RESEARCH REPORT

Linking Family Change, Parents' Employment and Income and Children's Economic Well-Being:  A Longitudinal Perspective

An Analysis of the National Longitudinal Survey of Children and Youth

Phase 1 of a three-part project commissioned by the Department of Justice Canada on: The Impact of Parent's Family Transitions on Children's Family Environment and Economic Well-Being: A Longitudinal Assessment

2003-FCY-2E

Prepared by:
Heather Juby
Céline Le Bourdais
Nicole Marcil-Gratton

TABLE OF CONTENTS

EXECUTIVE SUMMARY

1.   INTRODUCTION

2.   FAMILY TRANSITIONS, LABOUR FORCE PARTICIPATION AND INCOME CHANGE

2.1   Setting the score
2.2   From two-parent to one-parent family
2.3   From lone-mother to stepfamily: a way out of poverty?
2.4   Summary

3.   THE IMPACT OF INCOME AND WORK ARRANGEMENTS BEFORE SEPARATION ON CUSTODY AND CHILD SUPPORT ARRANGEMENTS

3.1   Pre-separation factors, custody and contact
3.2   Pre-separation factors and child support
3.3   Summary

4.   CHANGES OVER TIME: LIVING ARRANGEMENTS, FATHER-CHILD CONTACT AND CHILD SUPPORT

4.1   Changes in children's living arrangements and frequency of father-child contact
4.2   Changes in child support payments
4.3   Summary

CONCLUSIONS

REFERENCES

ENDNOTES

List of Tables

Table 1
Proportion of children experiencing at least one family transition between birth and Cycle 2, according to their age at the survey, and the type of parents' union at birth, NLSCY, 1996-97

Table 2
Intact and lone-parent family income, and percentage change for children whose parents separated after Cycle 1 and who were living with a lone parent at Cycle 2, according to the sex of the lone parent, NLSCY, 1994‑95 and 1996‑97

Table 3
Average family income and percentage increase for children whose parents separated after Cycle 1 and who were living with their mother at Cycle 2, according to the parents' pre-separation labour force participation, NLSCY, 1994-95 and 1996-97

Table 4
Family income and percentage increase for children living with their mother in 1994-95, according to the mother's conjugal situation in 1996-97, NLSCY

Table 5
Percentage increase in family income for children living with their mother in 1994-95, by income level in 1994‑95 and by the mother's conjugal situation in 1996-97, NLSCY

Table 6
Family income and percentage change for children living with their mother in 1994-95, according to her employment status in 1994-95, and her conjugal situation in 1996-97, NLSCY

Table 7
Living arrangements for children whose parents separated between Cycle 1 and 2, according to the parents' pre-separation labour force participation, NLSCY, 1994-95 and 1996-97

Table 8
Child support agreement and payments for children whose parents separated between Cycles 1 and 2, and who were living with their mother at Cycle 2, according to the family income before separation, NLSCY, 1994-95, 1996-97

Table 9
Contact with father in 1994-95 and 1996-97 among children of separated parents who were living with their mother in 1994‑95, NLSCY

Table 10
Distribution of children living with their mother, according to the existence of a support agreement and the regularity of payments in 1994-95 and in 1996-97, NLSCY

List of Figures

Figure 1
Household income category by family type, NLSCY 1994-95

Figure 2
Ratio of household income to the low-income cut-off, by family type, NLSCY 1994-95

Figure 3
Parents' labour force participation, for children living in two-parent families, NLSCY, 1994-95

Figure 4
Parents' labour force participation, for children living in one-parent families, by sex of parent, NLSCY, 1994-95

Figure 5
Household income category for children living in two-parent families according to their parents' labour force participation, NLSCY, 1994-95

Figure 6
Household income category for children living in one-parent families, according to the parent's sex and employment status, NLSCY, 1994-95

Figure 7
Type and frequency of contact with father in 1996-97 for children whose parents separated between Cycles 1 and 2, NLSCY

Figure 8
Living arrangements at Cycle 2, for children whose parents separated between Cycles 1 and 2, according to the household income before separation, NLSCY, 1994-94 and 1996-97

Figure 9
Level of contact with non-resident father of children living with their mother at Cycle 2, among children whose parents separated between Cycles 1 and 2, according to the household income before separation, NLSCY, 1994-95 and 1996-97

Figure 10
Child support arrangements by 1996-97 for children whose parents separated between Cycles 1 and 2, according to the household income before separation, NLSCY, 1994-95 and 1996-97

Figure 11
Distribution of children with parents separated before Cycle 1, according to living arrangements and contact with father, NLSCY, 1994-95 and 1996-97

Figure 12
Distribution of children with parents separated before Cycle 1, according to whether a child support agreement exists and, when it does, the regularity of support payments, NLSCY, 1994-95 and 1996-97

EXECUTIVE SUMMARY

This report was commissioned by the Family, Children and Youth Section of the Department of Justice Canada to continue analyses of data from the Family History and Custody section of the National Longitudinal Survey of Children and Youth (NLSCY). The study is based on data collected on approximately 15,000 children, aged 2 to 13 years at the time of Cycle 2 (1996‑97), who were part of the first cycle of the NLSCY (1994‑95). This is one of Canada's foremost data sources for studying the development and characteristics of Canadian children and their families.

Having these two cycles of data allows the examination of cases in which families had broken apart between 1994‑95 and 1996‑97. This enables, for the first time, analyses based on the situation “before” and “after” certain family transitions, such as parents' separation or family recomposition, thereby providing new insight into the relationship between family change, income and labour force participation. It also makes possible the examination of how custody (viewed as physical custody), father-child contact and child support payments change over time for those parents who were already separated at the time of Cycle 1.

MAIN FINDINGS

Family type and work and income patterns

In two-parent families, 95 percent of children from birth to age 11 in 1994-95 benefited from the full-time employment income of at least one parent. The most common situation, shared by 43 percent of children, is for both parents to have full-time employment. Just over a quarter of children (27 percent) had one parent working full time, the other part time. Another quarter (25 percent) had a parent at home while the other was in the labour force full time.

