By Katherine Marshall
Full article in PDF
Debt is most often associated with younger adults, as they borrow to finance their education and purchase housing and vehicles. However, research from the United States suggests that an increasing number of people, both in pre-retirement and retired, are now struggling with debt, as both the percentage with debt and debt levels have risen for those age 55 and over (Copeland 2009, and Draut and McGhee 2004). This paper investigates the debt-holding situation of older Canadians.
With funding for retirement shifting onto the individual,1 most Canadians believe they must take an active role in planning for the event. Indeed, among those age 25 to 64, 81% reported they were preparing financially for their retirement. However, only 46% of those preparing for retirement knew how much savings they would need to maintain their standard of living (Schellenberg and Ostrovsky 2010).
Debt management is a recurring theme in retirement planning literature. Debt may be problematic for older workers if not paid off before retirement since repayment can be more difficult on a reduced income. On the other hand, carrying debt into retirement may not necessarily be an issue if repayment is manageable and the household is financially sound.
Financial planning is particularly important for women for several reasons. Compared with men, women have a longer life expectancy and they usually retire having spent fewer years in the labour market with less earned and saved (Glass and Kilpatrick 1998, and Marshall 2000). Divorce or widowhood during this period can also affect economic well-being.
The 2009 Canadian Financial Capability Survey (CFCS) provides information on the income, wealth and debt of retired Canadians, combined with self-assessments of their financial situation and indicators of financial literacy. This article examines the proportion, type, and level of debt among almost 5 million retirees age 55 and over (see Data source and definitions). Logistic regression is used to examine the personal, demographic and economic factors that influence the likelihood of carrying consumer or mortgage debt in retirement. The financial situation of indebted retirees is also examined along with three indicators of financial security among retirees.
Debt can include mortgages; student, payday or other loans; outstanding balances on credit cards or lines of credit; or any other unpaid debt or liability. Among those age 55 and over, one-third of the retired and two-thirds of the not-yet-retired report having some form of debt (Chart A). The proportion holding debt in retirement is about one-third for both fully retired couples (where both partners are retired) and for people living on their own. The proportion of couples with debt is higher when one spouse is retired (61%) or when neither spouse is retired (75%). Research has found that dual-earner couples tend to have higher average debt per person, relative to income—possibly due to their sense of security from having two incomes (Toronto–Dominion (TD) Economics 2010).
Those retired and in debt owe a median amount of $19,000 compared with their non-retired counterparts who owe $40,0002 (Chart B). Retired singles owe less than fully retired couples—$13,000 versus $20,000. Not-retired couples were the most likely to have some form of debt and they also had the highest median value owing ($50,000).
The remainder of the paper focuses on retired individuals and couples where both partners are retired. Of those owing money, the type and level of debt varies little between those in a couple and those on their own. Roughly 60% of retired couples and singles carry consumer debt only, with the remaining 40% almost equally split between those with mortgage debt only and those with both forms of debt (Table 1).
Among retirees, average debt was $60,000, while the median (the value where half owe more and half owe less) was $19,000. This large difference between the median and the mean is characteristic of a skewed distribution: one wherein a small group carries a high debt load while most owe smaller amounts.
Of retirees with debt, 1 in 4 owes less than $5,000. Debt in this range may simply be related to using credit as a convenience or as promotional financing. For example, some big-ticket items can be purchased on credit with no payments or interest for up to one year.3 One-third of households with debt owe between $5,000 and $24,999, and another one-quarter owe between $25,000 and $99,999. The remaining 17% of households carry a debt of $100,000 or more.
This section assesses which socio-economic and demographic factors are associated with the likelihood of carrying mortgage or consumer debt. Age is a primary factor since it reflects the evolution of personal finances over the life cycle. Level of education, sex, family and immigrant status, urban living and region are included as they may be associated with different housing choices and costs, as well as variations in personal financial behaviours. Finally, income, net worth and home ownership may also be linked to the incidence of debt, either positively, as income and collateral enable further borrowing, or negatively, since they may allow individuals to pay down debt faster or avoid it altogether.4
In order to determine which factors better predict the incidence of holding debt, the variables mentioned above are included in a logistic regression model. This technique estimates the relationship of each variable with the probability of having debt while holding all other variables constant. Descriptive statistics and regression results of the explanatory variables appear in Table 2.
One possible factor associated with holding debt in retirement is age. The further along the life cycle, the more time someone has had to repay any outstanding debts. Indeed, among retirees age 75 and over, only 20% had some form of debt, compared with 48% of those age 55 to 64 (Table 2). Older retirees were found to be significantly less likely to carry debt even after controls for other factors in the regression model were applied.
