생애 이행 관점에서 본 여성의 가족형성 과정: 졸업·초혼·출산 전이의 분석 = Life-Course Transitions in Family Formation: Education, Marriage, and Childbearing

다국어 초록 (Multilingual Abstract)

This dissertation examines the timing of marriage and childbearing among women in South Korea from a life-course perspective. Rather than treating low fertility simply as a decline in the total number of births, this study conceptualizes it as a timing problem embedded in a marriage-centered sequence of family formation. In South Korea, where childbearing remains strongly linked to marriage, the timing of first birth cannot be fully understood without considering the timing of entry into first marriage. Likewise, the transition to a second birth depends not only on current household conditions but also on the timing of prior transitions. This dissertation therefore analyzes the linked sequence from graduation to first marriage, from first marriage to first birth, and from first birth to second birth.
Using the Korean Longitudinal Survey of Women and Families （KLoWF）, the original person-wave data are reorganized into person-period data suitable for discrete-time event-history analysis. The empirical analysis proceeds in two stages. First, a sequential stacked discrete-time multi-state hazard model is estimated to examine whether delays in prior transitions remain associated with subsequent transition hazards. This model combines the three transitions into a single sequential risk-set structure, allowing only women who have experienced the preceding event to enter the risk set for the next transition. Second, transition-specific discrete-time hazard models are estimated to identify how the correlates of marriage and childbearing timing differ across stages of the family formation process. The models use a complementary log-log specification and standard errors clustered at the individual level.
The empirical findings are as follows. First, delayed family formation begins most clearly at the transition into first marriage. The interval from graduation to first marriage is an important starting point of delayed family formation because, in a marriage-centered fertility regime, postponement of first marriage shifts the timing of potential first birth. However, the results do not support a strong general cumulative-delay interpretation. The stacked model shows that a longer delay from graduation to first marriage is positively associated with the hazard of first birth after marriage. This pattern suggests a catch-up process: women who enter first marriage later tend to move more quickly into first birth once they are married. At the same time, the delay from graduation to first marriage does not show a statistically significant association with the transition from first birth to second birth.
Second, the delay from first marriage to first birth is negatively associated with the hazard of second birth. This is the clearest evidence of selective persistence in the family formation process. The findings indicate that delays in prior transitions do not accumulate uniformly across all subsequent stages. Instead, only the delay between first marriage and first birth persists into the next stage by lowering the hazard of second birth. Thus, the results support a stage-dependent pattern of delay propagation rather than a general cumulative-delay mechanism.
Third, the transition-specific models show that the correlates of family formation differ by stage. In the graduation-to-first-marriage transition, duration structure, age, education, and conditions related to adult economic establishment are relatively important. In the transition from first marriage to first birth, household time resources and care resources become more salient. In particular, grandparental care is positively associated with the first-birth transition, whereas the spouse’s working hours are negatively associated with transitions to both first and second births. In the first-birth-to-second-birth transition, household time allocation and the spouse’s labor supply conditions become more directly relevant to the possibility of additional childbearing.
This dissertation contributes to the literature by reframing South Korea’s low fertility not merely as a decline in fertility levels but as a sequence of delayed and selectively constrained life-course transitions. It connects previously separate research on first marriage, first birth, and second birth within a single linked-transition framework. The findings suggest that policy discussions should distinguish between the starting point of delay and the stage at which delay becomes a persistent bottleneck. While delayed entry into first marriage is an important starting point of delayed family formation, the more persistent constraint appears after marriage, in the timing of first birth and its implications for second birth. Therefore, fertility-related policy discussions should consider stage-specific conditions: employment and housing stability before marriage, work–care compatibility around first birth, and household time resources and employment continuity after the first birth.
The results should be interpreted as conditional associations rather than causal effects. Employment, housing, income, care, and household labor allocation may be jointly determined with marriage and childbearing plans, and unobserved preferences or family background may affect multiple transitions simultaneously. Nevertheless, by reconstructing family formation as a sequence of linked transitions, this dissertation provides an empirical framework for understanding where delays begin, where they persist, and how different forms of constraint operate across stages of women’s life courses in South Korea.

목차 (Table of Contents)

• 제1장 서 론 1
• 제1절 연구의 배경 및 의의 1
• 제2절 연구의 범위 및 방법 4
• 제3절 연구질문 9
• 제4절 논문의 구성 12
• 제2장 선행연구 13
• 제1절 저출산과 혼인·출산 지연에 관한 이론적 논의 14
• 제2절 한국의 혼인·출산 관련 실증연구: 단일 이벤트 중심 접근 24
• 제3절 경로 중심 접근의 확대: 연쇄（sequence） · 궤적（trajectory） · 다상태（multistate） 문헌 29
• 제4절 선행연구의 공백과 본 연구의 위치 33
• 제3장 이론적 배경 39
• 제1절 한국의 생애 이행 동태분석과 현황 39
• 제2절 해외현황 57
• 제4장 분석자료 및 연구방법 75
• 제1절 분석자료와 생애전이 시점의 구성 75
• 제2절 Person-period 자료의 구축과 사건 정의 및 시간축 80
• 제3절 설명변수의 구성 및 사전 설정된 모형 명세 88
• 제4절 추정모형 94
• 제5절 표본 구성과 모형 진단 98
• 제6절 분석의 순서와 해석 전략 및 한계 112
• 제5장 실증분석 결과 115
• 제1절 적층 이산시간 다상태 해저드 모형（stacked discrete-time multi-state hazard model） 115
• 제2절 전이별 이산시간 해저드 모형 129
• 제6장 결론 및 정책적 시사점 154
• 제1절 핵심 발견의 요약 154
• 제2절 이론적 해석 156
• 제3절 정책적 시사점 164
• 제4절 연구의 한계와 향후 연구 168
• 제5절 소결 169
• 부록 171
• 참고문헌 182
• ABSTRACT 197