In one-parent families, slightly more than 40 percent of children lived with a parent who brings in a full-time employment income. Seventy-eight percent of lone fathers and thirty-nine percent of lone mothers worked full time.

When parents separate

Mothers who had full-time employment before separation are better off financially afterwards than those who did not. Their average annual income of nearly $32,000 is substantially higher than that of other lone mothers (median amount of $20,000) in Cycle 2.

When lone mothers form a conjugal union

The average income of lone mothers rose between Cycles 1 and 2 whether or not women formed a conjugal union, and the lower the income in 1994‑95, the larger the increase two years later.

Characteristics of the two-parent family prior to separation, and custody arrangements when parents separate

The higher the family income before separation, the more likely children are to live in their father's custody or in shared custody afterwards. Income seems to play a more direct role in shared custody arrangements than when father has sole custody, which is to be expected given the need for two family homes in shared custody situations.

Shared custody arrangements are most often chosen by separating couples who are both in the labour force. About 20 percent of children in double-income families were alternating between parents' homes when the parents separated, compared to less than 6 percent of those in families in which one or neither parent was employed.

Children from single-earner families are more likely than other children to be in their father's custody after separation. One fifth of these children (20 percent) remained with their father when their parents separated, compared to less than 8 percent of children with both parents or neither parent in the work force.

The chances are very high (89 percent) that children will live with their mother when parents separate if neither parent had full-time employment.

Family characteristics and child support

Support agreements are more common in higher income families, and payments are also more reliable. If a support agreement existed, payments were made regularly for the majority of children in the highest income families (55 percent); this was the case for only one third of children in the lowest income category (31 percent).

Changes in living arrangement and father-child contact over time

Living arrangements for children in their father's custody are very durable. Although relatively infrequent (7 percent of children in 1994-95), almost all children living with their father in 1994‑95 were still in his care two years later.

Living arrangements for children in their mother's custody are also stable, but the frequency of contact with their father varies over time. Of these, two fifths of the children who had some contact with their father at the start of the period had a different amount of contact by the end of it.

Shared living arrangements appear to be more flexible. Nine tenths of the children with shared living arrangements at Cycle 1 had a different arrangement two years later. More than two fifths (41 percent) lived with their father and half (50 percent) with their mother; almost all of the latter maintained regular contact with their father.

The absence of father-child contact is not necessarily permanent. More than one fifth of children with no contact with their father in 1994-95 had some contact (generally “irregular”) by 1996-97.

Changes in child support payments over time

The absence of child support payments is not necessarily permanent. More than a quarter of children who had received no support for at least six months in 1994‑95 received some support during the following two-year period. Almost half of these payments were being made regularly by 1996-97.

The absence of a child support agreement is not necessarily permanent, although the chance of coming to an agreement later is relatively low. Two thirds of the children (65 percent) for whom there was no agreement in 1994-95 were in the same position two years later. Among those for whom an agreement had been reached during the period, less than half were receiving payments regularly in 1996-97.

IMPLICATIONS

Studies of family break-up typically focus on the arrangements made and their formality, and the functioning and adherence to the arrangements. The NSLCY provides an opportunity to examine various pre-break-up characteristics of families, to determine whether there is any relationship to the choices and arrangements made after the break-up.

Specifically, are families' pre-separation earning opportunities, employment choices and employment patterns predictors of post-separation outcomes? Understanding how Canadians organize and live as families in the context of their income and employment choices allows policy makers to plan programs and services, and to consider legislative reforms with a greater appreciation of the implications. Further questions can be answered, with additional cycles of data including how do family arrangements evolve, and in what circumstances? How fixed or flexible are the outcomes? What do the changes typically involve, and how frequently are they made? Furthering research in this area is a high priority, and the value of the NLSCY in gathering these vital data remains of the highest order.

1.   INTRODUCTION

This report builds on an earlier analysis of custody, access and child support data gathered during Cycle 1 of the National Longitudinal Survey of Children and Youth (NLSCY) in 1994-95 (Marcil-Gratton and Le Bourdais, 1999). The NLSCY is a panel survey conducted jointly by Human Resources Development Canada and Statistics Canada. More than 22,000 children, aged 0 to 11 years were first surveyed during the winter of 1994-95; however, because of financial constraints, not all of these children were retained in the sample for the second cycle.[1] This study is based on approximately 15,000 children present for both cycles, and aged between 2 and 13 years when surveyed in 1996-97. Many topics are covered in the NLSCY, but this report draws largely on data from the Family and Custody History section of the survey. This section provides complete retrospective conjugal and parental histories of each child's biological parents up to the time of the survey. For children whose parents separated or never lived together, additional information is given related to custody arrangements, contact with the non-custodial parent and regularity of child support payments, both at the time of the separation and the survey.

An analysis of retrospective questions on family history and custody arrangements at Cycle 1 revealed the complexity of the family life course of Canadian children born near the end of the 20th century (Marcil-Gratton and Le Bourdais, 1999). With parental separation more frequent and occurring earlier in children's lives, the children's family environment expanded as their parents “went their separate ways”—remarrying or forming a common-law union with a different partner and having children within these new unions, thus adding stepparents, stepsiblings and half-siblings to their children's family network. Each change in the parental life course entails a “family transition” in that of their children. By Cycle 2 of the survey (1996-97), when these children were 2 to 13 years of age (with an average age just over 8), almost one quarter had experienced at least one transition in their family environment. This proportion varies depending on the age of the children at the time of the survey (see Table 1), from 14 percent at the age of 2 to 3 years, rising to almost 30 percent of children among those reaching adolescence.

