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 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. 58:

Measuring Differences in the Effect of Social Resource Factors on the Health of Elderly Canadian Men and Women

Steven G. Prus (Sociology and Anthropology, Carleton University) and Ellen Gee (Sociology and Anthropology, Simon Fraser University)

While gender differences in the exposure to social (i.e., socio- economic, lifestyle and psychosocial) resources play a role in influencing gender inequalities in health in old age, it is less clear in the literature if social factors have a differential impact on the health of older men and women. For example, do elderly men and women with similar levels of chronic stress in their lives have comparable health status? This paper thus looks at gender differences in the vulnerability to the health consequences of high/low socio-economic status, good/bad health behaviours and high/low psychosocial resources.

The data used are from Statistics Canada's National Population Health Survey (NPHS), specifically the cross sectional component of the 1998-1999 (Cycle 3) NPHS. Separate multivariate regression analyses of the relationship between socio-economic, lifestyle and psychosocial variables and the health of older (65 and over) males and females are carried out.

Dependent variables include both subjective and objective health measures. Self-rated health status is based on the question, "In general, how would you say your health is?", and has a five-point scale (collapsed to "positive" and "negative" in the analysis). To measure chronic health, the respondent was asked to list all long-term chronic health problems that have lasted or are expected to last 6 months or more and that have been diagnosed by a health professional. The authors also use a measure of disability based on responses to activity limitation/restriction and activity dependence items. Finally, the authors employ a Health Utility Index (HUI) based on a combination of eight self- reported characteristics of a respondent's health. Independent variables include socio-economic variables such as income, age and visible- minority status, lifestyle variables such as level of physical activity, Body Mass Index and years of smoking, and psychosocial indexes to measure levels of social support, distress and coherence.

Among the findings of the research are:

• Elderly women compared to elderly men have a higher positive perception of their health yet are more likely to suffer from physical health problems.
• While there is a steady decline in health with age for women, growing old does not necessarily mean a continuous decline in health for men.
• Income adequacy has a greater effect with respect to health for older men than for older women.
• Exercise (in either moderate or high amounts) has a greater positive health effect for elderly women than for elderly men. The adverse effect of daily smoking is generally greater for males.
• Social support has a marginally positive effect on health for elderly females but is a significant negative predictor of health for elderly men.

The authors conclude that this research reveals the importance of considering social resources in improving the health of Canadians and they emphasize the need for health-care planners to take into account the varied effects of social forces by gender when designing and implementing health policies.

SEDAP Research Paper No. 59:

APOCALYPSE NO: Population Aging and the Future of Health Care Systems

R.G. Evans (Economics, University of British Columbia), K.M. McGrail (Centre for Health Services and Policy Research, University of British Columbia), S.G. Morgan (Centre for Health Services and Policy Research, University of British Columbia), M.L. Barer (Health Care and Epidemiology, University of British Columbia), and C. Hertzman (Health Care and Epidemiology, University of British Columbia)

For the average individual, aging is associated with declining functional capacities. The inevitability of this individual experience has been combined with demographic trends to yield predictions of a larger and larger share of collective resources being devoted to an ever-older, ever-sicker population. The authors posit that a more detailed analysis of these projections shows they are not sufficient in themselves to sustain a scenario of "apocalyptic demography" and argue that, while an "apocalypse" cannot be definitively ruled out, it appears to be a dubious proposition.

The authors examine detailed data on health care utilization and costs over time for the province of British Columbia. They find that age- specific use rates over the period of analysis moved downward for hospitals but increased for physicians' services and, apparently, for pharmaceuticals. Closer examination, however, reveals that the trend for pharmaceuticals is in large measure price inflation rather than increased utilization. The authors conclude that changes in the age structure of the overall population have not in the past been major contributors to trends in the per capita utilization of health care services, and they will not be in the future; the source of change is elsewhere.

The authors argue that a number of factors other than population aging have contributed to increases in health care expenditures. They note that despite falling drug prices due to the entry of generic brands, average drug costs per patient go up, most commonly because of the introduction of newer, more expensive therapeutic agents that are prescribed in preference to older products in the same therapeutic class. The authors' evidence suggests that this more costly mix of drugs has not yielded corresponding improvements in therapeutic outcomes. The paper also argues that there are cost-increasing biases in the pricing and in the rate of uptake of new technologies, with prices less likely to go down in the face of cost-reducing innovations but more likely to increase to cover more costly innovations. It is also noted that the increase in physician supply in Canada from the late 1960s to the early 1990s was matched by an increase in physician utilization rates.

As to why the "apocalyptic" scenario has prevailed despite what the authors feel is evidence to the contrary, they cite a number of reasons. The most obvious beneficiaries of the "aging population" story are providers of care, for whom health care expenditures are incomes. The demographic story also provides an apparent justification for past expenditure trends: cost escalation is attributed to purely external factors - changes in the demographic mix - for which no one in the health care sector could be held accountable. The authors also argue that predictions of the "unsustainability" of the public health care system may serve the interests of those who advocate a private system.

