Forthcoming in the International Review of Law and Economics
Why Are Married Women Working So Hard?
Professor of Management
Anderson Schools of Management
University of New Mexico
Albuquerque, NM 87131
* This research was funded by a summer research grant from the Anderson Schools of
Management Foundation. I am grateful to Micha Gisser and Ron Johnson for valuable comments.
Recent developments might cause one to conclude that the welfare of women has been improving. Certainly, their relative wages and labor force participation have increased.(1) However, one disconcerting development has been the increase in the total hours worked by married women at home and on a job. Victor Fuchs found that between 1960 and 1986, married women increased the total hours they worked per week by four hours, while their husbands were decreasing theirs by two and a half hours.(2) Sarah Fenstermaker Berk notes, "Thus, the practical effects of employment on wives' household labor are limited primarily to the 'shortcuts' that wives can apply to their work; a reapportionment of domestic chores among household members is very unlikely."(3) A similar pattern has been observed by other researchers.(4)
This paper provides an explanation for some of this increase based on the introduction of no-fault divorce. In the first section, the change in the divorce grounds from fault to no-fault is discussed. Then the impact of this change on the allocation of married women's time is discussed and a hypothesis about the impact of no-fault divorce is developed. This hypothesis is tested using data from the Time Use Longitudinal Panel Study, 1975-1981. The empirical results suggest that the introduction of no-fault divorce has increased the hours worked by married women. If hours worked are a measure of their welfare, no-fault divorce has made them worse off.
II. The Divorce Laws
During most of the history of the United States, divorce to the extent that it was allowed at all was based on fault with the grounds being adultery, cruelty and desertion. Because of the gains from specialization during marriage, essentially all adults benefitted from being married with marriages usually lasting until the death of one spouse. The increase in the employment opportunities for women and domestic labor-saving devises for men eventually reduced the gains from marriage for some adults increasing the attraction of divorce.(5) The divorce rate rose gradually during the 20th Century as it became more common for the spouses to agree--not necessarily enthusiastically--to divorce using fabricated testimony to establish the fault grounds. Because most spouses who wanted to dissolve their marriage did not have evidence to support the fault grounds, they had to obtain their spouse's cooperation. This process usually required concessions by the party who more strongly wanted the divorce as an inducement for the other spouse to be the plaintiff in the divorce action. Eventually, public concerns about the hypocrisy of the perjured testimony along with the self-interest of some divorced men and career women lead to the introduction of no-fault divorce based on incompatibility, irretrievable breakdown or separation for an extended period.(6) Between 1969 and 1985, all the states either replaced or augmented their fault grounds for divorce with no-fault grounds. In most cases, the no-fault grounds for divorce permitted a spouse to obtain a divorce unilaterally.
The laws controlling the financial and custodial arrangements at divorce, which often had been ignored in the prior negotiated divorces, were not addressed in any systematic way during these deliberations. Since a no-fault divorce usually did not require the agreement of the spouses, the need for concessions from a divorcing spouse were reduced and the settlements shifted from those based on negotiations--that often ignored the applicable laws--to ones based on those legal standards.(7) These laws usually have not been generous to individuals who have increased their specialization in domestic work during marriage. The property settlement usually consists of a return of any separate property and an equal division of the marital property--generally ignoring the largest source of the couple's wealth, human capital. In addition, a divorced spouse could expect short term spousal support along with child support for and joint custody of any minor children.
Gradually, it was recognized that the no-fault grounds for divorce reduced the negotiating power and, therefore, the settlements of spouses who did not want to divorce.(8) Often these individuals were women who had increased their specialization in domestic work during marriage. These women were particularly vulnerable because the contributions of men and women to marriage can be asymmetric.(9) The primary child rearing obligations often assumed by wives occur before the peak earnings period of the husbands making divorce especially attractive to middle-age husbands.(10)
II. The Impact of No-fault Divorce on Work by Married Women
As these effects of no-fault divorce were recognized, married women started taking steps to protect themselves from the potentially adverse consequences of divorce. Prior to 1970, much of the increase in the labor force participation rate of women can be explained by the increase in real wages.(11) That pattern stopped in the 1970s as the labor force participation rate of women accelerated as the real wage growth rate slowed. No-fault divorce has provided an explanation for some of the recent increase in women working outside the home during marriage.(12) The labor force participation rate of women has increased dramatically since World War II with the rate of married women rising from 23.8 percent in 1950 to 58.5 percent in 1991. This increase was particularly evident for the women who traditionally had the strongest attachment to the home, mothers with young children. The labor force participation rate of married women with children less than six years of age rose from 11.9 percent in 1950 to 30 percent in 1970, the year in which no-fault divorce was introduced in California. It continued to rise to 59.9 percent in 1991 having accelerated after 1970.
