When Parents Separate: Further Findings from the National Longitudinal Survey of Children and Youth

2004-FCY-6E

IV  PHYSICAL CUSTODY ARRANGEMENTS AND CHILD SUPPORT: NEW INFORMATION

Only limited information on custody and child support was collected at the first two NLSCY cycles. An attempt was made to introduce more detailed questions in subsequent cycles. Unfortunately, problems with the complex questionnaire meant that this information was not collected exhaustively and, as a result, the analyses possible from the data are not as broad as had been hoped. The information is complete, however, for all parental separations occurring between Cycles 2 and 3 among children in the longitudinal cohort, and the analyses in this section are based on these children. The following analyses refer, therefore, to separations occurring during the two years preceding Cycle 3, among parents of children aged 4-15 years in 1998/99 (i.e. to children aged between 2 and 15 years when their parents separated.) Using only recent separations has certain advantages. First, it improves reliability, as the information about decisions taken at separation is still fresh in the respondent's mind. Second, the information is "up-to-date," describing the situation right at the end of the 1990s.

New information on custody arrangements

The distribution in Figure 4.1 provides a general picture of the custody and visiting arrangements in place by 1998-99 for children whose parents had separated in the two previous years. Most parents had maintained close ties with their children within the relatively short period of time since the separation:

  • The parents of 13% of the children were living together again by 1998-99.
  • Almost two-thirds (42%+14%+7%) were living with their mother; of these, two-thirds saw their father on a regular weekly or two-weekly basis.
  • Only one child in fourteen (7%) was living with their father full-time; of these, just under half saw their mother on a regular weekly or two-weekly basis.
  • One child in eight (12%) was living in shared custody, alternating between their mother's and father's home.
  • A minority of children (5%) fell outside the standard categories for a variety of reasons (e.g. in the care of someone other than a biological parent; a parent had died; a different form of contact).

Figure 4.1  Living arrangements and contact with other parent in 1998-99 for children, aged 4-15 years at Cycle 3, whose parents separated in the previous two years, NLSCY, Cycle 3 (N=867)

In other words, 70% of the children were still in close contact with both parents at Cycle 3; under one-fifth (14%+4%) had irregular contact with the other parent, and only 7% of children had no contact at all. But how does this compare with the amount of time parents had actually arranged to spend with their children? For the first time at Cycle 3, the parent responding to the survey was asked about the amount of time the child had spent with the other parent during the last year (or since the separation, if it happened less than a year ago) relative to what had been agreed; results show a high degree of flexibility in arrangements.

Do parents keep to the visiting arrangements agreed on at separation?

Overall, a little under half (46%) of these children spent the amount of time agreed upon with their other parent (see Figure 4.2). Parents not keeping to the agreement were divided fairly evenly between those who saw their child much more, a little more, a little less and much less frequently than agreed. Only 2% of children had not seen the other parent in the last year (or since the separation, if it was under a year ago); another 1% had no visitation rights. Clearly, arrangements made when parents separate are not fixed, and evolve in very different ways. Why should this be so? Additional information gathered for the first time at Cycle 3 throws some light on this question.

Figure 4.2  Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the frequency with which the child saw the "other parent" relative to the frequency agreed, NLSCY, Cycle 3 (N=856)

The multivariate analysis presented in Table 4.1 evaluates the impact of various factors on the likelihood that parents see their children more or less frequently than agreed.[11] Findings from the model show that two variables in particular are linked to whether or not agreements are kept to:

  • Seeing children either more frequently or less frequently than agreed is significantly less likely in shared custody arrangements than in sole custody. In other words, when parents agree to share living arrangements, the timetable is more strictly adhered to than when children live exclusively with their mother or father.
  • Families living in Quebec are also more likely to keep to the arrangements made, even after controlling for the greater incidence of shared custody in the province.

