due in 5 hours

Using longitudinal data covering 25 years from 1979 to 2004, the author examines the relationship between wives’ economic resources and the risk of marital dissolution. The author considers the effects of labor force participation, income, and relative income while accounting for potential endogeneity of wives’ economic resources. The extent to which wives’ economic resources are differentially related to marital disruption for Whites and Blacks is also ascertained. The author finds that the economic resources of women are tightly linked to the risk of divorce, both negatively and positively, for Whites but not for Blacks.

Keywords

divorce, economic resources, NLSY, labor force participation

Much has been written about the relationship between the economic resources of wives and the risk of divorce. On an empirical level, simultaneous increases in rates of divorce and rates of female labor force participation over the past 60 years, and longer, has generated interest in this relationship (Greenstein, 1990; Ruggles, 1997). On a theoretical level, models of marital structure and function- ing have stimulated continued concern for the link between wives’ economic resources and marital dissolution (Brines & Joyner, 1999; Oppenheimer, 1997),

1Western Washington University, Bellingham, WA, USA

Corresponding Author: Jay Teachman, Department of Sociology, Western Washington University, Bellingham, WA 98225-9081, USA Email: Jay.Teachman@wwu.edu

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1306 Journal of Family Issues 31(10)

even as divorce rates have stabilized in recent decades. In this article, I reexam- ine the nature of the relationships between the economic resources of married women and the stability of their marriages. Using longitudinal data from the 1979 National Longitudinal Study of Youth (NLSY), I pay attention to the dynamics of two components of wives’ economic resources, intensity of labor market participation (time spent working outside the home) and returns to labor market participation (income), while simultaneously considering the possibility that the relationships are endogenous. I also examine how the linkages between these two components of economic resources and risk of divorce vary according to race.

Prior Literature An extensive literature has investigated the link between wives’ economic resources and risk of marital disruption (see the review in Sayer & Bianchi, 2000; for more recent examples of such research, see Kalmijn, Loeve, & Manting, 2007; Rogers, 2004). Although numerous exceptions can be found, much of the past research has found that marriages in which women earn more, mar- riages in which women earn a greater share of total income, and marriages in which women work more weeks in the paid labor force are more likely to end. At first glance, the evidence would seem to indicate that women’s economic resources are positively linked to the risk of marital disruption.

A closer look at the literature, however, suggests some caution in accept- ing these findings without further consideration. There are several reasons for caution. First, a number of studies have failed to simultaneously consider the effects of both income and time spent in the labor force (e.g., Heckert, Nowak, & Snyder, 1998; Rogers, 2004; Sayer & Bianchi, 2000; South, 2001), making it difficult to disentangle the effects of these obviously cor- related, yet distinct, components of economic resources. For example, a study including a measure of labor force participation but not income will yield estimates that encompass the effects of working outside the home as well as income. Second, despite the fact that much of the prior research has made use of longitudinal data, little effort has been expended to sort the effect of economic resources on marital dissolution from its converse—that is, the effect of (potential) marital dissolution on economic resources. Third, prior research has generally ignored important race differences in the role played by women’s economic resources in marital stability. Failure to account for these differences runs the risk of confounding substantial race differences in the overall risk of divorce with race differences in the impor- tance of women’s economic resources to marital stability.

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1307

Recognizing these limitations of prior research, I seek to ascertain the relationship between wives’ economic resources and risk of marital dissolution while taking into account the possibility of endogeneity of wives’ resources. I consider these relationships separately for Whites and Blacks. I find that wives’ economic resources are strongly related to marital disruption for White women but much less so for Blacks.

Theoretical Concerns The perspectives most commonly used to link wives’ economic resources to risk of divorce focus on income and use some form of the specialization and trading model (Becker, Landes, & Michael, 1977; Brines & Joyner, 1999; Lundberg & Pollak, 1996; Oppenheimer, 1997). In this model, marital stability is generated through interdependence of spouses created by a traditional divi- sion of labor in which husbands specialize in labor market activities and wives specialize in home activities. When this division of labor is upset by growing equality in the economic resources of husbands and wives (and especially when the economic resources of wives, as measured by the ratio of their incomes, exceed those of their husbands), the marriage is destabilized, increasing the risk of disruption (Cooke, 2006; Heckert et al., 1998; Ruggles, 1997; Sayer & Bianchi, 2000). A related perspective, economic indepen- dence, yields expectations consistent with those derived from the trading and specialization model. In this perspective, though, higher absolute income (in contrast to relative income) allows wives the economic resources neces- sary to leave marriages in which they may have otherwise remained because of economic necessity (Becker et al., 1977; Sayer & Bianchi, 2000).