Not only have married women increased their labor force participation, they have also increased the total hours that they work. The decision by married women to work outside their home should not necessarily result in their working more total hours. Gary Becker asserts that married couples benefit from an increase in their specialization of labor based on comparative advantage with women normally increasing their specialization in domestic work and men in income earning.(13) This arrangement is in the best interests of the spouses, so they agree to it voluntarily. Labor market conditions may change increasing the opportunity cost of domestic work either because real wages have risen or the marginal productivity of some labor in the home has declined because of domestic labor-saving devices. Under those circumstances, it is attractive to families for wives to shift more time to employment. These women only would volunteer to make the shift if the benefit to them exceeded the cost. Some of their benefit may come from other family members who agree that the shift is also in their best interests and assume responsibility for some goods and services that had been provided by the women in the home.(14) The husband or children might assume some responsibility for cooking or cleaning. Overall, the total hours worked by married women should not necessarily increase.
Alternatively, married women can pursue employment due to the decline in the net benefit from increasing their specialization in domestic work during marriage. When divorce essentially required a negotiated settlement that could include favorable financial and custodial arrangements for the party who did not want the divorce, women could anticipate that the sacrifice of human capital to increase their specialization in domestic work had some protection. The no-fault grounds for divorce and the resulting reduced transfers at divorce provided by the law reduced this protection and, thereby, reduced the expected net benefit from specialization in domestic work during marriage. The likelihood increased that decisions by married women to become employed outside the home were based on the women's desire to protect themselves from the potentially adverse effects of no-fault divorce rather than to improve their family's welfare. The value of the goods and services available to their families due to the women's employment might be less than that of those produced during a corresponding period of domestic work. A shift of hours from household production to employment reduced the families' welfare. Under those circumstances, the other family members would be less likely to assume the responsibility for providing the goods and services that had been provided by the wives and mothers.
While women might feel compelled to take steps to protect themselves from the potentially adverse effects of divorce, their marriage will continue only as long as it is attractive to both spouses. In additional to physical and emotional attraction, their marriage should continue to be attractive to many wives because their husband's earnings have usually increased during the marriage.(15) Meanwhile, if wives' transfers of hours from domestic work to employment would lower their family's welfare, they have to reduce their domestic work by less than the increase in their hours of employment to continue to make this marriage attractive to their husband. Consequently, we would expect the total hours worked by married women to be greater in no-fault divorce states in which they had little protection than in fault divorce states in which they had some protection.
III. Empirical Analysis
The hypothesis developed above has been tested using 1981 data from the Time Use Longitudinal Panel Study, 1975-1981.(16) In 1975 and 1981, individuals were asked to complete diaries detailing their time spent in 223 activities during four days over a one year period. In addition, these individuals responded to demographic and financial questions. The 1981 survey consisted of 620 respondents and 376 spouses, although not all of them completed all four diaries. This survey is particularly attractive because it provides a cross-sectional sample of individuals living in both fault and no-fault divorce states. Less detailed demographic information was collected in 1981 than in 1975, so the only way to identify the state of residence in 1981 was to restrict the sample to the individuals who had not moved since 1975. This sample also had the advantage that it was more likely to consist of people familiar with that state's laws than new residents in a state. To provide for a lagged effect of the switch from fault to no-fault grounds for divorce, the sample was restricted to fault divorce states and the no-fault divorce states that had adopted no-fault divorce grounds by 1978. States that switched from fault to no-fault divorce between 1978 and 1981 were eliminated from the sample.
The primary focus of this study is on married women. Still, a concern was whether all married individuals worked more hours in no-fault divorce states. Regressions were also estimated for married men to address that concern. The sample consists of 172 married women and 159 married men.
