected reference periods and the intensity at which these changes were applied during the first 10 years of the projection (Table 3.2). As is notable in Figure 3.3 for Canada as a whole, in the low assumption, the reference period (2008 to 2011) was characterized by a diminution of fertility rates at ages 10 to 29 and a general stability at ages 30 and over. In the high assumption, the reference period (2002 to 2008) was characterized by increases of fertility rates at ages 30 and over and a general stability at ages 10 to 29. The medium assumption, based on the period 2001 to 2011, shows a slight diminution of fertility rates at ages 10 to 29 and a more substantial increase of those at ages 30 and over. Thus, all assumptions imply an increase in the mean age at childbearing (this appears to be also true for all individual provinces and territories; see Table 3.2).

References

Beaujot, R., C.J. Du and Z. Ravanera. 2013. “Family policies in Quebec and the rest of Canada: Implications for fertility, child-care, women’s paid work and child development indicators”, Canadian Public Policy, volume 39, issue 2, pages 221 to 239.

Bélanger, A. and S. Gilbert. 2002. “The fertility of immigrant women and their Canadian-born daughters”, Report on the Demographic Situation in Canada, Statistics Canada Catalogue no. 91-209-X, pages 129 to 165.

Bongaarts, J. and G. Feeney. 1998. “On the quantum and tempo of fertility”, Population and Development Review, volume 24, issue 2, pages 271 to 291.

Bongaarts, J. and G. Feeney. 2000. “On the quantum and tempo of fertility: Reply”, Population and Development Review, volume 26, issue 3, pages 560 to 564.

Bongaarts, J. and T. Sobotka. 2012. “A demographic explanation for the recent rise in European fertility”, Population and Development Review, volume 38, issue 1, pages 38 to 120.

d’Addio, A.C. and M.M. d’Ercole. 2005. "Trends and determinants of fertility rates in OECD countries: The role of policies", OECD Social, Employment and Migration working paper.

Désplanques, G. 1993. “Mesurer les disparités de fécondité à l’aide du seul recensement”, Population, volume 6, pages 2011 to 2023.

Easterlin, R.A. 1980. Birth and Fortune: The Impact of Numbers on Personal Welfare, Basic Books Publishers, 205 pages.

Edmonston, B., S.M. Lee and Z. Wu. 2010. “Fertility intentions in Canada: Change or no change?”, Canadian Studies in Population, volume 37, pages 297 to 337.

Frejka, T. 2010. "Cohorts overlays of evolving childbearing patterns: How postponement and recuperation are reflected in period fertility trends", MPIDR working paper WP 2010-026.

Gauthier, A. 2007. “The impact of family policies on fertility in industrialized countries: A review of the literature”, Population Research Policy Review, volume 26, issue 3, pages 323 to 346.

Gauthier, A. 2008. “Some theoretical and methodological comments on the impact of policies on fertility”, Vienna Yearbook of Population Research, volume 6, pages 25 to 28.

Goldstein, J., W. Lutz and M.R. Testa. 2003. “The emergence of sub-replacement family size ideals in Europe”, Population Research and Policy Review, volume 22, pages 479 to 496.

Goldstein, J., T. Sobotka and A. Jasilioniene. 2009. “The end of lowest-low fertility?”, Population and Development Review, volume 35, issue 4, pages 663 to 700.

Goldstein, J.R., M. Kreyenfeld, A. Jasilioniene and D.K. Orsal. 2013. “Fertility reactions to the ‘Great Recession’ in Europe: Recent evidence from order-specific data”, Demographic Research, volume 29, issue 4, pages 85 to 104.

Grabill, W.H. and L.J. Cho. 1965. “Methodology for the measurement of current fertility from population data on young children”, Demography, volume 2, pages 50 to 73.

Kim, Y.J and R. Schoen. 2000. “On the quantum and tempo of fertility: Limits to the Bongaarts-Feeney Adjustment”, Population and Development Review, volume 26, number 3, pages 554 to 559.

Kohler, H.P. and D. Philipov. 2001. “Variance effects in the Bongaarts-Feeney formula”, Demography, volume 38, pages 1 to 16.