Contrasting with these perspectives, however, are suggestions that wives’ greater income relieves economic stress on marriages (sometimes called an income effect), reducing the likelihood of marital disruption. A variety of studies have found that wives’ economic contributions tend to make marriages more stable (Conger et al., 1990; Greenstein, 1990; Voydanoff, 1990), at least until they begin to earn more than their husbands (Ono, 1998). Other research also suggests that wives’ income has become normative, acting to stabilize mar- riages, with both husbands and wives expecting joint contributions to household finances (Nock, 1998; Sayer & Bianchi, 2000; White & Rogers, 2000).

Other investigators have called for greater attention to the effects of working outside the home that may be separate from the effects of income (Cooke, 2006; Ono, 1998; Schoen, Rogers, & Amato, 2006). Although income largely flows from labor market participation, and thus the two are positively corre- lated, they may have unique effects on marital stability. Distinct from the

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1308 Journal of Family Issues 31(10)

impact of income, the specialization and trading model would suggest that mar- riages in which women remain at home and specializes in home production are less likely to dissolve because the interdependence of spouses is increased. South and Lloyd (2001) suggest another mechanism reducing the risk of divorce when wives do not work outside the home. They argue and find support for the notion that working wives are more exposed than stay-at-home wives to attractive opposite-sex romantic partnerships that destabilize marriages.

Countervailing arguments, again separate from the effects of income, can be made to support the notion that specializing in home production on the part of wives will increase the risk of marital disruption. As the roles and responsi- bilities of spouses have become less typed along gender lines (Bianchi, Milkie, Sayer, & Robinson, 2000; Robinson & Godbey, 1997), married couples have come to expect that both spouses will work for pay. For example, men cite labor market participation as a desired characteristic of a spouse (Buss, Shackelford, Kirkpatrick, & Larsen, 2001). Thus, women who remain at home may risk violating increasingly strong norms for labor market participation irrespective of the amount of money they earn, increasing the risk of marital disruption. Goldscheider and Waite (1991) suggest that the basis of marital negotiation and success has shifted, with greater expectation of equal contri- butions to home and labor market on the part of both spouses. Other research finds that employment provides a sense of balance and achievement among women who have increasingly pursued higher education and have filled more demanding and satisfying jobs, resulting in less psychological stress (Ross, Mirowsky, & Goldstein, 1990) and greater marital satisfaction (Rogers & DeBoer, 2001). Consistent with this argument, some researchers have found that the risk of divorce is elevated when wives do not work in the paid labor market (Cooke, 2006; Ono, 1998; Schoen, Rogers, & Amato, 2006).

Sorting between the unique effects of income and labor market participation is important, because they may not operate in the same direction in affecting risk of divorce. Because they tend to occur together, failure to consider both components of economic resources simultaneously, as most previous research has failed to do, could therefore lead to finding null or reduced effects for the included component because of offsetting effects associated with the other component. For example, inconsistencies in the reported relationship between wives’ income and risk of marital dissolution may be at least partially because of the failure to distinguish between the (positive) effects of income and the (negative) effects of employment (Ono, 1998). Without making this distinc- tion, therefore, theoretical progress may be hampered by inadequate attention to measurement.

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1309

The link between wives’ economic resources (either income or labor force participation) and marital stability is further confounded by the possibility of endogeneity—that is, the possibility that women enter the labor market or increase their work efforts and earn income because their marriages are already in trouble. For example, Rogers (1999) found that increases in marital discord led to increases in wives’ income and the likelihood that nonemployed wives entered the labor force. Similar findings are reported by Schoen, Rogers, and Amato (2006). The instability of wives’ participation in the labor market may mean that marital problems act as a stimulus to changes in employment and income rather than the converse. Endogeneity also means that unless specifically measured, it will be difficult to untangle the effects of long- term, and thus more stable, components of employment from shorter-term changes in employment. Longitudinal data on marriages and wives’ economic resources are necessary to reduce the likelihood that results are because of endogeneity.

Race also confounds the relationship between wives’ economic resources and risk of marital dissolution. For the most part, salient research has simply included race as a covariate, recognizing the fact that levels of marital dis- solution are higher among Blacks. However, other research has found that the predictors of divorce vary considerably according to race (Phillips & Sweeney, 2005, 2006). In general, predictors of divorce are weaker and less common among Blacks than among Whites. For example, Phillips and Sweeney (2005) found that premarital cohabitation is positively associated with mari- tal dissolution among White women but not among Black women. These findings suggest that the marriages of Blacks and Whites may function along different dimensions. For example, Chadiha, Veroff, and Leber (1998) find that Black couples are more likely to mention religion and relationships as important components of a marriage, whereas White couples are more likely to mention achievement and work. In a similar vein, Furdyna, Tucker, and James (2008) report a negative relationship between wives’ marital happi- ness and her relative contribution to household income among Whites but not Blacks. These findings raise the possibility that the effects of economic resources on marital stability will vary according to race, something that prior literature has not considered. In particular, prior research suggests that wives’ economic resources are more important among Whites than Blacks.

