Michael Hatfield, Social Development Canada
Michael Hatfield is Senior Economist, Knowledge Synthesis, Liaison and Applications at Social Development Canada.
Experiencing low income over several years can mean exclusion from meaningful participation in community life, long-term mental and physical health risks, living in bad housing, and restricted opportunities for the children whose parents must cope with inadequate incomes on a persistent basis.
As many of the findings in this article indicate, exclusion and persistent low income often reinforce one another.1 Many of the factors associated with persistent low income reflect absent, disrupted, or ineffective social networks, particularly those types of networks that enable access to earnings from stable paid employment.
This article focuses on five large groups among the non-elderly population that were highly vulnerable to persistent low income over the 1996 to 2001 period. It begins by defining persistent low income and introducing the five high-risk groups. It then summarizes the main findings of recent published and unpublished research studies of these groups with the following questions in mind.
Why are members of the highrisk groups so disproportionately vulnerable to persistent low income?
Why do most members of those groups avoid persistent low income while large minorities within each group experience it over several years?2
FIGURE 1 Percentage of Population 0-59 in 1996 Experiencing Persistent Low Income between 1996 and 2001
* Aged 45-64 between 1996 and 2002.
Source: Survey of Labour and Income Dynamics Main File.
FIGURE 2 Percentage of Persons Aged 0-64 in Low Income in 1996 – Group Membership Based on Main Income Recipient
Persistent Low Income in the Context of Overall Low Income
There are three dimensions of low income: incidence, depth, and duration. The low-income rate is the percentage of a population experiencing low income in a specified period of time (e.g., one year). The depth of low income measures how far below the low-income threshold the income of a low-income household falls. The persistence of low income, the focus of this article, tells us what percentage of the population experiences low income over a period of years.
In 1996, 14.7% of the total population under age 65 was considered to be in low income using Statistics Canada’s post-transfer, post-income tax lowincome cut-offs. However, using the definition of persistent low income described in endnote one, 8.0% of the population aged 0-59 in 1996 experienced low income persistently over the years from 1996 to 2001.3
Some groups in the non-elderly population were far more vulnerable to low income both on an annual and on a persistent basis. As shown in Figure 1, the five groups among the non-elderly population most vulnerable to persistent low income between 1996 and 2001 were4:
unattached individuals aged 45-64;
persons with a work-limiting disability;
lone parents; and
Aboriginal people living off reserves.
Persons in at least one of these five groups accounted for 23.8% of the total population aged 0-59 in 1996 but for 67.6% of those experiencing persistent low income over the next six years.
Over this period, members of these groups were 4.6 to 8.6 times more likely to experience long-term low income than the rest of those aged 0-59 in 1996.
As Figure 2 shows, these five groups also had very high annual rates of low income in 1996, but those rates were only 3.7 to 5.4 times those of the rest of the non-elderly population, and they accounted for only 57.6% of the total non-elderly low-income population. Thus, persistent low income was more concentrated among the five groups than annual low income.
A few paragraphs are now in order to explain why the high-risk groups are defined as they are.
Statistics Canada does not estimate low-income cut-offs for reserves, and their Survey of Labour and Income Dynamics, the source of income data for this article, does not survey Aboriginal people living on reserves. Thus, it is not possible to make a meaningful estimate of persistent low income for all Aboriginal people or for Aboriginal people living on reserves. Those Aboriginal people living on reserves almost certainly have a higher incidence of persistent low income than self-identified Aboriginal people living off reserves. According to 2001 Census data, Registered Indians living on reserves had median individual incomes 20% lower than those living off reserves.
Persons over age 65 are also not on this list of those most susceptible to persistent low income. The annual low-income rate for the elderly group as a whole in 2002, the latest year for which data are available, was relatively low (6.9%, as opposed to 9.7% for persons aged 18-64 and 10.2% for children under age 18).5 Even the worst-off sub-group, elderly unattached persons, had a 1996 low-income rate 13.9 percentage points lower than for unattached persons aged 45-64 (24.0%).
