Canadian Retirement Planning Behavior: How Reliance on a Government Pension Relates to Pre-Retirement Savings and Planning Behaviours

The Canadian Retirement System comprises three pillars: Old Age Security (OAS), the Canada Pension Plan (CPP), the Quebec Pension Plan (QPP), and employer-sponsored pension plans. Old Age Security (OAS) is a non-contributory plan, funded by general tax revenue with age and residency as the only eligibility requirements. OAS offers some additional benefits for very low-income Canadians to alleviate poverty. The CPP and QPP are contributory plans that are administered by federal and provincial governments. Quebec residents do not receive the CPP, but instead are entitled to the QPP. Both employees and employers pay into the plan through payroll deductions. To be eligible to receive the CPP/QPP pension you must contribute. The benefit is determined by a formula that considers the length of the contribution period, amount contributed, and age of the contributor. Average CPP 2023, is $717.00 per month. Canadians are eligible to draw CPP as early as age 60 (actuarially reduced) and as late as age 70 (actuarially increased). Maximum OAS is $687.00 (which most Canadians receive) and can be drawn as early as age 65 or delayed until age 70 (actuarially increased). Combined, these benefits provide a total of $1,404 per month or $16,848 per year, while the median retirement income need in Canada, 2023 is $55,911 (StatsCan 2023).

Employers can offer defined benefit defined contribution plans, deferred profit-sharing plans, as well as group registered retirement savings plans (RRSPs). Defined benefit (DB) plans and defined contribution (DC) plans work similarly to those in the U.S. DB plans provide a predetermined benefit amount at retirement based on years of service, contribution limit, and age of plan member. Employees may not have to make contributions, but employers must contribute 50% of the required plan contributions. Defined contribution pension plans have a pre-determined contribution amount or formula that is usually matched by the employer; employers do not commit to providing a guaranteed pension amount. Plan members are prohibited from making cash withdrawals while employed.

In the private sector, approximately 6.6 million Canadians were covered by pension plans in 2020 (Mercer 2022). Pension plans in Canada represent approximately 23% of the average pre-retirement income needed, while CPP and OAS account for approximately 26% of the medium retirement income of $65,300 (Statscan 2023). These three pillars account for approximately 49% of the median retirement income of Canadians, meaning Canadians need to save independently of government and employer retirement plans to maintain their standard of living in retirement (Statscan 2023). Similar findings are echoed by Clavet et al., (2022) who find public pensions and mandatory retirement accounts replace 46% of average pre-retirement income. It is difficult to know how much income is needed in retirement as many pre-retirement expenses are not needed or are lowered in retirement. Most Canadians (78%) believe they have sufficient retirement income when accounting for all sources of retirement income. The portion of income replaced by public pension plans put Canada 23rd out of 35 OECD countries in 2017 replacing 41% of average earnings compared to 48.4% on average for OECD countries (OECD, 2017).

To measure retirement preparedness, the National Retirement Risk Index (NRRI) is constructed using the Survey of Consumer Finance (SCF) every three years to measure the share of American households at risk of being unable to maintain pre-retirement standards of living in retirement. This percentage has typically been around 50% (Munnell, et al. 2021). MacDonald et al. (2011) constructed a Canadian NRRI using a similar methodology and found that even Canadians doing moderately or well economically were more at risk to have insufficient resources after retirement. There are many assumptions that must be made when calculating how much one needs in retirement, such as stock market performance, home value, and interest rates. These factors are often out of one’s control, but there are actions one can take to improve retirement preparedness, such as voluntarily saving and calculating retirement needs.

Non-retirees with retirement savings are more likely to perceive their retirement savings as adequate. In one study, those who had DC plans, DB plans, IRAs, or other savings were more likely to perceive their savings were adequate compared to those without retirement savings (Lim & Lee 2021). In looking at how people acted on these perceptions, Santen (2019) explored how uncertainty in future pension benefits affected household savings and found that a one dollar-increase in pension benefits resulted in a 32-cent decrease in household savings. In contrast, Witvorapong et al. (2021) found that, in Thailand, those that expected the government to provide the primary income source in retirement did not alter their savings patterns. It should be noted that Thai government assistance tends to be small.

Participation in DB and other government pension plans is often automatic and mandatory and the reduction in take-home pay may lead to some financial distress. Automatic enrollment in the U.S. Army’s Thrift Savings Plan did not increase non-mortgage debt, change credit scores, or cause adverse credit outcomes (Beshears et al. 2022). There is, however, evidence of a crowding-out effect. Parada-Contzen (2020) studied how mandatory retirement savings plans crowded out other forms of investments including homeownership and non-retirement savings. In this sense, homeownership and participation in retirement plans are substitute goods.

