Bonding After Divorce: Comments on Joint Custody: Bonding and Monitoring Theories

Ann Laquer Estin⁽¹⁾

I am happy to have this chance to comment on a very interesting paper by Margaret Brinig and F.H. Buckley.⁽²⁾ Peg and Frank propose a number of new theories about joint-custody arrangements following divorce, and then present an empirical analysis of joint-custody laws. Their conclusions are provocative: divorce rates are negatively and significantly correlated with joint-custody laws; in fact, a change to joint custody is associated with a two- to eleven-percent reduction in divorce levels.⁽³⁾ In addition, they conclude that joint-custody laws are associated with higher rates of child-support payment.⁽⁴⁾

Brinig and Buckley present their empirical findings as a test of two theories concerning the benefits of joint custody. Their bonding theory suggests that where parents anticipate sharing custody in the event of divorce, they will invest more heavily in, or bond more closely with, their children during a marriage. Brinig and Buckley propose that these stronger bonds may help to deter divorce under a joint-custody regime.⁽⁵⁾ Their monitoring theory suggests that joint custody allows a nonprimary parent to monitor the other parent's child-rearing decisions and how child-support payments are being spent. Thus, they hypothesize that joint-custody laws will promote better payment of child support.⁽⁶⁾

Part I of my Response addresses several concerns and observations I have about the empirical aspects of the piece, particularly the relationship the authors claim between joint-custody laws and divorce rates. Part II of my Response looks more closely at the monitoring theory utilized in this paper. By way of disclaimer, I should say at the outset that I am not an economist, and I have only a rudimentary appreciation for the quantitative methods that Professors Brinig and Buckley have employed here. I have spent some time trying to understand what economists have to say about family life, and that affects my reading of this piece. My approach is also influenced by my experiences as a lawyer working with mothers and fathers in divorce cases.

I. Joint-Custody Laws and Divorce Rates

Brinig and Buckley have developed impressive evidence of a correlation between the spread of joint-custody norms and declining divorce rates. Their study does not prove, however, that joint-custody laws are responsible for the reduction in divorce rates. A different possibility is that both of these trends are the result of other forces. I am also troubled by the treatment of "joint custody" in this analysis as a single, unitary category, which appears in law at a discrete moment in time. In practice, of course, there are a wide range of types of joint custody, beginning with the crucial difference between joint physical custody and joint residential custody, and these norms have evolved in our society over several decades.

From a law-and-economics perspective, the finding that joint-custody laws are not associated with an increase in divorce rates is very significant. By making custody divisible and negotiable, joint-custody rules create more bargaining room for divorcing couples. Law-and-economics theory suggests that this greater bargaining room should have the effect of promoting divorce, by making it easier to work out terms acceptable to both parties.

A. Do Joint-Custody Laws Cause a Drop in Divorce Rates?

Professors Brinig and Buckley suggest at several points in their paper that joint-custody laws cause a reduction in divorce rates.⁽⁷⁾ While their empirical work demonstrates a correlation between the adoption of joint custody and a decline in divorce, it is not clear that the relationship is a causal one. Brinig and Buckley also found a strong negative correlation between the divorce rate and the year being considered.⁽⁸⁾ This is not surprising, given the fact that divorce rates across the country began decreasing at the same time as joint-custody laws were becoming more widespread,⁽⁹⁾ and it suggests to me the possibility that other forces may be at work.

The first part of the Brinig and Buckley analysis covers a twelve-year period (from 1980 through 1991), and includes data for forty-seven states.⁽¹⁰⁾ This is the time during which joint-custody norms were popularized and dispersed across the country.⁽¹¹⁾ At the beginning of the period, just nine of these states had recognized joint custody by statute or case law, while by the end of the period, thirty-nine states had.⁽¹²⁾ According to this data, only three of the fifty states recognized joint custody before 1979.⁽¹³⁾ Eighteen of the study states enacted or adopted joint custody at the beginning of the period, from 1981 and 1983.⁽¹⁴⁾

During this period, the statistics demonstrate both a decreasing enthusiasm for divorce and an increasing enthusiasm for joint custody. It remains possible, however, that both of these trends are the result of independent social or cultural factors. While such forces are difficult to measure or quantify, there is evidence to indicate significant shifts in popular attitudes.

