Social and Economic Dimensions of an Aging Population


Vol. 3 No. 3

Autumn 2001

SEDAP (Social and Economic Dimensions of an Aging Population) is a multidisciplinary research program studying a wide range of aging-related issues and is funded by the Social Sciences and Humanities Research Council of Canada. SEDAP is centred at McMaster University and involves researchers from that institution as well as from the University of British Columbia, Université de Montréal, Queen's and the University of Toronto.


I. SEDAP Post-Doctoral Fellow Appointment

SEDAP is pleased to announce the appointment of Vincent Hildebrand as a Post-Doctoral Fellow in the Department of Economics at McMaster University. Vincent holds a Matrise in Monetary Economics and Banking from the Université Paris IX-Dauphine and an MA in Economics from York University. His doctoral thesis, entitled "Three Microeconomic Studies of Pensions and Retirement Savings", will be defended at York University on October 30, 2001. During his two-year tenure as a fellow, Vincent plans to extend his work in the area of pension portability, labour mobility, and retirement income in Canada. Some of his work will involve the use of the SLID data set at McMaster's Research Data Centre. His other research interests include the study of immigrant vs. native-born differences in asset holdings in the United States.

II. SEDAP Research Papers

SEDAP Research Papers are available on the SEDAP website at no cost. Paper copies may be obtained for a nominal charge. Please contact Mrs. Gail Kalika, Department of Economics, KTH-426, McMaster University, Hamilton, Ont., Canada, L8S 4M4.

Brief descriptions follow of the most recently released papers in the SEDAP series.

SEDAP Research Paper No. 47:

Local Planning for an Aging Population in Ontario: Two Case Studies

Lynda M. Hayward (Centre for Gerontological Studies, McMaster University)

Planning for an aging population involves a dynamic system of inter- related policy issues associated with housing, transportation, community support services and land use. Many of these issues are particularly sensitive to local factors. This paper discusses the context in which local planning for the elderly in Ontario is conducted and then examines the planning activity in two contrasting upper-tier municipalities, Simcoe County and Metropolitan Toronto (prior to amalgamation), as case studies.

The provision of health and social services for the elderly has largely remained the prerogative of the Ontario Ministry of Health. Local planning and resource allocation of long-term care services and facilities for the elderly is delegated to District Health Councils. Responsibility for the planning of land use and transportation systems is delegated to local municipalities under the Ministry of Municipal Affairs and Housing. Some municipalities may also have local planning bodies or social planning councils which act in an advisory capacity. In addition, many municipalities have a two-tiered municipal system: for example, the Municipality of Metropolitan Toronto examined in this paper includes the cities of Toronto, Etobicoke, York, North York and Scarborough and the Borough of East York.

The Municipality of Metropolitan Toronto (Metro) and Simcoe County were chosen for comparative case studies. In both municipalities, the political boundaries of the upper-tier municipalities correspond with those of the District Health Council, which means that the major local planning bodies are contained within the same geographical area. The proportion of the population 65 years of age and over in 1991 was similar for both (12.8% for Metro and 12.9% for Simcoe County) and only slightly above the provincial average (11.7%).

Metro is a highly urbanized community which contains about a quarter of Ontario's elderly population in a relatively small geographical area. It is a "sending" community, characterized by a net outflow of older migrants. If, as the migration literature suggests, the healthy and wealthy elderly are those migrating out of Metro, then those who remain could be relatively needy. Community capacity and resource constraints will require some prioritizing of issues; to accomplish this, data at the appropriate level of aggregation are required, particularly at the neighbourhood level.

Simcoe County is largely a rural area offering a high level of recreational amenities which attract older migrants. It has about 3% of Ontario's elderly population, distributed over an area of land 7.5 times larger than Metro. As a "receiving" community, a common planning issue is the development of retirement communities. The great geographical diversity has meant that issues associated with an aging population are localized, presenting major challenges to those planning for the County as a whole.

The author concludes that, given the considerable variation in the geographical distribution of the older population in the two areas studied, the proportion of the elderly in the population per se can be a poor indicator of the specific planning issues which develop. The lack of information at the local level of detail is emphasized as well as the difficulties posed by sectoral barriers at both the provincial and local level in taking an integrated approach to planning. There is little reason to believe that other local planning districts in Ontario would be much different, especially when the geographical boundaries of the District Health Council do not correspond to upper-tier municipal boundaries.

