Introduction

Fifteen years ago in the pages of this journal, Cancian and Meyer (1998) documented a significant shift in the living arrangements of children following divorce: between 1986 and 1993/1994, the proportion of mothers granted sole physical custody had fallen from 80 % to 74 %, and the proportion of parents sharing custody had doubled from 7 % to 14 %. In this short article, we update those results, demonstrating that the shift captured in the earlier article was just the beginning of a major transformation. Using newly available data from court records that provide detailed information on custody outcomes in divorce cases in Wisconsin through 2008, we document a sustained and dramatic pattern of change in mother–sole and shared custody, even while father–sole custody remained relatively stable. Because marriage and divorce patterns have changed considerably over this period, we then explore whether the changes in custody can be explained by changes in characteristics of the families experiencing a divorce or by changes in the process by which custody is determined.

Policy Context and Literature Review

Divorce decrees almost always include provisions prescribing with whom any minor child will live, called “physical custody.” (Divorce decrees also include provisions for legal custody, or decision-making responsibility.) When paternity is formally established for nonmarital children, the paternity establishment action can also include provisions for where the child will live. For most of the twentieth century, in both divorce and nonmarital cases, the most common outcome when children did not live with both parents was for them to live with their mother (e.g., Buehler and Gerard 1995). This conformed with gender norms in which mothers were seen as better caretakers of children, especially young children. However, more recently, policies with explicit gender preferences have been overturned, and a more general guiding principle of the “best interest of the child” has replaced an explicit statement that children should live with their mother. For example, the current Wisconsin statute (767.24(5)) states explicitly, “The court may not prefer one potential custodian over the other on the basis of the sex or race of the custodian.” Late in the twentieth century, some states changed their custody policy further so that legislation was not merely gender-neutral, but even had provisions encouraging the involvement of both parents. Wisconsin again provides an example, with its 2000 statute (767.24(4)(b)) stating, “A child is entitled to periods of physical placement [custody] with both parents unless, after a hearing, the court finds that physical placement with a parent would endanger the child’s physical, mental or emotional health.”

Definitions

In this article, we consider legal decisions regarding physical placement or custody (i.e., with whom children are to live following divorce). Several physical custody arrangements are possible. “Sole custody” means that children primarily or exclusively live with one parent; the other parent may have a schedule for seeing the children, in some cases including a limited number of overnight visits. As used in this article, “shared custody” refers to cases in which the children spend a significant number of overnights with each parent. For some analyses, we differentiate between “equal” shared custody and “unequal.” States have different thresholds for what distinguishes sole custody from unequal shared custody; here, we use the current Wisconsin distinction in which unequal shared custody involves a child staying with one parent 25 % to 49 % of the time and the other parent 51 % to 75 % of the time,Footnote 1 with “time” being defined by the number of nights spent with the parent. This then means that the five main custody types for a single child who lives with at least one parent can be differentiated by the percentage of overnights spent with the mother: mother–sole (76 % to 100 %), shared with mother primary (51 % to 75 %), equal shared (50 %), shared with father primary (25 % to 49 %), and father–sole (0 % to 24 %). In multiple-child families, another possibility is “split” custody, defined here as at least one child having primary residence with the mother and at least one child having primary residence with the father. Finally, some children do not live with either parent; and, in a multiple-child family, there may be other combinations (one child with sole custody and another with equal shared time, for example).

Custody Outcomes and Characteristics Associated With These Outcomes

No recent nationally representative data set has detailed information on custody outcomes. The most recent detailed national data are dated, containing information from divorces in the late 1970s and 1980s (Kelly and Rinaman 2003). Some studies of divorces that cover at least part of the 1990s provide information on custody outcomes, but these studies are often limited by an inability to differentiate physical custody (where children live, our interest here) from legal custody (who makes decisions about children) (e.g., Clarke 1995; Donnelly and Finkelhor 1993). The 2011 Current Population Survey-Child Support Supplement (CPS-CSS) does provide information on the number of custodial mothers and custodial fathers, based on where individuals say their children live, with 81.7 % of custodial parents being mothers and 18.3 % being fathers (Grall 2013). However, only a crude categorization of custody outcomes is possible with these data; some of these parents have sole custody, others have unequal shared custody, and others have equal custody.Footnote 2 New research is needed.