Five regression equations for married women and men were estimated.(17) Table 1 contains the definitions of the variables used in the regressions and their means. The dependent variables used in these regressions are the minutes per week spent in four activities: regular work, housework, child care, and leisure plus total work that is the sum of regular work and housework. Domestic work was separated into housework and child care because the former was viewed as more onerous than the later. The independent variables consist of variables associated with labor force participation: age, family assets, religion, number and age of children, race, education, spouse's earnings, whether the family lived in a SMSA, and regional variables for West, North Central and South. The influence of no-fault divorce was introduced by a dummy variable for states that permitted unilateral divorce within two years in 1978.(18)
The results of the regression analysis for married women are presented in Table 2. Four of the no-fault divorce coefficients have the anticipated signs and they are statistically significant in three of the five equations. The coefficient in the regular work equation is positive and significant corroborating the earlier research that living in a no-fault divorce state tends to increase the employment of married women. The coefficient in the housework equations is not statistically significant, but it has the anticipated negative sign and the sum of it and the child care coefficient is substantially less than the increase in regular work implied by its coefficient. The no-fault coefficients in the leisure and total work equations are statistically significant at the 10 percent level suggesting that living in a no-fault divorce state results in married women having four and a half hours less leisure and approximately the same amount of additional time devoted to work. These results support the hypothesis that married women in no-fault divorce states have been forced to take steps to protect themselves from the potentially adverse effects of no-fault divorce.
It is particularly noteworthy that the four and a half hour change in the total work of married women in no-fault divorce states noted in this study is very similar to the change observed by Victor Fuchs in the total work of married women between 1960 and 1986, before and after the no-fault divorce became the norm in the United States. While this study investigates the effect of living in a no-fault divorce state, by 1986 all states had some form of no-fault divorce, so all married women were affected by no-fault divorce, thereby, providing an explanation for the change observed by Fuchs.
Among the other variables that are statistically significant, the signs are usually the ones we would expect--total work decreases with age and increases with education. However, there is a substantial reduction in total work for married women with young children that probably does not conform to our usual expectations. This is probably due to child care not being treated as work and many leisure activities such as reading a newspaper, talking on the telephone or walking can occur when someone is caring for children, but not when they are at regular work.
Table 3 contains the results from the regression analysis for married men. The no-fault divorce term is not statistically significant in any of the equations. Since these regression equations imply that living in a no-fault divorce state does not affect the hours worked by married men, one has to conclude that the greater diligence of married women in those states is not likely to be due to all adults working harder in those states.
This paper investigates the reasons why the hours worked by married women have increased. The introduction of no-fault divorce that caused a decline in married women's net benefit from specialization in domestic work provides an explanation. Because of these laws, an increase in employment by married women often is motivated by a desire for personal insurance against the potential costs of divorce rather than to increase their family's welfare. Consequently, the other family members have been reluctant to assume responsibility for the goods and services produced by the women in the home. Still, to continue to make their marriage attractive to their husband, they have to continue to provide substantial hours of domestic work. The result has been an increase in the total work hours by married women.
Table 1--Variable Definitions and Mean Values
Housework minutes per week associated with housework in 1981 2,173.3 1,180.9
Child Care minutes per week in child care in 1981 260.1 104.1
Leisure minutes per week spent in leisure in 1981 6,535.2 6,507.1
Total Work minutes per week in regular work and housework in 1981 3,279.5 3,464.7
Age age in 1981 47.9 51.2
Assets assets owned in 1981 in thousands of dollars 69.1 70.8
Catholic 1 if Catholic .24 .24
Children 0-3 number of children in the household age 0-3 in 1981 .11 .11
Children 4-18 number of children in the household age 4-18 in 1981 1.36 1.30
White 1 if white .9 .9
Years of Education number of years of education completed by 1981 12.3 12.8
SMSA 1 if the person lived in a standard metropolitan statistical
area in 1981 .35 .36
Husbands Earnings husband's earnings in 1981 in thousands of dollars 20.2 NA
Wive's Earnings wive's earnings in 1981 in thousands of dollars NA 4.7
No-Fault 1 if a divorce can be obtained in state within two years
based on no-fault grounds in 1978 .66 .68
West 1 if the person lived in West geographic division .14 .13
North Central 1 if the person lived in the North Central geographic
division .37 .36
South 1 if the person lived in the South geographic division .22 .23
Source: Time Use Longitudinal Panel Study, 1975-1981, Survey Research Center, University of Michigan.
