2007 General Social Survey Report
The retirement plans and expectations of older worker
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.
By Grant Schellenberg and Yuri Ostrovsky
About half of near-retirees plan to retire at ages 60 to 65
What you should know about this study
Six in ten near-retirees are certain they will be able to retire when planned
Most expect their retirement income to be adequate
Expected adequacy of retirement income related to greater certainty of plans
Men and people without a spouse or partner plan to work longer
Immigrants face challenges
Those in good health more certain of their retirement plans
Retirement plans vary considerably across employment characteristics
Results of models confirm importance of employment characteristics
Retirement plans and expectations vary by industry
Income and RRSP contributions associated with more certainty regarding retirement plans
Retirement from two perspectives
After many years of public discussion about Canada's aging population, the leading edge of the baby boom generation is now on the cusp of retirement. Individuals born in 1946 are turning 62 this year—the average retirement age according to the Labour Force Survey—and over the coming decade more Canadians will retire than ever before. How do these individuals view this transition? When do they intend to retire and how firm are their plans? Do they believe they've made adequate financial preparations? These are the sorts of questions that are addressed in this series of Canadian Social Trends, dedicated to the 2007 General Social Survey Report.
Focusing on Canadians aged 45 to 59, we examine the age at which individuals intend to retire, the certainty of their plans, and their expectations regarding their retirement income. All three are subjective assessments—shaped by people's hopes and concerns, their circumstances, and the information they have. Whether their plans and expectations will come to fruition cannot be said, but evidence from the 2007 General Social Survey (GSS) shows that they are related to demographic, employment and financial characteristics. These are the focus of this article.
"Near-retirees"—defined as non-retired Canadians aged 45 to 59—who responded to the 2007 GSS were asked several questions regarding their retirement plans and expectations (see "What you should know about this study"). When asked, At what age do you plan to retire? three-quarters of them state a specific age. The other quarter either say they don't know when they'll retire (14%) or that they don't intend to retire (11%).
Across all near-retirees (including those who don't intend or don't know when they will retire), 22% plan on leaving the workforce before age 60 and 25% plan on doing so between the ages of 60 and 64. Age 65 is still an important reference point for retirement, with 25% of near-retirees planning to leave the workforce at that age. Another 4% plan on retiring at age 66 or older.
There has been a recent increase in the labour force participation rates of older workers in Canada1, raising the question of whether the retirement plans of working Canadians are also being pushed back. We draw on the 1991 Survey of Ageing and Independence (SAI) and the 2002 and 2007 General Social Surveys (GSS) to address this question. Because of differences in the designs of these surveys, our analysis is limited to individuals aged 45 to 59 who were employed throughout the 12 months prior to each survey. Since data are drawn from three different surveys, estimates may be influenced by differences in questionnaire content and survey design. A cautious interpretation of the results is thus warranted.
Between 1991 and 2007, the proportion of near-retirees aged 45 to 49 planning on retiring before age 60 decreased by about 4 percentage points, while the share planning on retiring at age 65 or older increased by about 7 percentage points (Table 1).
Similar patterns are evident among near-retirees aged 50 to 54, with the share planning on retiring before age 60 also decreasing by about 4 percentage points. These patterns are not evident among near-retirees aged 55 to 59.
Overall, these data suggest that Canadians in their late forties and early fifties have pushed back their retirement plans. Evidence from the Labour Force Survey points in the same direction, as the average retirement age of male employees in the private sector reached a low point of 61.4 in 2000 and subsequently rose to 62.3 years by 2007. Likewise, the average retirement age of female employees in that sector increased from 60.7 to 61.6 years over that period.2
Table 1 also shows the uncertainty that Canadians in their forties and fifties have about when they intend to retire. In each survey year, about one in four individuals said they either do not intend to retire or don't know when they will retire. Perhaps this is to be expected, given that retirement may be 10, 15 or even 20 years ahead for these individuals and much can happen in the intervening period. The 2007 GSS provides additional information on retirement uncertainty and suggests it may be even more prevalent than Table 1 indicates.