Table 4.1  Impact of given variables on the probability that children, aged 4-15 years in 1998-99, whose parents separated in the two previous years, saw the "other parent" more or less frequently than agreed, NLSCY, Cycle 3 (Multinomial logistic regression odds ratios1—N = 807)

Living arrangements
Variables More frequently Less frequently
· (With mother) 1.000 1.000
· With father 0.701 0.581
· Shared 0.281 *** 0.339 ***

Type of custody agreement
Variables More frequently Less frequently
· (Court order) 1.000 1.000
· No court order 0.679 + 0.958

Child support agreement
Variables More frequently Less frequently
· (Yes) 1.000 1.000
· No 1.159 0.993

Child's age at separation
Variables More frequently Less frequently
· (2-5 years) 1.000 1.000
· 6-11 years 1.078 0.764
· 12-15 years 0.508 * 0.890

Sex of child
Variables More frequently Less frequently
· (Boy) 1.000 1.000
· Girl 1.832 1.121

Region of Canada
Variables More frequently Less frequently
· (Canada, excluding Quebec) 1.000 1.000
· Quebec .304 * .242 ***

Reference category given in parentheses.

1 Coefficients significant at: ***p<.001; **p<.01; *p<.05; +p<.1.

In addition, children who are older when parents separate are less likely to see the other parentmore frequently than agreed than are younger children. This finding may have no relation to the separation itself; it may simply reflect the fact that children generally spend less time with parents and more time with their peer group as they enter their teens. More frequent contact than agreed is also less likely (at marginal significance) when custody was settled privately.

Why do parents see their children less often than agreed?

Interestingly, none of the other variables contained in the analysis are associated with less frequent contact, probably because the explanation is to be found elsewhere. A question was added at Cycle 3 aimed at providing some insight into this question: parents who stated that contact over the last year with the other parent was less frequent than agreed, or absent, were asked the main reason, or reasons, for this.

The distribution of replies is as follows (Figure 4.3):

  • The responding parent cancelled the visit—4%
  • The non-resident parent cancelled the visit—38%.
  • The non-resident parent has not tried to see the child over the last year—34%.
  • The child did not want to see the non-resident parent—24%.

Figure 4.3  Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the reason given for less frequent contact than agreed with the other parent, Cycle 3, NLSCY (N=259)

Respondent rarely claimed responsibility for visits being cancelled. The child's wishes were stated as the cause for close to one-quarter of cases, but mostly the visiting parent was declared directly responsible for the lower levels of contact. Evidently, these responses need to be interpreted within their context. This question is highly subjective, particularly given the relatively short time since separation. Many parents may still have problems relating to one another (and to their children), with access to children being a particularly touchy issue, and their perception of the source of problems relating to this is likely to differ. Unfortunately, the question could not be put to the other parent, making it impossible to assess the extent of any differences between separated parents. In addition, the choice of possible responses is very general, revealing little or nothing about the dynamic lying behind the stated cause.

Do parents consult their children about custody and visiting arrangements?

Children rarely participate directly in formal divorce, custody and access proceedings in Canada. A recent report commissioned by the Department of Justice recommended that they should be given more of a voice (Bessner, 2002). Most parents settle custody privately, however. Do children have a more direct say in these arrangements? Does it depend on their age when parents separate? For the first time in 1998-99, parents were asked about their child's involvement in the decisions made about custody and visiting arrangements. Unfortunately, the question was put only for children in the sole custody of mother or father; how far children were involved in arranging shared custody arrangements remains a mystery. Overall, among children aged 4-15 years, whose parents separated in the second half of the 1990s and who were placed in sole custody:

  • The majority (71%) had not been consulted, more than half (56%) because they were considered too young.
  • Among those who had been consulted (29%), their wishes prevailed the majority of the time (18%); for the others, parents asked their opinion, but took the final decision themselves.

That children's participation in the decision-making process is closely related to their age at the time of separation is clearly illustrated in Figure 4.4. The likelihood that children are involved in the discussions about their future rose dramatically with age, and the proportion of children considered too young to be consulted fell just as rapidly, with no parent declaring a teenage child too young. Nonetheless, many parents decide not to involve their children in the decision irrespective of age. Almost one-quarter of teenagers had no say in the custody arrangements made by their separating parents. On the other hand, when parents decide to consult their children on this issue, their opinion counts. Almost half the teenagers were not only consulted, they also had the final say in the decision. This was true for one-third of children aged 10‑12 years, and one-fifth of those aged 7‑9 years. In any case, even when parents make the final decision themselves, their children's wishes are likely to have been taken into account.