As the level of schooling goes up, so does the incidence of holding debt. While 26% of those with less than a high school diploma had some form of debt, 36% or more with at least a high school diploma had debt. The regression results indicate that only the difference between retirees without a high school education and those with high school or some postsecondary schooling was statistically significant. Other research has shown that higher education is associated with an open or positive attitude towards borrowing (Lee et al. 2007).
As a person's financial knowledge score increases, so does the likelihood of having debt (scores were calculated by adding the number of correct answers reported on a 14-question financial knowledge quiz). Understanding of financial concepts may be associated with borrowing to finance investments or smooth consumption.
Of the almost 5 million retirees in the study, 3.0 million lived with a spouse, 1.2 million were widowed, 500,000 were divorced and 300,000 had never married. Divorcees, at 43%, had the highest rate of holding debt followed by 35% for couples, 30% for the never-married and 28% for widows/widowers. Retired divorcees were found to be significantly more likely (1.3 times) to carry debt than those living in a couple after controls for other factors were applied. The financial cost of divorce, including legal fees, the division of properties and assets, and the setting up of separate households, has short-term economic consequences for both partners, which likely contributes to increased borrowing.5
Among retirees, 8 in 10 own a home. Homeowners are 1.4 times more likely to hold debt than non-owners, but not simply because of a mortgage. Overall, 37% of homeowners carry some debt, including 9% who have only a mortgage, 18% who have only consumer debt, and 10% who have both a mortgage and consumer debt. Overall, 28% of homeowners have some consumer debt compared with 23% of non-owners. A similar association was found among older Americans: "As expected, having mortgage debt was a significant and positive factor in the probability of holding consumer debt" (Lee et al. 2007, p. 316). Although houses can be expensive to maintain, homeowners have invested in a valuable, often-appreciating asset. Furthermore, there are financial gains in mortgage-free home ownership since rent does not have to be paid.
Individuals in households with an annual income of less than $25,000 were less likely to hold some form of debt compared with those in higher income groups. However, high net worth was also associated with a lower probability of debt.6 After controls for other factors were applied, retirees with a household net worth of $400,000 or more were found to be less likely to hold debt than the reference group with a net worth of less than $75,000.
Several other demographic and geographic factors were not significantly related to debt among retirees, including immigrant status, region and Census Metropolitan Area.
Holding debt is often assumed to affect financial security. One way to determine perceptions of financial well-being is to assess responses to questions about personal finances. The next section examines the personal and financial characteristics of all retirees in conjunction with indicators of financial security (see Data source and definitions).
From a subjective point of view, the vast majority of retirees give a positive report when asked about their economic well-being. Almost 8 in 10 believe that their financial situation is as expected or better than before they retired; similarly, 86% say their income is sufficient to cover monthly expenses; and 82% report that keeping up with bills and other financial commitments is not a problem (Table 3). When all personal and financial characteristics of retirees are considered together in logistic regression models, certain variables consistently increase or decrease the positive response rate to at least two, and usually all three, financial security questions.
Retirees age 75 and over are significantly more likely to report positively on their economic situation than those age 55 to 64 even after controls for variables such as level of income and debt are applied. For example, 83% of the older cohort felt they were financially as well off as or better than expected before retirement compared with 76% of the younger group. Similarly, the older cohort is 1.3 times more likely than the younger one to say their income is sufficient for their expenses. These findings may be related to the change in consumption and spending patterns as people age (Chawla 2005).
Family status is also a factor related to financial security. As opposed to those who never married or are widowed, the divorced are the only group to be significantly less likely than couples to say their income is sufficient for monthly expenses. Their relatively lower reporting of financial security may be associated with their overall financial situation. Although all categories of individuals not living with a partner were financially less well-off than couples, the divorced had the lowest median annual income and lowest median net worth7 (see Wealth, income and debt indicators). The findings suggest that marital dissolution by divorce may have a long-term negative effect on financial security in retirement. "Whether divorce interrupts the savings process or destroys assets, it is unlikely that most individuals will be able to save enough in later life to overcome the loss" (Fethke 1989, p. S121).
Overall, 82% of retirees without debt report their financial situation to be as expected or better than before retirement, whereas only 70% of those with debt report the same outcome. Similarly, 89% of debt-free retirees claim that their income is sufficient for monthly expenses, compared to 79% or less for those with any level of debt. Also, keeping up with bills and other financial commitments is not a problem for 89% of debt-free retirees, a figure that falls to 68% for those with any amount of debt.
