Main article
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.
1 Introduction
The economic well-being of retiring Canadians has been an important public policy item for quite some time. From the 1950s to the 1970s, discussions took place about whether Canadians were adequately prepared for retirement with policy analysts questioning the adequacy of the retirement system for retired Canadians. Their focus was both on low income rates, which were high by western standards at that time, and on income replacement rates, that is, the extent to which income earned during the working years would be replaced in retirement. 1
In Canada, these discussions resulted in the implementation of the Canada Pension Plan in 1965. With the implementation of this program and other universal income security programs for seniors, and with the maturation of private pension plans, low-income rates among Canadian seniors progressively declined.
Recently, the issue of the economic security of retirees reappeared on the policy agenda. Population aging, of course, likely explains a good deal of the interest. More recently, however, the recent stock market decline and declining interest rates sparked renewed discussion on the issue of retirement income adequacy.
Recent work by the same authors has focused on the size of the replacement rate. LaRochelle-Côté, Myles, and Picot (2008a) asked whether family income during the working years was in fact maintained during the senior years. In this paper, a longitudinal data set was employed to estimate the extent to which family income around age 55 was "replaced" by the time the individuals in the study turned 65 to 75. It found that the family income of individuals in their mid-70s (for the median worker in the sample) was about 78% of that registered around age 55 while he or she was still working and still had a strong attachment to the labour market. Among low-income individuals, this "replacement rate" was 100%; among middle-income individuals, it was 80%; and among individuals with a high family income, this rate stood at about 70%. Furthermore, income during the retirement years increased among more recent retirees.
"Median" replacement rates are summary statistics that capture central tendencies of a population. The full distributions are required in order to describe how individuals at the low end and at the high end fare. For example, among middle-income individuals, about one-quarter had replacement rates below 60% by the time they reached their mid-70s.
In the previous study, the focus was on individuals with a "strong attachment to the labour market" while in their mid-50s. More specifically, individuals in the sample had to have had wages and salaries of at least $10,000 at age 55 in order to be included in the study. The primary concern during the 1970s was whether Canadians with significant earnings during their working years would see that income replaced as they entered their senior years.
One key question is whether similar results are obtained when all Canadians are considered, whether they are strongly attached to the labour market or not. For example, spouses who have full-time employed partners, but who themselves are not working or are working part-time, would have been excluded from the earlier study. Yet, the extent to which pre-retirement living standards are maintained in their older ages for this group is an important issue. Other individuals would also have been excluded from the earlier study, such as those working part-time and those who were not in the labour force during their mid-fifties. In all, about 50% of the population was excluded from the earlier study; only those with a strong labour force attachment were included.
This study expands the earlier one to include Canadians as a whole and measures the extent to which family income levels are maintained in senior years. As in the earlier study, the focus is not on low-income in retirement, but rather on replacement of pre-retirement income. Owing in large part to data constraints 2 , individuals with very low family incomes at age 55-below $14,000 for a family of two, or below $20,000 for a family of four-remain excluded. Overall, approximately 80% to 85% of the Canadian population is included in this study, depending on the cohort examined, compared to about half the population in the earlier study (see Table 1). 3
2 Results: Income sources among retirees
Measuring income
This paper focuses on the change in the economic welfare of individuals as they age, in particular, with how their welfare changes relative to that experienced prior to the retirement years (starting around age 55). Family income is a better indicator of welfare than is individual income. Hence, when this paper refers to the income of an individual, it refers to the income of the family to which that individual belongs. Income components such as investment income and pension income are reported in the same manner; the values represent the income of the family to which the individual belongs.
Incomes are reported in 2007 constant dollars. To account for differences in family size, both among families at a given point in time, and over time as the size of the family to which individuals belong changes, all incomes and income components are adult-equivalent-adjusted (AEA). The AEA family income is a per capita measure of family income, after taking into account economies of scale available to individuals who live in larger families. All individuals in the same family have the same AEA family income. To obtain a sense of what the family income would have been prior to adjustment (i.e, the unadjusted family income), the adjusted income should be multiplied by two for a family of four or by 1.4 for a family of two. Hence, for an individual who has an AEA family income reported here of $25,000 for example, if that individual belonged to a family of four, that family's total unadjusted income would be $50,000. 4 If the individual were part of a couple, the unadjusted family income would be $35,000.
