Child Custody Arrangements: Their Characteristics and Outcomes



The practical requirements of shared and other forms of custody are not well researched.  What Maccoby and her associates wrote more than 10 years ago still applies today:

There is still relatively little information concerning the details of inter-parental co-operation—that is, the logistics of managing visitation and alternation, the division of responsibilities, the frequency and nature of communication, the amount of mutual undermining versus mutual backup—that prevails under different custodial arrangements (Maccoby et al., 1990: 142).

5.1 Stability of Custody Arrangements

5.1.1 Canada

There are three sources of Canadian data on child custody arrangements:  a small research study of court files in two communities, the Survey of Child Support Awards, and a national survey of parents and children.[12]

A pilot study of 1992 court files in Hull, Quebec, and Hamilton, Ontario, found that 86 percent of the arrangements involved sole custody, 9 percent joint legal custody, 5 percent split custody, and 1 percent joint physical (shared) custody (Ellis, 1995).  The Hull sample involved a higher percentage of sole custody cases than did the Hamilton sample (94 versus 79 percent) with the difference explained by the lack of use of joint legal custody in Quebec.  Nine out of ten sole custody parents were mothers.  Both separation agreements and divorce orders were included in these data.

The Survey of Child Support Awards (1998 to 1999 cases) provided data on custody arrangements as found in court files in selected courts across Canada (Bertrand et al., 2001).  The mother had sole custody in 80 percent and the father in 9 percent of cases.  Shared custody, defined in the Federal Child Support Guidelines as the children spending at least 40 percent of the time with each parent, was reported in 5 percent of cases.  There was a fairly large variation by jurisdiction in the incidence of shared custody, from 1 to 8 percent.  Split custody was found in 5 percent of the total cases, with a range of 3 to 7 percent of divorce cases depending on the province and territory.

As expected, at the time of separation, most children younger than 12 years are in the sole custody of their mother, according to data from the 1994-95 cycle of the National Longitudinal Survey of Children and Youth (Marcil-Gratton and Le Bourdais, 1999).  Table 1 shows that among all families sampled and according to court orders, 79 percent of these children were in their mother's custody, 7 percent were in the custody of their father, and 13 percent were in a shared custody arrangement.  There was a small difference in court-ordered arrangement by the type of union, with common-law couples reporting a lower proportion of shared custody arrangements.

Table 1
Court-ordered Custody Arrangements at Separation, by Type of Broken Union, NLSCY, Cycle 1 (1994-95)

Common-law relationship Marriage, common-law before marriage Marriage, no common-law before marriage Total sample
Mother exclusive custody 84.1 74.3 82.0 79.3
Father exclusive custody 6.2 7.9 5.3 6.6
Shared physical custody 8.7 16.8 10.9 12.8
Other 1.0 0.9 1.8 1.2
 Total percent 100.0 99.9 100.0 99.9
 Weighted number 328 489 409 1,239

Source: Marcil-Gratton and Le Bourdais, 1999.  The survey is confined to children under 12 years of age.  The unit of count is the child.

Still using the NLSCY as the data source, the actual arrangements at separation varied from what was in the court order.

Court-ordered custody status therefore frequently does not reflect the reality of physical placement.

The NLSCY analysis also revealed that shared residence arrangements decreased with time.  Among children of married couples who had separated, 13 percent had shared custody in the first two years after separation, compared to 10 percent two to four years after separation and 7.5 percent five or more years after separation (Marcil-Gratton and Le Bourdais, 1999: 27).

5.1.2 United States

Only about 15 to 20 percent of children live in dual residences, even in those U.S. states "that are at the forefront of the joint custody movement," such as Washington and California (Pruett and Santangelo, 1999: 391).  The authors suggest that rates of shared custody in other parts of the country are likely to be lower.  This assumption is supported by estimates from two national surveys in the United States that found that 12 to 13 percent of all households with formal custody arrangements had a shared custody arrangement (Donnelly and Finkelhor, 1993).

A longitudinal study in California found that actual custody arrangements did not correspond to the court order.  Twenty percent of divorce cases in the Stanford Child Custody Project resulted in joint legal and physical custody according to the court order.  Of the couples with this type of order, only one half had a dual residence arrangement with time shared between the parents a few years after the divorce (Maccoby and Mnookin, 1992: 198-199).  In most of the remainder of the families, children lived in a primary residence and the other parent had access.