Using data taken from the 1979 NLSY, I ask several questions generated by prior research. First, what is the link between wives’ income and risk of marital dissolution? Is the relationship positive or negative? Second, is the effect of wives’ income distinct from the effect of their labor market participation?

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1310 Journal of Family Issues 31(10)

Third, do the effects of wives’ economic resources on risk of divorce remain even after controlling for potential confounding because of endogeneity? Fourth, do the linkages between wives’ economic resources and divorce vary according to race? In particular, are Whites more sensitive to variations in wives’ economic resources than Blacks? I further minimize the likelihood of confounding effects by including controls for a wide range of covariates known to be related to the risk of marital dissolution, including age at mar- riage, mental aptitude, marital duration, religion, parental education, child- hood living arrangements, marriage cohort, premarital cohabitation, number of siblings, education, children, and school enrollment (Bumpass, Martin, & Sweet, 1991; Teachman, 2002).

Data and Method In 1979, the NLSY-79 interviewed 12,686 men and women between the ages of 14 to 21 years. The women in my sample were interviewed a maximum of 21 times between 1979 and 2004 (interviews were annual through 1994, biennial thereafter). In my analysis, I consider women who married for the first time between the years 1979 and 2002 and could have divorced by 2004. Because I employ a discrete-time event history model to examine the risk of marital dissolution, I create a database consisting of person years where women contribute a person year for each round of the NLSY-79 in which they were married and interviewed. Women exit the sample when they expe- rience marital disruption (either separation lasting more than 1 year or a divorce). To the extent possible, if a respondent was not interviewed in a particular year but was interviewed in a subsequent year, I used retrospective information collected by the NLSY-79 to complete information for the missing person year. Women who are permanently lost to follow-up, either because they cannot be tracked or because of changes in the sampling frame of the NLSY-79, contribute person years until they exit the survey.

Because I expect that economic resources are linked to subsequent marital disruption differently by race, I create two samples, one for Whites and one for Blacks. The sample for Whites consists of 21,210 intervals (2,166 unique individuals). The sample for Blacks consists of 6,375 intervals (561 unique individuals).

The dependent variable is a binary measurement indicating whether a woman divorced or separated in the interval between survey rounds (0 = did not divorce or separate, 1 = divorced or separated). Women who divorced or separated in an interval are dropped from subsequent intervals. In the discrete time procedure, I analyze this dependent variable using a logistic regression

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1311

procedure. The resulting coefficients, as shown below, indicate the effect of the covariate in question on the odds of divorcing in any given interval.

I use time-varying indicators to measure both the absolute and relative incomes of wives and husbands. Absolute income for both spouses is mea- sured as a 2-year average (lagged 1 year to account for the fact that spouses are not likely to react instantaneously to a given level of income; thus, at time t, the average of income at time t - 1 and t - 2 is used), from all sources (e.g., wages, transfer payments, interest on investments), and adjusted for inflation using an average of 1983 to 1984 dollars. To correct for a strong positive skew, the natural logarithm of average income is used.

Relative income is measured as a set of four time-varying dummy variables using information on the ratio of wife’s to husband’s (nonlogged) income over two consecutive years (2 years prior to the current interval and 1 year prior to the current interval). Two of these dummy variables indicate stability in relative income (the other two dummy variables indicate change in income ratio). Following prior research (Rogers, 2004), the first dummy variable is coded 1 when the wife made less than 40% of the couple’s total income in both years, and the second dummy variable is coded 1 when the wife made more than 60% of the couple’s total income in both years. The omitted cate- gory represents marriages where the wife made between 40% and 60% of the couple’s total income in both years.

The measure of the wife’s labor force attachment that I use is straightforward. I simply count the cumulative number of weeks she has been employed between first entry into the labor market and the beginning of the current interval. This measure is intended to tap the overall stability of labor market commitment on the part of the wife.

To obtain some leverage over the issue of endogeneity, I use recent changes in income and labor market participation as time-varying variables. If women respond to a poor marriage and the potential of an eventual marital disruption by increasing the economic resources on which they can draw after separation, then a (lagged) positive change in labor market participation or income ratio should be positively linked to the risk of marital disruption net of stable components of income and cumulative labor market participation. Similarly, if economic resources affect marital stability beyond the likeli- hood of endogeneity, then, net of recent change in wives’ economic resources, the risk of divorce should be linked to the stable components of her income, income ratio, and labor market participation.

An upward change in labor force participation is indicated by a dummy vari- able coded 1 for wives who were employed at least 4 weeks more in year t - 1 compared with year t - 2, 0 otherwise. A decrease in labor force participation

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1312 Journal of Family Issues 31(10)

is indicated by a dummy variable coded 1 for wives who were employed at least 4 weeks fewer in year t - 1 than in year t - 2, 0 otherwise. The baseline category consists of women whose employment across the 2 years did not vary by more than 1 month. Change in income ratio is coded as two dummy variables. The first dummy variable is coded 1 for women whose income ratio declined over the 2 years preceding the current interval (where decline is measured as a change downward across the three categories of income ratio: <40%, 40% to 60%, >60%), 0 otherwise. The second dummy variable is coded 1 for women whose income ratio increased over the 2 years preced- ing the current interval (again using the three categories of income ratio), 0 otherwise. As indicated previously, the omitted category represents mar- riages where the income ratio of spouses remained constant.