Empirical research has used frameworks in which custody is related to economic factors, bargaining, and the characteristics of children and parents. Standard economic theory of the family posits that within marriage, spouses can benefit from specialization (Becker 1981). Thus, given higher market wages for men, a traditional arrangement would have husbands focus on the labor market and wives on caregiving. In a traditional family, if mothers were awarded sole custody at divorce, this would most closely parallel predivorce roles (Weiss and Willis 1985). Consistent with this perspective, couples in which both parents are employed are generally more likely to have shared custody than mother–sole custody (Cancian and Meyer 1998; Juby et al. 2005).

Another economic consideration arises from the fact that shared custody is more expensive than sole custody in that parents need to have sufficient resources for both their households to accommodate children being present a substantial portion of the time (Melli and Brown 1994). Consistent with this perspective, shared custody is more likely among couples with more income (Bartfeld 2011; Cancian and Meyer 1998; Cook and Brown 2006; Donnelly and Finkelhor 1993; Juby et al. 2005).

Another model of custody outcomes posits that custody is determined through a bargaining and negotiation process (Mnookin and Kornhauser 1979; Teachman and Polonko 1990); consequences could include that the parent with more power in the relationship or the parent who initiates divorce proceedings would be more likely to get the custody outcome he or she desires. In support of this framework, as a mother’s share of the couple’s total income rises, she is more likely to have sole custody (Cancian and Meyer 1998). Also, if only the father has legal representation, father–sole custody and shared custody outcomes are more likely (Cancian and Meyer 1998; Cook and Brown 2006).

A final framework stems from studies of child development and suggests that the court should consider a child’s own perspective and/or select a custody outcome thought to be most advantageous based on a child’s characteristics. Consistent with this framework, older children are often asked their perspective in the custody determination process. If courts believe that mother–sole custody is most appropriate for younger children and girls, or believe that a parent with more relational complications (previous marriages or children from other relationships) might create some difficulties for children, they might give sole custody to a parent who had fewer complications. Again, there is some empirical support: father custody is more likely when children are older or boys (Cancian and Meyer 1998; Fox and Kelly 1995; Juby et al. 2005). New partners, previous partners, and previous children in general decrease the likelihood of sole custody (Cancian and Meyer 1998; Juby et al. 2005).

Trends

The CPS-CSS shows that the number of custodial fathers increased from 2.18 million in 1993 to 2.64 million in 2011; during this period, the number of custodial mothers also increased but at a slower rate (from 13.69 million to 14.44 million). As a result, the proportion of all custodial parents who are fathers increased during this period, from 16.0 % to 18.3 %. However, because of the limitations of the CPS-CSS, it is unclear whether this increase reflects a greater proportion of father–sole custody or instead more fathers reporting that they are custodial parents because they have some type of shared custody. Indeed, state-level empirical research that has examined actual custody arrangements shows substantial increases in shared custody without much increase (if any) in father–sole custody (Cancian and Meyer 1998; Cook and Brown 2006), and some international research suggests that a significant proportion of children thought to be living with their father only are actually sharing time with both parents (Toulemon and Pennec 2010).

The review of the legal environment suggests trends toward policies that encourage (or no longer discourage) shared custody or father custody. A trend in custody might also be related to changes in the characteristics of cases that get divorced. To the extent that rigid gender roles have changed over time, with fathers making greater contributions to caregiving (Livingston and Parker 2011; Pleck 1997) and mothers making greater contributions to earnings (Blau and Kahn 2007), one might expect decreases in mother–sole custody. Moreover, during this period, cohabitation has increased markedly; and marriage, especially among those with lower education, has declined and now occurs later (McLanahan 2004). To the extent that couples with lower socioeconomic status are not getting married (or are marrying later, after childbirth), perhaps couples getting divorced increasingly have moderate or high income. This then could be linked to an increase in the likelihood of shared custody. These changes over time increase the importance of new studies with more recent data, especially studies that consider the extent to which changing characteristics of cases are related to custody outcomes.