Data for this paper were drawn from Statistics Canada's 2007 General Social Survey (GSS). The target population for the 2007 GSS was all persons 45 years of age and over residing in Canada, excluding residents of Nunavut, the Yukon and Northwest Territories, and full-time residents of institutions.
The 2007 GSS used a subjective definition of retirement. Individuals who said their "main activity" during the previous 12 months was "retired" were identified as retirees, as were individuals who provided a positive response to the question "Have you ever retired from a job or business?" A definition of retirement was not provided.
Our analysis of near-retirees is limited to GSS respondents who 1) are aged 45 to 59, 2) have not previously retired and, 3) are either employed or had employment during the 12 months preceding the survey.
Based on GSS results, there were 7.2 million Canadians aged 45 to 59 in 2007. Of these individuals, 80% were either currently or recently employed at the time of the 2007 GSS and had not previously retired. Virtually all of these individuals (over 99%) answered the GSS questions regarding their retirement plans.
Of the 45- to 59-year-olds excluded from our analysis, about one-quarter were working at the time of the survey, but said they had already retired at least once before (accounting for 4.9% of all 45- to 59-year-olds). Just over one-quarter had retired from the workforce and were no longer working (accounting for 5.6% of all 45- to 59-year-olds). About half were no longer working but said they had never retired—mostly women who left the labour force earlier in life (accounting for 9.7% of all 45- to 59-year-olds). Adjustments have not been made to account for any possible selection bias introduced into our sample by the exclusion of individuals who have already retired. Overall, our sample of 9,241 respondents is representative of approximately 5.7 million non-retired Canadians aged 45 to 59.
In addition to their planned age of retirement, 2007 GSS respondents were asked whether they're very certain, somewhat certain or not at all certain that they will be able to retire at that age. Most near-retirees express confidence about when they will leave the workforce, with 28% stating a planned retirement age about which they are "very certain" and 33% stating an age about which they are "somewhat certain" (Chart 1).
Altogether, just over 60% of near-retirees are reasonably certain about when they'll leave the workforce. Among the remaining near-retirees who are less sure, about 14% state a planned age of retirement but are "not at all" certain their plans will come to fruition, while the rest either do not know when they'll retire (14%) or don't intend to retire (11%).
Individuals who express certainty regarding their planned retirement age generally expect to leave the workforce sooner than those who are uncertain. For example, of the near-retirees who are "very certain" of their plans, about one-third expect to leave the workforce before age 60; about one-third expect to leave between age 60 and 64; and the final third expect to leave at age 65 or older (Table 2). In contrast, about one-half of near-retirees who are "not at all" certain of their plans expect to work until at least age 65.
Turning to expectations regarding their financial future, GSS respondents3 were asked how adequate they thought their household income and investments will be to maintain their standard of living when they retire.
Two-thirds of near-retirees expect their retirement income to be adequate or more than adequate to maintain their standard of living
Most respondents have a positive outlook, with 62% expecting their retirement income will be "adequate" and another 7% expecting it will be "more than adequate" to maintain their standard of living (Chart 2). Others are less confident, with 19% expecting their retirement income to be "barely adequate" and 9% expecting it to be "inadequate" or "very inadequate". About 3% simply don't know.
Readers are reminded that while these assessments may be informed by careful planning and calculation, they may also be formed in the absence of reliable information and reflect concerns about a financial future that is still many years away. This issue is addressed in more detail in the article "The retirement puzzle: Sorting the pieces."4
Certainty regarding one's planned age of retirement and positive expectations of one's retirement income are interrelated. Over 90% of near-retirees who expect their retirement income will be adequate or more than adequate to maintain their standard of living are very or somewhat certain of their planned retirement age (Table 3), while this is the case for about 43% of near-retirees who expect their retirement income will be less than adequate.
Overall, plans and expectations about retirement timing, certainty and income adequacy often cluster together. Individuals who are most uncertain about their retirement plans often expect to continue working until age 65 and are often uncertain about their retirement income. Those who are certain of their plans are more likely expect to retire around age 60 and express confidence in their financial future.
Differences in the retirement outlooks of 45- to 59-year-olds are correlated with a variety of demographic, labour market and financial characteristics. Such relationships can be examined in two ways.