The incomes reported here, whether total income or income components, are intended to capture a 'permanent' income concept, that is, to smooth out transitory short-run fluctuations. Since income levels and their components such as earnings and investment income can vary dramatically from year to year, income replacement rates-family income at any given age compared to that around age 55-can also vary for any given individual. To ensure that the results present a more stable "permanent" income picture, income figures are all expressed in three-year moving averages. For example, the family income of an individual in 1983 (say, for an individual aged 55), is actually his or her average family income over 1982, 1983 and 1984 inclusively. Similarly, investment or pension income in 2006, for example, is the average for these income components, at the family level, for that individual in 2005, 2006 and 2007, inclusively.
Outcomes for the 1983 cohort
Just like the approach used in the earlier study (LaRochelle-Côté, Myles and Picot, 2008), the 1983 cohort consists of all people who were aged 54 to 56 in 1983. The reason for including all people aged 54 to 56, and not just the 55 years old, is to benefit from a larger sample size—a necessary condition for the analysis of replacement rates and income level across quintiles. Therefore, any reference to individuals aged "around" 55, in fact, comprise those aged 54 to 56. Because the LAD has longitudinal properties 5 , a 20% sample of tax filers aged 54 to 56 in 1983 was followed until they were 77 to 79 years of age, in 2006. 6 Given the use of permanent income figures, these 24 years of longitudinal data represent the longest period available; hence, these results are reported first. This paper also examines whether the outcomes for more recent cohorts improve or deteriorate relative to the 1983 cohort.
Before tax family income, along with its components, provide a sense of how the shares of various income components change as individuals age from their mid-50s to their late 70s. After tax income, a better measure of disposable income, is used subsequently to compute replacement rates and other measures.
For the 1983 cohort, average before-tax family income (AEA family income) falls from about $50,000 around age 55 to about $42,000 in their late 60s, and remains relatively stable until around age 77, the latest observation in our data. When individuals are 54 to 56 years of age, three-quarters of the family income comes from earnings. By age 77 to 79, when most members of the cohort are likely retired, private pensions account for about one-third of all income, public pensions (including CPP/QPP and OAS/GIS) for about one-third, and investment income for 14%; earnings continue to generate about 10% of family income (Table 2).
The extent to which income changes in the retirement years and the contribution of its components, both depend on whether the family is at the bottom or the top of the income distribution. Public pensions are more important to low-income families; private pensions and investments are more important to higher-income families. To assess these differences, similar results are also examined for individuals in the bottom, middle, and top family income quintiles. The idea is to examine how incomes developed over time, given that this family had a given level of income at the beginning of the period. Hence, individuals are assigned to income quintiles on the basis of their AEA family income around age 55 (i.e., average income over 1982, 1983 and 1984). Under this approach, each person's quintile remains fixed as he or she ages.
Unlike average income for the population as a whole, average before-tax family income increases with age for people in the bottom quintile, rising from about $19,000 around age 55 to from $23,000 to $24,000 during their mid-60s, and remaining stable to their late 70s (Table 3). This rise takes place when public pensions (CPP/QPP and OAS/GIS) replace earnings as the main source of income. When individuals are aged 54 to 56, earnings constitute about two--thirds of their family income; by age 77 to 79, public pensions account for 63%, private pensions for 14%, and earnings for about 10%. Interestingly, the reliance on earnings as a source of income in one's 70s is about the same for individuals at the bottom of the income distribution as for individuals at the top of the income distribution: earnings contribute roughly 15% of income around age 70, and decrease to about 10% around age 78. It is important to remember, however, that these are family, not individual, earnings. It may be the case that it is the individual aged 78 who is providing the earnings, or it may be some other family member. It is not known whether the earnings are generated out of necessity or because the individual chooses to continue working; nor is it known which family member provides the earnings. However, the results clearly indicate, that low-income individuals do not turn to earnings any more than do high-income individuals in their 70s. As people in lower-income families age, their average family income rises and becomes more stable (LaRochelle-Côté, Myles, and Picot 2008a) as public pensions replace the more unstable stream of earnings.
