Profiles and Transitions of Groups at Risk of Social Exclusion: Lone Parents - November 2002

2. Literature Review

Previous

Contents

Next

2.1 The Focus on Lone Mothers

Most of the literature on low income lone parents has focused on lone mothers. The most obvious reason is that over 90% of low income lone parents are lone mothers. Also, in general, lone fathers tend to leave low income and social assistance (SA) more rapidly, further justifying the focus on lone mothers. Finally, because low income lone fathers are a relatively small population, it is often difficult to study their low income experience using survey data because of sample size limitations.

In general, lone mothers have attracted particular attention in the literature on low income because of their high incidence of low income. For example, the annual low income profiles produced by the National Council of Welfare (2000) show that in 1998, 54.2% of lone mothers had incomes below Statistics Canada's pre-income tax Low Income Cut-Offs (LICO) the highest low income rate of any family type.

Lone mothers, along with individuals with disabilities, recent immigrants, individuals of aboriginal ancestry, and unattached persons aged 45 to 64, have been identified in the literature as high risk groups. They have a much higher probability of being in low income than the rest of the population, a risk that is particularly high when combined with low education (Hatfield 2001a and 2001b).

2.2 The Dynamic Nature of Low Income

Analysis of the dynamics of low income based on income tax data shows that low income is not a static phenomenon; instead, there are significant flows into and out of low income. For example, the work by Finnie (2000) shows that the experience of low income lone mothers is far from homogeneous:

• one-third did not experience low income during the five years covered by the study (1992-96); and
• among those who experienced low income, about one-third were in low-income for a short period (1-2 years) while, at the other end of the spectrum, just over one-third were in low income in all five years.

An objective of our study is to take advantage of the detailed characteristics available in the SLID to understand why some lone mothers experience low income while others do not, and why some lone mothers stay in low income for a long time while others do not.

2.3 Probing Low Income and Social Assistance Spells

A related issue is the possible "scarring" effect of prolonged low income or reliance on SA. This issue has been investigated looking at hazard exit rates i.e. the probability of exiting low income or SA after a certain stay period. The evidence from the literature is that the longer individuals stay in low income or on SA, the less likely it is that they will exit low income or end reliance on SA.¹

However, the above finding may be of limited policy value, since there are two competing interpretations with opposite policy implications:

1. Declining hazard rates may mean that the longer individuals stay in low income the more difficult it is for them to break away, possibly because employment skills deteriorate when not working for pay. The policy implication of this interpretation is that we should try to assist those entering low income as soon as possible to avoid the "scarring" effect of low income.
2. The competing explanation is that those with good employment prospects exit low income quickly. The policy implication of this interpretation is that we should wait for some time before intervening, otherwise we will be helping individuals who would have exited low income on their own after a short stay.²

Unfortunately, the longitudinal SLID sample is small and, as a result, most of the independent variables of the hazard models tested here, including the effect of the duration of spells, turned out to be statistically insignificant. As a result, in addition to the more conventional hazard analysis, we used OLS regression analysis to assess the effect of various factors on in-progress spells. This approach is methodologically less satisfactory than hazard analysis but, in the face of sample size limitations, it provides a simple way of complementing the hazard analysis results.

2.4 Factors Associated with Longer Low Income and Social Assistance Spells

Previous research has identified some of the factors associated with low income and SA durations. For example, Lacroix (2000a) concluded that "single men leave welfare more rapidly than single women. The more educated exit a little sooner than the less educated, and re-entry occurs faster for the less educated."

Gascon (2000) concluded that for non-aged families of two or more persons, the problem of low income "appears to have been particularly linked to the problem of inadequate access to the labour market." However, with respect to lone parents he concludes that "the problem (of low income) may be more linked to a lack of human capital, or at least to a lower income return for it in the labour market."

Among other factors that have been investigated in the literature is the correlation between low income or SA dependence and the rate of unemployment. Both longitudinal (Lacroix, 2000a) and time series data (National Council of Welfare, 2000) show that there is a close correlation between the two. This correlation is often interpreted as an indication that low income and reliance on SA are driven primarily by labour market trends and, therefore, the primary policy focus should be on keeping unemployment rates low.

However, while the unemployment rate is an important determining factor, it does not fully explain trends in SA caseloads or work effort among SA recipients. For example, the design of SA programs and the level of SA benefit rates also have an important effect, both on the number of SA beneficiaries and the level of work effort of SA recipients (Kapsalis, 1997).

In this study, we will take advantage of the wealth of variables available in the longitudinal SLID database, to assess the impact of various factors on low income and SA spells. The study will provide basic measures of the extent of low income or reliance on SA, and flows into and of out of low income and SA. In addition, the study will investigate the following key questions: to what extent is low income among lone mothers the result of no work for pay or low hours of work, as opposed to low hourly earnings? What factors are behind low hours or low hourly earnings? Which characteristics are more closely associated with short spells on low income or SA and what are the policy implications?

• ¹See for example Finnie (2000); Lacroix (2000b).
• ²This, for example, was the rationale for the long-term unemployed program under the old Canadian Job Strategy.