First, the proportions of people in different groups reporting a specific plan or expectation can be compared—for example, the shares of paid employees and self-employed workers planning on retiring before age 60. Comparisons of this sort do not take into account the fact that plans and expectations may be influenced by other characteristics that systematically vary between groups—for example, whether or not individuals have a pension plan. Nonetheless, these comparisons provide a useful overview of differences and encapsulate the varying circumstances of people's lives. These descriptive statistics are presented in Tables A.1 through A.9.
A second approach is to use analytical techniques to "control for" observable characteristics, such as pension coverage, income and health status in order to estimate how much of the observed differences in plans and expectations are attributable to specific characteristics. The results from this approach, based on a series of models, are presented in Appendix Tables 10 through 12 (See "Multivariate models"). Highlights from both approaches follow.
Multivariate results on certainty regarding planned age of retirement Ordered probit model, Canada, 2007
Multivariate results on planned age of retirement near-retirees who stated a planned age of retirement only, Canada, 2007
Multivariate results on certainty regarding retirement income near-retirees who stated a planned age of retirement only, Canada, 2007
Three outcome variables are used in the multivariate models. First, an ordered probit model is used to identify the factors associated with the likelihood of being very certain, somewhat certain or not at all certain/don't know when one plans to retire. Second, a linear regression model is used to identify the characteristics associated with the planned age of retirement. And third, an ordered probit model is used to identify the factors correlated with the likelihood of expecting one's retirement income to be very adequate or adequate, barely adequate or less than adequate.
These models are run on somewhat different samples of GSS respondents. The first model focusing on certainty excludes the 11% of respondents who said they don't intend to retire. Information is not available to determine if this intention reflects uncertainty about the timing of retirement or a firm plan. The other two models are limited to the 75% of near-retirees who answered the questions about their planned age of retirement and their expectations regarding their retirement income. Individuals who said they don't intend to retire or don't know when they plan on retiring were not asked these questions.
A consistent set of predictor variables is included in the three models. Demographic variables include sex, age, age squared, marital status, educational attainment, immigration status and health status. Employment characteristics include whether the person is self-employed or not, unionization, job tenure and occupation. Financial characteristics include household income, RRSP contributions in the previous five years and the value of accumulated RRSP assets, and housing tenure. All models are calculated using bootstrap weights to correct variance estimates for survey design.
Results from the multivariate models presented in Tables A.10 and A.12 are shown as "marginal effects" for ease of interpretation. The marginal effects show how the predicted probability of an outcome (e.g. expecting retirement income to be less than adequate, barely adequate or adequate) changes between categories of an independent variable. when a specific characteristic is changed by a small amount. For example, the model in Table A.10 predicts, holding all variables at their mean, that 29% of individuals are very certain about their planned age of retirement. It also shows that relationship status is related to certainty about age of retirement. The model predicts that those not in a married/common-law relationship are 4.5 percentage points less likely than those in a relationship to state that they are very certain about their age of retirement.
There are modest differences in the retirement plans and expectations of women and men. Women are slightly more likely to express uncertainty regarding the timing of their retirement as well as concerns regarding the expected adequacy of their retirement savings (Table 4). When other characteristics are taken into account, male-female differences in uncertainty about the timing of retirement remain, but differences in concerns about retirement savings do not. Net of other factors, the model shows that women plan on leaving the workforce almost 9 months earlier than men.5
While most Canadians (75%) approaching retirement are married or in a common-law relationship, one in four are not.6 Retirement plans and expectations differ across marital status. Individuals without a spouse or partner are less likely than their married or common-law counterparts to be certain about the timing of their retirement (54% and 63% respectively) and confident in their retirement savings (55% and 72%).
Part of this difference is attributable to lower household incomes among non-married individuals.7 Yet even when household income and other characteristics are taken into account, non-married individuals are still less likely than their married/common-law counterparts to express confidence in their retirement plans (a difference of almost 5 percentage points) and in their retirement savings (a difference of over 7 percentage points). Furthermore, non-married individuals plan on retiring almost 7 months later than their married/common-law counterparts, net of other factors.