The likelihood of experiencing family change also depends on whether children were born within a conjugal union and according to the type of this union. Most children (93 percent) were born within a conjugal union, to married or cohabiting parents; for these children, their parents' separation is the first family transition. This experience is much more usual among children born within a common-law union (44 percent), than among those born within marriage, particularly when the marriage was not preceded by a period of cohabitation (12 percent). Of the minority (7 percent) of children born outside a conjugal union, more than three quarters (76 percent) had experienced at least one family transition. This higher proportion is largely a product of the nature of the first transition, however. Born to essentially “separated” parents, the first transition involves union formation rather than dissolution, as their mother or father enter into a union either with each other or with a new partner.

Table 1
Proportion of children experiencing at least one family transition between birth and Cycle 2, according to their age at the survey, and the type of parents' union at birth, NLSCY, 1996-97

The transitions parents make into and out of conjugal unions generally translate, in their children's lives, as a movement from one type of family to another, from an intact family to a lone-parent family, or from a lone-parent family to a stepfamily, for instance. These types of transitions are the basis of the analyses in this report. Family mobility is just one of the processes affecting children, however. Families are in a state of constant flux at many levels. Economically, for example, family income rises and falls as parents are laid off or promoted at work, or as mothers enter or return to the labour force. However, while changes in income and employment may occur independently of family change, family transitions rarely occur without triggering important changes in these two areas. In a society in which discontinuity is on the rise, both in the workplace and the family, it is important to explore the links between the diverse facets of children's environment. With Cycle 1 data alone, family transitions could not be linked to changes in other areas because most relevant socio-economic and child development data referred only to the situation at the time of the survey. The addition of information collected from the same children two years later, at the time of Cycle 2 of the survey (1996-97), removed many of these limitations, making possible the longitudinal approach used in much of the present research.

This report has three main sections, each of which uses “before” and “after” data (from Cycles 1 and 2) to throw new light on specific questions.

• The first section focusses on the link between family type, income and the way income earning is shared by parents within the family. It looks, in particular, at the effect that separation and stepfamily formation have on family income levels, and examines how this relates to the parents' labour force involvement.
• The second section tackles the question of how couples organize the physical care and economic support of their children when they separate. For the first time, it is possible toexplore how “pre-break-up” family characteristics, such as income levels and labour force participation, influence the decisions parents make about custody arrangements and child support when they separate.
• The third section focusses on the changing nature of custody, contact and child support over time. It examines how living arrangements and father-child contact, on the one hand, and child support agreements and payments, on the other, evolved during the two-year period between Cycles 1 and 2 for children whose parents were already living apart at Cycle 1.

2.   FAMILY TRANSITIONS, LABOUR FORCE PARTICIPATION AND INCOME CHANGE

Much has been written about the crucial importance of income for children's well-being, and research suggests that some of the negative effects associated with separation and divorce are actually caused by the drop in income that generally accompanies it (Amato, 2000; Duncan and Brooks-Gunn, 1997; McLanahan, 1990). When a couple separates, resources that formerly maintained a single household have to support two and, even when those resources are distributed fairly between family members, each individual is by definition financially less well off. Most parents are obliged to reduce their standard of living, sometimes dramatically, and different strategies may be adopted to deal with this situation. Some separated parents may seek to reduce costs by lodging, at least temporarily, in a household maintained by someone else (often their parents), while others attempt to increase the resources available by taking paid employment or increasing the number of hours they work.

This leads us to the second important way that family transitions influence income. In Canada, families have maintained living standards over the last three decades largely through the increased labour force participation of mothers. The dual-earner family has become the norm. In other words, the standard of living enjoyed by a family is closely linked to the number of adults available to bring in an income—which generally means the number of parents within a household. Separation reduces the number of parents; family recomposition increases it. Having two potential wage earners also gives more choice in terms of the employment strategies devised by parents to balance earning a living with the other demands of family life; one can concentrate on earning an income, and the other on caring for the children, for example, or both parents may choose to work full time and use the extra income to pay for child-care services.

In this first part of the report, we focus on the direct impact that family transitions have on income levels and work patterns. The section starts by setting the scene, using Cycle 1 cross-sectional data to explore how the different family types (intact, step and lone-parent) compare, with regard to the two main variables in this part of the analysis: income and employment strategies. This is followed by a longitudinal analysis of the way in which income and work are affected by two specific family transitions occurring between the first two cycles of the survey.

• From an intact family to a lone-parent family. Using the sample of children whose parents separated between Cycles 1 and 2, we examine changes in family income before and after parental separation, and explore how this affects the labour force participation of women becoming lone mothers during the period.
• From lone-mother family to stepfamily. Using the sample of children living with a lone mother in 1994-95, we examine the extent to which forming a conjugal union modifies family income levels compared to remaining in a single-parent family. The small number of cases prevented a similar analysis of lone-father families.

2.1   Setting the score

Family type and income

At the time of Cycle 1 of the survey in 1994-95, children were living in different types of families (intact, step, and lone-father and lone-mother families), and the type of family they belonged to was closely linked to the financial resources available for their upbringing. Figure 1 compares family income levels across these four family types for six income categories, from less than $20,000 per year to $80,000 or more.[2] The income gap between two-parent and one-parent families, and between lone-father and lone-mother families, headed by fathers and those headed by mothers is clearly visible.[3]

Figure 1
Household income category by family type, NLSCY 1994-95

A few points can be taken from this figure.

• Intact families and stepfamilies have similar income levels, although incomes are slightly higher in intact families. A smaller proportion of children in intact families are in the lowest income brackets (less than $30,000) and a larger proportion in the highest.
• Two-parent families are considerably wealthier than lone-parent families, even male-headed lone-parent families. Fewer than one in eight children (4 percent and 7 percent) with their mother, and less than one third of children (9 percent and 23 percent) with their father, were in households with an annual income of over $40,000 (the case for more than two thirds of the children in two-parent families).
• Although lone fathers are, on average, much better off than lone mothers, many support their family on a very low income. More than one quarter of children (26 percent) living with their father were being raised on an annual income of less than $20,000.
• The majority of lone mothers are raising children on a very limited budget. For three fifths of the children in lone-mother families (61 percent), the annual family income was less than $20,000, and for almost four fifths it was under $30,000.

Below the low-income cut-off

No matter how informative they may be, income levels are not always the best indicator of living standards, as they do not take into account the number of individuals being supported by the income. The low-income ratio, on the other hand, is estimated as the ratio of economic family income to the low-income cut-off, a measure derived from a number of indicators including family size. A ratio of 0.75, for instance, means that the household income is 25 percent lower than the low-income threshold estimated for a family with a given set of characteristics (including the number of adults and children present). Low-income ratios are given for the four family types in Figure 2.

The two lower categories represent the proportion of families with incomes below the low-income cut-off. The figure confirms the general impression given by income levels in Figure 1, with two-parent families less often below the low-income threshold than one-parent families, and single fathers less often than single mothers. The difference between family types is clearly illustrated for incomes falling below the low-income cut-off by more than 25 percent; while only 1 in 12 children (8 percent) in intact families are living on these very restricted budgets, this is the case for more than half (53 percent) of the children in lone-parent families headed by mothers. In fact, less than 30 percent of the children in lone-mother families are not in “low-income” families. The proportions are inversed for children in lone-father families: almost 70 percent are above the low-income cut-off and slightly more than 30 percent are below.

Figure 2
Ratio of household income to the low-income cut-off, by family type, NLSCY 1994-95

Patterns of income earning

Two work-related factors explain much of the difference in income levels between two-parent and lone-parent families, and between lone-father and lone-mother families. These two factors are the number of potential wage earners and the number of hours worked. Couples have greater flexibility than lone parents in terms of balancing work and family life, since one or both parents can be employed, full- or part-time. The different arrangements adopted by parents in two-parent families are shown in Figure 3.

• Having both parents in full-time employment is the most common situation, shared by 43 percent of children.
• Slightly more than one quarter of children (27 percent) had one parent working full time, and the other part time.
• Another quarter had a parent at home while the other was in the labour force full time.

Figure 3
Parents' labour force participation, for children living in two-parent families, NLSCY, 1994-95

Overall, 95 percent of children in two-parent families benefited from the full-time employment income of at least one parent, a situation that contrasts strongly with the situation in one-parent families (Figure 4). Slightly more than 40 percent of children in one-parent families (78 percent in lone-father families, and 39 percent in mother-headed lone-parent families) live in a family supported by a full-time employment income, a figure that is all the more disquieting given that the vast majority of children in lone-parent families are with their mother. Having a parent at home, however, is more common, particularly among children in lone-mother families (43 percent); in fact, these children are the most “traditional,” in the sense that they are more likely than other children to have their mother at home.

The number of wage earners and the number of hours worked are closely linked to the level of family income, as shown in Figures 5 and 6 for two-parent and one-parent families respectively. The general pattern that emerges from these figures is predictable: the more time parents invest in paid employment, the higher their family income. However, a closer look highlights three important features of the relationship between work organization and income.

Figure 4
Parents' labour force participation, for children living in one-parent families, by sex of parent, NLSCY, 1994-95

Figure 5
Household income category for children living in two-parent families according to their parents' labour force participation, NLSCY, 1994-95

Figure 6
Household income category for children living in one-parent families, according to the parent's sex and employment status, NLSCY, 1994-95

First, there is a considerable income disparity for any given employment pattern. Consider the situation when both parents are employed full time as an example. The principal challenge for these parents is to “find the time” to meet the demands of family life, given the number of hours they spend working. The family income on which more than one quarter of children in these double-income families are raised exceeded $80,000 in 1994-95; these parents could probably compensate to some extent for the hours spent in the workplace by purchasing services that lighten the load at home. However, although having two employed parents raises the likelihood of a family being financially at ease, it is certainly no guarantee of it. The parents of 16 percent of children in double-income families earned less than $40,000 between them; with scarcely enough to cover the basics, paying for services is not an option.

Second, one full-time employment income does not have an equivalent “value” across the different family types, reflecting differences in earning power. More than half the time (33 percent and 13 percent and 7 percent), single earners in two-parent families bring in $40,000 per year or more; in lone-father families, the percentage drops to 39 percent, and in lone-mother families to 27 percent. At the other extreme, 9 percent of single-earner two-parent families, and 12 percent (7 percent and 5 percent) of employed fathers, bring in less than $20,000 per year, compared to 31 percent of lone mothers. In other words, not only do lone fathers have full-time employment twice as often as lone mothers, they are also better paid, two factors responsible for the huge disparity between the incomes of lone fathers and lone mothers.

Third, at the opposite end of the scale, except for a few families of “independent means”, two unemployed parents means that families have to live on a very tight budget. Nearly 60 percent of these children were being raised on an annual income of less than $20,000, a proportion almost identical to that among children of unemployed single mothers (57 percent) with an income of less than $15,000. The extra adult to feed means that the budget in unemployed two-parent families is just as stretched as among single mothers who are not working. On the other hand, two adults in a family increases the chances of having at least one parent employed at some point and, as a result, only 3 percent of children in two-parent families are in this situation compared to 45 percent of children in mother-headed lone-mother families. With only one in five lone fathers not fully employed, the numbers were too small to include this group in Figure 6. The data nonetheless suggest that the level of poverty of these fathers surpasses even that of lone mothers who are not employed.

Evidently, employment patterns and income are far from static, changing in response to many factors, such as market forces, educational opportunities and personal choices. They may also change as a result of family transitions. In the next section of this report, we consider the impact of separation and stepfamily formation on income and employment from a longitudinal perspective.