Individuals in the middle quintile saw average before-tax family income fall from $43,200 around age 55 to around $34,000 in their late 60s, and remain stable at that level through their late 70s (Table 4). Since income among lower-income families rises with age, and since income falls within the middle quintile, the income gap between individuals in the bottom and the middle quintile decreases as the cohort ages, from $24,000 at age 55 to $10,800 at age 70.
Among individuals in the middle quintile, earnings constitute 82% of total family income around age 55, but by age 78 public pensions (CPP/QPP plus OAS/GIS) also play an important role. They constitute 45% of before-tax family income (compared to 62% among bottom-quintile individuals); an additional one-third of individuals' income comprises private pensions. Of note, once individuals are in their late 60s and 70s, the composition of average family income changes very little. This was also the case for individuals in the bottom quintile.
Individuals who were in the top quintile saw their average AEA family income fall as they moved from age 54 to 56 to their late 70s—from $99,200 to around $87,000 (Table 5). At any age, investment income is more important to this higher-income group. Around age 55, earnings represent 73% of family income, and investments comprise 16%. Past age 70, private pensions are the major contributor (about 40%), followed by investment income (about 20%), public pensions (CPP/QPP and OAS/GIS, about 18%), capital gains (from 7% to 14%), and earnings (from 16% to 11%). Interestingly, even individuals in the top quintile of family income rely to a significant degree on public pensions as a source of income in their 70s (for one-fifth), although private pensions clearly play a more prominent role.
Outcomes for more recent cohorts
Generally, more recent cohorts have improved their income positions at all ages relative to the 1983 cohort, whether before the retirement years (e.g. age 54 to 56) or in the later retirement years (age 70 and over). When capital gains are excluded from the total, 7 the 1983 cohort possessed an average family income of about $49,300, while the 1986 cohort (i.e., those who were aged 54 to 56 in 1986) saw that rise to $51,100. The 1995 and 1998 cohorts had before-tax AEA family incomes of $54,500 and $58,100, respectively, around age 55. Generally, the income advantage of younger cohorts is also evident at other ages (Chart 1). By age 65, average total family income (excluding capital gains) had risen from around $40,000 for the cohorts of the 1980s to around $50,000 for the 1995 cohort.
This improvement was driven by both higher earnings and higher private-pension income (Charts 2 and 3). Of the $9,100 increase in total family income seen among those aged 65 to 67 between the 1983 and 1995 cohorts two-thirds was due to higher earnings, and one-third was due to higher private pensions. Wage rates for more mature workers rose over the 1980s and 1990s, while they fell among the young (Beaudry and Green 2000). This would have contributed to higher earnings. It may also be that more people in the early retirement years (possibly women) were working, or those employed were working longer hours. Whether this propensity among the 1990s cohorts to generate higher earnings than did the 1980s cohorts remains to be seen.
In conclusion, recent retirement cohorts have more economic resources than their predecessors on average, as a result of the higher earnings that they received while working and the higher private-pension income that they draw in the retirement years.
3 Results: Replacement rates
A replacement rate measures the extent to which the economic resources available to the individual through income flows (mainly earnings) are "replaced" by various sources of income (public and private pensions, investments, as well as earnings) as the individual moves from age 55 to any given retirement age, such as 78. 8 In this paper, replacement rates are obtained by dividing the individual's AEA family income at any given age, 78 for example, by that of the same individual at age 55. Note that, if the size of the family to which the individual belongs changes for any reason, such as divorce, marriage, widowhood, etc., the family income of that individual is adjusted in order to account for the change in family size. Since after-tax income is the best measure of "disposable" income available, it is the most appropriate concept to use in the calculation of replacement rates.
Median replacement rates across all individuals
The AEA family income available during the retirement years to the "median" individual is about 80% of that observed when that same person was in his or her mid-50s (Chart 4).
