par Costanzo, Claudio 
Président du jury Gassner, Marjorie
Promoteur Gobbi, Paula Eugenia
Publication Non publié, 2026-04-23
Président du jury Gassner, Marjorie
Promoteur Gobbi, Paula Eugenia
Publication Non publié, 2026-04-23
Thèse de doctorat
| Résumé : | This dissertation investigates how structural changes in the labor market—specifically industrial automation, import competition, and artificial intelligence (AI)—transmit to the household, reshaping fundamental life-course decisions regarding fertility, time allocation, and family formation. While the literature has extensively documented the productivity and wage effects of these shocks, less is known about how they alter the internal economics of the family. Across three essays, this thesis combines causal identification strategies with structural economic modeling to demonstrate that households are active agents, adjusting their demographic and labor supply choices in response to shifting opportunity costs, wage structures, and income risks. The first chapter, Robots, Jobs, and Optimal Fertility Timing (published in the Journal of Population Economics), examines the tempo effect of industrial robots on fertility across European regions. Using an instrumented shift-share design, the study reveals that a statistically negligible average effect on age at first birth masks sharp heterogeneity by skill level. In regions with high robot exposure, low- and high-educated women accelerate childbearing, while women in mid-skill occupations delay it. To rationalize these asymmetric responses, the chapter develops a dynamic optimal-stopping model where childbirth is an irreversible investment subject to stochastic opportunity costs. Simulations demonstrate that automation-induced employment reallocation acts as a cost shift, driving these timing variations and reshaping childlessness rates without necessarily reducing total cohort fertility. The second chapter, Automation, Import Competition, and Intra-Household Labor Supply, turns to the internal division of labor within U.S. couples. Linking time-use diaries to exogenous shocks from robot adoption and Chinese import competition, the analysis finds that while these forces often raise women’s relative wages, households reallocate time in a manner consistent with breadwinner social norms. Rather than increasing market work, women in exposed households shift time toward childcare and leisure, while men increase their market hours. A calibrated semi-cooperative household model attributes these dynamics to a norm-based utility wedge that becomes salient near income parity. The findings suggest that identity-based frictions can amplify gender gaps in time use and reduce welfare, even when economic incentives favor greater gender equality. The third chapter, AI, Risk Sharing, and the Marriage Market, pivots to the current wave of technological change, asking how AI diffusion affects marriage and divorce. Unlike earlier waves that eroded men’s comparative advantage in manual labor, AI targets cognitive tasks performed by both genders, thereby increasing idiosyncratic earnings risk for high-skilled workers. Using a novel Bartik index of AI exposure, the study finds that in AI-exposed labor markets, marriage rates rise and the divorce-to-marriage ratio falls. These results support an insurance view of the family: as labor market volatility increases, the value of marital risk pooling rises, strengthening family formation. This contrasts sharply with previous technology waves (ICT, robots), suggesting that AI may reverse long-standing trends in marital decline by reinforcing the economic gains from partnership. |