Other groups one might have expected to find on such a list include young people, women, persons who have not graduated from high school, and persons belonging to a visible minority group.
One category of youth – full-time, post-secondary students – experiences disproportionate rates of persistent low income. However, these youths are not considered a high-risk group in this article. Full-time, post-secondary students are not likely to continue to experience long-term low income once they complete their studies. Moreover, there is good reason to believe their incidence of low income while fulltime students is inflated by the fact that two of their main sources of income – student loans and gifts from their parents – are not counted as income in the SLID. While youths who have completed their education may have high rates of low income in any given year, the vast majority experience steady growth in their earnings and soon escape from low income.
The other groups either did not experience a disproportionate rate of persistent low income over the 1996 to 2001 period or experienced it, because they also belong to one of the high-risk groups already identified. For example, persons who have not graduated from high school have high rates of persistent low income. But the less well-educated among the non-elderly who do not belong to a high-risk group do not have excessively high rates of persistent low income.6
The 30% of visible minority Canadians who were born in Canada do not experience disproportionately high rates of persistent low income. However, those who are recent immigrants (33.7% of all visible minority Canadians in 2001) are much more likely than other recent immigrants to do so.7 This article thus considers them as a sub-group of recent immigrants, rather than as visible minorities.
Recent Studies of the High- Risk Groups and Key Findings
While many important studies of members of the high-risk groups have been published and are available to the public, others remain unpublished. One purpose of this article is to bring key findings of both sets of studies to a wider audience.
Two important recently published studies of these groups are by analysts at Statistics Canada: The Rise in Low- Income Rates Among Immigrants in Canada by Garnett Picot and Feng Hou (June, 2003), and “Groups at High Risk of Being in Low Income” in Low Income in Census Metropolitan Areas, 1980-2000 by Andrew Heisz and Logan McLeod (April, 2004). A third study, Economic Performance of Off- Reserve Aboriginal Canadians: A Study of Groups at Risk of Social Exclusion, authored by Dominque Fleury of the Social Policy Research Group in Social Development Canada, was published in 2002 in the Applied Research Branch working paper series.
Challenging Some Myths About the Rise in Low Income Among Recent Immigrants
The Picot-Hou paper uses census data to document a secular rise in low-income rates among recent immigrants between 1980 and 2000 and 1990 and 2000.8 All three years were at or near peaks in the business cycle. It had been commonly believed that this trend resulted from changes in the source countries, official language skills, and suitability for the Canadian labour market of more recent immigrant cohorts. But Picot and Hou found that the increase was only partially explained by such factors. Most of it resulted from a rise in low-income rates among recent immigrants from all age groups, family types, language groups, education levels, and most source regions.
Two possible explanations require further research. The first is that a progressively lower value has been placed on the work experience abroad of highly educated immigrants by Canadian employers. The second is that the large numbers of highly educated Canadian-born women who entered the paid labour market after 1980 may have reduced the employer demand for recent immigrants. Picot and Hou point out that the education levels of recent immigrants rose rapidly in the 1990s, and that they were increasingly likely to enter Canada as members of the economic immigrant class, the group that traditionally has had the best labour market outcomes.
A sharp decline in low-income rates among almost all recent immigrant sub-groups was registered during the economic recovery between 1995 and 2000, but this only partly offset the sharp rise in low-income rates for all categories of recent immigrants between 1990 and 1995.
Trends Are Diverging for Immigrants and Lone Parents in Large Urban Centres
The Heisz-McLeod paper confirmed that the low-income rate among recent immigrants in Canada’s 27 census metropolitan areas (CMAs consist of the largest cities and their suburbs) rose from 23.1% in 1980 to 35.0% in 2000, while it fell for all other residents of these areas from 18.0% to 16.0%.9 The low-income rate among persons in lone-parent families living in Canada’s large urban centres moved in the opposite direction falling from 54.2% in 1980 to 46.6% in 2000.
