*In this paper, we examine the effect of women’s empowerment on husband’s split migration (migration without his wife) in Indonesia. We conduct a difference-in-differences strategy exploiting a national policy experiment that exogenously fostered access to justice for women, as well as ethnical variation in traditional post-marriage residence norm. We first document that married women originating from traditionally matrilocal ethnic groups experienced a relatively stronger empowerment following the reforms than women originating from traditionally patrilocal ethnic groups. We also find that matrilocal husband’s propensity to split migrate relatively increased following the reforms by 2-2.7 percentage points (an increase of 56-69%). This result suggests that women’s empowerment fosters husband’s split migration. We also point to an additional mechanism whereby husband’s split migration is facilitated by spousal monitoring and communication. Finally, we find that husband split migrants are more likely to (1) originally live in a village having experienced a natural disaster recently, (2) originate from a liquidity constraint household and (3) be less educated. We therefore suggest that the relative increase in husband’s split migration was mainly driven by ex-post risk coping motives, rather than (ex-ante risk diversification) investment motives.

**Job polarization refers to the disproportionate increase in employment in top and bottom-paying occupations, at the expense of middling-paying jobs, that most Western economies have witnessed since 1990. At the same time, a large increase in female labour market participation took place in the same economies, raising the question of whether female employment may explain part of the observed polarization. This paper uses German data to investigate the role of high-skilled females in the development of a market for home production substitutes. To provide causal evidence on the demand for services, I exploit exogenous variation in top paying-occupation labour market incentives, captured by the dispersion of hourly wages at the occupational level. I find that when top-employed females work more hours, low-skilled women are more likely to be employed. These results highlight the role of female employment in explaining job polarization.