Early Fertility Desires: Hidden Drivers of Union Formation?

While many studies have investigated how fertility preferences affect childbearing behavior, not much is known yet on whether and how early fertility desires may impact union formation over the subsequent life course. However, the formation of co-residential unions is an important precursor for the birth of children; the large majority of children are still born to co- residential couples today despite sweeping changes in rates and timing of marriage over the last decades. In the literature, it has been argued that union formation timing and the transition to parenthood may be both driven by the same individual preference structures, but little direct evidence testing this hypothesis exists. Drawing on data from the NLSY79 and focusing on the US context, this project aims at closing these research gaps in three steps. First, we will investigate how fertility desires early in life impact the occurrence and timing of first co-residential union formation/marriage later in life. We will look at this process for both, men and women, and also address how it may vary by education and parental background. Next, we will explore a larger array of early desires and expectations for life trajectories and look at how they relate to subsequent family formation behaviors. In the second step we will use measurements on desired childbearing, union-formation, and educational- and occupational life outcomes and examine whether separable clusters of envisioned life trajectories are present in this sample. Then, third, we will investigate how membership in these clusters a) may be linked to the formation of unions, and b) whether it is indeed a common driver of both, first-co-residential union timing and the transition to parenthood. The project will use a variety of methods, ranging from more basic methods for analyzing longitudinal data to cutting-edge structural equation methodology for longitudinal data and modeling more than one outcome at the same time. We hypothesize that early childbearing desires have a significant impact on union formation behaviors specifically and family formation (including childbearing) more generally. We also expect that this linkage between preferences and outcomes may differ between men and women. This project will advance our knowledge on peoples family formation process by uncovering how preferences early in life may drive these later life outcomes. We will shed light on the question of whether the desire for children may act as a hidden driving force behind whether and when men and women commit to co-residential unions and marriage. Also, the project will address all questions for men and women separately, thereby contributing to our still spotty knowledge on mens life courses and how drivers of family formation may affect men and women differently.

The project has produced new insights into patterns of family formation in the US in the 20th century and some of the factors which are driving them. Three distinct findings arose. First, and most importantly, our results showed that early life fertility desires, measured during late adolescence, predict family formation pathways across the life course in complex ways among US individuals from the National Longitudinal Study of Youth 1979 born 1960-1964. While it was known that fertility desires play a role for whether, when and how many children people will have, we uncovered that they are also associated with people's marital behavior. Specifically, US men and women who desired three or more children married later than individuals desiring two or fewer children. Nonetheless, around age 40, men desiring large families had caught up with those desiring fewer children, meaning that similar proportions of men were married regardless of early life fertility desires. However, for college educated women desiring large families, delaying the first marriage translated into remaining unmarried later in life more often. At age 40, their probability of ever having been married remained about 10% lower compared with that of college educated women who had desired two or fewer children. Second, we found that this group of college educated women, who had desired to have three or more children in life, had significantly lower chances of ever becoming a mother, compared to lower educated women who desired a large family. We also found that their later entry into marriage, which we had uncovered in the first project paper, was not the responsible factor for their increased probability to remain childless. About half of eventually college educated women of this birth cohort had desired three or more children during late adolescence, so this finding pertains to a sizeable group of women. Future research is needed to illuminate the reasons for the lowered marriage and motherhood chances they experienced. Third, we examined the link between the postponement of parenthood and fertility outcomes among tertiary educated women in the USA born in 1920-1986, using data from the CPS June Supplement. Women with postgraduate education differed from women with college education in terms of first birth timing, childlessness, and completed fertility. While the postponement trend, which began with the cohorts born in the 1940s, has continued among highly educated women, its associations with childlessness and completed parity have changed considerably over subsequent cohorts. Tertiary educated women born in the 1950s and early 1960s experienced record childlessness levels, underscoring that our findings on fertility desires and family formation need to be interpreted in light of unique circumstances these birth cohorts faced in the domains of education, employment, and family.