Steady Paid Employment the Key Factor for Off-Reserve Aboriginal People in Avoiding Persistent Low Income
The Fleury study examined why Aboriginal people living off reserves, of all the high-risk groups, had the lowest rate of annual and persistent low income. Using data from the first longitudinal panel (1993-1998) of the Survey of Labour and Income Dynamics, she investigated four potential explanations for this outcome. Three of these explanations – higher educational levels, higher exit rates from their high-risk group, and membership in only one high-risk group – had to be rejected because Aboriginal people living off reserves did no better or worse than the other high-risk groups in these areas.
|High-Risk Group||1998 Low-Income Rate – No Paid Work |
|1998 Low-Income Rate – Any Paid Work* |
|1998 Low-Income Rate – All Persons 16-64 |
|Persons with Work-limiting disabilities||Not Available||Not Available||27|
|Aboriginal people living off reserve||43||13||23|
|All persons 16-64||Not Available||Not Available||12|
* Any paid work means at least one hour of paid work in the year.
Source: Kapsalis and Tourigny, 2002, 2003a, 2003b, 2004a, 2004b.
However, Fleury found that off-reserve Aboriginal people were more likely than all other high-risk groups to have stable paid employment. This single advantage explained their lower incidence of persistent low income.
The main sources for the following analysis are a series of detailed studies on each of the high-risk groups using regression analysis done by Constantine Kapsalis and Pierre Tourigny under contract to the Income Security and Labour Market Studies group of the Applied Research Branch at Human Resources Development Canada.10
Avoiding Persistent Low Income
a) Have Steady Attachment to Paid Work
Not surprisingly, as Table 1 illustrates, Kapsalis and Tourigny confirmed Fleury’s finding that the most important factor associated with avoiding low income for members of all the high-risk groups was attachment to paid work.
With the exception of lone parents, the low-income rate for persons in the high-risk groups examined by Kapsalis and Tourigny who had any paid work in 1998 was similar to that for all persons. Even among lone parents, the low-income rate for those who had at least 750 hours or more of paid work during the year was only 15%.
Low hourly wages for the primary adult earner in a household are also a significant factor in explaining the low income of many households. Nevertheless, Kapsalis and Tourigny demonstrated that for the lone-parent group, low or no annual hours of paid work are the predominant explanation for high annual low-income rates. Low hourly wages and low annual hours of paid work are also correlated. Most low-paying jobs are precarious, parttime, temporary, or part-year. Overall, members of the high-risk groups with a steady attachment to paid work were no more likely to experience persistent low income than other Canadians.
This naturally leads to investigation of other factors correlated with persistent low income that might be connected to the lack of steady paid employment among significant minorities of each high-risk group. These are examined later in the article.
b) Exit from the High-Risk Group
Another important way of reducing the odds of experiencing persistent low income is to leave one’s high-risk group. As Table 2 illustrates, members of some high-risk groups between 1996 and 2001 were more likely to be able to do this than were others.
c) Draw on a Spouse for Support
Having at least two potential adult earners in a family is clearly an advantage in avoiding persistent low income. By definition, this is not possible for lone parents and the unattached aged 45-64. Having a second potential adult earner is particularly important for persons with work-limiting disabilities. On a cross-sectional basis in 1998, persons with a work-limiting disability had a low-income rate of 8% if they were not the main income recipient in their family and 43% if they were.