Immigrants to Canada, particularly those who arrive as adults, may face unique challenges preparing for retirement. Their careers and earnings trajectories are often disrupted, reducing their financial capacity to save. Furthermore, the length of time they reside in Canada has implications for Old Age Security eligibility and the contributions they are able to make to public and private pensions. A growing body of research also shows that immigrants who arrived during the 1990s have fared worse in the labour market than immigrants who arrived during the 1970s and 1980s.8
The retirement outlooks of immigrants are indeed different from those of persons born in Canada. For example, near-retirees who immigrated since 1990 are far less likely than the Canadian-born to express certainty regarding their retirement plans (44% and 64%), to have confidence in their retirement savings (50% and 71%), and to expect to retire before age 60 (9% and 32%).
Immigrants who arrived during the 1980s have less favourable retirement outlooks than the Canadian born as well. Much of this difference is attributable to the employment and financial characteristics of immigrants. Yet even when these factors are taken into account, immigrants are still less likely than their Canadian-born counterparts to expect their retirement income to be adequate—7 percentage points less for those who arrived in the 1990s and 8 percentage points less for immigrants who arrived between 1975 and 1989.
Health is an important consideration in retirement transition and is strongly correlated with plans and expectations. While almost two thirds of individuals who rate their health as very good or excellent express certainty regarding the timing of their retirement, this is the case for fewer than half of those who rate their health as fair or poor. Likewise, individuals who report better health have more favourable expectations of their retirement income and are more likely to plan on retiring before age 60, compared with individuals who reported their health as fair or poor.
Substantial differences remain when other factors, such as employment and financial characteristics, are taken into account. For example, compared with near-retirees who rate their health as excellent, those who rate their health as good, fair or poor are less likely to expect their retirement income to be adequate and to express certainty regarding their planned age of retirement.
Finally, of the near-retirees who state a planned retirement age, about 45% of those in good, fair or poor health plan on working until at least age 65, although many of these individuals are uncertain about the adequacy of their retirement income.
Retirement plans and expectations are associated with several employment characteristics. Across employment status, the plans and expectations of self-employed workers are markedly different from those of paid employees.
Quite striking is the fact that almost one in four self-employed individuals (22%) say they do not intend to retire compared with fewer than 1 in 10 paid employees (8%—Table A.2). Among both groups such intentions are most prevalent among individuals with fewer financial resources (Table A.3). Of those self-employed who state a planned age of retirement, almost half (47%) expect to keep working until at least age 65 (Table 5).
Retirement plans and expectations differ across a variety of other employment characteristics as well. For example, compared with unionized employees, non-unionized workers are less likely to plan on retiring before age 60 and to express certainty regarding their plans.9 Similarly, certainty regarding plans, confidence in retirement savings and plans to retire before age 60 are all more prevalent among individuals with longer job tenures and among those with an employer-sponsored pension plan.
These findings reflect a variety of interrelated factors. For example, among paid employees aged 45 to 59, union members are about twice as likely as non-members to have a pension plan (at about 90% and 45% respectively). Likewise, length of time in one's job is related to whether one works in a unionized workplace and to the likelihood of having an employer-sponsored pension plan.
To disentangle the importance of these factors we turn to the models. The first model predicts that about 29% of near-retirees are very certain of their planned age of retirement (Table A.10).
Those with pension coverage are about 10 percentage points more likely to be certain about their planned age of retirement than those with no pension coverage. As well, employees belonging to a union are bout 5 percentage points more likely to be very certain about their planned retirement age than their non-union counterparts. Furthermore, every additional year of job tenure also increases the likelihood of being very certain about one's retirement age.
Findings are similar for expectations regarding retirement income. According to the model in Table A.12, it is predicted that about 73% of near-retirees expect their retirement income to be adequate or more than adequate. Having pension coverage, being in a union, or being in a job longer increase the likelihood that people perceive their income to be adequate.
In terms of planned retirement ages, the second model shows that pension plan members expect to retire about 13 months earlier than non-members, while unionized employees expect to retire about 11 months sooner than their non-unionized counterparts.
Wages, pension coverage, job tenure, unionization and other characteristics vary across industries10 and so do the retirement plans and expectations of individuals in those industries.