I control for a number of additional covariates well-known to be related to the risk of divorce (Bumpass, Martin, & Sweet, 1991; Teachman, 2002). These variables include three time-varying covariates: highest grade of edu- cation completed as of the beginning of each interval, the number of children residing in the household at the beginning of each interval, and a dummy variable indicating whether the respondent was enrolled in school during May of each interval (0 = no, 1 = yes). I also control for a number of fixed covariates, including mother’s education measured as years of schooling completed as of 1979, mental aptitude of the respondent measured as her score on the AFQT measured in 1980 (see Holley, Yabiku, & Benin, 2006, for justification of using intelligence as a covariate for divorce), and number of siblings measured in 1979. A series of dummy variables control for whether the respondent was raised by both parents until age 18, was born in the 1960s versus earlier, cohabitated prior to marriage, or was raised in a rural area (in all cases 0 = no, 1 = yes). Religion is measured as a series of dummy variables (0 = no, 1 = yes): Catholic, none, and other. Protestant con- stitutes the omitted category. Because rates of marital dissolution may vary significantly, and nonlinearly, by duration, I include measures of marital duration (measured as years) and its square.

Descriptive Statistics Descriptive statistics (based on weighted data) for the analysis are presented in Table 1, separately by race. Values for all variables included in the analysis are shown for three points in time: the end of the 1st year of marriage, the end of the 5th year of marriage, and the end of the 15th year of marriage. Consistent with the much higher rates of marital dissolution for Blacks, only about 37% of Black marriages survived 15 years compared with about 53% of White

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1313

Table 1. Descriptive Statistics for Variables Used in Analysis of Risk of Marital Disruption: NLSY-79

Year 1 Year 5 Year 15

Variable Mean SD Mean SD Mean SD

Whites Cumulative proportion .94 .77 .53 of marriages surviving Income and labor force participation

Log income 6.94 3.28 6.54 3.71 6.85 3.75 Log husband’s 5.22 3.30 7.78 3.45 8.64 3.21 income Cumulative labor 210.14 188.09 364.70 202.24 636.31 252.04 market experience Income ratio 0.36 0.51 0.30 decreases Income ratio 7.08 2.70 2.89 increases Income ratio 38.87 50.29 56.32 consistently low (<40%) Income ratio 18.48 14.23 9.90 consistently high (>60%) Increase in weeks 28.53 27.44 77.62 worked Decrease in weeks 23.18 22.63 1.21 worked

Control variables Age at marriage 23.40 4.47 23.33 3.91 21.59 2.67 AFQT score 52.69 26.48 54.34 26.78 56.11 25.23 Catholic 32.96 32.92 32.72 No religion 3.55 3.36 3.50 Other religion 13.22 14.23 11.72 Mother’s education 11.82 2.44 11.94 2.41 11.77 2.44 Lived with biological 77.11 80.58 81.88 parents at 18 Born in 1960s 51.78 52.26 42.47 Ever cohabited prior 25.72 24.23 12.79 to marriage

(continued)

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1314 Journal of Family Issues 31(10)

Table 1. (continued)

Year 1 Year 5 Year 15

Variable Mean SD Mean SD Mean SD

Number of siblings 3.21 2.08 3.15 2.05 3.03 1.89 Highest grade 13.12 2.30 13.57 2.39 13.61 2.25 completed Number of children 0.35 0.65 1.25 0.97 2.06 living in household Enrolled in school 8.27 5.47 4.41

Blacks Cumulative proportion .91 .65 .37 of marriages surviving Income and labor force participation:

Log income 6.05 3.65 6.50 3.57 7.23 3.65 Log husband’s income 3.81 3.31 5.54 3.97 6.79 4.17 Cumulative labor 201.99 204.27 321.91 219.92 597.78 256.89 market experience Income ratio 0.78 0 0.56 decreases Income ratio 9.67 6.87 4.47 increases Income ratio 26.87 29.62 29.61 consistently low (<40%) Income ratio 25.96 24.88 25.13 consistently high (>60%) Increase in weeks 32.90 34.36 83.80 worked Decrease in weeks 20.31 18.48 2.23 worked

Control variables Age at marriage 25.25 5.58 24.37 4.50 21.82 2.66 AFQT score 24.65 20.01 25.63 21.20 27.55 19.97 Catholic 8.70 8.29 7.82 No religion 3.24 2.37 2.79 Other religion 10.77 11.37 12.84 Mother’s education 10.86 2.49 10.84 2.61 11.02 2.20 Lived with biological 51.65 54.50 56.98 parents at 18