Table 2--Activity Regression Equations: Married Women
Variable Work Work Care Work
(3.11) (2.54) (2.30) (7.88) (6.03)
Age -40.95*** 10.42* -8.36*** 38.91*** -30.53***
(4.83) (1.83) (4.26) (5.84) (4.37)
Assets .15 .02 .05 -.21 .16
( .24) ( .03) ( .31) (.43) (.32)
Catholic 155.41 -113.69 -26.71 -15.18 41.72
(.71) (.71) (.53) (.09) (.23)
Children 0-3 -648.09 *** 167.70 343.68*** 135.82 -480.39**
(2.81) (.99) (6.43) (.75) (2.53)
Children 4-18 -144.69* 102.20* 49.09*** -6.85 -42.49
(1.89) (1.82) (2.77) (.11) (.67)
White -256.38 216.88 141.04 -102.96 -39.49
(.54) (.63) (1.29) (.28) (.10)
Year of Education 72.61* -15.05 -2.44 -55.32* 57.56*
(1.96) (.55) (.29) (1.90) (1.89)
SMSA -122.63 -3.26 -10.35 136.57 -125.89
( .61) (. 02) (.22) ( .87) (.77)
Husband's Earnings -12.35*** 3.65 .70 9.41** -8.71**
(2.62) (1.06) (.64) (2.54) (2.24)
No-Fault 392.52** -122.42 1.74 -271.94* 270.10*
(2.01) (.85) (.04) (1.77) (1.68)
West -325.77 -155.42 101.22 379.56 -481.19*
(.99) (.64) (1.33) (1.46) (1.77)
North Central 78.15 -105.80 -56.14 83.62 -27.65
(.32) (.59) (.99) (.43) (.14)
South -109.75 -24.71 4.26 130.14 -134.46
(.38) (.12) (.06) (.57) (.57)
Adjusted R-Square .16 .03 .45 .26 .14
N 172 172 172 172 172
* Significantly different from zero at the 10-percent level.
** Significantly different from zero at the 5-percent level.
*** Significantly different from zero at the 1-percent level.
Table 3-Activity Regression Equations: Married Men
Variable Work Work Care Work
(4.13) (1.13) (1.33) (6.12) (5.28)
Age -53.14*** 10.71* -3.42** 45.85*** -42.43***
(5.55 ) (1.93) (2.59) (5.30) (4.89)
Assets -.03 -.23 -.01 .27 .26
( .04) (.58) (.13) (.43 ) (.41)
Catholic 121.08 48.06 -50.20 -119.54 169.14
(.49) (.33) (1.47) (.53) (.75)
Children 0-3 -218.60 88.00 118.88*** 11.33 -130.61
( .94) (.66) (3.73) (.05) (.62)
Children 4-18 49.73 -1.56 10.83 -59.00 48.17
(.57) (.03) (.90) (.75) (.61)
White 69.70 285.11 116.83* -472.64 354.81
(.15) (1.03) (1.78) (1.10) (.82)
Year of Education 46.45 -14.50 1.84 -33.97 31.95
(1.40) (.76) (.40) (1.14) (1.06)
SMSA -21.37 -190.17 -15.28 226.96 -211.53
(.10) (1.52) (.51) (1.16) (1.08)
Wive's Earnings 16.86 -3.99 2.48 10.44 12.86
(1.13) (.46) (1.21) (.78) (.95)
No-Fault 127.32 -65.21 -40.73 -21.22 62.11
(.56) (.50) (1.31) (.10) (.30)
West -131.26 156.69 17.48 -43.88 25.43
(.37) (.77) (.36) (.14) (.08)
North Central 189.14 56.81 -31.02 -215.85 245.95
(.66) (.34) (.79) (.84) (.95)
South 53.70 76.63 -34.06 -97.01 130.33
(.17) (.41) (.76) (.33) (.44)
Adjusted R-Square .28 .02 .22 .28 .23
N 159 159 159 159 159
* Significantly different from zero at the 10-percent level.
** Significantly different from zero at the 5-percent level.
*** Significantly different from zero at the 1-percent level.
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
1.See Robert T. Michael, "Consequences of the Rise in Female Labor Force Participation Rates: Questions and Probes," (1985) 3 J. Lab. Econ. S117; and Kevin M. Murphy and Finis Welsh, "Inequality and Relative Wages," (1993) 83 A. E. R 104.
2.Victor Fuchs, Women's Quest for Economic Equality, Harvard (1988), p. 77.
3.Sarah Fenstermaker Berk, The Gender Factory, Plenum (1985), p. 8,
4.For example, see Joseph H. Pleck, Working Wives/Working Husbands, Sage (1985) and Beth Anne Shelton, Women, Men and Time Greenwood (1992).
5.See June Carbone and Margaret F. Brinig, "Rethinking Marriage: Feminist Ideology, Economic Change, and Divorce Reform," (1991) 65 Tulane L. R. 953 and Lloyd R. Cohen, "Rhetoric, the Unnatural Family, and Women's Work," (1995) 81 Va. L. R. 2275.