Focusing on three industries for illustrative purposes,11 individuals in public administration (i.e. government) are most likely to report being certain of their retirement plans, followed by those in manufacturing and utilities, and then by those in consumer services. And while almost half of individuals in public administration expect to retire before age 60, this is the case for about one-quarter and one-fifth of those in manufacturing and utilities, and in consumer services respectively.
Inter-industry differences in expectations of retirement income are also evident, with 80% of workers in public administration expressing positive expectations compared with 67% of those in manufacturing and utilities and 59% of those in consumer services.
Financial preparations are central in retirement decisions and it is not surprising to find strong associations between financial characteristics and plans and expectations. For example, three-quarters of individuals who have contributed to an RRSP in the previous five years and have more than $100,000 in accumulated RRSP assets express certainty regarding their retirement plans. This is the case for about half of individuals who have not made a recent contribution12 (Table 6).
RRSP characteristics are also strongly related to planned age of retirement and expectations regarding the adequacy of retirement income. The same patterns are evident across household income and home ownership. For example, 85% of near-retirees with household incomes of $100,000 or more expect their retirement income to be adequate compared with 53% of those with household incomes of $40,000 to $60,000. The multivariate models confirm the strength of these relationships.
The fact that retirement outlooks are related to characteristics such as health, pension coverage and household income is likely not much of a surprise to most readers. However, the size of the differences across these characteristics – often 20 or 30 percentage points – is an indication of how different the retirement future looks to Canadians in different circumstances. In this respect, the "baby boom generation" is far from homogeneous.
To put a finer point on this, the statistical models are used to estimate how two hypothetical individuals with specific sets of characteristics might view retirement. Unlike the results presented above, the cumulative effects of differences in various characteristics are tallied.
Our hypothetical individuals are both men, 52 years of age, married, and in very good health. Both were born in Canada, both work as technicians, and both own their homes on which they're making mortgage payments.
Here the similarities end. Person no. 1 has worked in the same job for 20 years and belongs to a union. He has a pension plan and also contributes to an RRSP, although his accumulated assets are less than $50,000. His household income is over $100,000. Person no. 2 has been in his job for 10 years and does not belong to a union. He does not have a pension, but contributes to an RRSP and has accumulated assets between $50,000 and $100,000. His household income is between $60,000 and $80,000.
Overall, these hypothetical individuals are not exceptional. Nor are the differences between them large. If we insert their characteristics in our statistical models, what would we expect their retirement plans and expectations to look like? The results are shown in Table 7.
The planned retirement age of Person no. 1 is predicted to be 59 years, while that of Person no. 2 is predicted to be 62 years – an additional 3 years in the workforce. Person no. 1 is fairly sure of his retirement age—a 46% likelihood of being "very certain." Person no. 2 is less sure, with a 29% likelihood of being "very certain." And while Person no. 1 is predicted to be quite confident about his retirement income, with an 81% likelihood of expecting it to be adequate to maintain his standard of living, Person no. 2 is less so, with a 67% likelihood.
Overall, the reasonably modest differences in the employment and financial characteristics of these hypothetical individuals translate into appreciable differences in their retirement outlooks.
The focus of this paper has been on the plans and expectations of Canadians approaching retirement. These must be viewed with a degree of caution given that they are based on the assumptions and best-guesses of survey respondents and may change over time. Nonetheless, there is a close relationship between retirement plans and expectations and characteristics such as health and finances. Retirement plans and expectations also matter as they are likely to influence behaviour.
The prevalence of uncertainty is a strong theme in the results. Many Canadians, particularly those with health concerns or modest financial resources, are unsure about when they will retire. Results from the 2007 GSS show that such uncertainty is more widespread than past surveys indicated.
There is some evidence suggesting that Canadians in their late forties and early fifties have pushed back their planned age of retirement. Nonetheless, age 65 remains an important reference point, with few Canadians saying they plan to work later than that. Whether such expectations will come to fruition cannot be said, but it is worth noting that labour force participation rates among 65- to 69-year-olds have returned to levels not seen since the mid-1970s.