Individuals automatically enrolled in DC plans are less likely to make additional voluntary savings and contribute less than voluntary participants (Burtica and Karamcheva 2019). Andersen and Bhattacharya (2021) looked at retirement savings across several countries and found that for young people, compulsory pension savings reduced voluntary savings and led to increased debt. This could be due to the belief that mandatory contributions are sufficient or simply that income is finite, and a limited amount can be dedicated to retirement savings. However, Andersen and Bhattacharya (2021) found that after a certain mandatory savings rate, this substitution effect ceased, and additional mandatory savings led to a higher lifetime benefit.

Financial knowledge has many definitions and has been used interchangeably with financial literacy or capability (Xiao & Huang 2021). Individuals with high levels of financial knowledge may be better equipped with skills to plan and save for retirement. Financial knowledge or understanding recognizes that an individual must possess both financial knowledge and the ability to use personal finance-related information to be financially literate and effectively make financial decisions (Schuchardt et al., 2009; Huston, 2010). Objective financial knowledge pertains to understanding financial concepts and relates to a person’s ability to engage in effective financial decision-making (Lee et al. 2019). For example, Anderson, Baker, and Robinson (2017) found that financial literacy was positively related to financial behaviours such as planning for retirement and saving for emergencies. Objective financial literacy scores also corresponded to higher retirement account contribution rates and higher investment returns (Clark, Lusardi & Mitchell 2015; Lusardi, Michaud & Mitchell 2017).

Subjective financial knowledge concerns a person’s self-confidence relating to knowledge about and ability to make sound financial decisions (Lee et al. 2019). Subjective financial knowledge is associated with better savings behaviours and higher retirement plan account balances (Tang & Baker 2016). Anderson, Baker, and Robinson (2017) noted that the perception of one’s financial literacy was more important to financial decision-making in many instances than objective financial literacy. For example, among those with low objective financial literacy, their perception of their financial knowledge was more important in determining whether they planned for retirement than their actual knowledge score (Anderson, Baker & Robinson 2017). Increasing financial knowledge alone may not be enough to improve financial behaviour if individuals do not have confidence in their own abilities (Tang & Baker 2016), which helps explain attitude and behaviour gaps in practice.

Individual and familial characteristics have implications for retirement preparedness. Women tend to live longer, earn less, and be less prepared for retirement than men (Burn et al. 2020; Williams, Elizabeth & Spencer-Rodgers, 2010). Individuals nearer to retirement have different needs than those entering the workforce, especially when considering their saving behaviours (Gough & Niza 2011). Millennials’ retirement planning behaviour is driven by income, financial planner use, job tenure, and motivation to save for retirement (Yao & Cheng 2017). Younger individuals may not be making enough to put away money for retirement or be eligible to participate in employer plans (Cole & Liebenberg 2008). Education also seems to be positively related to retirement preparedness, perhaps because more education is typically associated with greater levels of income and wealth (Campos et al. 2016; Grishina 2019; Kaplan & Rauh 2013).

Indigenous Canadians (First Nations, Métis, Inuit) and immigrants (Canadian newcomers) face barriers to access relating to financial services and mainstream financial institutions, albeit the barriers may differ. Indigenous Canadians often reside in remote areas or on reserved lands severely limiting opportunities for financial education and literacy (Kremer & Mah, 2021). Canadian immigrants may not be eligible for the same retirement benefits that Canadians are entitled to, which may affect their primary source of retirement income and find it difficult to find well-paying jobs and establish economic security (Ahmed et al., 2016). Indigenous Canadians face limited job opportunities and higher costs of goods and services if living in rural or remote locations(Kremer & Mah, 2021). Most of the research surrounding Indigenous people in Canada has focused primarily on health (e.g., Horrill et al., 2018; Nelson & Wilson, 2017), but more research is needed to understand how both immigrant and Indigenous populations are preparing for retirement.

Household composition may dictate both saving motivations and constraints. For example, marital status and the presence of children may affect one’s ability and motivation to prepare for retirement. Married individuals tend to have economic advantages when compared to single individuals including having dual incomes and/or reduced childcare costs, both of which can be advantageous in saving for retirement (Joo & Grable, 2005; Zissimopoulos et al., 2015). The effects of marital status on retirement preparedness have often been shown to be influenced by gender. Single women accumulate significantly less wealth than single men (Gornick & Sierminska, 2021) and widows tend to be less likely to be prepared for retirement than widowers (Stone & Neumann, 2012). The presence of children is also an important consideration, as children represent a large economic investment. Segel-Karpas and Werner (2014) found that being married and having children were negatively related to financial preparedness.