Data and Methods

Data and Sample

To update the earlier results and trace changes in custody arrangements, we use two decades of Court Record Data (CRD) for Wisconsin, which include information collected from the court records of more than 10,000 divorce cases with minor children. The original analysis (Cancian and Meyer 1998) included divorces through 1993 (although it included a handful of cases in 1994); here, we update the results with comparable information on divorces through 2008. These data are unique in including detailed information on physical custody and child support arrangements over an extended period, as well as information on family’s and parents’ characteristics (Brown et al. 1994). Data come from 21 Wisconsin counties, including the largest urban county in the state, Milwaukee. We also use information from administrative records of earnings with these records matched to the divorce records through Social Security numbers. All results presented here are weighted to adjust for different sampling probabilities across counties.

We use data from all 13 waves, or cohorts, of the CRD for cases coming to court between 1987 and 2007; divorces were finalized during the period from 1987 through 2008.Footnote 3 Cases were collected by the date of the court petition, not by the date of the final divorce itself; for ease of interpretation, we show cases based on the year of divorce. However, data were not collected in every year; for example, there is a gap in the mid-1990s. In addition to tracking changes in who gets custody over the entire period, we also consider changes between the early period (1987 to 1993), the middle period (1997 to 2002) and the most recent period (2003 to 2008).Footnote 4 The early period generally matches the period studied by Cancian and Meyer (1998); the later periods are set to minimize difficulties caused by years without data collection and to cover approximately five years in each period. We analyze cases that have minor children at the time of the final divorce judgment, and for whom we can determine custody.Footnote 5 In the descriptive analysis, we have 9,873 cases (3,611 from the early period of 1987 to 1993; 2,407 from the middle period of 1997 to 2002; and 3,855 from the recent period of 2003 to 2008).

Measures

We use information contained in the divorce record’s final judgment to categorize cases into different physical custody types. Following the earlier analysis, we distinguish a series of divisions between mother and father that range from: (1) mother–sole custody (more than 75 % of overnights with mother); (2) shared custody (including mother–primary shared custody, equal shared custody, and father–primary shared custody); and (3) father–sole custody (more than 75 % of overnights with father). For the basic descriptive analysis, we also differentiate equal shared from unequal shared custody and consider split custody (in which at least one child lives with the mother and at least one lives with the father), but there are too few cases with father primary shared custody or with split custody to use all six categories in the multivariate analysis.

The court record contains information on a variety of characteristics of the court case, the parents, and the children. When these characteristics vary within a case, we use measures taken at the final divorce judgment. The court record typically contains information on each parent’s income. When income is not available in the court record, we use a measure of earnings from the Unemployment Insurance system. We adjust income for inflation to 2012 U.S. dollars, using the Consumer Price Index for All Urban Consumers (CPI-U).

Methods

In addition to presenting descriptive information on trends in custody, we examine the factors related to different custody outcomes, following the previous analysis and estimating multinomial logit models of whether a case has mother–sole custody, shared custody, or father–sole custody (excluding the split custody cases). We consider a range of factors measured in the court record that may be related to the likelihood of different custody outcomes, including parents’ economic characteristics (their employment, total income, and the share of income attributed to the mother); the legal environment (which parent petitioned for the divorce, whether they were represented by counsel); and characteristics of children (their number, age, and gender). We control for other factors as well: the length of the marriage, whether either parent had had previous children with a different partner, whether the father was substantially older than the mother, and the county.Footnote 6 We estimate the model separately for the early and most recent periods, although we discuss the robustness of our findings to a pooled model. We exclude the split custody cases and fewer than 60 cases in each period missing information on selected independent variables (final N = 3,427 for the early period and 3,739 for the recent period).

We then use the results to conduct simulations that illustrate the extent to which cases with particular characteristics would be predicted to have different custody outcomes in the two different periods. We conduct a simple thought experiment, asking what the predicted distribution of custody outcomes in the most recent period would be if the characteristics of cases had not changed from the early period. We then ask what the predicted custody outcomes would be in the most recent period if the characteristics of cases had changed but the custody process (the coefficients) had not changed from the early period. These simulations are not the equivalent of a formal decomposition model because they do not derive from an underlying linear model (see, e.g., Fairlie 2005) but do provide an indication of the relative magnitudes of the shifts associated with the two types of changes.

Results

We begin by examining the trend in physical custody. Figure 1 shows significant change over time and demonstrates an acceleration in the trends documented in the earlier research. For example, Cancian and Meyer (1998) documented that between 1986 and 1993/1994, the share of cases awarded mother–sole custody fell from 80 % to 74 %. Here we show that by 2008, mother–sole custody declined further to 42 %. This decline is largely mirrored by a dramatic increase in shared custody: equal shared custody increased from 5 % to 27 % of all cases, and unequal shared custody increased from 3 % to 18 % of all cases. Most of the unequal shared custody cases—more than 80 %—have children staying with mothers the majority of the time (mother–primary shared custody). There is little change in the share of cases that are awarded father–sole custody: 11 % in 1988 and 9 % in 2008. Overall, the trend away from mother–sole custody and toward shared custody is dramatic, representing a substantial change in the living situations of children of divorce over a relatively short period. Figure 1 also shows that this change occurred throughout the period and does not seem to be greatly affected by the 2000 change in the custody statute explicitly stating a preference for placement with both parents.

Fig. 1
figure 1

Child custody at divorce in Wisconsin, 1988–2008. aIncludes a small number of cases from 1994 (12) and 1995 (1). bIncludes a small number of cases from 1996 (22). cIncludes a small number of cases from 2009 (20) and 2010 (2)

Full size image

In the analysis of the early period, Cancian and Meyer (1998) showed that shared custody is more common when both parents (rather than just the father) are employed, among parents with higher incomes, when the mother has a prior marriage, when the father does not have children from a previous relationship, when all the children are boys, and when only the father has a lawyer. They also showed an increase over time, controlling for these and other covariates. (As we show later, the analyses reported here for the early period, with one less year of data, largely show the same results.)

The distribution of some of these characteristics has changed substantially over time, but others have been more stable, as shown in Table 3 in the appendix. For example, mean total family income increased from about $65,000 in the early period to $80,000 in the recent period (in constant 2012 U.S. dollars), which was just prior to the height of the Great Recession. This is perhaps related to marriage being increasingly concentrated among higher-status individuals (see, e.g., McLanahan 2004). On the other hand, the share of that total income from the mother was fairly stable, increasing from 39 % to 41 %. Mothers are older in the more recent cohort (mean age of 36, compared with 33), and both parents are somewhat more likely to have had a prior marriage. The court process has also changed over the period: in 53 % of cases, both parents had a lawyer in the early period, declining to 40 % in the recent period.

Table 1 examines whether family characteristics are related to custody outcomes, showing the results of the multinomial logit models. Table 2 shows predicted probabilities of the three custody outcomes, given the coefficient estimates in Table 1.

Table 1 Multinomial logit models of custody, two periods
Full size table
Table 2 Simulation of estimated probabilities of custody for selected cases in two periods (percentages)
Full size table

Predicted probabilities that are derived from the estimates of nonlinear models, such as the multinomial logit, and that use mean values for independent variables need not match the observed probabilities (Borooah 2002). Thus, the first row of Table 2 need not match the observed data, which are shown in the bottom row of Table 2 (and derived from Fig. 1).Footnote 7 Nonetheless, the base estimates in the first row of Table 2 are similar to the observed custody outcomes: cases in the early period were very likely to be predicted to have mother–sole custody (87.1 %), this prediction declined to 52.3 % in the recent period, and the proportion of cases estimated to have shared custody increased dramatically from 7.5 % to 44.2 %. Father–sole custody is predicted to show a slight decline between the periods. Although the levels of predicted probabilities do not precisely match the observed custody outcomes, the trends are quite similar.