Adults aged 16-64 without a worklimiting disability had a low-income rate of 10%. Among Aboriginal people living off reserves, those who were unattached or lone parents had a predicted probability of being in low income in 1998 of over 60% compared to a probability of 24% for all persons in that group.12
Other family status characteristics associated with limited or disrupted family networks were also associated with persistent low income for groups not living with a spouse. Among lone parents, those with a dependent child under age five or who were not a partner in a marriage or a common-law union at the time of the birth of their first child were much more likely (4.4 times and 2.2 times respectively) to experience persistent low income than other lone parents. Among the unattached aged 45-59, those who had been in a union in the past were 28% more likely to experience low income than those who had never married or been in a common-law union.
d) Belong to Only One High-Risk Group
As would be expected, persons who belong to more than one high-risk group are extremely vulnerable to annual and persistent low income. Among off-reserve Aboriginal people in 1998, for example, the predicted incidence of low income (holding all other factors constant) was 56% if the person had a work-limiting disability and 60% if a lone mother. For all Aboriginal people living off reserves, the predicted incidence was only 24%. Similarly, among unattached individuals aged 45-59 in 1993, the incidence of persistent low income was 66% for those who also had a work-limiting disability in that year, but only 23% for those who did not.
e) Graduate from High School
Within all high-risk groups (except for Aboriginal people living off reserves), not graduating from high school was a very important variable identified by regression analysis as being correlated with persistent low income over the 1993-1998 period.13 (This finding is consistent with what we know about the growth in demand for high-skilled workers.)
|Aboriginal people living off-reserve||100|
|Persons with work-limiting disabilities||20.411|
Source: Survey of Labour and Income Dynamics Main File.
f) Live in a Region with a High Employment Rate
Living in an economic region with a lower than average employment rate was a significant variable associated with persistent low income among lone parents and recent immigrants, but not for the other groups.
Factors Important for Specific High-Risk Groups
Several significant factors associated with persistent and annual low income were unique to individual groups. Being an older unattached person (55 and over as opposed to 45-54) was a significant variable associated with persistent low income within the older unattached group. Being a member of a visible minority (particularly being of Arab origin) among recent immigrants and being a Registered Indian among the Aboriginal population living off reserves were also significant factors associated with being in low income in 1998 among those groups.
What insights do these findings provide in response to the two key questions posed at the beginning of this article?
The answers appear to be strongly linked to factors that encompass absent, disrupted, or ineffective social networks. Those most vulnerable to persistent low income in Canada are those most likely to belong to a group whose circumstances have sharply reduced their potential for extensive social relationships. Being a member of a high-risk group, in itself, does this. But some persons within each group are more likely than others to have supports through which they can obtain access to stable earnings from the labour market. For example, as previously noted, being married or in a common-law union gives many persons with work-limiting disabilities access to earned income through their spouse. In one way or another and, in many cases, in several ways simultaneously, the members of the five high-risk group in persistent low income lack such social support and networks.
|High-Risk Group||Incidence of Market Low Income15 1996 |
|Incidence of Market Low Income 2002 |
|Incidence of Post-Income Tax Low Income 1996 |
|Incidence of Post-Income Tax Low Income 2002 |
|People with a disability||52.0||36.7||34.2||21.4|
|Aboriginal people living off reserve||43.6||29.4||31.8||17.0|
Source: SLID Main File.
The factors statistically correlated with a disproportionate incidence of persistent low income within each group all relate to absent, disrupted, or ineffective social networks.14 Persons who belong to multiple high-risk groups, persons with a disability who live alone, older unattached individuals who were once in a union, lone parents who had their first child outside a union, those off-reserve Aboriginal people most likely to have grown up on a reserve, and recent immigrants who are members of a visible minority group and from countries with cultures most distinct from that of Canada are the persons within the high-risk groups most likely to experience persistent low income.
This is also the case for off-reserve Aboriginal people and recent immigrants living in depressed local labour markets. Those within the high-risk groups who do not have such characteristics and have established and maintained family and community networks are much more likely to avoid low income on both an annual and persistent basis.
A Final Note – How Each High-Risk Group Has Been Doing
Persons in households where the main income recipient belongs to a high-risk group have benefited from the strong labour market conditions prevailing in Canada since 1996, experiencing sharp reductions in their annual low-income rates. However, these rates remain well above those for persons aged 0-64 in households where the main income recipient does not belong to one of the five groups.
Among all groups, except for lone parents, the percentage point decline in the low-income rate based on market income is greater than the decline experienced once transfer payments and the income tax system are taken into account. The impact of the introduction and enrichment of the National Child Benefit Supplement after 1996 is reflected in the larger percentage point decline in post-income tax low income for lone parents over this period (14.4 percentage points) than in market-income-based low income (11.0 percentage points).16
The percentage point gap in lowincome rates between each of the high-risk groups and the non-high-risk group also declined over this period. We would expect to see such a trend during a period of tightening labour markets and strong economic growth. However, it must also be underlined that these groups suffer disproportionately during periods of recession and rising unemployment.
A more important question and, unfortunately, one which data constraints make much more difficult to answer is whether these groups have experienced structural improvement between the two most recent peaks in the business cycle between 1989 and 2000. The picture is most clear, in opposite directions, for lone parents and recent immigrants. The former group’s incidence of low income fell from 39.0% to 29.5% over this period while the latter’s rose. Changes to the definitions of Aboriginal persons and persons with work-limiting disabilities over the period make it impossible to make valid comparisons for these groups. The incidence of low income among unattached persons under age 65 changed little over the 1989 to 2000 period. This probably also applies to those aged 45-64 within that group.
What has changed is the share of the population living in households where a member of one of these groups is the main income recipient. Of those persons under age 65 in households where the high-risk status of the main income recipient could be established, the share in households where the main income recipient was a member of a high-risk group increased from 25.2% in 1996 to 29.0% in 2002. This indicates that these groups may be increasing in numbers faster than the rest of the population. Over the same period, their share of the low-income population whose high-risk status could be identified rose from 57.6% to 63.3%.
Thus, the members of the high-risk groups and their dependants appear to be at the centre of the struggle against annual and persistent low income from both a demographic and a socio-economic perspective.
Fleury, Dominique. 2002. “Economic Performance of Off-Reserve Aboriginal Canadians: A Study of Groups at Risk of Social Exclusion.” Human Resources Development Canada, Applied Research Branch working paper series W-02-1E.
Heisz, Andrew, and Logan McLeod. 2004. “Groups at High Risk of Being in Low- Income.” Low-Income in Census Metropolitan Areas, 1980-2000. Statistics Canada. Catalogue 89-613.
Kapsalis, Constantine, and Pierre Tourigny. 2002. Profiles and Transitions of Groups at Risk of Social Exclusion: Lone Parents. Human Resources Development Canada, Working Paper SP-577-06-03E.
———. 2003a. “Profiles and Transitions of Groups at Risk of Social Exclusion: Individuals with Work-Limiting Disabilities.” Human Resources Development Canada, unpublished.
———. 2003b. “Groups at Risk of Social Exclusion: Unattached Individuals Aged 45 to 64.” Human Resources Development Canada, unpublished.
———. 2004a. “Groups at Risk of Social Exclusion: Recent Immigrants.” Social Development Canada, unpublished.
———. 2004b. “Groups at Risk of Social Exclusion: Off-Reserve Aboriginal Individuals.” Social Development Canada, unpublished.
Morissette, René, and Xuelin Zhang. 2001. “Experiencing Low Income for Several Years.” Perspectives on Labour and Income. Statistics Canada. (March): 5-15.
Picot, Garnett, and Feng Hou. 2003. The Rise in Low-Income Rates among Immigrants in Canada. Statistics Canada. Analytical Studies Branch research paper series No. 198.
In this article, a person is considered to experience persistent low income if the cumulative income of the economic family over the period 1993-1998 or 1996- 2001 fell short of the cumulative amount of that family’s post-transfer, post-income tax low-income cut-offs for those periods as measured by Statistics Canada.
This definition of persistent low income incorporates dimensions of incidence and depth of low income.
An economic family is either an unattached individual living alone or with others to whom he or she is unrelated, or a group of two or more persons who live in the same dwelling and are related to each other by blood, marriage, adoption, or a common-law union.