Three types of custody arrangements—sole maternal, sole paternal and shared—were tracked over two years in a Canadian study (Cloutier and Jacques, 1997a and 1997b).  One half of the children in a shared custody arrangement changed custody arrangements during this period.  Girls who were in the custody of their father changed their physical placement more often than other girls.  These girls typically went to live with their mother.  Boys who changed custody normally divided their time equally between their parents.

When sole maternal and sole paternal custody arrangements are compared, non-resident mothers are more likely to regain custody than are non-resident fathers (Stewart, 1999).

Thus, empirical data from both Canada and the United States show that custody arrangements are not necessarily stable and that the actual incidence of shared custody is considerably lower than what is found in court files containing the initial arrangements.  Maccoby comments on this phenomenon:

At the time of separation, many parents are not in a position to make good decisions concerning residence, visitation, and financial support for their children.  [As time passes] the language of the formal agreement concerning custody and visitation begins to fade into the background₀  The changes that families make in the residential and visitation arrangements for their children, or even in the amounts of child support mutually agreed upon, are usually made informally, without going back to court for formal modification of the terms of the divorce (Maccoby, 1999: 66-67).

The instability of shared custody arrangements need not be considered a negative outcome for children.  In some, perhaps even many, cases the children themselves have asked to move.  It would be valuable, however, to obtain quantitative data on the reasons for the adjustments to the initial arrangements.

5.2 The Logistics of Shared Custody and Other Arrangements

No research could be found that fully describes the logistical or practical requirements of different custody arrangements.  Even qualitative descriptions of the various types of custody arrangements are rare.

5.2.1 Amount of Parent-Child Contact

The analysis by Pearson and Thoennes (1990: 243) looks at visitation frequencies as specified in the court order (Table 2).  The proportion of time that the non-resident parent had with the children, according to court orders, was 20 percent for sole custody, 28 percent for joint legal custody, and 40 percent for shared custody.  Interviews revealed that complaints of missed or sporadic visits were much more common in sole custody cases than in joint legal and shared custody arrangements (bottom row of Table 2).

Table 2
Amount of Child-Non-resident Parent Contact by Type of Custody Arrangement in a Sample from the United States

Sole custody
with visitation
Joint legal
Mean number of days specified in court order
  Weekdays 19 45 188
 Weekends 30 55 49
 Overnights 43 87 137
Percentage of court-ordered time that non-resident parent had access to the children 20% 28% 40%
(most common)
Percentage of resident parents who reported sporadic access by non-resident parent 54% (maternal custody)
44% (paternal custody)
7% (maternal custody)
20% (maternal custody)
0% overnights and weekends
12% other missed visits

Source: Pearson and Thoennes (1990).

In her voluntary Canadian sample of 16 co-operating and 16 disagreeing ex-couples (16 in each group) Ehrenberg (1996) reported the percentages of the children's time spent with their mother versus their father.  She concluded that the percentage of time children spent with each parent after separation or divorce was influenced by whether their parents were able to co-operate about parenting rather than the type of custody arrangement (Table 3).

Table 3
Time Spent with Each Parent, Co-operating and Disagreeing Ex-couples

Type of custody arrangement Co-operating parents:
Percentage of time children spend with mother
Disagreeing parents: Percentage of time children spend with mother
sole (maternal) custody 64%  (n = 5) 87% (n = 11)
joint legal custody 66% (n = 5) 74% (n = 5)
shared custody 55% (n = 4)  
split custody 50% (n = 2)  

Adapted from Ehrenberg (1996).

The small sample size and the voluntary nature of the sample mean that caution should be used in generalizing these relationships.  In this sample, children with disagreeing parents are less likely to see their father frequently than those with parents who co-operate.  As Ehrenberg noted (1996: 112), this could mean either that the children's time is a source of conflict for disagreeing ex-couples or that co-operation facilitates equity in the amount of time children spend with each parent.  Other evidence, albeit from the United States (e.g. the Stanford Child Custody Project), suggests that the former explanation (that it is the children's time which is the source of conflict) may apply in many cases.

Sole custody arrangements that involve frequent access by the non-resident parent may not differ in many respects from equal or almost equal shared custody arrangements.  However, the content of parent-child interaction between families where the father has the children on the weekends and holidays and those where the father has responsibility during the week probably differs greatly (Pruett and Santangelo, 1999).  On weekdays, child care is "normalized" because the resident parent is responsible for supervising bedtimes, curfews and homework.  The time is much less likely to be spent in play and recreational activities.

5.2.2 Scheduling

Descriptive information suggests that shared custody typically involves the children spending four days a week with one parent and three days a week with the other, or spending one or two weeks alternatively with each parent.  Luepnitz (1986) reported oddities, such as families in which the children spend half a day with each parent, live half a year with each parent, or alternate years.  Other extremely rare arrangements include those where the children reside in the same house and the parents move in and out, and the parents live in such close proximity that both interact daily with the children. The most common arrangement was some form of split week.

Arendell (1995b) interviewed only a small number of shared custody fathers in her study (9 of 75 cases).  Most of these parents either divided the week in two halves or alternated weeks.  Scheduling entailed regular meetings with the other parent to review and plan the schedule.  Respondents emphasized that scheduling was the "key to success" and required a willingness to be flexible when circumstances warranted.  For younger children, calendars were kept so that the children could know where they would be.  Older children were included in the planning and discussions of needed adjustments and were allowed to request adjustments, but were kept out of parental disputes.  According to the fathers in the study, they and their ex-spouses maintained direct contact with each other and did not use the children as intermediaries.  Children telephoned freely between their homes.  The majority of these families had child care assistance from relatives.  In several cases, these relatives—grandmothers and in one case a stepmother—were involved in the planning of the children's schedules.

About one half of the shared custody parents interviewed by Irving et al. (1984) had divided time with their children equally, with the typical schedule being a one-week rotation and an equal division of holidays.  The next largest group (30 percent) chose a 75:25 split.  In this group the children would spend the school week with one parent and most weekends with the other, and the holidays are divided equally.  The remainder of the arrangements ranged from splitting the week equally, to having the children spend the school year with one parent and the summer holidays with the other.  Residential proximity and scheduling were closely related.  Parents who lived close together tended to have a more equal time-sharing arrangement.  In the total sample, 46 percent of the parents lived a short drive from each other and 32 percent lived within walking distance.

These schedules were not problem-free according to parental interviews.  Initial difficulties arose in about one half of the cases.  While many respondents reported early success in ironing out these problems, 32 percent said that the problems had lasted a year or more and 13 percent continued to experience scheduling difficulties (Irving et al., 1984: 131).  Scheduling of holidays and household transitions was also mentioned as an issue by the shared custody parents interviewed by Rothberg (1983).

5.2.3 Task Sharing

Arendell (1995a) enumerated the following activities that could be involved in shared parenting: sharing of major and daily decisions about child rearing and co-parenting, sharing responses to the children's school and medical problems, planning special events in the children's lives, discussing the children's adjustment to divorce and their progress and accomplishments, and examining and planning child-related finances.

One study was located that quantified the differences in parental activities and responsibilities by the type of custody arrangement.  In the Pearson and Thoennes study (1990), 90 percent of sole custody parents reported full responsibility for assisting with homework, driving the children to activities, attending school events, arranging visits with friends, and staying home with sick children.  In joint legal arrangements, 73 to 85 percent of the resident parents were responsible, depending on the task.  In shared custody, however, about one half of the mothers said that they shared these responsibilities equally with their former partner.  The exception was staying home with a sick child:  fewer than 35 percent of parents with shared custody shared this responsibility equally.

5.3 The Costs of Access and Shared Custody

Two research questions are addressed in this section.

5.3.1 Costs of Child-Non-resident Parent Visits

Australian data on the additional expenditures associated with access visits to the homes of non-resident parents are available (Murray Woods and Associates, 1999).  The data were collected by telephone survey with a volunteer sample of 252 non-resident fathers who had their children between 18 and 110 overnights annually; two thirds of the sample had between 55 and 110 nights of contact.  The sample was drawn from the Australian child support agency client list.  The fathers interviewed had much higher median incomes than did child support payors as a whole.  Interestingly, while this could be construed as invalidating the generalizability of the findings, , the analysis found that income was not related to the number of expenditures (see below).

The large majority of the parents in this sample (about 90 percent) provided a separate bedroom for each child.  The following is the frequency of items that the fathers purchased, "mainly because the child needed or wanted the item:"

Other expenditures were as follows.

Four out of five non-resident parents (81 percent) said that the other parent had not shared the costs of contact.

This research includes an index of the number of expenditures made by the non-resident parent (although the dollar value was not included).  When the figures in the index were correlated with other variables, it was found that there was a significant positive relationship between expenditures and the number of nights of contact.  That is, as the number of overnights increased, the number of items the non-resident parent purchased also increased.  There was also an increase in the number of items purchased as the children got older.  The income of the non-resident parent was not, however, associated with the expenditures, suggesting that parents purchase similar numbers of items for their children during their access visits regardless of income.

5.3.2 Efforts to Estimate the Costs of Shared Custody

A shared custody arrangement is widely believed to cost more than a sole custody arrangement (Carberry, 1998; Arendell, 1995a; Morrow, 1995; Melli and Brown, 1994; Zinner, 1998).  There is little quantitative research on this topic, perhaps in part because of the complex and controversial calculations that are required to estimate the costs of raising children.

In a shared custody arrangement, each parent pays part of the children's fixed and non-fixed expenses, and the total expenditures of both parents increase.  One parent's expenses do not decrease for every dollar that the other parent pays in expenses because each parent is responsible for fixed expenses that both parents must pay, such as a bedroom for the children, toys and utilities.  Morgan (1999) drew the comparison to child support guidelines that recognize that the amount to support two children is less than twice the amount needed to support one child because certain household expenses are shared.

Two studies were located that employ unsophisticated calculations to estimate the costs of shared custody in contrast to the costs of intact families.  Both originated in the United States and used statistics from the Department of Labor's itemization of expenditures by intact families in urban areas.  A third study, from Australia, estimated the costs of access by non-resident parents at different levels of annual parent-child contact.

Melli and Brown (1994) set out some of the complexities involved in estimating the costs of shared custody.  As mentioned above, some costs are fixed and must be borne by both parents.  This is because of the need to duplicate housing, utilities and other items.  Expenditures on these items make up from 24 to 34 percent of the total child-related expenditures for a child up to the age of 18, and must be duplicated when a child resides with both parents (Melli and Brown, 1994: 554-5).

There may be child-related costs that decrease with shared custody, such as child care expenses, because the parents are able to coordinate work schedules in a way that reduces the need for paid care.  This may be unlikely when both parents work full time and the children require child care during the day or before and after school.

Some child-related costs may or may not change with shared custody.  This category contains expenses that are not necessarily related proportionately to the amount of time the children spend with each parent, such as clothing, medical care and school expenses.  These items were estimated to constitute about 25 percent of child-raising expenses.  As the Australian survey of non-resident parents found, these parents frequently incurred expenses in this category, especially for clothing.

Expenses that are almost directly proportionate to the amount of time that the children spend with each parent include items such as food, recreation and some transportation costs.  However, additional transportation costs may be required to manage the changeovers from one home to another.[13]  It could be argued that because many parents with shared custody live close together, any additional transportation may be offset by the shorter distances travelled.  Melli and Brown (1994) estimated that these items make up between 40 and 50 percent of the budget for a child younger than 18.  These costs represent the largest reduction in the expenditures of the "primary" parent.

Thus, according to this analysis, from one quarter to one third of the total costs of child rearing must be duplicated in shared custody arrangements.  This would appear to be a minimum, given that some costs are duplicated for clothing and personal care items, and additional transportation costs are involved in many cases.  The costs of paid child care would not be duplicated, but it is possible that non-resident parents could be responsible for the costs of paid child care when the children are in their home.

The only other study from the United States that could be located is an inadequate and highly misleading paper written in the early 1980s (Patterson, 1984).

In Australia, Henman and Mitchell (2001) undertook the sole methodologically sophisticated attempt to estimate the costs of contact by non-resident parents.  This research used the budget standards method, which is a normative approach to determine living standards.  The estimates produced do not represent actual expenditures but rather costs required to meet a specified standard of living.  It is beyond the scope of this report to describe the complexities of the budget standard calculations or those of the other methods used to estimate the costs of children.  Suffice it to say, there are several methods utilized in the economics literature and they often reach quite different estimates (see, for example Harding and Percival, 1999).  The budget standard approach is, however, one of the main methods.

The researchers calculated two estimates of the cost of maintaining contact with children, one for non-resident parents with a low or frugal standard of living and another for parents with a modest but adequate standard of living.  All costs of contact with children were expressed in annual dollar amounts (Australian dollars) and as a percentage of the annual costs of children in intact families.  The costs refer only to the additional costs of contact to the non-resident parent.

The costs were calculated for parent-child contact of 15, 20 and 30 percent of the year.  Based on the researchers' assumptions, the costs of increased contact are not large.  For example, when contact increases from 20 to 30 percent of the year (a 50 percent increase), the cost of contact increases only by 8 to 12 percent, depending on the assumptions made.  This is because the costs of "infrastructure"—primarily housing but also furniture and toys—are only minimally affected by moderate changes in the amount of contact.

Another variable introduced into the calculations by the authors was the cost of transportation.  They found that a moderate variation in the distance between the non-resident parent and the children had only a modest effect on costs:  there was only a 3 percent difference in the cost of contact when parent had to travel 15 kilometres as compared to 50 kilometres.  However, the price of gasoline has risen since this research and the differences could well be higher today.

The costs of contact expressed as a proportion of costs of raising children in intact families were higher at the frugal than at the modest but adequate standard.

From these data, the authors conclude the following:

The higher proportional cost of contact at the low cost standard suggests that there is a basic set of unavoidable costs associated with contact that do not increase proportionately as the living standard rises.  Housing, transport and household infrastructure are clear examples of this unavoidable basic set of contact costs₀  This explanation of a basic set of costs ₀ may also explain the tendency for the proportional costs of contact to drop when the number of children increases (Henman and Mitchell, 2001: 32).

This research therefore quantifies the frequently made assumption that the total cost of raising children increases substantially when parents separate.  Household infrastructure, such as bedrooms, furniture and toys, and transportation costs are the primary reasons for the higher costs.  In order for these findings to be extrapolated to Canada, the normative assumptions would have to be confirmed by data on the behaviour of non-resident parents in this country.  In addition, data on the costs of raising children in intact families with different standards of living would have to be available.  From the perspective of Canadian social policy, perhaps the most important finding from this work is that the different frequencies of child-non-resident parent contact did not greatly affect the estimates of increased costs.

5.3.3 Costs of Shared Custody and Sole Custody with Frequent Access

There does not appear to be any research that explicitly addresses the differences in costs between shared custody arrangements using the 40 percent standard, as set out in the Federal Child Support Guidelines, and sole custody arrangements involving frequent access by the non-resident parent.  The studies described above do however provide some insight into the magnitude of the differences.  It appears that from the Australian budget standards research and the Australian survey of non-resident parents that the differences in expenditures may be incremental.  The costs of raising children in two households—regardless of the number of overnights involved and the income of the non-resident parent—are substantially higher than the costs of child rearing in intact homes.

5.4 Summary

In Canada, court files show that about 80 percent of divorce cases result in a sole maternal custody arrangement; 9 percent in a sole paternal custody arrangement, and 5 percent in each of shared and split custody arrangements.  A nationally representative survey involving children under 12 years of age found a similar proportion of sole maternal custody arrangements, but higher proportions of shared custody arrangements (court-ordered custody arrangements at separation).  The same survey also found that the actual arrangements differed from what was in the court order (e.g. 87 percent of children lived with their mother).  There is a good deal of evidence that custody arrangements change over time for a variety of reasons, including the wishes of the children.

There is little information on the logistics of shared custody arrangements.  Scheduling of shared custody can vary widely.  Qualitative data suggest that split weeks or children spending one to two weeks alternatively with each parent are typical.  Mothers in shared parenting arrangements are much more likely than other mothers to report that they share most child-rearing tasks equally with their ex-partner.  The exception was staying home with a sick child, a responsibility that tends to fall more often to the mother.

A survey of non-custodial fathers in Australia (a volunteer sample) found that these fathers made a number of expenditures for their children related to contact visits, including providing a separate bedroom for each child.  As the number of overnight visits increased so did the number of items purchased by the non-resident parent.  The income of the non-resident parent did not affect the number of expenditures.

Another Australian study used the budget standards approach to quantify the costs of raising children after separation or divorce.  The cost of raising a child who spends 30 percent of the year with the non-custodial parent is from 46 to 59 percent higher than the cost of raising the child in an intact household (with the variation depending on the standard of living of the parents).  Household infrastructure, such as a bedroom, furniture and toys, and transportation costs are the primary reasons for the higher cost.  There was little difference in the estimated costs for different frequencies of contact visits.

No Canadian studies were found that quantified the differences between the costs of raising children in a shared custody arrangement, based on the 40 percent definition used in Canada, and the costs of a sole custody arrangement involving frequent visitation by the non-resident parent.