2.2   FROM TWO-PARENT TO ONE-PARENT FAMILY

Data from consecutive cycles of the NLSCY make it possible for the first time to directly assess the drop in family income that accompanies the transition from an intact to a lone-parent family. Information on family income gathered both before separation (at the time of Cycle 1) and after (at the time of Cycle 2) was available for approximately 500 children whose biological parents separated between the two surveys. Of these, only children living with a lone parent in 1996-97 were included in this analysis. Children in shared custody are treated as belonging to the family of the parent responding to the survey in 1996-97, since information on family income was not gathered for the other parent.[4] Children were first classified in four groups according to the level of family income before separation in 1994-95, and according to whether they were living with their mother or father two years later. For each of these subgroups, the average annual family income before and after separation, as well as the percentage change, is presented in Table 2. The two lowest income categories have been combined for lone fathers due to the small number of children from families with an annual income of less than $40,000 who reside with their father after separation.

Table 2
Intact and lone-parent family income,¹ and percentage change for children whose parents separated after Cycle 1 and who were living with a lone parent at Cycle 2, according to the sex of the lone parent, NLSCY, 1994‑95 and 1996‑97

Lone mothers versus lone fathers

First, Table 2 provides clear evidence of a much sharper overall decline in relative income for children living with their mother (45 percent) than father (27 percent) after separation. The disparity in income decline is even greater when calculated using the median value rather than the mean, at 50 percent for mothers and only 18 percent for fathers. The median income indicates that half the children remaining with their mother when parents separate are being raised on an annual income of less than $19,000; the comparable figure for children raised by their father is $49,200. The difference between mothers and fathers holds for each pre-separation income category, and reflects the fact that men, on average, bring in a higher proportion of family income than women do.

Second, there is some evidence that, among fathers, the relative drop in income is closely linked to family income levels before separation: the higher the pre-separation income, the greater the relative drop at separation. For mothers, this pattern is less clear since the relative drop in income is high at all income levels, nearly 50 percent for all but the lowest income group. We observed earlier that higher incomes were most common in families with two full-time incomes. In other words, at higher income levels the mother's contribution to family income is greater. On the one hand, this offsets the decline in income for lone mothers who continue bringing a full-time income into the family after separation. On the other hand, the loss of the former spouse's earnings is responsible for the steeper decline in family income among fathers as pre-separation family income rises. We will be able to verify this with a greater number of cases in Cycle 3.

Perhaps one of the most unexpected and interesting findings is the considerable gap between the overall average pre-separation family income for children who continue living with their mother ($46,500) or father ($68,100) after separation. Are children from relatively well-off families more likely to remain with their father after separation than those from poorer families? We will return to this subject in a later section dealing with the impact of family income before separation on custody arrangements.

Mothers' employment status and income after separation

Findings in the previous section suggest that pre-separation earning patterns explain much of the difference between mothers' and fathers' households in the relative decline of family income after separation. Table 3 presents similar figures on family income change for the transition from two-parent to lone-mother family when parents separate, but this time according to the parents' labour force participation before separation. In families with only one full-time income, the mother is normally the parent at home or she has part-time work.

• Lone mothers from double-income families who were working full-time before separation are better off financially than other lone mothers. Their average annual income of close to $32,000 is significantly higher than that of mothers from the three other situations.
• Lone mothers from families with only one full-time income were in a similar financial position whether they had worked part-time before separation ($22,500) or not ($21,800), despite the higher pre-separation income of the former. With a higher relative decline in income among lone mothers who had worked part time (54 percent) than among those who had not (47 percent), many of these mothers were raising children on an income below the low-income threshold after separation.
• Families with neither parent fully employed were already struggling financially, thus, given that a bare minimum is necessary for survival, the relative decline after separation for lone mothers is small (9 percent).

• Average income rose between Cycles 1 and 2 whether or not women formed a conjugal union, and the lower the income in 1994-95, the larger the increase.
• Among mothers with an annual income of less than $12,000 in 1994-95, even women who were still alone two years later had seen their income rise by close to 60 percent. Mothers who became part of a couple (with a new partner or their children's father) more than tripled their income. However, the scale of these increases are more a reflection of how poor these lone-mother families were in 1994-95 than of how rich they had become two years later.
• At slightly higher 1994-95 income levels (over $16,000 annually), mothers who did not re-unite with their children's fathers or acquire a new partner scarcely registered any rise in the average income.
• Even at the highest income category, mothers reuniting with their children's father or forming a new conjugal union saw a sizeable increase in family income (46.8 percent).

2.4   Summary

In this section, we have explored the relationship between family type, income levels and employment patterns. First, the cross-sectional analysis of Cycle 1 data emphasized the close association between the family type in which children live, on the one hand, and the financial resources available for their upbringing, on the other. Also highlighted was the close relationship between the number of potential income-earners in a family, the level of income and the strategies available to parents in terms of balancing income-earning with the other responsibilities of family life. The longitudinal analyses of Cycle 1 and 2 data that linked the passage from a) an intact family to a lone-parent family and b) from a lone-parent family headed by a mother to a two-parent family, confirmed the importance of the close association between family, income and employment, and clearly illustrated the impact of couple formation or dissolution on a family's financial circumstances.

In addition to the general trends, however, these analyses also reveal the great diversity in the experience of Canadian families, whether in relation to income levels, patterns of employment or the impact that family transitions have in these two areas. For example, while it is true that, generally speaking, the more parents are involved in the labour force, the higher their family income, many double-income families earn scarcely enough to make ends meet. Likewise, not all lone mothers living on a tight budget will remain in this situation for an extended period; a rise in family income will result, for some, from the labour market and, for others, from a change in their conjugal situation.

Analyzing the transition from intact family to lone-parent family suggests that the lone parent's financial circumstances after separation are closely linked to the intact family income before separation and to the way the couple shared income-earning. The wealthiest lone mothers, for instance, are those from the wealthiest intact families, and who were already fully involved in the labour force before they separated.

The next section continues this line of investigation, examining how the socio-economic status of the intact family influences the decisions that separating parents make about the physical and financial care of their children.

3.   THE IMPACT OF INCOME AND WORK ARRANGEMENTS BEFORE SEPARATION ON CUSTODY AND CHILD SUPPORT ARRANGEMENTS

The relationship between family change and income is not unidirectional. In less direct ways, income also has an effect on family transitions, playing a role in the decision to separate or to form a new conjugal union. The stress of financial hardship, for instance, may contribute to the demise of a conjugal union. At the same time, the threat of even greater financial hardship may be a powerful inhibitor to separation, although it may have a different effect on low-income families than on those who are better off financially. A subject that has received little attention so far, probably because of the dearth of information about it, is whether intact family characteristics, such as the level of income and the way parents organize their work schedules, influence how separating parents divide responsibilities for children. Longitudinal data from consecutive cycles make it possible for the first time to relate post-separation arrangements to characteristics of the family and parents prior to separation, information not available for those who were already separated at Cycle 1. This is the subject of this part of the report: the impact of pre-separation family income and employment schedules on the type of custody arrangements, and the level of contact and child support payments established by couples separating between Cycles 1 and 2.

Findings in the previous section suggest that pre-separation family income may be related to whether children live with their mother or father after parental separation. It is also possible that the parents' employment, particularly that of mothers, influences decisions made about living arrangements and child support when the parents separate. With data collected both before and after separation, it is possible to address these questions directly and gain new insight into the relationship between economic factors and adapting to family change. The following analysis is based on approximately 500 children whose parents separated between Cycles 1 and 2 of the NLSCY and who were still living apart at the time of Cycle 2.

3.1   Pre-separation factors, custody and contact

Looking at children whose parents separated between the two cycles provides interesting insight into living arrangements and father-child contact during the period immediately following the separation. That fathers remain closely involved with their children in the first year or two is apparent (see Figure 7). At the time of Cycle 2, almost a quarter of children were living full time (10 percent) or part time (14 percent) with their fathers; and more than half saw them at least once a week (36 percent) or once every two weeks (15 percent). Of the rest, most had irregular contact (17 percent) and only 8 percent no longer had any contact with their father.

Figure 7
Type and frequency of contact with father in 1996-97 for children whose parents separated between Cycles 1 and 2, NLSCY

(Click here to view Figure 7)

Pre-separation family income and custody arrangements

Figure 8 relates children's living arrangements after parental separation (whether living with their mother, father or in shared custody) to the family's income before the parents separated.[5] The distributions suggest that pre-separation family income is linked to the type of custody arrangement adopted by separating parents. The relationship is particularly clear in the case of shared custody, with proportions rising steadily as family income rises. Children in families with an annual income greater than $60,000 are four times (25 percent) as likely to share time between both parents than are children in families with an annual income of less than $20,000 (6 percent). Shared custody appears to be an increasingly common arrangement among separating parents with the financial means to support two family homes. The link between father custody and income is less clear, however. Although children in families with an income of $60,000 or more are twice as likely to live with their father than other children, below this level, there is little variation in the proportions (7 percent and 8 percent). Nonetheless, it is clear that the higher the family's income before parental separation, the more likely children are to go on living at least part of the time under the same roof as their father.

Figure 8
Living arrangements at Cycle 2, for children whose parents separated between Cycles 1 and 2, according to the household income before separation, NLSCY, 1994-94 and 1996-97 (Click here to view Figure 8)

Pre-separation family income and contact with non-resident fathers

Another question, related to this, concerns the frequency of contact that non-resident fathers have with their children. Does family income before separation affect the amount of time fathers spend with children who live with their mother after separation? If the payment of child support is linked both to a father's ability to pay and to the frequency of contact he has with his children, we would expect non-resident fathers from higher income families to stay more closely involved with their children after the separation than fathers from lower income families. However, the findings presented in Figure 9 support this hypothesis only with regard to the father actually losing contact with the children: almost one in five non-resident fathers (19 percent) from the lowest income families compared to only 3 percent of non-resident fathers from high-income families lost contact with their children within the relatively short period between the separation and the survey. Beyond this, the association is less clear. Non-resident fathers at both extremes of the income scale, for instance, are virtually indistinguishable in terms of the proportions remaining closely involved with their children; approximately half the children in these groups see their father at least once a week. Clearly, income alone cannot explain variations in the level of contact that non-resident fathers maintain with their children.

Figure 9
Level of contact with non-resident father of children living with their mother at Cycle 2, among children whose parents separated between Cycles 1 and 2, according to the household income before separation, NLSCY, 1994-95 and 1996-97

(Click here to view Figure 9)

Women's employment and custody arrangements

The connection between family income before parental separation and children's living arrangements afterwards is not necessarily a causal one, however. We have seen that families with higher incomes are also families in which both parents are employed. In families with mothers who work, fathers tend to be more involved in caring for their children; they may be more competent parents, with more confidence in their ability to look after their children, and they may more likely wish to have their children living with them after separating from their spouse. For their part, working mothers may also be more willing to continue sharing the everyday responsibilities for children with fathers after the separation. The information in Table 7 supports this hypothesis.

• Shared living arrangements are most commonly chosen by parents who both had paid employment, full-time or part-time, before separation. About 20 percent of children with both parents in the labour force were alternating between parents' homes, compared to less than 6 percent of those with only one or neither parent employed.
• Children in single-earner intact families are much more likely than other children to be living with their fathers after their parents' separation. One fifth of these children (20 percent) were living with their father at the time of Cycle 2, compared to less than 8 percent of children with parents in other types of work-sharing patterns.
• The chances are very high that children remain with their mother (89 percent) when neither parent had full-time employment before the separation.

In other words, it is possible that the apparent association between income and custody arrangements is due to this relationship between women's employment and custody arrangements, in the sense that family income is higher when mothers also contribute to the household income. To get a better understanding of how these two factors are related to decisions about children's living arrangements when the parents separate, we conducted a multinomial logistic regression analysis, controlling for a number of other elements measured in Cycle 1 that are likely to influence custody arrangements. Concretely, we wished to test how income, mother's work and other potentially important variables influence whether children live in shared custody or with their father, rather than remaining with their mother as is most often the case. The other variables in the analysis included the parents' level of education, the mother's employment status, the type of union into which the child was born, whether or not the parents had children from an earlier union, the number of children in the household, the age of the parents and children at separation, the child's sex, and the region (Quebec versus the rest of Canada).

The regression analysis confirmed the significant role played by both income and the mother's employment in the choice of custody arrangement, even when the effect of other characteristics are taken into account.[6] The higher the income, the more likely children are to be living with their father or in shared custody. However, beyond this similarity, a very different picture emerges of the circumstances favouring shared rather than the father's sole custody (compared to living with the mother). First, income seems to play a more direct role in shared custody, which is to be expected given the need for two family homes. The mother's employment also is crucial, with shared custody significantly more likely among women who were in the labour force before the separation. All in all, the profile of separating couples choosing shared living arrangements for their children is one of relatively well-off couples, with older, educated fathers and working mothers—couples who may well have relative “equality” of gender roles in the family. This profile is also more common in Quebec, and in families with no more than two non-infant children.

Table 7
Living arrangements for children whose parents separated between Cycle 1 and 2, according to the parents' pre-separation labour force participation, NLSCY, 1994-95 and 1996-97

The context encouraging the relatively uncommon arrangement in which children live with their father after separation appears to be rather different, and more influenced by mother-related factors. The multivariate analysis confirms the importance of the mother's labour force status, with fathers far more likely to keep children with them after separation if the mother was not employed, and especially if she had not completed high school. The number of children is also important, with children in two-child families more likely to remain with their father than single children or those with more than one sibling.

3.2   Pre-separation factors and Child support

Many factors influence the payment of child support, but one might expect a close association with two factors central to this report: whether the non-resident parent can afford to pay, and how financially dependent the custodial parent is on the child support payments. The ability to support two households rather than one depends largely on the level of pre-separation family income; for example, non-resident parents without the means to keep a roof over their heads are hardly in a position to pay child support. The financial dependence of the custodial parent, on the other hand, is determined largely by how parents organized income-earning; mothers who were fully integrated in the labour force before separation and who continue to work afterwards, for example, are less dependent on child support payments than are mothers who were not.

Reaching a child support agreement

The distribution of children whose parents separated between the two cycles, according to whether or not parents had a child support agreement or whether an agreement was in progress, is shown in Figure 10. It lends support to the hypothesis that the higher the family income prior to separation, the more likely separating parents are to reach a support agreement. In 1996-97, nearly half of the lowest income families (49 percent) had no agreement; this proportion decreases with each income increment, to the extent that slightly more than one fifth (21 percent) of the highest income category families were without a support agreement, despite the relatively short period since the separation. Another interesting feature of these figures is the proportion of highest income families with an agreement in progress at the time of Cycle 2: 17 percent, far higher than in the other categories. Do wealthier families take more time than others to reach an agreement? It is possible that they not only have more financial issues to resolve, but that they are also less pressured to settle matters rapidly. We have seen, for instance, that high pre-separation family income generally depends on mothers' full-time employment. At separation, therefore, reaching a support agreement may be a matter of less urgency for working mothers for whom child support income represents a smaller proportion of total income than it does for women without paid employment.

Figure 10
Child support arrangements by 1996-97 for children whose parents separated between Cycles 1 and 2, according to the household income before separation, NLSCY, 1994-95 and 1996-97

(Click here to view Figure 10)

Paying child support

Having a support agreement does not necessarily mean that money changes households. At times, payments may not be made because of the nature of custody arrangements. For instance, while a child support agreement exists for more than three quarters of children in shared custody, the agreement may not entail money passing from one parent's household to the other. In the following analysis of the link between income and the regularity of child support payments, therefore, only children living with their mother have been included in the distributions shown in Table 8. For each pre-separation income category, proportions are given for those whose parents have no agreement, have an agreement in progress, or have reached an agreement on the payments to be made by non-resident fathers. Those with a support agreement are further subdivided according to whether payments are made regularly (if late at times), or irregularly or not at all.

Figure 11
Distribution of children with parents separated before Cycle 1, according to living arrangements and contact with father, NLSCY, 1994-95 and 1996-97

(Click here to view Figure 11)

Although research indicates that father-child contact tends to decline with time following the separation, many children living with their mother in 1994-95 were actually spending more time with their father two years later. The figures in bold on the diagonal show the proportion of children in each situation who had roughly the same amount of contact with their father at the beginning and end of the period. The percentages above and to the right of the diagonal represent a decrease in time spent with the father; those below and to the left, an increase.[8]

Overall, among children living with their mother at the time of Cycle 1, 16 percent had more and 23 percent had less contact with their father (data not presented). For instance, among children with regular weekly or biweekly contact with their father in 1994‑95, three quarters still had frequent contact in 1996-97, and a small percentage had moved in with him. However, about 10 percent of these children had lost touch with their father, though the data do not enable us to establish the reason for this. At the other end of the scale, more than one fifth of children (7 percent and 15 percent) who were not in touch with their father in 1994-95 had some contact with him by 1996-97, although in the majority of cases this contact was sporadic.

To summarize, this analysis of changes in living arrangements and contact with father indicates the following.

• Living arrangements for children in their father's custody are very durable. Nearly all children living with their father in 1994-95 were still in his care two years later. Although relatively infrequent (7 percent of children in 1994-95), the factors responsible for this type of arrangement appear to encourage its continuation.
• Living arrangements for children in their mother's custody are also stable; the frequency of contact with their father, however, varies over time. Two fifths of the children who had some form of contact with their father at the start of the period had a different level by the end of it.
• Shared living arrangements appear to be more flexible. Nine tenths of the children with shared living arrangements at the time of Cycle 1 had a different arrangement two years later. More than two fifths (41 percent) lived with their father, and half were with their mother; most of the latter maintained regular contact with their father.
• Changes in levels of father-child contact are not unidirectional: two fifths of the changes represented more contact and three fifths less contact.
• The absence of father-child contact is not necessarily permanent. More thanone fifth of children with no contact in 1994-95 had some form of contact (generally “irregular”) by 1996‑97.