After a period of dramatic increases in the divorce rate, Americans appear to have begun rethinking what is sometimes referred to as the "culture of divorce."⁽¹⁵⁾ The leveling off of divorce rates in the mid-1980s, along with other cultural and demographic evidence, indicates that we take a more sober approach today both to marriage and divorce than was true twenty years ago.⁽¹⁶⁾ For example, demographers note that young adults are much more hesitant to marry than was true of earlier cohorts: age at first marriage and the rate of cohabitation prior to marriage have both increased.⁽¹⁷⁾ In a number of states, recent legislative sessions have seen efforts to reinstitute fault-based divorce laws.⁽¹⁸⁾ In Louisiana, this has culminated in the adoption of a law giving couples the choice to enter into a "covenant" marriage.⁽¹⁹⁾

Additionally, the new norms of parenthood that evolved between the mid-1970s and the 1990s have made the old approach to custody after divorce unthinkable for many couples--and perhaps impractical as well.⁽²⁰⁾ As more married women enter the labor force, fathers are more fully engaged in child care. Shared parenting during marriage, and after divorce, reflect new social and economic circumstances and represent a shift away from the older norms with more clearly divided gender roles.⁽²¹⁾

Because there are also good reasons to be cautious about joint custody,⁽²²⁾ it is important not to oversell the results of this empirical work as a basis for major policy changes. While it seems plausible that preferences for continuing marriage rather than divorcing are shaped in part by the legal regime,⁽²³⁾ including rules applied to financial and custodial aspects of divorce law, such preferences are certainly also shaped over time, and more powerfully, by our collective social and cultural experience of divorce. These significant shifts are not changes we would expect to see following immediately after a decision of a state appellate custody case, or even upon enactment of a new statute. There was considerable discussion of joint custody in the popular and academic literature during the late 1970s, well before most states recognized it in their law.⁽²⁴⁾ In their discussion of their bonding theory, Brinig and Buckley suggest a process of behavioral change that requires a significant period of time, but their quantitative work implies that divorce rates will respond immediately to new joint-custody law. Certainly, law can have an important expressive function, but I am not persuaded that it has such an immediate and powerful influence on human behavior as these conclusions suggest.⁽²⁵⁾

B. What Is "Joint Custody"?

A second reason to be cautious here is that the Brinig/Buckley paper does not distinguish different types of joint custody. The paper does not indicate whether these data are based on laws recognizing joint legal custody, joint residential custody, or both. This is problematic, because in practice "joint custody" is not a single, unitary category. If this work is intended as a basis for policy recommendations, the differences are very important.

Joint custody sometimes refers to sole legal custody in one parent combined with some form of shared residence. This arrangement allows parents to "share access to children and child-rearing responsibilities,"⁽²⁶⁾ and, depending on the time-sharing provisions, may permit frequent and prolonged contact. While some monitoring can occur with this pattern, it does not give the nonprimary parent a right to control or even to participate in decisions concerning the children. Alternatively, divorced parents might have joint decisionmaking authority, while the children reside primarily (or almost exclusively) with one of them.⁽²⁷⁾ This allows the nonprimary parent a greater measure of authority, but not much opportunity for a relationship with the children. At the other extreme, joint custody is sometimes understood to imply an equal division of both decisionmaking and residence.

Joint physical custody is more pertinent to the bonding theories in this paper, and joint legal custody is more pertinent to the monitoring theories. From their paper, I assume that Brinig and Buckley intend both joint decisionmaking and residence.⁽²⁸⁾ But even with this specification, the category is extremely fluid. In one legal-aid case I handled a few years ago, the father sought joint custody. My client, the mother, was concerned about their ability to share decisionmaking--especially as they could not afford mediators to help them resolve disputes. The judge handled this by ordering joint legal custody, with the proviso that if the parents could not agree, Mom would be allowed to make final decisions for the child.⁽²⁹⁾ (To my surprise, Dad, and his lawyer, seemed satisfied with this.)

When all of these variations are included, the "joint custody" terminology becomes primarily symbolic. Of course, symbols are important in this setting. Many people care deeply about the words we use for these categories in family law, and these concerns have led legislatures and academics to replace the old vocabulary of "custody" and "visitation" with a series of new terms: "parental functions,"⁽³⁰⁾ "access,"⁽³¹⁾ "parenting time,"⁽³²⁾ and so on. In this area, words matter so much that "joint custody" in name only seems to be sufficient in many cases.

In their quantitative work, Brinig and Buckley treat joint custody as a single category, one which appears in law at a discrete moment in time. Thus, the study does not indicate whether shared residence and shared decisionmaking are equally significant. This approach also blurs the line between states that merely recognize or permit some type of joint-custody award, and those that affirmatively promote such a practice with a statutory presumption.⁽³³⁾ Therefore, the results tell us only that it is beneficial for states to have some type of joint-custody law on the books.

A more nuanced approach to joint-custody orders may not be possible to test empirically.⁽³⁴⁾ But if the goal is to develop a theory of joint custody, and to offer policy recommendations, then greater clarity is essential. Because it is very unlikely at this stage that any state will move to eliminate its joint-custody laws, the important policy questions concern how they can be most effectively utilized.

C. Joint Custody and Divorce Bargaining

From the perspective of law and economics, the most interesting aspect of the Brinig & Buckley paper may be simply the fact that joint-custody laws are not associated with an increase in the divorce rate. As Brinig and Buckley note in their paper, under the Coase theorem, a change in the legal rules governing custody would be expected to have primarily distributive consequences, without changing the substantive outcomes.⁽³⁵⁾ Although some writers believe that nonprimary parents routinely use custody demands for bargaining leverage in financial divorce negotiations,⁽³⁶⁾ a number of studies suggest that this is not a common problem.⁽³⁷⁾

Beyond the question of distributive effects, there is another issue. By making custody divisible and negotiable, joint-custody rules seem to create more bargaining room.⁽³⁸⁾ Greater room to bargain should promote divorce, by making it easier to work out terms acceptable to both parties.⁽³⁹⁾ Yet the evidence Brinig and Buckley have marshalled here points in the opposite direction, toward a reduction in divorce rates.

As any practicing divorce attorney recognizes, there are many different varieties of mothers and fathers in these cases. Families differ widely on the patterns of nurture they embrace during marriage, and in how they respond to the crises of divorce. Those who intentionally put their children's interests in jeopardy to secure a financial advantage are undoubtedly a very small minority. But in many families, divorce requires both parents to come to terms with new realities. A primary caregiver may need to begin to work outside the home, or to put in greater hours, and a breadwinner or workaholic parent may decide to make a different set of changes in order to have more time available to be an involved parent.

A different economics-influenced understanding of the joint-custody "default rule" might be that it promotes bargaining between parents over what is best for their children. In some respects, a joint-custody presumption may operate like a mutual-consent divorce rule: it gives both parents property rights in decisionmaking over their children's welfare. Just as some economists advocate a shift to mutual-consent rules in order to promote bargaining between the party seeking divorce and the party opposing it,⁽⁴⁰⁾ this paper might be used to argue for a presumptive joint-custody rule in order to foster cooperative behavior between parents.⁽⁴¹⁾

II. Monitoring Theories and Bonding After Divorce

At a more theoretical level, I would like to address the "monitoring" theory developed in the paper.⁽⁴²⁾ Brinig and Buckley discuss sole custody in economic terms as presenting an agency-cost problem, stating that "[u]nder sole custody, the noncustodial parent is asked to delegate exclusive child-rearing authority to the custodial parent at the same time that he is required to support the child."⁽⁴³⁾ They go on to describe the monitoring problem in these terms:

The noncustodial parent may be seen as a principal, and the custodial parent as an agent. The noncustodial parent will make child-support payments, but will not easily be able to monitor how the money is spent. He might simply disagree with the custodial parent's child-rearing decisions. Alternatively, he might worry that the custodial parent will shirk her duties--or worse still--misappropriate the funds. The noncustodial parent would want to condition support payments on child expenditures. But because he cannot verify actual expenditures on the child, this will prove unworkable. The noncustodial parent may react to the monitoring problem by cutting his support payments, to the child's detriment.⁽⁴⁴⁾

As the authors indicate, other economists have taken this approach to child-custody and support problems. Yoram Weiss and Robert J. Willis argued in a 1985 article that fathers' reduced interest in their children's welfare after divorce could be explained by two factors: the fathers' difficulty in monitoring spending by custodial parents,⁽⁴⁵⁾ and their reduced interest in their children because of the "loss of proximity."⁽⁴⁶⁾ During marriage, mutual trust, altruism, and proximity help to prevent opportunistic behavior by the parents, but after a marriage ends, these controls are less effective. After divorce, the custodian determines alone how much of her total household income to spend on the children, and how much to spend on herself.⁽⁴⁷⁾

What troubles me about the economists' monitoring theory is the implicit model of family life on which it is based. That model is a familiar one in economic analysis of the family. It assumes the efficiency of a highly traditional division of labor within the family, in which Dad is the provider and Mom the nurturer.⁽⁴⁸⁾ The father is conceived as altruistic head of the household, whose control over the financial resources coming into the family unit are an appropriate basis for asserting power to allocate those resources within the family.⁽⁴⁹⁾

Criticism of this conventional model is common, and both Peg Brinig and I have contributed to it.⁽⁵⁰⁾ Whatever may be the empirical or normative validity of the model when used to describe married couples with children, it is far more troublesome when it is extrapolated into the period after divorce. The discussion of monitoring in the economic literature seems not to question whether the noncustodial parent's wish for control is appropriate, or why his decisions would be preferable to the custodial parent's. It also does not question the extension of a traditional, gender-based division of household responsibility into the postdivorce family setting. This is especially remarkable given the evidence to suggest that there is very little division of labor after divorce, as custodial parents, usually mothers, take on primary responsibility for both financial support and nurturance of their children.⁽⁵¹⁾

Economists need to develop more useful models of family relationships after divorce. This will require fresh thinking about fundamental questions. What is the nature of parental responsibility after divorce? Whose "authority" is being exercised here? Reflecting on this, it seems to me that the monitoring problem really moves in the other direction. Where one parent has had the primary caregiving responsibility during the marriage, we might conceive of the basic authority over child rearing as hers. These primary parents, as principals, have significant difficulty in monitoring the behavior of their parenting agents under any form of custody award. In many of the divorce cases I saw in practice, this was a very powerful dynamic. Mothers often experienced enormous difficulty in accepting the unorthodox parenting methods of their ex-spouses. For any practicing attorney, these complaints are familiar:⁽⁵²⁾ "All they had for dinner was Coke and potato chips!," "They never go to bed on time at his house, and they're always exhausted when they come back to me on Sunday night," "Can you believe it? He took a five-year-old to see Robocop! Can't I do anything about it?"⁽⁵³⁾

In practice, of course, I also had my share of "monitoring" complaints from support-paying, nonprimary parents concerned about a variety of issues including where all that support money was going. A few of these parents clearly had no idea how expensive it is to feed and clothe children. Others were trying to impose a new budgetary restraint after separation where none had ever existed during the marriage. A few were unhappy with particular purchases, or wanted the money spent a certain way. One of my favorite clients called me on Easter Monday, a year after his divorce was final, and launched into a complaint. He had had the children for Easter weekend, and planned to take them to church. Two weeks earlier, he had taken his teenaged son shopping for a suit, "so he would have something decent to wear." After breakfast, Dad sent the kids upstairs to get ready for church. When they came down, his son was wearing running shoes with his new suit. It turned out, of course, that he had no dress shoes. Dad was very unhappy about this, and he called me to complain about why, with all the child-support money he paid, his son had nothing better to wear. We had a long talk about it that morning. I was sympathetic, but I think the bottom line is that next time, Dad needs to buy his son a suit and shoes if it really matters to him.

There is another important point here. In a significant number of the divorce cases that I handled, the crisis that began with the parties' separation marked a turning point in the parents' relationship with their children. Undoubtedly, the trauma of divorce impairs the parenting abilities even of good parents, at least temporarily, and it introduces significant new complications into parent-child relationships.⁽⁵⁴⁾ But for some parents who had not been deeply involved in caring for their children's day-to-day needs, the separation made it clear that it would take an enormous commitment to continue to be a real parent after divorce. Some men recognized that they were not prepared to take on these responsibilities, and settled into a visiting-parent routine, sometimes fading almost entirely out of the picture.⁽⁵⁵⁾ There were others, however, who began to take their roles entirely differently. Typically, wives watching this transformation in a former partner were skeptical, cynical, and disbelieving: surely this was all a charade, intended only to impress the lawyer, custody evaluator, or judge. In several of the cases I saw, however, there was a genuine and profound change.

It is important to recognize that all forms of shared parenting permit, and indeed require, many more interactions between parents. The models we use should conceptualize this monitoring across a broad range of issues and as taking place in both directions. I tend to agree that this collaboration, even when it is somewhat against the parents' wishes, can be very important for children. The trick, of course, is to keep both parents' monitoring impulses in balance with a healthy respect for the other parent.

In some postdivorce families, there are enormous, ongoing power struggles over every conceivable issue concerning the children. In these cases, the parents may view their behavior as "monitoring," but it is clearly much less neutral and altruistic than that term suggests. Often, it is quite harmful.⁽⁵⁶⁾ Many of these couples have not come to grips with the divorce itself. The conventional wisdom teaches that joint custody is not appropriate for these high-conflict families,⁽⁵⁷⁾ but they often engage in escalating misbehavior even with a sole-custody arrangement.

In my community, some sophisticated family-law attorneys have had success in working with these couples using joint-custody awards combined with a skilled mediator-arbitrator as a referee.⁽⁵⁸⁾ In cases with chronic postdissolution conflict, the appointed mediator-arbitrator is given authority to decide all issues except a change in custody. At the outset, the parents agree to meet at least once a month with the mediator-arbitrator to assess how their children are doing and to facilitate better communication. When a dispute arises that the parties cannot resolve themselves, they submit written statements to the mediator-arbitrator, who schedules a time to meet with them, usually almost immediately. If the dispute cannot be resolved through mediation, the mediator becomes an arbitrator, and decides the issue for the parties. My colleagues tell me that this system works because it is cost effective and more responsive to the parents' needs than are repeated hearings in front of a judge. Ideally, it also works as a behavior-modification device. As parents begin to anticipate how the mediator-arbitrator will decide particular issues, they begin to limit their complaints and requests to those that are likely to be deemed reasonable.

As Brinig and Buckley point out, joint custody is often politicized, depicted in terms of a power struggle between women and men.⁽⁵⁹⁾ I agree that women's nonmonetary contributions to a marriage, and specifically to raising children, are often not recognized or are undervalued.⁽⁶⁰⁾ I believe that we need to give more moral and financial support to caregivers, as a way to help children, particularly after divorce. But, in my experience, there really are two sides to almost every divorce story, and many fathers have very compelling stories to tell. I saw many for whom the divorce was an awakening to the importance of a real relationship with their children. There are certainly many cases in which mothers share, or deserve, the primary blame for conflict and unhappiness to which children are subjected.

The question of how best to foster cooperative behavior between divorced parents is a perennial one, with many writers contributing to the debate. Perhaps the real value of joint custody lies in encouraging bonding between parents and children after divorce. What formerly married parents may need, more than opportunities for "monitoring" each other, is encouragement toward keeping their control issues under better control.

1. Associate Professor, University of Colorado School of Law.

2. See Margaret F. Brinig & F.H. Buckley, Joint Custody: Bonding and Monitoring Theories, 73 Ind. L.J. 393 (1998).

3. See id. at 420. Brinig and Buckley also look at divorce rates in connection with a series of socioeconomic variables. See generally id. at 414-16. Their study finds a positive and significant correlation between divorce rates and variables based on the percentage of married women in the labor force, urbanization, and some regional differences. See id. at 417-19 tbls.IV-V. These findings are in some respects ambiguous, as there were inconsistencies between the results obtained when different statistical methods were applied. The paper indicates, however, that the equations are more properly modeled with "fixed effects" rather than with cross-sectioned specifications, see id. at 417, 423, and so the "fixed effects" results are the ones I have considered.

4. See id. at 420-23. The study also found positive correlations between child-support payments and variables reflecting state spending for child-support enforcement and the percentage of working-age women in the labor force. See id. at 422-23 tbl.VII.

5. See id. at 403.

6. See id. at 410.

7. See, e.g., id. at 403 ("We therefore hypothesize that a move to joint custody will decrease divorce rates."); id. at 407 ("There would be a strong reason to promote joint custody if it significantly reduced divorce levels in the manner predicted by bonding theories."). At other points, however, the conclusion drawn is more modest: "Our principal result is that divorce levels are negatively and significantly correlated throughout with joint-custody laws. These results are consistent with the hypothesis that joint-custody laws reduce divorce levels." Id. at 417.

8. See id. at 417-18 & tbl.IV. This test involved a "dummy variable" for joint custody --one that takes a value of either 1, if a state recognizes joint custody, or 0, if it does not. See id. at 414. Because of this dummy variable, it is not possible to compare the correlation between divorce rates and joint custody with the correlations Brinig and Buckley found between divorce rates and the other independent variables.

9. Divorce rates in the United States peaked in about 1980 at 5.2 per 1000 population. See Bureau of the Census, U.S. Dep't of Commerce, Statistical Abstract of the United States 104 tbl.146 (1996). Andrew Cherlin locates the peak in divorce rates in 1979. See Andrew J. Cherlin, Marriage, Divorce, Remarriage 22 (rev. & enlarged ed. 1992). One might also argue that divorce rates are not dropping when we look separately at different age cohorts. See id. at 20-24 (estimating that 51% of those marrying in the 1980s will eventually divorce).

10. These are the 48 continental United States, excluding Nevada. See Brinig & Buckley, supra note 1, at 413.

11. Brinig and Buckley also compared divorce rates in a group of 19 states for which data were available on the actual percentage of divorce cases in which joint custody was ordered. Here, as well, they found that the correlation between divorce rates and the frequency of joint-custody awards was negative and significant. See id. at 417, 420 & tbl.VI. In this test, they used data from a later time period; here, the relationship between the year considered and the divorce rate was positive rather than negative. See id. at 420 & tbl.VI.

12. See id. at 396-97 tbl.I. Of the 47 states studied, five have still not recognized joint custody in case law or statute; these are Alabama, Arkansas, South Carolina, Washington, and West Virginia. This may be another point at which the data are subject to Type II errors. See id. at 414.

13. See id. at 396-97 tbl.I. These were Indiana, New Hampshire, and Iowa. See id.

14. See id. This was also a period which saw "explosive growth" in the use of joint legal custody in some jurisdictions. See Eleanor E. Maccoby & Robert H. Mnookin, Dividing the Child 108 (1992).

15. See, e.g., Maggie Gallagher, The Abolition of Marriage (1996); Barbara Dafoe Whitehead, The Divorce Culture (1997). For additional sources see Brinig & Buckley, supra note 1, at 404 n.50.

16. A number of popular books reflected an upbeat approach to divorce in the 1970s. See, e.g., Susan Gettleman & Janet Markowitz, The Courage to Divorce (Ballantine Books 1978) (1974); Mel Krantzler, Creative Divorce (1975).

17. See Cherlin, supra note 8, at 10-13.

18. See Dirk Johnson, Attacking No-Fault Notion, Conservatives Try to Put Blame Back in Divorce, N.Y. Times, Feb. 12, 1996, at A10; see also Ira Mark Ellman, The Place of Fault in a Modern Divorce Law, 28 Ariz. St. L.J. 773, 775 (1996); Deborah L. Rhode, To Fault or Not to Fault, Nat'l L.J., May 13, 1996, at A19.

19. See Act of July 15, 1997, No. 1380, 1997 La. Sess. Law Serv. 2358 (West) (to be codified at La. Rev. Stat. Ann. �� 9:224-:225, :234, :245, :272-:275, :307-:309, and at La. Civ. Code Ann. arts. 102-103). Parties to a covenant marriage can seek divorce only on the basis of a spouse's adultery, felony conviction, abandonment, physical or sexual abuse, habitual intemperance or cruelty, or two years separation. See id. � 4, 1997 La. Sess. Law Serv. at 2362-63 (amending La. Rev. Stat. Ann. � 9:307).

20. See, e.g., Diane Ehrensaft, Parenting Together (1990); Arlie Hochschild & Anne Machung, The Second Shift (1989).

21. Indeed, the new norms may be better for children. See Brinig & Buckley, supra note 1, at 402.

22. For important questions and criticism, see Anne Opie, Ideologies of Joint Custody, 31 Fam. & Conciliation Cts. Rev. 313 (1993); Elizabeth Scott & Andre Derdeyn, Rethinking Joint Custody, 45 Ohio St. L.J. 455 (1984); Jana B. Singer & William L. Reynolds, A Dissent on Joint Custody, 47 Md. L. Rev. 497 (1988).

23. See Brinig & Buckley, supra note 1, at 401.

24. See, e.g., Miriam Galper, Co-Parenting (Peter John Dorman ed., 1978); Isolini Ricci, Mom's House/Dad's (1982); Mel Roman & William Haddad, The Disposable Parent (1978); Charlotte Baum, The Best of Both Parents, N.Y. Times, Oct. 31, 1976, at 44; Georgia Dullea, Joint Custody: Is Sharing the Child a Dangerous Idea?, N.Y. Times, May 24, 1976, at 24; Emily Jane Goodman, Joint Custody, McCall's, Aug. 1975, at 34; Loretta Kuklinsky Huerta, Joint Custody: Co-Parenting After Divorce, L.A. Times, Jan. 30, 1979, � IV, at 1; Mary Alice Kellogg, Joint Custody, Newsweek, Jan. 24, 1977, at 56. In the legal academic literature, see H. Jay Folberg & Marva Graham, Joint Custody of Children Following Divorce, 12 U.C. Davis L. Rev. 523 (1979); Carole S. Bruch, Making Visitation Work: Dual Parenting Orders, Fam. Advoc., Summer 1978, at 22. See also Joint Custody Handbook (Ann L. Milne ed., 1979).