SEDAP Research Paper No. 60:

The Education Premium in Canada and the United States

J.B. Burbidge, L. Magee and A.L. Robb (Economics, McMaster University)

In the United States, the education premium - the ratio of the earnings of those with more education compared to those with less education - has risen sharply in the last twenty years. The authors of this paper question whether the same is true for Canada. They find that this premium has indeed risen sharply in the U.S. but has been constant or fallen in Canada.

Data on weekly wages are taken from the Current Population Survey (CPS) for the U.S. for the period 1981 to 1999. Canadian data most comparable to the CPS are found in the Survey of Consumer Finances (SCF). However, since the final SCF was for the calendar year 1997, the authors also make use of two additional data sets, the Survey of Labour and Income Dynamics (SLID) and the Labour Force Survey (LFS), to extend the time frame of the analysis. The focus of the analysis is real weekly earnings of men and women aged 25 to 64 who work 30 or more hours per week, who are not self-employed and for whom the major source of income is wages and salaries.

The authors find that the ratio of weekly earnings for those with a university degree compared to those without a university degree rose from 1.3 for U.S. males in 1981 to 1.8 in 1999. For females, the corresponding ratios are 1.5 and 1.7. Using the SCF, weekly earnings of Canadian males with a university degree compared to Canadian males without a university degree were 1.4 in 1981 and 1.3 in 1997. The corresponding ratios for females were 1.7 and 1.5. Extending the analysis using SLID, the male university/non-university ratios were 1.4 in 1997 and 1.4 in 1998. The female ratios were 1.5 and 1.5 for 1997 and 1998. Using the LFS, male ratios were 1.4 in 1997 and 1.4 in 2000; female ratios were 1.5 and 1.5 for 1997 and 2000.

SEDAP Research Paper No. 61:

Student Enrolment and Faculty Recruitment in Ontario: The Double Cohort, the Baby Boom Echo, and the Aging of University Faculty

Byron G. Spencer (Economics, McMaster University)

The growth in the population of student age, associated with the baby boom echo and exacerbated by the "double cohort" (which will see two classes of secondary school graduates enter university in the same year), and the trend toward higher enrolment rates will increase demand in the Ontario university system. Combined with the retirement of the large fraction of faculty hired in the late 1960s and the 1970s to meet the demands associated with the baby boom, this has led some to suggest that the need to recruit new faculty is of crisis proportions. This paper develops a model to provide enrolment projections for the Ontario university system and also projects how many current full-time faculty will be available over the remainder of the decade.

Detailed descriptions of the projections of student enrolment and faculty retention are given in an appendix to the paper. Enrolment projections incorporate anticipated reductions in the typical age at which new students enter university as the double cohort is phased in and also reflect a possible increase in the proportion of the age group that will attend university. The current downward trends in part-time enrolment are assumed to continue. Projections of faculty retention are bounded on the one hand by the assumption that current early retirement rates will continue and on the other hand by the assumption that no one will retire before age 65. Under the latter assumption, it is projected that about 4,300 of the current 12,000 full-time faculty will be lost through retirement; under the former assumption, about 7,000 will be lost.

The paper concludes that recruitment of new faculty of between 600 and 1,000 per year will be essential if the university system at the end of this decade is to have standards even close to those now in place. Eliminating mandatory retirement could reduce the need for recruitment by about one-sixth, based on experience in the U.S. where mandatory retirement in post-secondary institutions has been prohibited since 1994.

SEDAP Research Paper No. 62:

The Social and Demographic Contours of Contemporary Grandparenthood: Mapping Patterns in Canada and the United States

Candace L. Kemp (Sociology, McMaster University)

The purpose of this paper is to establish the demographic patterns and social trends which characterize contemporary grandparenthood in Canada and to provide some comparative analysis of Canadian and American trends. The data for this study were derived from Statistics Canada's 1990 and 1995 General Social Surveys of Canada (Cycle 5: Family and Friends, and Cycle 10:Family). American data are drawn from M.E. Szinovacz's 1998 research.

Among the findings of this paper are:

• In Canada, nearly 80 percent of those aged 65 and over are grandparents. More women than men experience grandparenthood, reflecting the fact that women marry at a younger age than men and also live longer.
• Close to 70 percent of Canadians in their twenties and one-third of those in their thirties have at least one living grandparent. Very few Canadians aged 40 and over have living grandparents, but among those who do, it is most likely a maternal grandmother.
• Grandparent-grandchild co-residence is not a typical experience for Canadians: only 3 percent of all Canadian grandparents reside in a three-generation household.
• Grandparenthood is more prevalent among Americans than among Canadians at all age groups. Roughly 3 percent of those aged 21 to 39 in the U.S. and 33 percent of those aged 40 to 54 report having grandchildren, compared to 0.4 percent and 20 percent, respectively, in Canada.
• Older American children have a greater number of living grandparents than do similar Canadian children. These differences are especially evident in the 23 to 26 year age group where approximately 30 percent of Canadians are grandparentless compared to about 18 percent of Americans.