SEDAP Research Paper No. 48:

Management Experience and Diversity in an Ageing Organisation: A Microsimulation Analysis

Ted Wannell and Martin Gravel (Statistics Canada)

The first post-war birth cohort will reach age 55 this year - the age at which many public service pension plans offer unreduced pension benefits for employees with 30 years of service. Thus many organizations are likely to see retirement rates turn upward and keep climbing for some time. Since long-tenured workers have, on average, advanced further up the organizational hierarchy, the highest level occupational groups will be the first affected. This paper presents a case study of the likely numerical outcomes of increasing retirements on the organizational hierarchy of Statistics Canada.

The model used is a demographically driven micro-simulation model which simulates at the individual employee level, conditioning the simulated events on the individual's demographic characteristics. Four types of events are simulated: retirements, other exits, promotions and hiring. The simulated occurrence of these events is governed by conditional probability tables. The probability tables can be generated from historical data and modified by the user to create alternative scenarios.

The authors run a number of different scenarios with their model. Given assumptions based for the most part on recent trends, they find that increasing numbers of retirements in the next 10 years will generally not reduce management experience to below recently observed levels. They also find that given equal promotion rates for men and women, the representation rate of women among Statistics Canada managers is likely to increase rapidly in coming years as increasing promotion rates move more gender-balanced cohorts into management. On the other hand, visible minority representation among managers will likely stall for several years, even with proactive recruitment and advancement policies.

SEDAP Research Paper No. 49:

Resilience Indicators of Post Retirement Well-Being

Elsa Marziali (Faculty of Social Work, University of Toronto) and Peter Donahue (Faculty of Social Work, University of Calgary)

The purpose of this study was to examine the effects of internal and external resources on response to adverse life events (widowhood or involuntary retirement) in a cohort of elderly men and women. The quantity and adequacy of external resources (e.g. financial resources, accommodation, social support) can be readily defined. In contrast, defining what constitutes "internal resources" is a challenging undertaking.

The research involved secondary analyses of 26 transcripts of interviews from two existing data sets, the first consisting of interviews with 40 older women who were widowed in retirement and the second involving 100 older adults who had retired involuntarily. The secondary analyses of the selected 26 cases were intended to examine the degree of convergence of two rating systems for tracking external resources used to classify cases as good versus poor outcomes in terms of well-being.

One rating system focused on financial status and positive versus negative appraisal of the experienced life event. The second system focused on the dimension of social support. Upon comparison of the two classification systems, it was found that the coders agreed on the categorization of eighteen of the cases. On eight cases, the raters disagreed in a consistent fashion: the first rater classified all eight cases as unsuccessful in contrast to the second rater who rated the same cases as successful. Because of each rater's consistency in classifying the eight discrepant cases, it was decided that a re-analysis of these interview transcripts with a broader list of indicators of resilience response might explain the outcomes more fully.

The authors found that for the eight recoded cases, 5 to 10 indicators of resilient response to life stresses were found in each interview transcript. The initial and subsequent analysis of the study cases showed that resilient responses consisted of an array of adaptive behaviours that contributed to well-being despite limited financial resources and negative appraisals of the life events. The authors conclude that the development of methods for distinguishing more versus less resilient older adults would yield useful assessment tools for service planning and the reformulation of policies that address the needs of an aging population.

SEDAP Research Paper No. 50:

Continuity or Change? Older People in Three Urban Areas

Judith Phillips, Miriam Bernard, Chris Phillipson and Jim Ogg (Centre for Social Gerontology, Keele University, England)

Family life has changed rapidly in the years since the Second World War, and brought with it dramatic changes to the communities within which older people live. Three areas of England provided the locations for a number of community studies undertaken in the 1940s and 1950s: Bethnal Green (a deprived, ethnically diverse, inner city area of London with a history of transient populations), Wolverhampton (an industrial and multi-cultural Midlands borough, which experienced substantial re- development and slum clearance) and Woodford (a relatively affluent, ageing suburb in northeast London). Some forty years on from the original studies, this paper explores what changes and continuities could be observed in the family and community lives of older people.

The present-day research was carried out in two main phases: first, a questionnaire survey of 627 older people (over pensionable age) chosen randomly in the three urban locations and second, semi-structured taped interviews with people over the age of 75 (62 persons); nominated members of the younger generation (19 persons); Bangladeshi and Punjabi elders in Bethnal Green and Wolverhampton (35 persons); and two group interviews (a Bangladeshi carers' group in Bethnal Green and a group of Punjabi interpreters and social workers in Wolverhampton). The original studies of the 1940s and 1950s had each used different approaches and rather differently formulated questions. Trying to replicate this research was therefore not a viable option but some of the original questions were retained where possible.