Median replacement rates for the cohort of Canadians who were aged 54 to 56 in 1983 fell from 1.0 (by definition) to 0.8 around age 68, and remained stable at this level through to their late 70s. More recent cohorts display the same general pattern: a slow decline from the mid-60s to the late 60s, then stability at around 0.8. Data for more recent cohorts who turned age 55 during the 1990s suggest that their replacement rates may be somewhat higher than those of the 1980s cohorts. However, the data for these groups extend only into their mid-60s, and it is difficult to say whether this pattern will hold as they move into their 70s. As noted earlier, rising family incomes among these groups were related to higher earnings when they were in their 50s and 60s as well as to the associated higher private-pension incomes.
Results across the income distribution
Replacement rates vary depending upon where an individual is located in the income distribution. In general, the higher the income at age 55, the lower the replacement rate during the retirement years (LaRochelle-Côté, Myles, and Picot 2008a). On average, as noted earlier, individuals in the bottom quintile (around age 55) find that the public-pension system more than replaces the earnings and other income they had in their mid 50s; this results in replacement rates that rise above 1.0 (Chart 5). Among the 80s cohorts, for example, replacement rates rose to slightly over 1.1 when cohort members reached their mid-60s, and remained at around 1.1 to age 77 to 79, the latest observation. Data on the cohorts' income during their 60s suggest that replacement rates for the cohorts of the 1990s may be marginally higher than those of the 1980s cohorts.
However, these results refer to median replacement rates among lower-income individuals. Table 6 shows that, while they were in their late 60s, when replacement rates had stabilized, approximately 9% of the members of this bottom quintile had replacement rates below 0.8. Since eligibility for OAS and GIS is almost universal for lower-income individuals, these shares are relatively small. However, some people may not receive these incomes, even though they are eligible for them (Luong 2009), and some may not receive CPP/QPP.
On the other hand, some individuals who were in the bottom income quintile in their mid-50s moved up the income distribution, sometimes producing replacement rates well above 1.0. Two-thirds of bottom-quintile individuals had replacement rates above 1.0 during their late 60s and mid-70s, and as many as 23% of these had replacement rates above 1.5.
Members of the middle income quintile (around age 55) saw their replacement rates fall to about 0.75 during their late 60s, and again remain stable through to age 77 (Chart 6). As in other cases, there is preliminary evidence to suggest that the rates for the cohorts of the 1990s may be marginally higher than those of the 1980s cohorts.
The distribution of the rates for middle-income individuals indicates that 22% had replacement rates below 0.6 during their late 60s and mid-70s.
Finally, people in the top income quintile displayed the lowest replacement rates (in comparison to the other quintiles) while they were in their early 70s (Chart 7). Median rates for this group with respect to the 1980s cohorts fell to around 0.65 when they were in their mid- to late 60s, recovering to about 0.7 during their late 70s. Furthermore, some 34% among this group experienced replacement rates below 0.6 in their mid-70s. However, it is also interesting to note that nearly one in five (17%) of top quintile individuals had replacement rates above 1.0— indicating that a significant portion of top-quintile individuals benefited from even higher income levels as they aged.
4 Conclusion
In LaRochelle-Côté, Myles and Picot (2008a), the focus was on replacement rates for Canadian individuals who had a substantial attachment to the labour force—about 55% of Canadians. This paper extends the analysis to consider a larger group (some 80-85% of Canadians). Despite these changes, the results remain broadly similar. The AEA family income available during the retirement years to the "median" individual is about 80% of that observed when that same person was age 55. The replacement rate for individuals with a strong attachment to the labour force reported in the previous paper was 78%. 9
As in the earlier study, the lower the income in the mid-fifties, the higher the replacement rates in the senior years. Individuals in the bottom quintile typically achieved a 110% replacement rate by their mid-60s, while individuals in the top income quintile had replacement rates in the 0.7 range. There was some variation within quintiles. For example, more than 20% of middle-income Canadians had replacement rates below 0.6 in their mid-70s.
- Date modified: