Unilateral Divorce and Time Allocation in the United States

ABSTRACT

Using time-diary data from the Time Use in Economic and Social Accounts (TUESA) 1975–76, which covers heterosexual couples in the United States, this paper analyzes the relationship between a state’s adoption of unilateral divorce and couples’ time allocation. Married women in states with unilateral divorce spend less time on core housework than those in states without unilateral divorce, and married men contribute to a greater share of housework. This paper also uses cross-state and time variation in divorce law adoption by including additional data from the early 1990s to estimate the effect of a state’s adoption of unilateral divorce on daily time use. The analysis confirms the findings for women in the 1970s: the availability of unilateral divorce substantially decreases married women’s time spent on housework. The results suggest that the adoption of unilateral divorce law shifts the relative bargaining power within heterosexual married households to women.

KEYWORDS:

• Time use
• intrahousehold bargaining
• family law
• housework

JEL Codes:

• D1
• J1

ACKNOWLEDGMENTS

This work was supported by the Minnesota Population Center (P2C HD041023), the Time Use Data Access System (R01HD053654), and IPUMS Redesign (R01HD043392), all funded through grants from the Eunice Kennedy Shriver National Institute for Child Health and Human Development (NICHD).

SUPPLEMENTAL DATA

Supplemental data for this article can be accessed at https://doi.org/10.1080/00343404.2017.1390318

Notes

1 The National Conference of Commissioners on Uniform State Act used the California Family Law Act as a guideline for the creation of the Uniform Marriage and Divorce Act. It eliminated fault from marital dissolution, property allocation, and child placement decisions. It also established a set of rules for a court to use in deciding whether a marriage had broken down (Kay 1987).

2 Dates for the adoption of unilateral divorce are based on Gruber (2004).

3 Internal threat point or “separate spheres” bargaining models include threat points within a marriage as opposed to external threat points (Lundberg and Pollak 1993). The internal threat point bargaining models predict little change in household allocation of time with unilateral divorce because changing the ability to get a divorce does not affect the internal threat points.

4 Child custody and support reform occurred shortly after the first round of the data used for this analysis, but nearly every state had adopted it by the second round of data used in the difference-in-differences analysis. Recent research suggests that the reforms increased women’s time spent on market work and increased fathers time spent on household production (Altindag, Nunley, and Seals 2015).

5 This study performed the couple-level analyses using the 1975 data while accounting for the three types of property rights. The statistically significant results are the same in sign and vary slightly in magnitude for each type of property law. Results can be obtained from the author.

6 The states that adopted unilateral divorce between 1975 and 1992 are Massachusetts, Rhode Island, South Dakota, Utah, Wisconsin, and Wyoming. The respondents in these states are 5.63 percent of the total weighted sample, 6.42 percent in the 1975 sample, and 4.16 percent of the 1990s sample.

7 For each respondent, this study imputed marital status twenty times using the multiple imputation program in Stata. The respondents predicted to be married at least 70 percent of the imputations are coded as married.

8 This sample did not use the spouses’ diaries. For comparability with the NHAPS data, this study included only the primary respondent.

9 The inability to identify recent marriage could impact the results of this analysis. The adoption of unilateral divorce laws will likely have a different impact on couples marrying prior to or following the law change.

10 As shown in Table 1, three states had already adopted a unilateral divorce law: Alaska, New Mexico, and Oklahoma.

11 Standard errors account for correlation within households.

12 This study uses OLS for the analysis of time allocation outcomes for two reasons. The first is that previous research has found that, with time-use data, OLS produces unbiased results when compared to Tobit and two-part models (Stewart 2013). The second reason is that these broad time allocation activities have few respondents reporting no time spent on them. This study performed Tobit models as robustness analysis, showing similar results.

13 The first model did not include race because the race of the spouse was not collected. However, the author estimated the models for only respondents with and without race control variables and showed similar findings. These results can be obtained from the author.

14 The author also ran the cross-sectional model without state-level fixed effects. Without the inclusion of state-level fixed effects, the results are similar to the mean comparisons where there are no differences in time allocation by unilateral divorce law status. However, many of the state-level fixed effects are significant in the analyses when they are included, suggesting there are state-level differences in time allocation. Results can be obtained from the author.

15 The author also performed the couple-level analyses replacing the divorce law indicator with a set of indicators identifying state property classifications (equitable distribution, common law, and community property). Significant differences in the relationship between unilateral divorce law and time allocation were not found by property law classification. Results can be obtained from the author.

16 In order to make the data compatible with the following time-diary studies, this study selected the first single-day diary in the TUESA data set.

17 Data limitations preclude the use of family income and children under age 5 as control variables.

18 The standard errors are clustered at the state and year level to control for serial correlation of error terms, allowing for nonindependent errors by the state, year, and household groups (Bertrand, Duflo, and Mullainathan 2004).

19 The standard errors are reported in the parentheses below the coefficients.

20 This study estimated the time allocation outcomes separately for respondents in states with and without unilateral divorce law. Differences in predicted means from the models between the states confirm the general findings of this analysis. Full results can be obtained from the author.

21 An analysis of total household time in home production shows that total time spent on core housework by both spouses decreases by 58 minutes in states with unilateral divorce compared to states without unilateral divorce. This is 48 minutes less than the difference found for women in states with unilateral divorce (shown in Table 4). Thus while there is a large difference in time spent by the whole household in housework, it does not fully explain the differences found for women between the states.

22 There are many more married women without a college degree in the sample, making the sample of women with college degrees very small; thus, the results for the women with a college degree should be interpreted with caution.

23 This study performed dynamic analyses accounting for time since the unilateral adoption with both data sets. The majority of states in the 1975–76 sample adopted the law within the five previous years, and there were no significant differences found using time since adoption. For the 1992–95 sample, the results look to be similar across time since adoption, with larger magnitudes occurring for four to six and seven plus years out. Results can be obtained from the author.

Additional information

Notes on contributors

Katie R. Genadek

Katie R. Genadek is Research Associate at the Institute of Behavioral Science at the University of Colorado - Boulder. She earned her PhD in Applied Economics from the University of Minnesota in 2012. Her research interests include the fields of labor economics, family economics, and demography.