SEDAP Research Paper No. 63:

Changing Income Inequality and the Elderly in Canada 1991- 1996: Provincial Metropolitan and Local Dimensions

Eric G. Moore and Michael A. Pacey (Geography, Queen's University)

Between 1991 and 1996, inequality in Canadian pre-tax household incomes increased by 3.56 percent when measured by a generally accepted measure, the Gini coefficient. In this paper, the authors explore this relatively strong though complex trend by geographic location and by age. The degree of income inequality has important policy implications, both in the calculation of the overall need for public assistance and services and in the redirection of resources to less well- off areas.

Grouped income data for 1991 and 1996 were obtained from Statistics Canada's Public Use Microdata Files (PUMF) at the national, provincial and Census Metropolitan Area level, and, for census divisions, from the Statistics Canada website. These data consisted of raw counts of households by income category. For individuals, pre-tax household income was defined in 25 categories, with the highest being incomes greater than $200,000. For census divisions in 1991, there were eight categories, ranging from less than $10,000 per year to greater than $70,000. In the 1996 census, the range was extended to 11 categories, with an upper category of greater than $100,000. The centre point of each income class was used to estimate income; for the highest and lowest categories in the income distribution, a mean value was derived from PUMF for each province in both 1991 and 1996.

The authors find that income inequality increased on a national scale between 1991 and 1996, but that the significant increases in inequality were confined to the working-age population. Inequality among seniors actually declined during that period as those components of income which are particularly relevant to seniors, such as private pensions, CPP/QPP and OAS payments, increased substantially more than wage and salary incomes and employment insurance payments.

At the provincial level, inequality increased for all provinces except Saskatchewan, and the increases ranged from a low of 0.94 percent in Manitoba to more substantial changes in Newfoundland (7.4 percent), Ontario (4.9 percent) and Quebec (4.8 percent). At the sub-provincial level, the Census Metropolitan Area of Toronto experienced the largest increase in income inequality.

SEDAP Research Paper No. 64:

Mid-life Patterns and the Residential Mobility of Older Men

Lynda M. Hayward (Clinical Epidemiology and Biostatistics, McMaster University)

Although there are numerous ways for local planners to integrate the elderly into communities, such plans may be contingent on whether the elderly will decide to age in-place, make a local move, or migrate to a different community. Within the residential mobility literature, there is an increasing awareness that moving decisions in later life reflect a lifetime of experiences. This research paper examines the relationship of a number of life trajectories at mid-life to residential mobility in later life.

The data used are from the Ontario Longitudinal Study of Aging (LSA), which began in 1959 with a stratified quota sample of 2000 employed, 45-year-old men. The subjects were interviewed in person once a year (except in 1977) until 1978, at which time they were 64 years of age. In 1990, a follow-up telephone survey was conducted with the remaining respondents or their survivors. The sub-sample of the LSA data used here is the 1063 respondents who were still in the study in the year of retirement or upon turning age 65, whichever came first.

The primary independent variable used was the time to the first move after retirement or upon turning age 65. Independent variables included number of mid-life moves, years of residence, housing tenure, marital status just prior to retirement/age 65, marital status change during mid- life, the existence of children, period of empty nest, contact with a non- household relative during mid-life, continuity of social relations, income patterns, occupation, years of education, health rating over mid-life and the proportion of years during mid-life with declining health. Cox proportional hazards modeling was used for the analysis.

The author's findings include:

• Respondents who had children were more likely to make a first move in later life. Widowed, separated or divorced respondents were increasingly more likely to move as they aged in later life, even when the existence of children was taken into consideration.
• Those who had monthly contact with kin during mid-life were more likely to move than those with either more or less contact.
• Subjects who were blue collar workers were less likely to move in both mid- and later life than those in white collar occupations.
• Those with an early onset of poor health were more likely to move in general and those with a high proportion of mid-life years with declining health were less likely to move shortly after retirement.
• An unexpected finding was that the higher the educational attainment of respondents, the more likely they were to age in-place. This was contrary to the expectation that education would have a slightly positive or no effect on residential mobility in later life. However, men making moves out of Ontario during mid-life were dropped from the study, resulting in a longitudinal sample with a history of shorter-distance moves.
• A second unexpected finding was that early retirement had a relationship with residential mobility which reversed with time. Early retirees were more likely to move shortly after retirement but they were also increasingly less likely to move as they aged, compared with those retiring at a later age.

The author notes that as a study of a single gender and birth cohort in a specific period of history, the generalizability of the findings from the analysis is limited.