In other words, the most consistent father-child contact is found at the two extremes, among children who live with their father and among those who have no contact with him. In between, there is a lot of movement, particularly among those with shared living arrangements. This situation is chosen by a growing minority of parents at separation (rarely entered into at a later date) and often evolves into a different form of custody after a number of years. Overall, the more frequent the contact with the father at the time of Cycle 1, the more frequent this contact remained two years later. The majority of children who lost contact with their father during the period had only intermittent contact with him at the start, although an almost equal proportion of the latter had a more solid relationship with their father by the end of the period.

4.2   Changes in child support payments

Child support is an important component in children's well-being following parental separation. Research shows a positive relationship between the payment of child support and children's educational attainment and other behaviour (Amato and Gilbreth, 1999; Argys et al., 1998; Bartfeld, 2000; McLanahan et al., 1994). However, only limited information is available about child support for the first two cycles of NLSCY. Respondents were asked whether or not they had a private or court-ordered support/maintenance agreement; those who had an agreement were asked how regular the support payments had been. In the absence of an agreement, or if the agreement was in progress at the time of survey, no direct information is available on the existence or regularity of support payments,[9] which makes changes in child support payments during the period difficult to assess. Already limiting the analysis of Cycle 1 data (Marcil-Gratton and Le Bourdais, 1999), this problem is magnified when attempting to assess the evolution between the two cycles. For instance, a sizeable proportion of individuals with an agreement in 1994-95 (and therefore a response to the question on the regularity of payments) stated two years later that they had no agreement or an agreement in progress (and therefore were not asked about the regularity of payments). For these cases, it is not possible to assess whether, and in what way, support payments changed during the period, because the absence of a support agreement does not necessarily imply the absence of child support payments. Among mothers reporting an agreement in progress, a relatively high proportion declared receiving some income from child support (34 percent and 48 percent at Cycles 1 and 2 respectively).[10] This proportion was much smaller for those without an agreement (6 percent and 8 percent at Cycles 1 and 2, respectively).

This problem reflects the changeable nature of child support arrangements over time, and although it imposes certain limitations, it does not prevent analysis of the question. Figure 12 compares the distribution of children whose parents were separated at the time of both cycles in terms of whether or not there was a support agreement, and how regularly payments had been made (four categories). In 1994-95, a support agreement existed for three fifths (60 percent) of the children of separated parents; for 9 percent, an agreement was in progress, and 31 percent were without an agreement. Payments were made regularly and punctually for one third of the children (33 percent), i.e. for more than half of the children with an agreement. In just under a quarter of cases in which an agreement existed, no payments had been made for at least six months before the survey. By 1996-97, the overall situation had improved to some extent, with the proportion of children with a support agreement rising from 60 percent to 68 percent. This did not always translate into maintenance support payments, however. While the percentage of children for whom payments were made regularly increased during the period, so did the proportion of those for whom no payment had been made in the past six months, from 14 percent to 17 percent. Moreover, almost one third of children were still declared either without a support agreement (28 percent) or with one in progress (4 percent).

Figure 12
Distribution of children with parents separated before Cycle 1, according to whether a child support agreement exists and, when it does, the regularity of support payments, NLSCY, 1994-95 and 1996-97

However, a detailed look at the evolution of child support arrangements during the period reveals a much more complex and changing picture, largely concealed by these cross-sectional images. Table 10 shows the movement of children between child support categories during the two years separating the survey waves. As child support arrangements do not work in the same way for children in shared custody, and as mothers are much less likely to provide child support for children in their father's custody (Seltzer, 1994), this table includes only data related to children reported living with their mother at the time of both surveys.[11] In this table, the figures in bold on the diagonal show the proportion of children for whom the situation remained stable throughout the period. Among children for whom payments were regular and on time in 1994‑95, for example, this was still the case two years later for 71 percent of them. The proportions within the rectangle refer to children with a child support agreement at the time of both surveys, and for whom information on the regularity of payments was therefore available at the start and end of the period. Within this rectangle, figures above and to the right of the diagonal indicate a decline in the regularity of payments, while those below and to the left indicate greater regularity.

Table 10
Distribution of children living with their mother, according to the existence of a support agreement and the regularity of payments in 1994-95 and in 1996-97, NLSCY

• Once the payment of regular and punctual support is established, it tends to continue. For more than 70 percent of children in this category in 1994-95, regular payments continued throughout the period. Only 11 percent no longer received regular support after two years (although this may also be true for some or most of the 8 percent of children without an agreement in 1996-97).
• Receiving any kind of support, even if payments are late or irregular, is a positive sign. Late or intermittent support payments often become more regular with time. For almost one third of children receiving irregular payments in 1994-95, for example, payments became more reliable during the period (22 percent and 10 percent).
• The absence of support payments is not necessarily permanent. More than a quarter of children (9 percent and 3 percent and 14 percent) who had received no support for at least six months in 1994-95 received some sort of payment in the interval. In almost half the instances, these payments were made regularly by 1996-97.
• The absence of a support agreement is not necessarily permanent, although the chance of coming to an agreement later on is relatively low. Two thirds of children (65 percent) for whom there was no agreement in 1994-95 were in the same position two years later. Moreover, among those who had reached an agreement in the interval, fewer than half (12 percent and 2 percent) were receiving payments regularly in 1996-97.

• the reason for the absence of a child support agreement;
• the type of agreement regarding child support for those with a private arrangement;
• the means of payment of support (directly, through the court, enforcement program, etc.); and
• the proportion of the awarded payments that were actually received.

[13]   Problems currently emerging with these data may reduce their usefulness, however.

Table of Contents