In the last several decades, a shift has occurred in the relative contributions of married women to household earnings. Yet we know little about the impact of relative earnings of husbands and wives on the likelihood of marital disruption. This study estimates a discrete-time hazard model using data on first married couples from the 19861989 waves of the Panel Study of Income Dynamics. The relative earnings of husbands and wives are a significant predictor of marital disruption, although the relationship is nonlinear. We suggest this nonlinear effect is linked to the varying economic circumstances of different groups of couples.
Key Words: divorce, marital disruption, Panel Study of Income Dynamics, relative earnings.
Broad changes in the United States economy and society have produced shifts in the relative contributions of husbands and wives to family income. The transformation from a manufacturing to a service-based economy has resulted in a decline in men's participation in the labor force and a continuing increase in women's participation in the labor force through the 1980s and into the 1990s (Menaghan & Parcel, 1990). A majority of families now rely on more than one earner to maintain their standard of living.
Married women are more likely to be yearround, full-time workers than in the past. Between 1960 and 1988, Wilkie (1991) found a marked decrease in the percentage of married couples in which the male householder was the sole earner. Additionally, Wilkie found a marked increase in the percentage of married couples in families with multiple earners and in families in which someone other than the male householder was the sole earner. Nowak and Snyder (1994) found that the mean percentage of combined couple labor income earned by prime-aged husbands (25-55 years) declined substantially between 1968 and 1988. Further, using a nationally representative sample, they found that the wife was the primary earner in fully 26% of the couples during at least one year in the 1983- 1988 time period. Among prime-aged men, the "traditional" household in which husbands consistently earn more than their wives still predominates. Nevertheless, the percentage of households in which women are continually the primary earner tripled between 1968 and 1988. Consequently, the sheer percentage of couples in which husbands no longer earn all or most of the couple's labor income makes the relative earnings of husbands and wives an important phenomenon to examine.
In spite of these shifts, we know little about what consequences the relative earnings of husbands and wives have on marriages, including their effect on the likelihood of marital disruption. Previous research has focused primarily on wives' participation in the labor force, their occupational characteristics, or absolute earnings, with equivocal results. (See Greenstein, 1990, and White, 1990.) The few studies that have included measures of relative earnings use older data and often operationalize relative economic power differently. Additionally, they examine outcomes other than marital disruption, outcomes such as thoughts of divorce, commitment, or the division of household labor. In this article, we analyze the impact that the relative earnings of husbands and wives in a given year has on the likelihood of marital disruption. We use data from the 1986-1989 waves of the Panel Study of Income Dynamics (PSID) for prime-aged couples in first marriages. We argue that, theoretically, relative earnings of husbands and wives should be an important predictor of marital disruption. We examine four types of couples with regard to earnings: traditional couples in which husbands earn 75%-100% of family income, new traditional couples in which husbands earn 50%-75% of the family income, nontraditional couples in which wives earn 50%-75% of the family income, and reverse traditional couples in which wives earn 75%-100% of the family income. Our preliminary analysis shows that relative earnings are a significant predictor of marital disruption, although the relationship is more complex than previously surmised. …