6.Allen M. Parkman, No-Fault Divorce: What Went Wrong? Westview (1992).
7.Under both the fault and the no-fault grounds for divorce, private negotiations have been important, as negotiations are often an attractive alternative to litigation. George L. Priest and Benjamin Klein, "The Selection of Disputes for Litigation," (1984) 13 J.Leg.Stud.. 1. This is especially true in divorce cases. Robert H. Mnookin and Lewis Kornhauser, "Bargaining in the Shadow of the Law: The Case of Divorce," (1979) 88 Yale L. J. 950. Under fault divorce, approximately 90 percent of divorces were uncontested. Max Rheinstein, Marriage Stability, Divorce, and the Law, University of Chicago (1972) 248. The major change with no-fault divorce was that it was easier for a divorcing spouse to dissolve a marriage, thereby, reducing the concessions that they were willing to make to obtain the agreement of the other spouse.
8.Lenore J. Weitzman, The Divorce Revolution, Free Press (1985).
9.Lloyd Cohen, "Marriage, Divorce, and Quasi Rents; or `I Gave Him the Best Years of My Life,'" (1987) 16 J. Leg. Stud. 267.
10.While more divorces are initiated by women than men, the greatest injustices tend to occur when middle aged women in long duration marriages are divorced. For a discussion of who initiates a divorce, see Sanford L. Braver, Marnie Whitley, and Christine Ng, "Who Divorces Whom? Methodological and Theoretical Issues," (1993) 20 J. of Divorce and Remarriage 1.
11.James P. Smith and Michael P. Ward, "Time-Series Growth in the Female Labor Force," (1985) 3 J. Lab. Econ. S59. The increase in the labor force participation of women reduced the gains from marriage for some adults by decreasing the gains from specialization during marriage. As marriage became less attractive, support for easier divorce processes developed that contributed to the introduction of no-fault divorce. See Parkman, No-Fault Divorce, op. cit. note 6 at 58.
12.H. Elizabeth Peters, "Marriage and Divorce: Informational Constraints and Private Contracting." (1986) 76 A. E. R. 437 and Allen M. Parkman, "Unilateral Divorce and the Labor-Force Participation Rate of Married Women, Revisited," (1992) 82 A. E. R. 671 reach this conclusion with slight different frameworks using 1979 data. Using data from 1972, William R Johnson and Jonathan Skinner, "Labor Supply and Marital Separation," (1986) 76 A. E. R. 455 found that the probability of divorce increases the labor force participation rate of married women, but they found that living in a no-fault divorce state had a negative effect on a married woman's labor supply. However, 1972 was probably too early in the no-fault divorce era to provide a reasonable test of its effect.
13.Gary S. Becker, A Treatise on the Family, Harvard (Enl. ed.--1991).
14.Becker has argued that people do not maximize their utility by just buying goods, but instead they accomplish that goal by combining the goods that they buy with their time to produce "commodities." Their budget constraint is their "full" income that would be realized if all time were devoted to market work. Gary S. Becker, "A Theory of the Allocation of Time," (1965) 75 Econ. J. 493.
15.Another reason may be that women have a stronger attraction to their children than their husbands and they feel that an intact family is a preferred environment for raising those children. Fuchs, Women's Quest, op. cit., note 2.
16.The data used in this paper were made available by the Inter-university Consortium for Political and Social Research. The data for Time Use Longitudinal Panel Study, 1975-1981, were originally collected by F. Thomas Juster, Martha S. Hill, Frank P. Stafford, and Jacquelynne Eccles Parsons of the Survey Research Center, Institute for Social Research, the University of Michigan. Neither the collector of the original data nor the Consortium bear any responsibility for the analyses or interpretations presented here. For other uses of time budget surveys, see Jonathan Gershuny and John Robinson, "Historical Changes in the Household Division of Labor," (1988) 25 Demography 537 and F. Thomas Juster and Frank P. Stafford, "The Allocation of Time: Empirical Findings, Behavioral Models, and Problems of Measurement," (1991) 29 J. Econ. Lit. 471.
17.Since the property division at divorce varied among states based on legal standards such as community property and equitable distribution and on the amount of judicial discretion, equations were estimated that included those variables. None of those variables were significant, so they are not reported here.
18.In these states, a divorce could be obtained based on incompatibility, irretrievable breakdown or living separate and apart for up to two years. The key to these grounds is the availability of a divorce based on the desires of one spouse and over which the other spouse has little or no control.