Individuals smooth consumption by borrowing when income is low, saving when income is high, and spending assets during retirement (Ando & Modigliani, 1963). Lowered needs for income in retirement relate to decreased needs for savings. In this way, public benefits can crowd out present savings. Fisher’s (1930) intertemporal choice model suggests that rational consumers make decisions now that will lead to maximum lifetime satisfaction. Read, Olivola and Hardisty (2017) applied intertemporal choice decision-making to how people weigh current versus future utility. They found that people viewed the opportunity cost of saving for larger future rewards as greater than that of current, smaller rewards and therefore were likely to choose current consumption over future consumption (and savings). Bleichrodt et al. (2015) discussed how different discount models can be applied to intertemporal choice. When weighing present versus future consumption, a discount rate applied to future utility determines present value which is compared to the present value of other options. Over-discounting the future value of present savings could lead to a retirement shortfall. Expectation of guaranteed retirement benefits in the future may increase discounting of future utility brought on by additional savings and therefore decrease savings beyond mandatory pension contributions.

Some argue those who are young or have lower incomes should not save for retirement at all. Scott et al. (2022) applied the life-cycle model to retirement savings and concluded that younger people should take care of their current needs and save for retirement when they are older and have a higher income. For example, in the U.S., Social Security may provide adequate income replacement for lower-income earners, further arguing in favour of current consumption. This attitude may include higher rates of discounting the utility of future income as they choose higher consumption today and lower income in retirement (Hanna & Lindamood, 2010). Those with lower incomes may rely more on public pensions due to reduced means to save and the fact they have a higher percentage of income replaced (90% for individuals classified as “lower income”) by public Canadian pensions as average lifetime earnings drop (Clavet et al., 2022; Horner, 2009).

Significant research has been conducted on the relationship between public retirement systems, retirement preparedness, and financial knowledge. However, little research exists on this impact in Canada, and this paper attempts to bridge that gap. This paper explores how the expectation to rely on a government pension as the primary source of retirement income affects other retirement preparedness behaviours, when controlling for financial knowledge, demographics, and socioeconomic status.. Understanding how individuals’ expectations may influence savings and planning informs policy makers, practitioners, and individuals on the importance of concretizing future plans to make apparent gaps that require changes to current behaviours.

The Canadian Financial Capability Study (CFCS) is funded by the Financial Consumer Agency of Canada and conducted by EKOS Research Associates. The purpose of the study is to better understand Canadians’ financial knowledge, skills, confidence, and well-being. The 2019 data had 7,169 responses that were collected online and by telephone. The focus of this study is retirement preparation and therefore excludes retired individuals (both those who are not working and reported they are still working) and individuals who responded, “do not know” or “prefer not to say”, reducing the sample to 4,047. To create a more representative sample of Canadian residents, oversampling for Indigenous and new Canadians was performed. Sample weights are then applied to align more closely with the 2016 census figures and make the sample more representative of Canadians.

The primary focus of this study is pre-retirement financial behaviours. There are two financial behaviours that will be used as dichotomous dependent variables in this analysis; saving for retirement and calculated retirement needs. Savings for retirement is the response to the question: “Are you financially preparing for your retirement either on your own or through an employer pension plan?” Respondents were provided with four possible choices: (1) Yes, (2) No, (3) Don’t know, and (4) Prefer not to say. Calculated retirement need was measured by asking respondents “Do you have a good idea of how much money you will need to save to maintain your desired standard of living when you retire?” Respondents were provided with four possible choices: (1) Yes, (2) No, (3) Don’t know, and (4) Prefer not to say.

Government pension reliance. The main explanatory variable — government pension reliance, is measured by asking respondents, “What do you think will be your primary source of income in retirement?” Individuals can select government pension benefits, workplace pension plans, retirement savings, taxable investments, residence, business or rental income, continued work, medical or disability pension, none of these, don’t know, prefer not to say, or other (specify). Those who selected government pension benefits were coded as “yes”, don’t know”, and “prefer not to say” were left the same; and all others were coded as “no.”