To examine the relationship between custody outcomes and income, in rows B–G, we show estimated probabilities for three levels of total income and vary each level by whether the father’s share of income is 80 % or 50 %. The coefficients on Table 1 and the estimated probabilities in Table 2 show that cases with higher total income are more likely to have shared custody in both periods; in the early period, the relationship was particularly strong, with the probability of shared custody about doubling when income rose from $30,000 to $75,000, and almost doubling again when it rose to $120,000. In the more recent period, shared custody remains more likely for higher income families, but the magnitude of the variation in outcomes by income is smaller. Turning to the mother’s share of total income, we see in Table 1 that when mothers have a higher share of income, all else equal, father custody is statistically less likely. However, the simulations show that estimated probability for father–sole custody is not greatly affected by the parents’ relative income in the most recent period; at each level of total income, father–sole custody declines by about 2 percentage points when we compare mothers whose incomes are one-half the total with those whose incomes are only one-fifth of the total.

In both periods, parents with previous marriages or previous children are less likely to receive sole custody. For example, when the mother has a previous marriage and previous children but the father does not (row H, Table 2), father custody is about twice as likely as in the base case. Similarly, if the father has a previous marriage and previous children but the mother does not (row I, Table 2), he is predicted to receive sole custody only about 1 % of the time.

Table 1 shows that child’s age and gender were important to custody outcomes in the early period: fathers were unlikely to get sole custody of young children, and shared custody was more likely than mother–sole custody when all children were boys. Most of these relationships are no longer statistically significant in the most recent period. The simulations in Table 2, rows J and K, show large differences in estimated custody outcomes in the early period between a case with one young girl and a case with two older boys; in the more recent period, the case with two older boys is more likely to receive father–sole custody, but the gender and ages do not matter as much in the later period.

The court process remains strongly related to custody outcomes. In cases in which only the mother has a lawyer and is the plaintiff, she was predicted in the early period to nearly always receive sole custody, as shown in row L of Table 2. In the recent period, the predicted probability of shared custody in this type of case is 22.7 %. In the cases in which only the father has a lawyer and is the plaintiff (row M), he was predicted in the early period to receive sole custody in about one-third of the cases, and the mother was predicted to receive custody in about one-half the cases; in the recent period, one-half of the cases are predicted to have shared custody. Although the probabilities would seem to suggest that the relationship between these variables and the outcomes differ between the two periods, the relationships are not measured precisely, and the coefficients on these variables in the early period are not statistically different from the comparable coefficients in the later period.

We have documented very large changes in the custody outcomes of the children of divorce over this 20-year period. These trends may be related to changes in the characteristics of cases that get divorced or to changes in the divorce process. To explore these issues, we conduct two additional simulations. In the first, we take the mean characteristics from the early period and apply the coefficients from the recent period. This results in an overall estimate of 51.9 % with mother–sole custody, 43.5 % with shared custody, and 4.6 % with father–sole custody. Comparing these estimates to the last columns of the top row of Table 2 highlights the role of changes in process: even if the characteristics had stayed constant, we would still see a dramatic change in custody outcomes—nearly the change that we observe.

In the second simulation, we explore a different counterfactual, examining the predicted outcomes if the characteristics of cases had changed to be what they are observed to be in the recent period but the coefficients were still those of the early period. This results in an estimate of 87.9 % with mother–sole custody, 7.8 % with shared custody, and 4.3 % with father–sole custody. These estimates are remarkably similar to the predicted outcomes in the early period. This again suggests that the trend in custody is less the result of changes in the characteristics of cases than the result of changes in the process of awarding custody.