Figure 4.4  Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the child's involvement in decisions regarding custody and contact arrangements, by age at separation, NLSCY, Cycle 3 (N=743)

What other factors have an influence on whether or not children are consulted? Are children consulted as often when custody is subject to a court order even though they rarely have a direct say in the proceedings? The logistic regression analysis presented in Table 4.2 confirms the importance of children's age, but it also indicates:

  • Children are significantly more likely to be included in the decision-making process when custody is settled privately rather than in court.
  • In Quebec, children are much more likely to be consulted than elsewhere in Canada.
  • Being consulted about living arrangements is related, statistically, to neither the gender of the child, nor that of the custodial parent.

Table 4.2  Impact of given variables on the probability that children in sole custody are consulted about living arrangements by separating parents, among children aged 4-15 years in 1998-99, whose parents separated in the previous two years, NLSCY, Cycle 3 (Logistic regression N = 731)

Variables Odds Ratio 1
Child's age at separation
(continuous variable)
1.504 ***

Living arrangements
Variables Odds Ratio 1
· (With mother) 1.000
· With father 1.149

Type of custody agreement
Variables Odds Ratio 1
· (Court order) 1.000
· No court order 2.574 ***

Sex of child
Variables Odds Ratio 1
· (Boy) 1.000
· Girl 1.296

Region of residence in 1998-99
Variables Odds Ratio 1
· (Canada, excluding Quebec) 1.000
· Quebec 3.512 ***

Reference category given in parentheses.

1 Coefficients significant at: ***p<.001; **p<.01; *p<.05; +p<.1.

New information on child support

Deciding on how to divide the financial responsibilities for children between parents who no longer share a household is a complex question and open to much dispute. The Federal Child Support Guidelines, introduced in 1997,[12] were an attempt to establish more adequate and more consistent child support payments, and to reduce disagreement about support, by creating a more objective basis on which amounts would be estimated. The separations under observation in this analysis occurred precisely during the period following the introduction of these guidelines, and provide some insight into their initial effectiveness.

Reaching a child support agreement

As Figure 4.5 indicates, a child support agreement had been reached for 60% of children aged 4‑15 years in 1998-99, whose parents had separated within the two preceding years. Almost three quarters of support agreements were reached privately, although the majority of these had been drawn up with legal assistance. Altogether, only 16% of children had support agreements contained in a court order, although there was one in progress for another 10%.

Not all separating couples reach an agreement on child support; among the children in this sample there was no child support agreement, and none in view, for three children in ten. Nonetheless, this figure is appreciably lower than that (38%) for separations occurring in the two years preceding Cycle 1 (Marcil-Gratton and Le Bourdais, 1999, Table 12). In other words, more couples appear to be reaching a child support agreement within a relatively short period after they separate. This result may support the claim that the 1997 Federal Child Support Guidelines have "reduced conflict and tension between parents, by making the calculation of child support orders more objective" and have "improved the efficiency of the legal process and most parents are now setting child support amounts without going to court" (Department of Justice, 2002, Volume 1, Minister's preface).

Figure 4.5  Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the existence and type of child support agreement, NLSCY, Cycle 3 (N=864)

Why no child support agreement?

For the first time at Cycle 3, parents were asked why there was no support agreement. However, the fact that the reason fell outside the eight options in close to one-third of cases illustrates the complexity of the issue (Figure 4.6). Among the possible responses, the other parent's inability to pay was the most common, given for 29% of children. Four other reasons were offered in almost equal proportions (close to 10%): that child support had not been asked for; that parents could not agree; that other financial arrangements had been made; and that custody and cost-sharing were split between the parents. In a few cases, the custodial parent did not want child support or had been unable to locate the other parent.