Financial knowledge. Objective financial knowledge is measured by asking three questions about inflation, life insurance needs, and leveraging debt. Individuals with higher scores on the first three questions are asked harder questions for the remaining four. The final four questions were selected from questions on comparing unit pricing of groceries, credit ratings, cost of housing, ATM cards, saving university money, inflation, credit report, and stock market insurance. From these seven responses, EKOS Research Associates created a score for financial literacy that ranges from -1.1703304 to 0.9150032. This measure is treated as a continuous variable. Subjective financial knowledge is a self-assessed measure where respondents respond to the question “How would you rate your level of financial knowledge?” Options for response include very knowledgeable, knowledgeable, fairly knowledgeable, not very knowledgeable, don’t know, and prefer not to say.

Demographics. Age is provided in the ranges of 18-24, 25-34, 35-44, 45-54, 55-64, and 65 and up. Gender is measured by asking respondents “Are you.…” to which they can respond male, female, prefer to self-describe, or prefer not to say. Immigration is measured by whether the respondents answered yes or no to the question “Were you born in Canada?” Finally, those who answered yes to “Are you an Indigenous person, that is, First Nations, Métis or Inuk (Inuit)?” are considered Indigenous persons. Marital status is measured by asking “What is your current marital status?” with married, living with a partner (common-law), separated, divorced, widowed, single (never married), don’t know, and prefer not to say. Financially dependent children was a binary variable that included whether the respondent was financially responsible for at least one child or no children.

Socioeconomic status. Household income is measured by asking “What was your approximate total household income in 2018?” Categories increase by increments of $20,000 up to $100,000 and then $100,000 to $150,000, $150,000 to $200,000, and $200,000 or more as the final income range. Don’t know and prefer not to say were also included as options. Education is measured by asking “What is the higher level of schooling that you have obtained?” with less than high school, high school or equivalent, some college, without degree, college diploma or certificate (includes completed trade, vocational, or technical school), university undergraduate degree, university graduate degree, don’t know, and prefer not to say. Employment is measured by whether respondents are employed (full- or part-time, self-employed, retired (and still working or not working), not working, don’t know and prefer not to say. Those who are retired were excluded from the sample, so only employed, or not working respondents are included in the analysis. Homeownership is measured by whether respondents own with a mortgage, own without a mortgage, or rent. Student loans have also been of interest to researchers to determine whether student loan balances inhibit saving behaviours. Therefore, student loans have been made into a binary variable from a question that asks respondents what types of debt they have, where 1 = those who have student loans and 0 = those who do not.

As both dependent variables for each model are dichotomous, a probit model was chosen to test the association between government pension reliance and behaviours of saving for retirement and calculating retirement needs using the following probit model via maximum likelihood:

Saving for retirement & calculated retirement needs* = β0 + β1govpenj + β2subfin + β3IRTj + β4incomej + β5genderj + β6agej + β7maritalj + β8findepj + β9immigrantj + β10indigenousj + β11educationj + β12employmentj + β13homeownerj + β14studentloanj + 3

Where saving for retirement and calculated retirement needs* are the estimated likelihoods that the individual is saving for retirement or has calculated retirement income needs.

Table 1 lists the descriptive statistics for the variables of interest in this study. In the analytical sample, 67% of respondents reported relying on an income source other than a government pension as their primary expected source of income in retirement. Respondents appear to score highly on the objective financial literacy questions and only 19% of respondents said they were not very knowledgeable about finances. About 68% of the sample respondents were between 35 years old and 64 years old, and 52% of the respondents were male. Slightly less than 20% of respondents were born outside Canada, and only 7% reported being an Indigenous person. Most respondents were married or single, did not have children, and made less than $150,000 a year. Only 5% of respondents reported not having a high school education, and more than half the sample were employed fulltime. Finally, 65% said they did not own a home while 88% said they had student loans.

Table 2 shows the marginal effects and standard errors of both models. Model 1 is focused on whether individuals are currently saving for retirement. Those who expect their primary income source in retirement to be a government pension and who were not working full time had a decreased probability of saving for retirement income needs. Those with higher subjective financial knowledge, household income, older than 34 (as opposed to 18 to 24-year-olds), widowed (as opposed to married), and homeowners all had an increased probability of saving for retirement.

Model 2 focuses on whether individuals have an idea of how much money they will need to maintain their standard of living. Those who expect that a government pension will be their primary income source in retirement have a decreased probability of having an idea of how much money they will need. Individuals who had higher levels of subjective and objective financial knowledge, more household income, and were older than 44 (as opposed to 18 to 24-year-olds) had an increased probability of having an idea of how much money they will need in retirement. Interestingly, those who were widowed also had a higher probability of having an idea of how much money they will need in retirement. Perhaps the financial stress of losing a spouse and/or the acquisition of investment and life insurance money can lead to the desire to develop a financial plan.