Retirement outlooks are associated with a variety of demographic characteristics. The importance of marital status, and the capacity of a household to put two rather than just one earner in the workforce, is certainly evident in the results. Health too is a critical factor and deserves more detailed analysis in future research.
Much has been said about the deteriorating labour market and financial outcomes of immigrants arriving in Canada during the 1990s and 2000s. Immigrants who were in their early forties when they arrived in the 1990s are now nearing retirement, raising questions about the adequacy of their financial preparations. Results from the 2007 GSS show that their retirement outlooks are far less positive than those of individuals born in Canada.
Retirement expectations across employment and financial characteristics are much as one would expect. The strong relationship between pension coverage and retirement certainty testifies to the confidence that pension members have in their pension plans.
Results from the 2007 GSS indicate that about one-third of near-retirees express concerns about the adequacy of their retirement income. Whether such concerns are justified is beyond the scope of this paper. However, one question it does raise is whether individuals have the information they need to accurately plan and forecast for a retirement future that may still be many years away? This issue is addressed in the article "The retirement puzzle: Sorting the pieces."13
Grant Schellenberg is a senior analyst and Yuri Ostrovsky is an analyst in Business and Labour Market Analysis Division, Statistics Canada.
- According to the Labour Force Survey, through the 1980s and early 1990s, the share of "older" men participating in the paid labour force declined steadily. Since the mid-1990s, however, this trend has been reversed. For example, after declining from 66% to 43% between 1976 and 1995, the labour force participation rate of men aged 60 to 64 subsequently rebounded to 54%— -- a gain of 11 percentage points in just over a decade. Similarly, one-in-four men aged 65 to 69 are now in the workforce—a share last seen in the mid-1970s. Increasing participation is also evident among older women. After having been stable at around 24% to 25% for more than twenty years, the labour force participation rate of women aged 60 to 64 increased from 25% to 40% between 1998 and 2007.
- The average retirement age of men and women in the public sector has remained fairly constant around age 58 to 59 since the mid-1990s.
- Individuals who do not intend to retire or do not know when they plan to retire were not asked this question. As shown in Table A.3, individuals with lower household incomes, no pension coverage, and no RRSP contributions or accumulated assets were most likely to say they do not intend to retire. Because these individuals were not asked about their expectations regarding the adequacy of their retirement income, the estimates of "barely" and "less than adequate" income in Chart 2 may be underestimated.
- Schellenberg, G. & Ostrovsky, Y. (2008). 2007 General Social Survey Report: The retirement puzzle: Sorting the pieces. Canadian Social Trends, 86. Statistics Canada, Catalogue no. 11-008-XIE.
- The regression coefficients in Table A.11 are expressed using "years" as the unit of measurement. Because "0.74 years" is not very intuitive, these estimates have been converted into months (i.e. 0.74 x 12 = approximately 9 months) in the text.
- Among all Canadians aged 45 to 64, three-quarters are married or in a common-law relationship while one-quarter is separated, divorced, widowed or never married (Labour Force Survey, 2007).
- "Marriage" in this context includes common-law relationships.
- Picot, G. (2004). The deteriorating economic welfare of Canadian immigrants. Canadian Journal of Urban Research. 13(1): 25-46.
- The 2007 General Social Survey does not include information on the size of the firm in which respondents are employed. Rates of unionization are higher among employees in larger than smaller firms. Consequently, the strength of the correlation between unionization and retirement expectations includes firm size as well as unionization effects.
- For example, among near-retirees in these three industries (paid employees only), the incidence of unionization is 70% in public administration, 32% in manufacturing and utilities, and 15% in consumer services. The proportions of near-retiree (paid employees only) in these three industries with personal incomes under $40,000 are 12%, 24% and 54% respectively.
- Retirement plans and expectations of individuals in all industries are included in the Tables A.2, A.5 and A.8.
- "Recent" RRSP contributions are defined as within the previous five years.
- Schellenberg, G. & Ostrovsky, Y. (2008). The retirement puzzle: Sorting the pieces. Canadian Social Trends, 86. Statistics Canada, Catalogue no. 11-008-XIE.
- Date modified: