136 American Law and Economics Review V2 NI 2000(126-169) carry off specific marital investments that the other spouse has sunk into the marriage is the one who leaves (and,presumably,files).In this con- text,filing for divorce is a method of acquiring a disproportionate8 share of the marital assets.For most financial investments,wives are vulner- able to appropriation because husbands have traditionally been the first to finish their formal education or obtain vocational training.However, husbands can also make specific investments up front and can therefore become vulnerable as well.Next to children,one of the most important forms of early investment by a couple is in education.When one party in- vests in the education of the other early on,the benefits(rents)are earned once the high-human-capital spouse starts working.As a result,however, the spouse with the lower income may fall victim to the opportunistic di- vorce.Hence,for example,a wife may leave her husband after he has put her through law school.19 2.3.Custody Until now our focus has been on the role of quasi-rents in a marriage Individuals file when filing stops them from being exploited within the marriage or when filing allows them to exploit the other spouse.Although quasi-rents in marriage can arise from a number of specific investments, children are likely to be the most serious and common source.In addi- tion to this,children are often the most valuable assets in a family.As such,custody is often a critical issue in divorce filing behavior.Property distribution laws that might allow one spouse to appropriate marital invest- ments apply to marital property,not children.For children,the standard isn't strictly measured by investment or even fairness,but by the"best in- terests of the child"(Altman,1996).Time spent with children or,even more,money lavished upon them does not necessarily equate to what will be best for them when parents no longer live together. Because of the differences in attachment we've discussed earlier,hav- ing a child in itself makes wives less likely to file for divorce.The presence of children also makes divorce less likely in the first place,as couples stay together "for the sake of the children"(Lacey,1992;Scott,1990). 18.Disproportionate,that is,compared to what the spouse had invested in the marriage. 19.See Srinivasan v.Srinivasan (1990)(wife exploited husband through graduate school and a lengthy probationary faculty period)
136 American Lawand Economics ReviewV2 N1 2000 (126–169) carry off specific marital investments that the other spouse has sunk into the marriage is the one who leaves (and, presumably, files). In this context, filing for divorce is a method of acquiring a disproportionate18 share of the marital assets. For most financial investments, wives are vulnerable to appropriation because husbands have traditionally been the first to finish their formal education or obtain vocational training. However, husbands can also make specific investments up front and can therefore become vulnerable as well. Next to children, one of the most important forms of early investment by a couple is in education. When one party invests in the education of the other early on, the benefits (rents) are earned once the high-human-capital spouse starts working. As a result, however, the spouse with the lower income may fall victim to the opportunistic divorce. Hence, for example, a wife may leave her husband after he has put her through lawschool.19 2.3. Custody Until nowour focus has been on the role of quasi-rents in a marriage. Individuals file when filing stops them from being exploited within the marriage or when filing allows them to exploit the other spouse. Although quasi-rents in marriage can arise from a number of specific investments, children are likely to be the most serious and common source. In addition to this, children are often the most valuable assets in a family. As such, custody is often a critical issue in divorce filing behavior. Property distribution laws that might allow one spouse to appropriate marital investments apply to marital property, not children. For children, the standard isn’t strictly measured by investment or even fairness, but by the “best interests of the child” (Altman, 1996). Time spent with children or, even more, money lavished upon them does not necessarily equate to what will be best for them when parents no longer live together. Because of the differences in attachment we’ve discussed earlier, having a child in itself makes wives less likely to file for divorce. The presence of children also makes divorce less likely in the first place, as couples stay together “for the sake of the children” (Lacey, 1992; Scott, 1990). 18. Disproportionate, that is, compared to what the spouse had invested in the marriage. 19. See Srinivasan v. Srinivasan (1990) (wife exploited husband through graduate school and a lengthy probationary faculty period)