The authors conclude that over the past decades, there has been a movement from an old age experienced largely within the context of family groups to one shaped by "personal communities". These suggest a more voluntaristic element in social relationships in old age. Whether or not there are kin nearby, there are now alternatives to kin and neighbours. All of this suggests variety in how older people experience family life. There are many more different types of families and different types of older people. But the family in some form is still central to support, even if this is often focused around a smaller number of network members.

SEDAP Research Paper No. 51:

Intracohort Income Status Maintenance: An Analysis of the Later Life Course

Steven G. Prus (Centre for Gerontological Studies, McMaster University)

There are varying findings in the literature about the pattern of economic outcomes over the later life course. This paper examines changes in the income status of individuals within a cohort, with the primary research question being whether individuals with early-life socio- economic status advantages, such as higher education, maintain their relative status position during old age.

Intracohort status maintenance is best addressed with panel data that follow the same individuals over time. Given that such data are not readily available in Canada, a synthetic cohort approach is used. Based on a series of cross-sectional data from Statistics Canada's Survey of Consumer Finances (SCF), the impact of background status characteristics on income is estimated at various points in the later life course of a given birth cohort (those born between 1924 and 1928). Income is measured as family income divided by the square root of family size. Socio-economic status is measured by education level, in six categories ranging from elementary schooling to university graduate.

The author finds that socio-economic status becomes less important to income in old age. He suggests that the primary reason for this is that government pension benefits become more important than labour market income, and the former are more equally distributed than the latter. Specifically, since the public pension system (on which seniors rely quite extensively) is not strongly tied to employment history, it is particularly generous to those at or near the bottom of the socio-economic ladder. Those persons consequently improve their position relative to those at higher socio-economic levels.

SEDAP Research Paper No. 52:

Tax-Preferred Savings Accounts and Marginal Tax Rates: Evidence on RRSP Participation

Kevin Milligan (Department of Economics, University of Toronto and Department of Economics, University of British Columbia)

In this paper, the author considers how marginal tax rates affect the decision to participate in Registered Retirement Savings Plans (RRSPs). The participation rate (measured by the number of tax filers making contributions divided by the number of tax filers with income from employment) rose from 2.4% in 1968 to 46% in 1996. The research in this paper concentrates on the period 1982 to 1996, to exploit the considerable variation in marginal tax rates over that time frame.

Data on marginal tax rates, RRSP contributions and household characteristics are taken from various Family Expenditure Surveys (FAMEX) from 1982 to 1996. Households from outside major urban centres are excluded from the analysis as are non-male-headed households. Only those respondents who reported income from employment are included as only they are eligible to make RRSP contributions. While income data are by individual, RRSP contribution data are by household. Marginal tax rates are calculated as the tax rate faced by the household head (and spouse if present) on the first dollar of RRSP contribution.

The author finds a statistically significant effect of marginal tax rates on RRSP contributions, but the magnitude is small: a 10 percentage point increase in marginal tax rates is estimated to increase participation by about 8%. He calculates that this magnitude of response can explain only 5% of the increase in RRSP participation over 1982 to 1996. The author concludes by suggesting alternative explanations for the increase in RRSP participation. For example, if perceptions have changed about the adequacy of future public pension benefits to meet the desired levels of retirement income, the increase in RRSP participation may be explained by a desire to offset decreases in other sources of retirement income. In addition, in view of a documented decline in employer-provided pension coverage during the 1980s and 1990s, broader RRSP participation may reflect this development.

SEDAP Research Paper No. 53:

Cohort Survival Analysis is Not Enough: Why Local Planners Need to Know More About the Residential Mobility of the Elderly

Lynda M. Hayward (Centre for Gerontological Studies, McMaster University) and N. Michael Lazarowich (School of Planning, University of Waterloo)

When planning for an aging population, especially for the large baby boom cohort, one cannot assume that it is simply a question of greater numbers. Too often, projections of future demand are based on the estimated number of older people who will be in the population and the assumption that current use patterns will continue. For example, with the recent shift to a community-based approach to meet the needs of seniors, local planners are challenged to find ways to integrate an aging population into communities which have not been designed for this purpose.

Different residential mobility choices (aging-in-place in pre-retirement homes, moving within the community, migrating into or away from the community) have different policy implications. The policy implications of aging-in-place are numerous. Although the community care model is gaining acceptance, issues concerning how this will be implemented, the relationship between informal and formal care, privatization of services and the co-ordination of service delivery remain in contention. The affordability, suitability and condition of homes for aging-in-place are also issues. Sources of supplementary income such as the rental of accessory apartments can be contentious for local planners.

Elderly residential mobility primarily involves local moves. In general, seniors make local moves at a later developmental stage associated with declining health and a need for assistance. With local mobility, different housing may be appropriate for people at different stages of the life cycle, suggesting that over-regulation for the benefit of one group may restrict the choices available to people in other stages. In addition, if housing is to be designed specifically for seniors, there has also been much discussion concerning whether it should be age-integrated or segregated at the building level and/or the neighbourhood scale.

The elderly who migrate tend to be younger, wealthier, healthier, better educated and more likely to be married than other senior mobility groups. Because of this differential selection of migrants in terms of health and economic status, planners are increasingly looking at the economic development potential associated with an affluent retired population. While the magnitude of income and asset transfers associated with elderly in-migration is argued to be a positive stimulus to economic growth, little is known concerning the long-term economic and social impact of these migrants as they age-in-place and require an increasing level of community support services. There is some evidence that elderly long-distance migrants can become more dependent on formal community services than other elderly residents, possibly because of a lack of local familial or social supports.

The authors conclude the paper by emphasizing that in planning for an aging population, it will become increasingly important to use a finer scale and to recognize local contextual differences.

SEDAP Research Paper No. 54:

Unemployment and Health: Contextual Level Influences on the Production of Health in Populations

François Béland (Faculty of Medicine, University of Montreal), Stephen Birch and Greg Stoddart (both of Centre for Health Economics and Policy Analysis, McMaster University)

Unemployment is consistently associated with poor health for individuals. While there is a large and growing literature investigating the relationship between an individual's employment status and his or her health, considerably less is known about the effect on this relationship of the context in which unemployment occurs. Few studies consider the effect of local context while those that do use the local unemployment rate as the only contextual variable. The purpose of this study is to examine the synergistic relationship between health status and individuals' own experience of unemployment, given the level of unemployment and other characteristics of the individuals' context.

Individual-level data are from the 1987 Sant‚-Qu‚bec health survey. Data were provided from an interviewer-administered questionnaire for 11,323 households covering 31,995 individuals. In addition, a pre-paid postal questionnaire was provided for every household member aged 15 and over and 19,724 of these were completed. Because the primary sample units used were also census units, individual-level data from the health survey could be linked to the 1986 Canadian census on the basis of census tracts. The 9422 subjects included in the present study were distributed among 624 census tracts with an average of 15 subjects per tract.

The variables of primary interest in the analysis were perceived health (healthy or unhealthy) and employment status (employed or unemployed). Other individual level variables were used, including stress (as measured by the number of stressful events excluding sickness), self- perceived stress generated by these events, occupational status, level of social support, household income, education as well as indicators of "locus of control" and "sense of coherence". Contextual-level variables in each census tract included unemployment rates, gender distribution, age group distribution, education levels, proportion of immigrants, family structure, household income, labour force participation rates and occupational status. The analysis involved the estimation of multi-level equations using five analytical stages.

The authors conclude that this study does not provide evidence to support the hypothesis that the association of unemployment with health status depends upon whether the experience of unemployment is shared with people living in the same environment. They note further that studies such as theirs demonstrate both the subtlety and complexity of individual and contextual-level influences on the health of individuals and caution against simplistic interpretations of the unemployment-health relationship.

SEDAP Research Paper No. 55:

The Timing and Duration of Women's Life Course Events: A Study of Mothers with At Least Two Children

Karen M. Kobayashi, Anne Martin-Matthews (both of Social Work and Family Studies, University of British Columbia), Carolyn J. Rosenthal (Gerontological Studies, McMaster University ) and Sarah Matthews (Sociology, Cleveland State University)

This paper examines the incidence and duration of women's life events by age structure, birth cohort, educational status and place of birth. Research on the timing of major life events may provide the basis for an understanding of how age structure affects family relationships in general and levels of support in later life families in particular. For example, the authors cite literature in which it is found that women who delay childbirth into their thirties are more likely to live with aging parents than those who experience child birth as an early life course transition.

The authors use data from the 1995 (Cycle 10) General Social Survey of Canada. The focus of their investigation is on age structure according to the first birth of children in families. The selected sample for this study is all female adults with at least two natural children at the time of the survey (sample size=1800). A female respondent in the sample is categorized as age condensed if she gave birth to her first child when she was 21 years of age or younger, normative if her first child was born between the ages of 22 and 29 years, and age gapped if she had her first child at 30 or older.

Four dependent variables are examined: age of respondent at birth of first child, age of respondent at first marriage, the interval between respondent's first and last births and respondent's total number of birth children. Independent variables include respondent's birth cohort (1915- 30, 1931-45, 1946-60, 1961-76), respondent's educational attainment (high school diploma or higher, less than high school diploma) and respondent's place of birth (Canada, other). Data are analyzed by Analysis of Variance (ANOVA).

The authors' findings include:

SEDAP Research Paper No. 56:

Age-Gapped and Age-Condensed Lineages: Patterns of Intergenerational Age Structure Among Canadian Families

Anne Martin-Matthews, Karen M. Kobayashi (both of Social Work and Family Studies, University of British Columbia), Carolyn J. Rosenthal (Gerontological Studies, McMaster University ) and Sarah H. Matthews (Sociology, Cleveland State University)

The authors examine intergenerational connections within families over time. Their specific interest is in the interval between generations: the gap in years that separates one generation from the next.

As in Working Paper No. 55 above, the data used are from the 1995 (Cycle 10) General Social Survey of Canada and respondents (Generation 2 or G2) and their mothers (G1) are categorized as age condensed, normative or age gapped according to the age at which they gave birth to their first child. The selected sample for this study (sample size=404) is all female adults with a living mother who are only or eldest children and who had at least one natural child at the time of the survey. As the survey did not ask the respondent how old her mother was when she (the respondent) was born, the only way in which the difference in age between a respondent adult daughter and her mother could be determined was to subtract the daughter's age from that of her mother. Mother's age was known only for living mothers, and thus the sample was reduced to those women who reported that their mother was still alive. Also, as the focus here was on the age interval at the point when a new generation was added to a family, sample selection was limited to eldest or only children because the survey did not collect age data for a respondent's siblings.

Among the authors' findings are:

SEDAP Research Paper No. 57:

The Relationship between Age, Socio-Economic Status, and Health among Adult Canadians

Steven G. Prus (Gerontological Studies, McMaster University)

The author posits that a strong relationship between socio-economic status (SES) and health is well documented, with social groups closer to the top of the socio-economic ladder, notably the well educated and higher income classes, having lower rates of morbidity and mortality compared to those at the bottom. This study examines the extent to which persons with early-life health advantages (i.e., those with high SES) maintain their health status relative to those with early-life health disadvantages (those with low SES) over the life course. The study also estimates the overall rate of inequality in the distribution of health outcomes over the adult life course.

Health changes over the life course are estimated by comparing age groups/cohorts in cross-sectional data from the master file of the third cycle (1998-99) of Statistics Canada's National Population Health Survey. Several measures of health status are used. Subjective health status is based on the question, "In general, how would you say your health is?" . Responses have been collapsed here to "positive" (from good, very good and excellent) and "negative" (fair or poor). The Health Utility Index (HUI) is based on respondents' answers to questions about their vision, hearing, speech, mobility, dexterity, cognition, emotions, pain and discomfort. Scores may range from 0 (completely unfunctional) to 1 (perfect functional health) and a score of 0.80 or higher is typically used to indicate a high level of functional health. Restriction of activities due to long-term conditions is also measured by the NPHS survey.

The author uses education level to measure SES, which is collapsed into age-specific quartiles to reduce the impact of cohort effects (i.e., younger adults are better educated than older adults). The study is restricted to those aged 25 and over (sample size=10778) since most persons this age have completed their formal education. Finally, personal resources are measured as follows: 1) multiple indicators are used to measure lifestyle : level of physical activity, number of years smoked and body mass index 2) level of coherence, distress and social support are used to measure psychosocial resources 3) household income (adjusted by household size) is used to measure economic resources.

Among the author's findings are the following (all findings adjusted for gender and marital status):

editor: Deb Fretz -
Return to SEDAP Home Page.
Last revised: Sept 17, 2001.