Figure 4.6  Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the reason why no agreement was made on child support payments, NLSCY, Cycle 3 (N=270)

Child support agreements and child support payments

Throughout Canada, child support is based on two principles: the joint financial responsibility of the parents and the sharing of responsibility in proportion to the resources available. Unless parents share custody time proportionally to their relative incomes, or unless the non-resident parent has insufficient income to make any contribution at all, it would be reasonable to assume that most child support agreements involve money exchanging hands. However, at Cycle 3, respondents who declared that a child support agreement had been reached were also asked whether payments were supposed to have been made for the child during the previous 12 months, or since the separation. Surprisingly, payments were expected for under two-thirds (63%) of the child support agreements reached for children, aged 4-15 years in 1998-99, whose parents separated at the end of the 1990s.

Why is it that over one-third of those with support agreements were not expecting child support payments to be made for the child? As no direct information is available at the NLSCY about the content of agreements, we do not know whether child support payments were actually stipulated in the agreement, or whether an alternative form of financial support, such as mortgage repayments, was agreed upon. It is possible, nonetheless, to glean some insight into the question. The figures in Table 4.3 represent the percentage of children for whom child support payments were expected among those for whom a child support agreement exists. This table shows:

  • Payments were expected for four-fifths (81%) of children whose child support agreement was contained in a court order. This was so for under two-thirds (64%) of formal private agreements, and under half (47%) of informal private agreements.
  • Whether or not support payments are anticipated depends largely on the child's living arrangements after separation:
    • With the mother: child support payments are most common when children are living all or most of the time with their mother after separation; child support was supposed to be received for almost three-quarters of those with a support agreement. On the other hand, support payments were not expected for a quarter of children in their mother's custody—a proportion that reaches almost 40% when the agreement is a private informal one.
    • With the father: payments are less common for children in their father's custody; approximately one-third of those with an agreement expected child support payments to be made. This supports other studies showing that non-custodial mothers are rarely required to contribute to their children's financial support. Seldom in paid employment before the separation, these mothers are unlikely to have the resources to make a contribution (Juby et al., forthcoming).
    • Shared living arrangements: money rarely moves between households when children are in shared custody.[13] Parents who share custody are often both in paid employment, and are more likely to each assume the costs of their child's financial support when living in their household. It is also possible that the cooperative parenting required to manage shared custody arrangements may also lead to more flexible arrangements over children's financial support.
    • When child support agreements concern children of primary school age rather than younger or older children, child support payments are less frequently expected. This is also the case for agreements made in Quebec.

These factors are linked in various ways and we conducted a logistic regression analysis to understand the issue more clearly (see Table 4.4). This model confirms that, even controlling for the other variables, payments are expected significantly less frequently:

  • In the case of private child support agreements, particularly if the agreement has not been formalized.
  • If children are living with their father, or in shared custody, rather than with their mother after separation.
  • When children are aged 6-11 years at separation, compared with younger children. This is also true for older children, though the coefficient is significant only at the 0.1 level.

An interesting result relates to differences between Quebec and the rest of Canada. Table 4.3 showed that payments are expected in a lower proportion of child support agreements made in Quebec than in the rest of Canada. In the model, however, once other factors are controlled, the expectation of receiving payments is actually higher in Quebec than elsewhere in Canada.This suggests that the greater incidence of shared and father custody partly explains why the expectation of child support payments is lower in Quebec, though it may also be linked to differences in the way the Quebec maintenance enforcement program is implemented.