Why Most Divorce Filers Are Women 137 However,given the presence of children,it is expectations of custody that drive divorce filing.By making a preemptive filing,the wife may be able to secure rights such as child or spousal support that require court enforcement.20 When the wife files,she is often given temporary custody of the children.Temporary custody,like possession,tends to be "nine-tenths of the law"and plays a role in the assignment of permanent custody,especially when the divorce does not occur for some time.21 When a spouse expects custody,filing may act as an appropriation of marital resources.For wives versus husbands,children are an unusual as- set because working men are unlikely to receive substantial child support from their wives.Children may hurt the earning potential of wives,but this very fact makes them more likely to seek custody.Mothers are less employable because they have spent the most time with the children,but because they have spent the most time,they probably have closer rela- tionships with them.Some women may also feel that by making career sacrifices,they have"earned"custody.Thus,although the presence of chil- dren,and especially more than one child,may be a damper on a wife's filing,realizing that the marriage is very troubled,she may nevertheless file to secure custody (Grillo,1991,p.1547). 3.Filing and the Welfare of Wives A spouse may thus file for divorce because he or she is escaping an ex- ploitative partner,leaving the marriage with the bulk or all of its enhanced human capital,or attempting to establish custody over children.Keeping in mind that wives are more likely to file for divorce,each of these reasons for filing has different implications for the welfare of wives and different implications regarding the policy recommendations for divorce law.Wives who file in an attempt to flee an exploitative relationship resulting from 20.Again,this makes the most sense in historical periods when husbands owned all the assets,and wives might not even be able to establish their own domiciles without court orders (Brinig and Carbone,1988,pp.863-64).A non-biological party unsuccessfully sought child support without such an arrangement(Larson v.Diveglia, 1997.) 21.Analogously,temporary child support is now being made under exactly the same guidelines as permanent support (Va.Code Ann.,1998).Strategic motivations would have been particularly important for wives in historical periods when they lacked resources themselves
Why Most Divorce Filers Are Women 137 However, given the presence of children, it is expectations of custody that drive divorce filing. By making a preemptive filing, the wife may be able to secure rights such as child or spousal support that require court enforcement.20 When the wife files, she is often given temporary custody of the children. Temporary custody, like possession, tends to be “nine-tenths of the law” and plays a role in the assignment of permanent custody, especially when the divorce does not occur for some time.21 When a spouse expects custody, filing may act as an appropriation of marital resources. For wives versus husbands, children are an unusual asset because working men are unlikely to receive substantial child support from their wives. Children may hurt the earning potential of wives, but this very fact makes them more likely to seek custody. Mothers are less employable because they have spent the most time with the children, but because they have spent the most time, they probably have closer relationships with them. Some women may also feel that by making career sacrifices, they have “earned” custody. Thus, although the presence of children, and especially more than one child, may be a damper on a wife’s filing, realizing that the marriage is very troubled, she may nevertheless file to secure custody (Grillo, 1991, p. 1547). 3. Filing and the Welfare of Wives A spouse may thus file for divorce because he or she is escaping an exploitative partner, leaving the marriage with the bulk or all of its enhanced human capital, or attempting to establish custody over children. Keeping in mind that wives are more likely to file for divorce, each of these reasons for filing has different implications for the welfare of wives and different implications regarding the policy recommendations for divorce law. Wives who file in an attempt to flee an exploitative relationship resulting from 20. Again, this makes the most sense in historical periods when husbands owned all the assets, and wives might not even be able to establish their own domiciles without court orders (Brinig and Carbone, 1988, pp. 863–64). A non-biological party unsuccessfully sought child support without such an arrangement (Larson v. Diveglia, 1997.) 21. Analogously, temporary child support is nowbeing made under exactly the same guidelines as permanent support (Va. Code Ann., 1998). Strategic motivations would have been particularly important for wives in historical periods when they lacked resources themselves