(continued)

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1315

Note: NLSY-79 = 1979 National Longitudinal Study of Youth

Table 1. (continued)

Year 1 Year 5 Year 15

Variable Mean SD Mean SD Mean SD

Born in 1960s 55.29 57.58 47.48 Ever cohabited prior 21.35 22.04 12.84 to marriage Number of siblings 4.78 3.14 4.90 3.23 4.74 2.58 Highest grade 12.88 1.97 13.20 2.10 13.34 1.94 completed Number of children 0.99 1.07 1.65 1.13 2.10 1.06 living in household Enrolled in school 0.84 6.87 5.59

marriages. In the early years of marriage, White women also enjoy higher incomes (although they have less cumulative labor market experience) and spouse incomes. For both Blacks and Whites, relatively few women increased or decreased their income ratios over time, although there is a greater like- lihood that this ratio increases rather than decreases (see also Winkler, McBride, & Andrews, 2005; Winslow-Bowe, 2006). Indeed, there were no Black women for whom there was a decrease in income ratio. Most women have relatively stable income ratios. White women are more likely to have a consistently low income ratio (<40% of family income), and Black women are more likely to have a consistently high income ratio (>60% of family income). In terms of changes in labor market intensity, there is a greater like- lihood of increasing weeks worked rather than decreasing weeks worked. The change in labor force participation evident between 5 and 15 years of marriage is particularly strong for both Blacks and Whites, likely reflecting the decreased burdens associated with rearing young children allowing women to enter the labor force and do so more actively.

Multivariate Analysis Results from estimating a discrete-time event history model for the risk of divorce is shown in Table 2, separately for Whites and Blacks. As is most often the case with the NLSY data, I use unweighted data to estimate the multivariate models (Teachman, 2007). Results based on weighted data do not differ substantively (results not shown). I also conducted Chow tests (using a chi-square distribution) to determine whether separate logistic regression

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1316 Journal of Family Issues 31(10)

Table 2. Relationship Between Measures of Economic Resources and Control Variables and Odds of Marital Disruption for Whites and Blacks: Odds Ratios Estimated From the NLSY-79

Whites Blacks

Variable Model 1 Model 2 Model 1 Model 2

Income and labor force participation Log of income 1.038** 1.032** 1.009 0.993 Log of husband’s 0.953** 0.968** 0.973 0.982 income Cumulative labor 0.998** 0.998** 0.999** 1.001 market experience Income ratio decreases 0.693 0.711 — — Income ratio increases 1.069 1.247 0.987 1.073 Income ratio 1.007 0.998 0.977 1.007 consistently low (<40%) Income ratio consistently 1.381** 1.492** 1.195 1.304* high (>60%) Increase in weeks worked 1.287** 1.256** 1.068 1.115 Decrease in weeks worked 0.947 0.911 1.074 1.065

Control variables Age at marriage 0.972 0.990 AFQT score 0.993** 0.997 Marital duration 0.981 0.966 Years married squared 1.001 0.997 Catholic 0.861* 0.894 No religion 1.237 1.203 Other religion 0.959 0.752* Mother’s education 1.033* 0.985 Lived with biological 0.848** 0.966 parents at 18 Married in 1960s 1.189** 1.042 Ever cohabited 1.345** 0.873 prior to marriage Number of siblings 0.974 0.9738 Highest grade completed 0.929** 0.918** Number of children 0.900** 1.050 living in household Enrolled in school 1.336** 1.163 Intercept 0.079** 0.400** .1066** 0.549** -2 log likelihood 7491.75 7015.30 3295.57 3117.73

Note: NLSY-79 = 1979 National Longitudinal Study of Youth. *p < .10. **p < .05.

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1317

models for each race were warranted. For Model 1, the chi-square result was 44.62 with 1 degree of freedom (statistically significant with p < .001). For Model 2, the chi-square result was 48.82 with 1 degree of freedom (statisti- cally significant with p < .001). These results imply that the models predicting marital dissolution for Whites and Blacks are indeed different and should be estimated separately.

The values shown in Table 2 are odds ratios and represent the multiplicative relationship between each variable and the risk of marital dissolution in an interval (period between rounds of the survey) controlling for the effects of other variables in the model. Values less than 1.0 indicate a decreased risk of divorce, and values greater than 1.0 indicate an increased risk of divorce. To estimate the percentage change in the odds of divorce in an interval the fol- lowing transformation can be applied: (eb - 1) * 100, where eb is the value shown in Table 2. Two models are shown for each race. Model 1 is a reduced- form estimate, indicating the overall effects of the primary independent variables measuring economic resources and change in these resources. Model 2 indi- cates the effects of the primary independent variables controlling for the host of measured covariates of marital dissolution.

The results for Model 1 indicate substantial differences between Whites and Blacks in the effects of economic resources on the risk of divorce. For Black women, only cumulative labor market experience is linked to the risk of marital disruption. Every 100 weeks of additional labor market experience reduces the risk of divorce in an interval by about 10%. For White women, the story is different. Not only does cumulative labor market experience lower the risk of divorce in an interval (about 10% for every 100 weeks of cumulative experience) but several additional indicators of economic resources are statistically significant.

For White women, the effect of husband’s income on marital disruption is negative. The effect of the wife’s income, however, is positive (a 3.8% increase in risk of divorce in an interval for every natural log unit increment in income), a finding consistent with the specialization and trading and inde- pendence perspectives. There is no effect of changes in wives’ income ratio. Wives who consistently make more than their husbands, however, are 38% more likely to experience a divorce in an interval than wives who make about the same or less than their husbands. An upward change in weeks worked is also linked to an increased risk of divorce. Consistent with the notion of endogeneity, wives who increased their time in the labor market are about 29% more likely to experience marital dissolution in an interval compared with wives who kept a constant presence in the labor market or decreased their participation.

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1318 Journal of Family Issues 31(10)

The results for Model 2 are generally similar to those for Model 1, particu- larly for Whites. The magnitudes of coefficient estimates are not substantially different across the two models and no variables gain or lose statistical sig- nificance. For Blacks, however, the effect of cumulative labor market experi- ence loses statistical significance, whereas the effect of having a consistently high income ratio gains marginal statistical significance (mostly because of the control for education). Specifically, net of the control variables, Black women who consistently earn more than 60% of total family income are about 30% more likely to experience marital dissolution in an interval com- pared with women who earn a lower fraction of family income. Consistent with prior research, the results for Model 2 indicate many fewer predictors of marital dissolution for Blacks than Whites (Phillips & Sweeney, 2006).

Discussion The results shown in Table 2 provide important evidence about the linkages between the economic resources of wives and the risk of marital dissolution, answering several questions left open by prior research. First, the results show that income and labor force participation act in different ways to affect marital stability. Wives’ higher incomes and higher income ratios act to destabilize marriages, whereas cumulative labor market participation acts to stabilize marriages. Second, consistent with the notion of endogeneity, there is a positive effect of an increase in weeks worked on divorce. Third, the effects of economic resources are much different for Whites and Blacks. Vir- tually none of the measures of economic resources were related to the risk of marital dissolution for Blacks.

The fact that women who earn more are more likely to experience divorce is consistent with an independence effect. More income allows otherwise unhappy women to afford alternatives to their marriage by reducing their economic dependence on their spouses. The fact that independent of this effect women who earn a greater share of family income are more likely to experience marital disruption suggests that breaking traditional gender roles is also costly to the stability of a marriage. This finding is consistent with the specialization and trading model. Women may earn as much as their husbands without negative consequences for marital stability, but once that ratio exceeds parity, destabilizing influences result. The results do not provide any evi- dence that supports the income model where wives’ income stabilizes mar- riage by reducing uncertainty in household income.

The fact that wives’ cumulative labor force participation has a negative effect on rates of marital dissolution suggests a multifaceted relationship

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1319

between wives’ economic resources and marital stability. Greater wives’ income is destabilizing to a marriage, but so is limited participation in the labor force. This pattern is consistent with expectations that couples might have about a more equal division of household of labor. Breaking these norms may act to destabilize a marriage (Brines & Joyner, 1999). In addition, the presence of these norms may mean that as a consequence labor force participation provides a source of fulfillment and accomplishment not avail- able to women who remain at home and which acts to increase marital hap- piness and thus reduce the risk of marital disruption.

It is also the case that a recent increase in work hours is associated with an increase in the risk of marital disruption. This effect suggests that women react to uncertainty in their marriages by increasing their work effort (it is also possible that increased work effort destabilizes a marriage by forcing the couple to renegotiate their division of labor). That is, the relationship between divorce and economic resources likely runs both ways. The threat of marital dissolution stimulates greater labor force activity on the part of wives and their higher income increases the likelihood of marital dissolution.

The relationship between wives’ economic resources and marital stability is different for Black women. As expected, there is little relationship between economic resources and the risk of marital disruption among Blacks. Net of the control variables, only earning more than 60% of family income is (positively) related to the risk of marital dissolution. Neither husbands’ income nor wives’ income is linked to divorce risk. In addition, the participation of Black wives in the labor market is not related to risk of marital disruption; neither cumulative labor force participation or change in labor force participation. These results speak to the more precarious economic position of Black families in the labor market, and the long history of Black wives’ participation in the labor market, where the economic activity of both spouses is needed to obtain the eco- nomic stability necessary for a stable union.