25. On the expressive function of custody law, see Katharine T. Bartlett, Re-Expressing Parenthood, 98 Yale L.J. 293 (1988); Elizabeth S. Scott, Pluralism, Parental Preference, and Child Custody, 80 Cal. L. Rev. 615, 656 (1992). See also Carol Weisbrod, On the Expressive Functions of Family Law, 22 U.C. Davis L. Rev. 991 (1989). Borrowing from Patrick Atiyah, Carl Schneider terms the role of law in shaping behavior as the "hortatory function" of law. See Carl E. Schneider, Rethinking Alimony: Marital Decisions and Moral Discourse, 1991 BYU L. Rev. 197, 203-09.

26. Brinig & Buckley, supra note 1, at 393.

27. This is apparently very common under statutes, like the California statute, which foster joint legal custody. In the sample studied by Maccoby and Mnookin, 80% of families had joint legal custody, but in the majority of these families, almost half of the sample, mothers had sole physical custody. See Maccoby & Mnookin, supra note 13, at 112-13. Even in those cases in which joint physical custody was ordered or agreed to, a significant percentage of families reverted to a de facto single residence, most commonly with the mother. See id. at 164-67.

28. See Brinig & Buckley, supra note 1, at 396 ("Joint custody means more than a sharing of physical custody, as parents must share the responsibility for the child's upbringing.") (citing Ark. Code Ann. � 61.13(b)(1) (West Supp. 1997), and Va. Code Ann. � 20-107.2 (Michie Supp. 1997)).

29. In other reported cases, joint decisionmaking provisions have been interpreted to give the physical custodian ultimate authority. See, e.g., Griffin v. Griffin, 699 P.2d 407 (Colo. 1985); Parrinelli v. Parrinelli, 524 N.Y.S.2d 159 (Sup. Ct. 1988); cf. Dolan v. Dolan, 548 A.2d 632 (Pa. Super. Ct. 1988) (holding that the best interests of the child predominate when parents cannot agree under a joint decisionmaking provision).

30. See In re Marriage of Kovacs, 854 P.2d 629, 631-37 (Wash. 1993) (applying Parenting Act of 1987, Wash. Rev. Code �� 26.09.002-.913 (1987)).

31. See William F. Hodges, Interventions for Children of Divorce (2d ed. 1991).

32. This term was adopted by the Colorado legislature in 1994 and is now used in place of "visitation" throughout the Colorado statutes. See, e.g., Colo. Rev. Stat. Ann. � 14-10-129 (West 1997).

33. See Brinig & Buckley, supra note 1, at 395 n.14.

34. The test based on actual court reports may come closer to the mark, yet there is still a wide range of variation among joint-custody orders.

35. See Brinig & Buckley, supra note 1, at 400. But see Scott, supra note 24, at 649 (arguing that the Coase theorem is not applicable to custody bargaining "because the delicate process of trading custody rights holds the potential for generating punishing costs").

36. See, e.g., Richard Neely, The Divorce Decision (1984); see also sources cited in Brinig & Buckley, supra note 1, at 400 n.32.

37. See, e.g., Maccoby & Mnookin, supra note 13, at 154-59; see also Scott Altman, Lurking in the Shadow, 68 S. Cal. L. Rev. 493, 540-43 (1995). Brinig and Buckley tend to discount the distributive problem, suggesting that if joint-custody laws have this type of distributive effect, a better solution would be to change the underlying financial entitlements. See Brinig & Buckley, supra note 1, at 400-02.

38. See Robert H. Mnookin & Lewis Kornhauser, Bargaining in the Shadow of the Law: The Case of Divorce, 88 Yale L.J. 950, 975-76 (1979) (arguing that if joint custody is not permitted, there may be no middle ground on which parties can compromise).

39. See, e.g., Yoram Weiss & Robert J. Willis, Children as Collective Goods and Divorce Settlements, 3 J. Lab. Econ. 268, 290 (1985) ("anything which improves the ability of couples to allocate resources efficiently upon divorce will tend to increase the probability of divorce").