138 American Law and Economics Review V2 NI 2000(126-169) the nurturing asymmetry that often arises between them and their hus- bands are clearly made better off by divorce and should by virtually all accounts be allowed to leave the marriage.Reformers who feel these cases of physical or emotional abuse predominate may consider ways of discour- aging the exploitative spouse from profiting from the bad behavior,such as punitive alimony or property shares (Brinig and Crafton,1994),and, if exploitative behavior dominated,might counsel law reformers to make divorce less difficult (or less painful),with transfer payments to mitigate any financial hardship (Brinig,1997,pp.116-18).22 Furthermore,govern- ment should emphasize improving marriage rather than making divorce difficult. However,wives (and husbands)who file for divorce in order to ap- propriate marital assets such as enhanced earning capacities are better off after divorce,but only at the expense of the other spouse.In this case,the opposite recommendation for divorce law is forthcoming.When divorce is used to exploit the spouse remaining behind,divorce should be made more difficult to obtain23 in order to protect the specific investments of each spouse (Cohen,1987;Ellman,1989). If it is custody outcomes that most influence divorce filings,changes in custody rules (or their likely outcomes)rather than in divorce grounds should most shape the patterns of both marriage and divorce(Brinig and Buckley,1998a).24 In particular,this could take the form of a presumption of joint custody or a rule that made post-divorce patterns mirror pre- separation time shares as closely as possible,with sole custody only in cases where one party can show the other parent unfit.25 An appropriate custody rule mitigates the incentive for one-party filing for the purpose of 22.Some would say to stigmatize the exploiter as well(Golden and Taylor,1987. pp.12-13). 23.Apparently this is the view of the law and economics writers who feel that those leaving are by and large men. 24.Chapter 2 of the American Law Institute Marital Dissolution project (1998), which would substantially change substantive child custody rules,may well have this effect. 25.The American Law Institute (1998)proposal rules can be found in Section 2.02(e)(continuity of existing parent-child attachments)and 2.09(1).which provides that unless otherwise resolved by agreement or manifestly harmful to the child,"the court should allocate custodial responsibility so that the proportion of custodial time the child spends with each parent approximates the proportion of time each parent spent performing caretaking functions for the child prior to the parents'separation
138 American Lawand Economics ReviewV2 N1 2000 (126–169) the nurturing asymmetry that often arises between them and their husbands are clearly made better off by divorce and should by virtually all accounts be allowed to leave the marriage. Reformers who feel these cases of physical or emotional abuse predominate may consider ways of discouraging the exploitative spouse from profiting from the bad behavior, such as punitive alimony or property shares (Brinig and Crafton, 1994), and, if exploitative behavior dominated, might counsel lawreformers to make divorce less difficult (or less painful), with transfer payments to mitigate any financial hardship (Brinig, 1997, pp. 116–18).22 Furthermore, government should emphasize improving marriage rather than making divorce difficult. However, wives (and husbands) who file for divorce in order to appropriate marital assets such as enhanced earning capacities are better off after divorce, but only at the expense of the other spouse. In this case, the opposite recommendation for divorce lawis forthcoming. When divorce is used to exploit the spouse remaining behind, divorce should be made more difficult to obtain23 in order to protect the specific investments of each spouse (Cohen, 1987; Ellman, 1989). If it is custody outcomes that most influence divorce filings, changes in custody rules (or their likely outcomes) rather than in divorce grounds should most shape the patterns of both marriage and divorce (Brinig and Buckley, 1998a).24 In particular, this could take the form of a presumption of joint custody or a rule that made post-divorce patterns mirror preseparation time shares as closely as possible, with sole custody only in cases where one party can show the other parent unfit.25 An appropriate custody rule mitigates the incentive for one-party filing for the purpose of 22. Some would say to stigmatize the exploiter as well (Golden and Taylor, 1987, pp. 12–13). 23. Apparently this is the viewof the lawand economics writers who feel that those leaving are by and large men. 24. Chapter 2 of the American LawInstitute Marital Dissolution project (1998), which would substantially change substantive child custody rules, may well have this effect. 25. The American LawInstitute (1998) proposal rules can be found in Section 2.02(e) (continuity of existing parent-child attachments) and 2.09(1), which provides that unless otherwise resolved by agreement or manifestly harmful to the child, “the court should allocate custodial responsibility so that the proportion of custodial time the child spends with each parent approximates the proportion of time each parent spent performing caretaking functions for the child prior to the parents’ separation
Why Most Divorce Filers Are Women 139 gaining unilateral control over the children and,to the extent both parents remain involved through visitation or child support,the other spouse. The question is,which type of issue drives most divorces?We now turn to an empirical analysis of divorce filing to sort out these theories and conclude that while we see exploitative,appropriative,and custody- related filing behavior,the custody patterns are the most important.We also note that no-fault(unilateral)divorce laws apparently do two things. They increase divorce filing rates generally (Allen,1992a;Brinig and Buckley,1998b;Friedberg,1998),and they increase the percentage of women who file.The fact that women,more than men,take advantage of the easier exit suggests that a return to a fault-based system will not advance women's goals. 4.Testable Implications There are three major individual characteristics that we can use to test our theory that filing behavior is driven by the two forms of quasi-rent appropriation outlined above.First,there are differences in age and age at marriage that generate specific quasi-rents for men and women.Second, the presence of children and their number create asymmetric quasi-rents that we can analyze.Third,there are differences in human capital and, in mixed-race couples,race between husbands and wives.In addition to these characteristics,a number of testable implications arise from specific grounds for divorce,information on the length of separation,and aggregate county-level demographic features. The hypotheses we develop are tested through data obtained from all the divorce certificates for 1995 (more than 46,000 with complete data) from Connecticut,Virginia,Oregon,and Montana.26 This data set is un- usually good for several reasons.First,it is large and inclusive:divorce law requires that people file where they are domiciled (Williams v.North Carolina,1942),so virtually all the years'divorces in the jurisdictions used are included.27 The data are relatively accurate,since they are com- 26.These states were selected based on the availability of data.Although all states except Louisiana collect divorce statistics,very few keep records of who files. 27.The exceptions would be spouses who left the state of their marriage to be divorced elsewhere or those who temporarily moved into one of the jurisdictions we are considering
Why Most Divorce Filers Are Women 139 gaining unilateral control over the children and, to the extent both parents remain involved through visitation or child support, the other spouse. The question is, which type of issue drives most divorces? We now turn to an empirical analysis of divorce filing to sort out these theories and conclude that while we see exploitative, appropriative, and custodyrelated filing behavior, the custody patterns are the most important. We also note that no-fault (unilateral) divorce laws apparently do two things. They increase divorce filing rates generally (Allen, 1992a; Brinig and Buckley, 1998b; Friedberg, 1998), and they increase the percentage of women who file. The fact that women, more than men, take advantage of the easier exit suggests that a return to a fault-based system will not advance women’s goals. 4. Testable Implications There are three major individual characteristics that we can use to test our theory that filing behavior is driven by the two forms of quasi-rent appropriation outlined above. First, there are differences in age and age at marriage that generate specific quasi-rents for men and women. Second, the presence of children and their number create asymmetric quasi-rents that we can analyze. Third, there are differences in human capital and, in mixed-race couples, race between husbands and wives. In addition to these characteristics, a number of testable implications arise from specific grounds for divorce, information on the length of separation, and aggregate county-level demographic features. The hypotheses we develop are tested through data obtained from all the divorce certificates for 1995 (more than 46,000 with complete data) from Connecticut, Virginia, Oregon, and Montana.26 This data set is unusually good for several reasons. First, it is large and inclusive: divorce lawrequires that people file where they are domiciled (Williams v. North Carolina, 1942), so virtually all the years’ divorces in the jurisdictions used are included.27 The data are relatively accurate, since they are com- 26. These states were selected based on the availability of data. Although all states except Louisiana collect divorce statistics, very fewkeep records of who files. 27. The exceptions would be spouses who left the state of their marriage to be divorced elsewhere or those who temporarily moved into one of the jurisdictions we are considering
140 American Law and Economics Review V2 NI 2000(126-169) piled by judicial personnel from facts alleged by the parties and proven based upon sworn testimony in court hearings.The data gathering pro- cess has been standard since 1960 because the National Center for Health Statistics has prepared national statistics based on these submissions.28 There are no obvious incentives for judicial personnel to exaggerate or distort any of these measures. Table 2 contains a list of definitions of variables,while Tables 3 and 4 list logit regression results from the four states.(A logit regression esti- mates the equation that will best produce the actual distribution between 0 and 1 for a two-option dependent variable [Maddala,1983,pp.59 et seq.][here,which spouse is the plaintiff].)The dependent variable in all regressions is 1 when the wife files for divorce.We now turn to each set of predictions in turn. 4.1.Age and Age at Marriage In a"traditional"marriage where the wife marries young and remains in the home,most of the appropriable quasi-rents are earned by the wife, and this puts her in a vulnerable position.In other words,she has in- vested more than the husband in the relationship at the beginning and has the most to lose if it ends.The longer these types of marriages last, the more the husband becomes advantaged,since wives are more likely to stay out of the labor force to perform caretaking and other household services (Fineman,1995a)29 and will perform such tasks longer.Further, when wives are older,they are relatively unlikely to remarry following di- vorce.Husbands,whose earning capacities are enhanced by marriage even if their wives do not put them through school,should be more likely to appropriate through divorce in these cases because most of the wife's con- tribution is up front.More subtly,the circumstances in which the spouses married reveal how much can be carried away at divorce.The woman who marries very young is unlikely to have a healthy earning capacity (and is more likely to have married because of an unplanned pregnancy).Hence, in these circumstances,husbands are more likely to opportunistically file. If the wife were to file,she would clearly be leaving a bad marriage. 28.The completed data set used here has not been compiled since 1995 because funding for the National Center for Health Statistics was reduced.Since that time,raw divorce rates have continued to be reported.Monthly vital statistics reports contain estimates from samples.For methods,see,for example,Technical Notes (1996,p.17). 29.They are less likely to remarry (Bumpass,Sweet,and Castro Martin,1990)
140 American Lawand Economics ReviewV2 N1 2000 (126–169) piled by judicial personnel from facts alleged by the parties and proven based upon sworn testimony in court hearings. The data gathering process has been standard since 1960 because the National Center for Health Statistics has prepared national statistics based on these submissions.28 There are no obvious incentives for judicial personnel to exaggerate or distort any of these measures. Table 2 contains a list of definitions of variables, while Tables 3 and 4 list logit regression results from the four states. (A logit regression estimates the equation that will best produce the actual distribution between 0 and 1 for a two-option dependent variable [Maddala, 1983, pp. 59 et seq.] [here, which spouse is the plaintiff].) The dependent variable in all regressions is 1 when the wife files for divorce. We now turn to each set of predictions in turn. 4.1. Age and Age at Marriage In a “traditional” marriage where the wife marries young and remains in the home, most of the appropriable quasi-rents are earned by the wife, and this puts her in a vulnerable position. In other words, she has invested more than the husband in the relationship at the beginning and has the most to lose if it ends. The longer these types of marriages last, the more the husband becomes advantaged, since wives are more likely to stay out of the labor force to perform caretaking and other household services (Fineman, 1995a)29 and will perform such tasks longer. Further, when wives are older, they are relatively unlikely to remarry following divorce. Husbands, whose earning capacities are enhanced by marriage even if their wives do not put them through school, should be more likely to appropriate through divorce in these cases because most of the wife’s contribution is up front. More subtly, the circumstances in which the spouses married reveal how much can be carried away at divorce. The woman who marries very young is unlikely to have a healthy earning capacity (and is more likely to have married because of an unplanned pregnancy). Hence, in these circumstances, husbands are more likely to opportunistically file. If the wife were to file, she would clearly be leaving a bad marriage. 28. The completed data set used here has not been compiled since 1995 because funding for the National Center for Health Statistics was reduced. Since that time, raw divorce rates have continued to be reported. Monthly vital statistics reports contain estimates from samples. For methods, see, for example, Technical Notes (1996, p. 17). 29. They are less likely to remarry (Bumpass, Sweet, and Castro Martin, 1990)