Moreover, all levels of debt appear to significantly affect perceptions of financial vulnerability. Debt repayment amounts vary depending on repayment schedules and interest rates, and can be independent of the total amount owing. In other words, it is not necessarily the size of debt that has the potential to strain a monthly budget, but the repayment amount in relation to other expenses and income. However, even if debt repayments are manageable, the monthly financial obligation and outstanding balance may increase the perception of financial insecurity.
Retirees with less than $25,000 in annual household income are significantly less likely than those with higher incomes to give positive responses to the three financial security questions. After controls were applied for net wealth, having a lower income (whether with or without debt) was found to increase the perception of financial insecurity. For example, 75% of those with a household income of less than $25,000 say their income is sufficient to cover monthly expenses compared with 85% of those in the $25,000 to $49,000 income range.
Similarly, retirees with a household net worth of under $75,000 are not as likely to express a high rate of financial security as those with higher levels of net worth.
Using data from the Canadian Financial Capability Survey, this study found that, in 2009, 1 in 3 retired individuals age 55 and over, whether single or in a couple, held mortgage or consumer debt. Since retirement usually coincides with a drop in income and an increased reliance on savings, debt management is a frequently cited component of retirement planning.
The median amount owing for retirees with debt was $19,000. At the high end of the debt scale, 17% owed $100,000 or more. The likelihood of holding debt decreased with age but increased with household income and financial knowledge. Individuals with some postsecondary education were more likely to hold debt than those with less schooling, while households with a high net worth were less likely to have debt. Being divorced was a strong correlate of holding debt among retirees.
The majority of retirees report that their finances are what they had expected them to be prior to retirement, that their income is sufficient to cover expenses, and that they are able to stay on top of bills and keep up with their financial commitments. After controls for personal and financial factors were applied, those with any level of debt were found to be more likely to respond negatively to these questions. For example, while 9 in 10 retirees without debt reported they had no trouble keeping up with bills and other financial commitments, 7 in 10 with debt reported this to be the case.
Roughly 1 in 10 retirees was divorced. This group had lower positive response rates for all three financial security questions, even after controls were applied for other factors, including debt. For example, 65% of the divorced reported that their financial situation was as good as or better than they expected before retirement, compared with rates of 77% or greater for those who were in a couple, were widowed or never married. The substantially lower-than-average income and net worth of the divorced coincide with their poorer perception of their financial condition.
Immigrants, as well as those having relatively low income or net worth, also report lower perceptions of financial security.
The incidence and level of debt among the pre-retired population age 55 and over were higher than among retirees. Two-thirds of pre-retirees carry mortgage or consumer debt with a median value of $40,000.
The Canadian Financial Capability Survey (CFCS), a new survey conducted between February and May 2009, collected information from Canadians age 18 and over in the 10 provinces. The content focuses on the financial situation of individuals and households as well as their financial knowledge, ability and behaviours concerning money management, budgeting and general financial planning and decisions. One of the goals of the survey is to understand Canadians' use of financial services and their knowledge of programs aimed at facilitating financial planning for retirement.
The target population for this article includes all respondents age 55 and over and who, when asked about their employment status, reported themselves to be "retired." Those who reported themselves to be not working and not looking for work and with no employment income in the past year were also counted as retired. If the respondent was part of a couple, his or her spouse would also have to report being retired in order to be considered living in a fully retired household. Approximately 3,730 respondents fit this definition, representing a weighted count of 4,869,000.
As noted above, retired is derived from self-reported information regarding employment status. Although households may report some of their income in the previous 12 months came from employment, for the purpose of this study their main activity remains "retired."
Respondents are considered to be in a couple relationship if they are currently married or in a common-law relationship and living with a partner, whereas single includes those who are divorced (including separated) or widowed, or those who never married, and are not living with a partner.
Assets are the monetary value of all personal or business goods owned, including, for example, real estate, vehicles, jewellery, stocks and bonds, registered retirement savings plans and savings in the bank. Unlike the Survey of Financial Security, the value of registered pension plans is not included among the assets. Respondents were asked to provide a total value for each type of asset and the combined total refers to a household's total assets. If just one of the individual asset questions is marked as "don't know," then the total household asset figure is marked as "not stated." For this reason, upwards of 50% of the total asset variable is unknown, which is a limitation of the survey. However, all calculations made using the asset, debt and net worth variables only use valid responses.
Debt is the amount the respondent and other family members still owe on mortgages; student, payday or other loans; outstanding balances on credit cards or lines of credit; or any other unpaid debt or liability. Although respondents were asked to identify different types of debt, only the value of all debts combined was collected. Debt was collected for the family as a whole since it cannot easily be assigned to just one person in the family. While most respondents knew whether they owned money, many were not able to provide a dollar figure for their total debt. Among the retired population, 21% of those with debt did not know the total value of their indebtedness.
Net worth is calculated by subtracting total debt from total assets. Since both the asset and debt variables had considerable proportions of "not stated" responses, more than one-half of the responses for the net worth variable are unknown.
The following three questions on financial self-assessment, appropriate for retirees and included in the CFCS, were selected to assess household finances and were used as indicators of financial security:
The survey also included a 14-question section to objectively assess an individual's financial knowledge. Scores were calculated by adding up the number of correct answers.
Logistic regression models were used to examine the probability of the retired having mortgage and/or consumer debt and the probability of giving a positive response to each of three questions pertaining to financial security. In order to retain as large a sample as possible, dummy "not stated" categories were created for any values missing from the annual household income and household net worth variables. Multicollinearity diagnostic tests were run for all models. Although test statistics indicate some correlation between household income and net worth, both were left in the models because of the problem with missing values. Alternative models were estimated, each including only wealth or income. The results remained the same when the variables were included separately in the models or when they were both in the models together. The entire analysis used 250 bootstrap weights to adjust for the survey design and produce a more accurate variance estimation.
The Canadian Financial Capability Survey also collected information on the wealth of respondents, although the rate of non-response was high (see Data source and definitions). This box presents information on the subpopulation of retirees with debt.
Overall, retirees with debt have a median annual household income of $42,000, a median net worth of $295,000, and a median debt of $19,000 (Table 4). Other indicators that help put household debt in perspective include the median debt-to-income ratio (D/I ratio) and the median debt-to-assets ratio (D/A ratio). The overall D/I for retirees is 0.47 and the D/A is 0.07. Households with high D/I ratios may have higher debt repayments, relative to income, compared with households with low D/I ratios.8 On the other hand, those with low D/A ratios are assumed to be more financially secure than those with high ratios.
Noteworthy findings from the financial indicators table include:
Chawla, Raj K. 2005. "Shifts in spending patterns of older Canadians." Perspectives on Labour and Income. Vol. 6, no. 12. December. Statistics Canada Catalogue no. 75-001-XIE. p. 17-28. (accessed March 22, 2011).
Copeland, Craig. 2009. "Debt of the elderly and near elderly, 1992–2007." Employee Benefit Research Institute (EBRI) Notes. Vol. 30, no. 10. October. Washington, D.C. Employee Benefit Research Institute. p. 2-14. (accessed March 24, 2011).
Draut, Tamara and Heather C. McGhee. 2004. Retiring in the Red: The Growth of Debt Among Older Americans. Briefing Paper. New York. Dēmos. 11 p. (accessed March 24, 2011).
Fethke, Carol C. 1989. "Life-cycle models of saving and the effect of the timing of divorce on retirement economic well-being." The Journal of Gerontology. Vol. 44, no. 3. May. p. S121-S128.
Glass, J. Conrad, Jr. and Beverly B. Kilpatrick. 1998. "Financial planning for retirement: An imperative for baby boomer women." Educational Gerontology. Vol. 24, no. 6. September. p. 595-617.
Lee, Yoon G., Jean M. Lown and Deanna L. Sharpe. 2007. "Predictors of holding consumer and mortgage debt among older Americans." Journal of Family and Economic Issues. Vol. 28, no. 2. p. 305-320.
Marshall, Katherine. 2000. "Incomes of younger retired women: The past 30 years." Perspectives on Labour and Income. Vol. 12, no. 4. Winter. Statistics Canada Catalogue no. 75-001-XPE. p. 9-17. (accessed March 22, 2011).
Schellenberg, Grant and Yuri Ostrovsky. 2010. Retirement-related Highlights from the 2009 Canadian Financial Capability Survey. Statistics Canada Catalogue no. 11-624-M – No. 026. Canadian Economy Research Paper Series. Ottawa. 20 p. (accessed March 22, 2011).
Wilmoth, Janet and Gregor Koso. 2002. "Does marital history matter? Marital status and wealth outcomes among preretirement adults." Journal of Marriage and Family. Vol. 64, no. 1. February. p. 254-268.
Katherine Marshall is with the Labour Statistics Division. She can be reached at 613-951-6890 or at firstname.lastname@example.org.