40. See Ann Laquer Estin, Economics and the Problem of Divorce, 2 U. Chi. L. Sch. Roundtable 517, 533-50 (1995).

41. In both contexts, there is an obvious concern with the possibility for abuse, suggesting that the more limited entitlement created by a presumption--a liability rule rather than a property rule--is more appropriate. See generally id. at 529-31, 571-76.

42. See Brinig & Buckley, supra note 1, at 408-11.

43. Id. at 409.

44. Id. (footnotes omitted).

45. See Weiss & Willis, supra note 38.

[I]t is costly and often impossible for the noncustodian, say the father, to control the allocation of resources by the custodian, say, the wife. The difficulty arises because it is costly for the husband (or a third party) to monitor (or verify) this allocation. Consequently, there is a loss of control that, in turn, leads to a less efficient allocation of resources; in particular, the amount spent on the maintenance of children is reduced.

Id. at 270. Another study that refers to this model is Andrea H. Beller & John W. Graham, Small Change (1993). Beller and Graham develop an economic model to explain why child-support awards and receipts after divorce fall short of the predivorce support levels. Noting that the well-being of both parents typically declines after divorce, they argue that the award and receipt outcomes are determined by "the ability and desire of the father to pay support" and the "expected costs and benefits of support to mothers," id. at 81, as well as by laws governing child-support determination and enforcement, see id. They argue that greater child-support payments have substantial beneficial effects beyond the money itself, and that increasing the incidence, value, and collection of awards should be a central policy goal. See id. at 246-47.

46. Weiss & Willis, supra note 38, at 288. On the relationship between visitation and child-support payments, see Jessica Pearson & Jean Anhalt, Examining the Connection Between Child Access and Child Support, 32 Fam. & Conciliation Cts. Rev. 93 (1994).

47. See Weiss & Willis, supra note 38, at 274. They also suggest, however, that "an added incentive to maintain the marriage" is "the difficulty of controlling the expenditure on children in the divorce state." Id. at 280. Weiss and Willis argue that with joint custody, each spouse is better able to affect the level of spending for children. They point out additionally that maintenance of close contact between both parents and the children following divorce is costly, and that these costs tend to increase with time after the separation. These costs reduce the contact between the noncustodian and the children, and decrease the amount that the noncustodian is willing to pay voluntarily. See id. at 288. As a solution, Weiss and Willis recommend a shift in legal rules to permit conditioning child-support payments on visitation and vice versa. See id.

48. See generally Ann Laquer Estin, Love and Obligation: Family Law and the Romance of Economics, 36 Wm. & Mary L. Rev. 989, 1001-13 (1995) (describing how economists' household-production theory is used to explain marriage and divorce and behavior within the family).

49. See Ann Laquer Estin, Can Families Be Efficient? A Feminist Appraisal, 4 Mich. J. Gender & L. 1, 14-16 (1996); Estin, supra note 47, at 1013-16.

50. See Margaret F. Brinig, Comment on Jana Singer's Alimony and Efficiency, 82 Geo. L.J. 2461 (1994); Estin, supra note 48.

51. See Maccoby & Mnookin, supra note 13, at 271.

52. Maccoby and Mnookin discuss aspects of this problem in their book. See id. at 227-31.

53. The children sometimes had their own complaints; the one I heard most commonly was the ponytail problem--Dad's inability to fix his daughter's hair properly in the morning before school.

54. See Maccoby & Mnookin, supra note 13, at 204-12.

55. There is a description of this pattern in Frank F. Furstenberg, Jr. & Andrew J. Cherlin, Divided Families 34-39 (1991).

56. See generally Carla B. Garrity & Mitchell A. Baris, Caught in the Middle (1994); Janet R. Johnston & Linda E.G. Campbell, Impasses of Divorce (1988).

57. See Garrity & Baris, supra note 55; Janet R. Johnston et al., Ongoing Post-Divorce Conflict: Effects on Children of Joint Custody and Frequent Access, 59 Am. J. Orthopsychiatry 576 (1989); Opie, supra note 21.

58. See Dale E. Johnson, Joint Custody with Mediation/Arbitration to Manage Chronic Cases (Dec. 2, 1994) (unpublished manuscript, on file with author).

59. See Brinig & Buckley, supra note 1, at 397-99.

60. See Ann Laquer Estin, Alimony, Maintenance, and the Rehabilitation of Family Care, 71 N.C. L. Rev. 721 (1993).