Examining this issue more closely, what about the process has changed? One way to examine this question is to pool the data across the three periods (including the intermediate period not yet analyzed) and allow the relationships between selected variables and custody to differ in different periods. Estimates from a pooled model suggest that the share of income attributed to the mother does not have a markedly different relationship with custody outcomes in the three periods, nor do the employment variables or the variables reflecting the legal process. In contrast, the relationship between total income and custody changes across the periods, as can also be seen in Tables 1 and 2. Moreover, in the pooled model, the variable representing the period has a very large coefficient, suggesting that the time trend itself is a very important part of the changing process.

These results suggest that although the characteristics of divorcing cases are changing over time, the change in composition does not seem to explain the majority of the observed changes in custody outcomes. Although we find some evidence of changes in the importance of some characteristics for custody outcomes, these changes also do not seem to explain much of the transformation of child living situations. Instead, the explanation most consistent with the patterns is a change in norms and the custody determination process, with the growing adoption of the innovation of shared custody, across the distribution of characteristics.

Conclusions

Fifteen years ago, Cancian and Meyer (1998) examined divorce cases and found a 6-percentage-point increase in shared custody over the late 1980s and early 1990s, with corresponding declines in mother–sole custody. These trends have accelerated. Shared custody continues to grow, even more dramatically than shown in our earlier article. Father–sole custody has not changed much, and mother–sole custody has declined dramatically. In fact, in the last decade, we reached a significant milestone: there are more divorce judgments without mother–sole custody than with it.

Our analysis of the factors related to custody shows some change and some continuity relative to earlier work. Shared custody is still more common among those with higher income. Children’s characteristics, however, which were strong predictors in the early period, are less important in the later period. In general, as shared custody cases become more common, they are becoming less distinctive. Moreover, changes in characteristics do not seem to explain much of the dramatic trends we see; the importance of the time trend of an increased legal and societal preference toward more shared custody, regardless of characteristics, is one conclusion. This suggests that changes in social norms are important.

Like all empirical research, this study has limitations. It has unknown generalizability, based on data from 21 counties in a single state. It is a study of court-prescribed living arrangements; children’s actual living arrangements may differ. It is based on administrative records, so we cannot account for the role of some variables that could be related to custody (e.g., education, race, the distance the parents live from each other, their commitment to co-parenting). Nonetheless, it is based on data that are unique in terms of the detailed custody information and the length of the period analyzed.

These findings have important implications for social policy. Most tax and transfer programs in the United States take the family as the unit of analysis, often determining eligibility and benefit levels based on the number of people in the family. The results here show that many children of divorce are supposed to live substantial amounts of time with both parents. Which parent gets to count them as part of their family, and whether they can count them as a whole child or as some fraction of a child, raises thorny policy issues (Hakovirta and Rantalaiho 2011). If this affected a small number of children, it might not be a consequential problem, but this research shows that shared custody has increased dramatically. This suggests a need for policymakers to reexamine how a variety of programs define families (Meyer and Carlson forthcoming).

These findings also have important implications for data collection and research. This article focuses on children of divorce; research on custody outcomes for children whose parents were never married would also be useful. Another area in which research would be useful is the relationship between custody and child support obligations. Perhaps some parents pursue shared custody so that they will have lower child support obligations. Recent research in Australia shows little evidence of this type of strategic bargaining (Smyth and Rodgers 2011; Smyth et al. 2012), but if it is occurring, this may suggest a need for reconsidering custody policy or child support policy.

Conducting research on shared custody probably requires new data given that shared-custody families cannot be separately identified in current national data collection efforts. This substantially limits our ability to address a range of important questions regarding the correlates and consequences of a distinctive and increasingly common living arrangement. How are children who have shared custody faring, emotionally and educationally, and how does this vary by the children’s ages or the details of living arrangements and schedules? Are changes in custody schedules over time experienced by family members as “instability” or as “flexibility”? Do children with shared custody get the needed financial support from both parents? The increasing prevalence of shared custody increases the importance of these questions.