In attempting to answer these questions, the article draws heavily on recent and largely unpublished studies on each of these groups prepared by Constantine Kapsalis and Pierre Tourigny for Human Resources Development Canada. Any conclusions drawn from these findings are those of the author and do not necessarily reflect the views of the authors of the studies cited or of the Government of Canada.
Different age groups are cited in different parts of this article. The main group examined is the population under age 65. However, to ensure all persons in this group remained under age 65 throughout the 1993-1998 and 1996-2001 periods, the sample for longitudinal analysis must be those aged 0-59 in the first year of the panel. In some cases, the article cites annual data from other authors who analyze only that part of the 0-64 population aged 16-59 or 16-64.
Persons are considered to be members of a group if the main income recipient in their economic family was a member of a group. In Figure 2, unattached indicates persons 45-59 who lived alone or with non-family members. Disabled indicates persons who, in 1996, had a long-term physical or mental condition that restricted the type and amount of work for pay they were capable of doing. Recent immigrants indicates persons who first came to live in Canada in the 10 years prior to 1996. Lone parents were persons raising at least one child under age 18 in a residence in which no other parent was present. Aboriginal off-reserve indicates Canadians who self-identified as a member of an Aboriginal group (Registered Indians, Inuit, Métis, or non-status Indians) and did not live on a reserve in 1996. Non-high risk designates all other persons aged 0-59 in 1996 who did not live in a household where a member of one of the five groups described above was the main income recipient.
Similarly, among lone parents, female lone parents have higher rates of annual and persistent low income than males. But male lone parents are also disproportionately vulnerable to persistent low income.
At the time of the 2001 Census, 73.4% of recent immigrants (those coming to Canada after 1991) were members of a visible minority group. Only 13.4% of all Canadians were members of a visible minority group.
These studies were done between November 2002 and May 2004 and were based on cross-sectional data from Statistics Canada’s Survey of Labour and Income Dynamics (SLID) for 1998, and from the 1993-1998 longitudinal panel of that survey.
Having less than a high school education is an important determinant of having a work-limiting disability. Kapsalis and Tourigny found that this is not because the onset of the disability reduced opportunities to graduate from high school. Even considering only persons who became disabled at age 30 or later, persons with less than a high school diploma were more than twice as likely to become disabled as the rest of the population. Two plausible explanations are that the percentage of persons with a healthy lifestyle increases with the level of education, and persons who have not graduated from high school are more likely to have to earn their living doing hazardous work.
Note that low levels of education are not a significant factor correlated with low income for Aboriginal people living off reserves. These individuals speak at least one official language fluently and have lived all their lives in Canada. They have social connections through their families and are not geographically isolated from healthy local labour markets. In at least one of these respects they have more extensive and useful social networks than members of each of the other disadvantaged groups. These factors appear to offset lower levels of educational attainment in finding and maintaining stable paid employment.
Note that these numbers are based on Statistics Canada’s post-income tax LICOs and thus are not comparable to the statistics based on pre-income tax low-income cut-offs cited in the studies by Picot and Hou, and Heisz and McLeod.
Bookmark - Combating Exclusion by Enhancing Financial Capability
Individuals with a good knowledge of economic and financial basics can make more informed decisions – to their benefit and to the benefit of the entire economy. Not only can they manage their personal affairs and their households better, they can be more effective and productive members of society. They can also be better citizens, capable of making reasonable judgments on public policy issues that have a bearing on their personal prospects and those of the nation (David Dodge, Governor of the Bank of Canada, 2003).
Ample evidence suggests that low financial capability (or financial literacy) is widespread. As yet, however, there is no comprehensive public policy framework in Canada for enhancing financial capability – even though there are many reasons for making such a framework a priority. A complete discussion of financial literacy and financial capability, including a review and evaluation of activities to enhance these attributes in the United Kingdom, the United States, and Canada, can be found in the discussion paper, Financial Capability and Poverty (2004), prepared for the PRI by Social and Enterprise Development Innovations (SEDI).
The paper is available on the PRI web site at <www.policyresearch.gc.ca>.