That women who earn the majority of family income are more likely to experience divorce indicates the continuing importance of traditional gender roles in Black marriages, however. Even though Black women are more likely to participate in the economic support of their families and their eco- nomic contributions do not appear to generate an independence effect, and there is evidence in support of endogeneity, Black women are no less likely than White women to subscribe to traditional gender-role ideologies (Furdyna et al., 2008; Taylor, Tucker, & Mitchell-Kernan, 1999). As was the case with White marriages, breaking with the strongly held breadwinner role of husbands appears to destabilize marriages. This is the only finding pertaining to eco- nomic resources that is consistent across both Blacks and Whites.

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1320 Journal of Family Issues 31(10)

These results suggest several lessons for subsequent research on the link- ages between wives’ economic resources and marital stability. First, both income and labor force participation have distinct effects on risk of marital dissolution. Because the two are obviously correlated (the majority of income comes from working outside the home), failure to consider each dimension separately will lead to biased results. Second, there is evidence that women might respond to the threat of future marital instability by increasing their labor market participation. Failure to account for this possibility will yield biased estimates of the effects of both income and cumulative labor market participation. Longitudinal data are necessary for sorting through this alter- native. Third, and perhaps most important, Black and White marriages react differently to wives’ economic resources. Thus, research that simply uses race as a covariate will miss much of the interesting story about the processes impacting marital disruption.

Conclusion I have used data taken from the NLSY-79 to ascertain the relationship between the economic resources of wives and the risk of marital dissolution. I have extended prior research in several ways. First, I pay attention to the distinct effects of both labor force participation and income. Second, I take into account the possibility that economic resources are endogenous to marital instability. Third, I consider the relationship between economic resources and divorce as it varies according to race.

The results indicate that the effects of wives’ income and labor market participation are distinct, particularly for White women. More income, both absolute income and relative income, tends to destabilize marriages, whereas labor market participation tends to stabilize unions. The results also show that the effects of change in economic resources are different from the effects of level of economic resources. In particular, whereas cumulative labor market experience acts to stabilize marriages, positive change in labor market activity is linked to an increase in marital instability. Finally, the results indicate that the relationship between economic resources and risk of marital dissolution is different for Whites than for Blacks. Only Black women who earn more than 60% of family income are more likely to experience marital disruption. No other economic resources of Black women are linked to the risk of divorce.

Although it has provided a more comprehensive coverage of the link between women’s economic resources and marital dissolution, this study is not without weaknesses. One of the most significant weaknesses is lack of information pertaining to marital quality and gender role ideology. Prior research indicates that the effect of wives’ income on marital disruption is heightened

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1321

when women hold nontraditional gender roles (Greenstein, 1995) or are in unhappy marriages (Schoen, Astone, Rothert, Standish, & Kim, 2000). Pos- sessing such information would better allow researchers to understand the nature of the linkages between women’s economic resources and divorce. Another limitation of the study is that it is entrenched in a particular historical period, largely covering women who married for the first time in the 1980s and early 1990s. As marriages continue to evolve so too might the relation- ship between wives’ economic resources and divorce. Finally, the data lack full information about the economic resources of husbands, which might allow better specification of economic stress faced by couples.

These limitations notwithstanding, the results show that the effects of wives’ economic resources are multidimensional, with effects likely running in both directions—from economic resources to marital stability and the reverse. Equally important is the finding that wives’ economic resources are only weakly tied to the risk of divorce among Black women, reflecting the different bases on which Black marriages are formed.

Declaration of Conflicting Interests

The author(s) declared no conflicts of interest with respect to the authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research and/or authorship of this article.

References

Becker, G., Landes, E., & Michael, R. (1977). An economic analysis of marital insta- bility. Journal of Political Economy, 85, 1141-1187.

Bianchi, S., Milkie, M., Sayer, L., & Robinson, J. (2000). Is anyone doing the house- work? Trends in the gender division of household labor. Social Forces, 79, 191-228.

Brines, J., & Joyner, K. (1999). The ties that bind: Principles of cohesion in cohabita- tion and marriage. American Sociological Review, 64, 333-355.

Bumpass, L., Martin, T., & Sweet, J. (1991). The impact of family background and early marital factors on marital disruption. Journal of Family Issues, 12, 22-42.

Buss, D., Shackelford, T., Kirkpatrick, L., & Larsen, R. (2001). A half century of mate preferences: The cultural evolution of values. Journal of Marriage and Family, 63, 491-503.

Chadiha, L., Veroff, J., & Leber, D. (1998). Newlyweds’ narrative themes: Meaning in the first year of marriage for African American and white couples. Journal of Comparative Family Studies, 29, 115-130.

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

1322 Journal of Family Issues 31(10)

Conger, R., Elder, G., Lorenz, F., Conger, K., Simons, R., Whitbeck, L., . . . Melby, J. (1990). Linking economic hardship to marital quality and stability. Journal of Marriage and the Family, 52, 643-656.

Cooke, L. (2006). “Doing” gender in context: Household bargaining and risk of divorce in Germany and the United States. American Journal of Sociology, 112, 442-472.

Furdyna, H., Tucker, M., & James, A. (2008). Relative spousal earnings and marital happiness among African American and white women. Journal of Marriage and Family, 70, 332-344.

Goldscheider, F., & Waite, L. (1991). New families, no families? The transformation of the American home. Berkeley: University of California Press.

Greenstein, T. (1990). Marital disruption and the employment of married women. Journal of Marriage and the Family, 52, 657-676.

Greenstein, T. (1995). Gender ideology, marital disruption, and employment of married women. Journal of Marriage and the Family, 57, 31-42.

Heckert, A., Nowak, T., & Snyder, K. (1998). The impact of husbands’ and wives’ relative earnings on marital disruption. Journal of Marriage and the Family, 60, 690-703.

Holley, P., Yabiku, S., & Benin, M. (2006). The relationship between intelligence and divorce. Journal of Family Issues, 27, 1723-1748.

Kalmijn, M., Loeve, A., & Manting, D. (2007). Income dynamics in couples and the dissolution of marriage and cohabitation. Demography, 44, 159-179.

Lundberg, S., & Pollak, R. (1996). Bargaining and distribution in marriage. Journal of Economic Perspectives, 10, 139-158.

Nock, S. (1998). Marriage in men’s lives. New York, NY: Oxford University Press. Ono, H. (1998). Husbands’ and wives’ resources and marital dissolution. Journal of

Marriage and the Family, 60, 674-689. Oppenheimer, V. (1997). Women’s employment and the gain to marriage: The spe-

cialization and trading model. Annual Review of Sociology, 23, 171-200. Phillips, J., & Sweeney, M. (2005). Premarital cohabitation and marital dissolution

among white, black and Mexican-American women. Journal of Marriage and Family, 67, 296-315.

Phillips, J., & Sweeney, M. (2006). Can differential exposure to risk factors explain recent racial and ethnic variation in marital disruption? Social Science Research, 35, 409-434.

Robinson, J., & Godbey, G. (1997). Time for life: The surprising ways Americans use their time. University Park: Pennsylvania State University Press.

Rogers, S. (1999). Wives’ income and marital quality: Are there reciprocal effects? Journal of Marriage and the Family, 61, 123-132.

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/

Teachman 1323

Rogers, S. (2004). Dollars, dependency, and divorce: Four perspectives on the role of wives’ income. Journal of Marriage and Family, 66, 59-74.

Rogers, S., & DeBoer, D. (2001). Changes in wives’ income: Effects on marital hap- piness, psychological well-being, and the risk of divorce. Journal of Marriage and Family, 63, 458-472.

Ross, C., Mirowsky, J., & Goldstein, K. (1990). The impact of the family on health: The decade in review. Journal of Marriage and the Family, 52, 1059-1078.

Ruggles, S. (1997). The rise of divorce and separation in the United States, 1880- 1990. Demography, 34, 455-456.

Sayer, L., & Bianchi, S. (2000). Women’s economic independence and the probability of divorce. Journal of Family Issues, 21, 906-943.

Schoen, R., Rogers, S., & Amato, P. (2006). Wives’ employment and spouses’ marital happiness: Assessing the direction of influence using longitudinal couple data. Journal of Family Issues, 27, 506-528.

Schoen, R., Astone, N., Rothert, K., Standish, N., & Kim, Y. (2002). Women’s employment, marital happiness, and divorce. Social Forces, 81, 643-662.

South, S. (2001). Time-dependent effects of wives’ employment on marital dissolu- tion. American Sociological Review, 66, 226-245.

South, S., & Lloyd, K. (2001). Changing partners: Toward a macrostructural-opportunity theory of marital dissolution. Journal of Marriage and Family, 63, 743-754.

Taylor, P., Tucker, M., & Mitchell-Kernan, C. (1999). Ethnic variations in percep- tions of men’s provider role. Psychology of Women Quarterly, 23, 759-779.

Teachman, J. (2002). Stability across cohorts in divorce risk factors. Demography, 39, 331-351.

Teachman, J. (2007). Race, military service, and marital timing: Evidence from the NLSY-79. Demography, 44, 389-404.

Voydanoff, P. (1990). Economic distress and family relations: A review of the 1980s. Journal of Marriage and the Family, 52, 1099-1115.

White, L., & Rogers, S. (2000). Economic circumstances and family outcomes: A review of the 1990s. Journal of Marriage and the Family, 62, 1035-1051.

Winkler, A., McBride, T., & Andrews, C. (2005). Wives who outearn their husbands: A transitory or persistent phenomenon for couples? Demography, 42, 523-535.

Winslow-Bowe, S. (2006). The persistence of wives’ income advantage. Journal of Marriage and Family, 68, 824-842.

at ARIZONA STATE UNIV on May 9, 2014jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/