Abstract:
This paper examines the role of remote work on child penalties in four Latin American countries, using a pseudo-event study around childbirth. At the extensive margin, work-from-home (WFH) may help mothers remain in the labor force. At the intensive margin, it can either help maintain or increase work hours by reducing commute time, or lead women to re-optimize and reduce hours (and earnings) to balance childcare demands. I find suggestive evidence of the latter, as mothers who work remotely after childbirth tend to reduce their working hours more than those who do not. Child penalties are smaller in households where partners can WFH, with mothers more likely to WFH after childbirth, especially if their partners cannot. Finally, I find suggestive evidence that child penalties in employment decreased more in regions where remote work increased.
Abstract:
In the context of a general sample selection model with a continuous, randomized treatment variable, we introduce a trimming procedure that provides sharp bounds on a “local average response” treatment effect parameter. The procedure can be understood as a generalization of that in Lee (2009) to the case of a continuous, rather than binary, treatment. We illustrate our procedure by considering the problem of estimating the effect of wealth on labor supply, presenting the bounds in a simulation loosely calibrated to the results of Cesarini et al. (2017), an empirical study which uses Swedish lottery winnings to estimate these wealth effects.