Kohler, H.P. and J.A. Ortega. 2002. “Tempo-adjusted period parity progression measures, fertility postponement and completed cohort fertility”, Demographic Research, volume 6, issue 6, pages 91 to 144.

Lapierre-Adamcyk, E. 2010. “L'évolution de la fécondité et la politique familiale québécoise�, Santé, Société et Solidarité, issue 2, pages 63 to 74.

Laplante, B., J.D. Morency et M.C. Street. 2010. « L'action politique et la fécondité: Aperçu d'une étude empirique du processus qui régit la première naissance au Canada », Santé, Société et Solidarité, issue 2, pages 75 to 84.

Lee, R.D. and L.R. Carter. 1992. “Modeling and forecasting U.S. mortality”,  Journal of the American Statistical Association, volume 87, issue 419, pages 659 to 671.

Lefebvre, P., P. Merrigan and F. Roy-Desrosiers. 2011. "Quebec’s childcare universal low fees policy 10 years after: Effects, costs and benefits", working paper 11-01, Centre Interuniversitaire sur le Risque, les Politiques Économiques et l’Emploi.

Lesthaeghe, R. and G. Moors, 2000. “Recent trends in fertility and household formation in the industrialized world”, Review of Population and Social Policy, volume 9, pages 121 to 170.

Li, N. and Z. Wu. 2003. “Forecasting cohort incomplete fertility: A method and an application”, Population Studies, volume 57, issue 3, pages 303 to 320.

Lutz,W., V. Skirbekk and M.R. Testa. 2006. “The low-fertility trap hypothesis: Forces that may lead to further postponement and fewer births in Europe”, Vienna Yearbook of Population Research, pages 167 to 192.

Martel, L. and A. Bélanger. 1999. “Relative income, opportunity cost and fertility changes in Canada”,  Report on the Demographic Situation in Canada, 1998-1999, Statistics Canada Catalogue no. 91-209-X.

Milligan, K. 2005. “Subsidizing the stork: New evidence on tax incentives and fertility”, The Review of Economics and Statistics, volume 87, issue 3, pages 539 to 555.

Morency, J.D. and B. Laplante. 2010. "L'action publique et la première naissance au Canada", Cahiers québécois de démographie, volume 39, issue 2, pages 201 to 241.

Myrskylä, M., J.R. Goldstein and Y.H.A. Cheng. 2013. “New cohort fertility forecasts for the developed world: Rises, falls, and reversals”, Population and Development Review, volume 39, issue 1, pages 31 to 56.

Ní Bhrocháin, M. 2011. “Tempo and the TFR”, Demography, volume 48, pages 841 to 861.

OECD. 2011. “Fertility trends: What have been the main drivers?”, Doing Better for Families, OECD Publishing.

OECD, 2013.Table SF2 1 Fertility rates, updated June 2013, http://www.OECD.org/social/soc/OECDfamilydatabase.htm, site accessed December 18, 2013.

Preston, S.H., P. Heuveline and M. Guillot. 2001. Demography: Measuring and Modeling Population Processes, Oxford, Blackwell.

Schmertmann, C., J. Goldstein, M. Myrskylä and E. Zagheni. 2012. "Fertility forecasting: Using Bayesian methods to extrapolate trends while preserving cohort features", MPIDR working paper WP 2012-003.

Sobotka, T. 2003. “Tempo-quantum and period-cohort interplay in fertility changes in Europe: Evidence from the Czech Republic, Italy, the Netherlands and Sweden”, Demographic Research, volume 6, pages 151 to 214.

Thévenon, O. and A.H. Gauthier. 2010. “L'influence des politiques d'aide aux familles sur la fécondité dans les pays développés�, Santé, Société et Solidarité, issue 2, pages 53 to 62.

Van Imhoff, E. and N. Keilman. 2000. “On the quantum and tempo of fertility: Comment”, Population and Development Review, volume 26, issue 3, pages 549 to 553.

Van Imhoff, E. 2001. “On the impossibility of inferring cohort fertility measures from period fertility measures”, Demographic Research, volume 5, pages 23 to 64.

Zeng, Y. and K.C. Land. 2000. "Adjustment for bias in observed changes in period tempo of fertility and an extension of Ryder’s basic translation equation", working paper, Center for Demographic Studies, Duke University, Durham N.C.

Notes