Table 4.3  Proportion of child support agreements that include child support payments, among children aged 4-15 years in 1998-99, whose parents separated in the two previous years, according to various characteristics, NLSCY, Cycle 3

Characteristics Proportion of child support agreements that include child support payments
N 513
Proportion of child support agreements that include child support payments 63%

Type of support agreement
Characteristics Proportion of child support agreements that include child support payments
· Court order 81%
· Private formal 64%
· Private informal 47%

Living arrangements
Characteristics Proportion of child support agreements that include child support payments
· With mother, full-time: 74%
· Court order 84%
· Private formal agreement 77%
· Private informal agreement 62%
· With father, full-time 35%
· Shared 14%

Child's age at separation (with mother)
Characteristics Proportion of child support agreements that include child support payments
· 2-5 years 82%
· 6-11 years 65%
· 12-15 years 77%

Region of residence in 1998-99
Characteristics Proportion of child support agreements that include child support payments
· Canada, excluding Quebec 66%
· Quebec 57%

Table 4.4  Impact of given variables on the probability that a child support agreement does not include child support payments, among children aged 4-15 years in 1998-99, whose parents separated in the previous two years, NLSCY, Cycle 3 (Logistic regression odds ratios1 N = 518)

Type of support agreement
Variables Odds Ratio
· (Court order) 1.000
· Private formal 1.807 *
· Private informal 4.719 ***

Living arrangements
Variables Odds Ratio
· (With mother) 1.000
· With father 4.282 **
· Shared 23.564 ***

Child's age at separation
Variables Odds Ratio
· (2-5 years) 1.000
· 6-11 years 2.808 ***
· 12-15 years 1.756 +

Region of residence in 1998-99
Variables Odds Ratio
· Canada, excluding Quebec 1.000
· Quebec .639 +

Reference category given in parentheses.

1 Coefficients significant at: ***p<.001; **p<.01; *p<.05; +p<.1.

Are child support agreements adhered to?

The question about whether or not child support payments were part of the agreement, included for the first time at Cycle 3, has improved significantly the utility of subsequent questions referring to whether or not agreements are adhered to. In earlier cycles, it was assumed that a child support agreement entailed support payments, and all respondents declaring a child support agreement were asked about the regularity of payments. In cases where the agreement did not include child support payments, it is impossible to know how parents responded to this question. Respondents may have quite rightly declared that no payments had been received, thereby swelling the ranks of "non-paying" parents, and reinforcing, erroneously, the stereotype of parents neglecting their obligations towards children with whom they no longer live.

At Cycle 3, only in cases where child support was expected were parents asked about the regularity of payments. Overall, agreements that include the payment of child support were generally respected, at least within a relatively short period after the separation (Table 4.5). Payments were made regularly for the vast majority (84%) of children, and were mostly made on time (73%). Support was irregular for one child in eight (12%); and payments had never been received, or not for the last six months, for only one child in twenty-five.

Table 4.5  Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the method and regularity of child support payments, and the proportion of payment received, NLSCY, Cycle 3

Regularity of payments
Characteristic %
N 323
· Regular and on time 73
· Regular but late sometimes 11
· Irregular 11
· No payments for at least 6 months 5
Total 100

Proportion of payment received
Characteristic %
· The whole amount 80
· Half the amount or more 6
· Under half 12
· None 2
Total 100

Although the question of the regularity of child support payments was included in all three cycles, in earlier cycles it was not possible to assess whether all, or only part of, the amount agreed upon was being paid. This information was collected for the first time in 1998-99, and showed that most child support was not only paid regularly, but was also paid in full. Over 80% of payments were for the total amount; at least half the sum agreed (and usually over three-quarters) was paid for an additional 6% of children. Less than half the amount was paid for one child in eight (12%) and no payment at all had been received for only 2% of the children for whom child support payments had been agreed upon. In addition, when payments were regular, they tended to be for the total amount: more than 90% of regular payments were paid in full, rarely the case for child support payments made at irregular intervals.

Cycle 1 data showed that child support payments agreed to privately were more often respected than those contained in a court order. This continues to be the case, as is evident in Figures 4.7 and 4.8. Private formal agreements are those in which support payments are both most regular and most often paid in full:child support is paid regularly for over 90% of children, usually on time, and the whole amount is paid in just under 90% of cases. The majority of child support contained in a court order is also paid regularly and in full, but the proportions drop to just over 70%. In other words, payments are intermittent and partial, or absent, for almost 30% of children whose financial support arrangements are included in a court order. Private informal agreements lie between these extremes, with child support less reliable than private formal agreements, but more so than settlements included in a court order.


Date modified: