Working papers
"Can you do the dishes? Intra-household time use, division of labor and fertility"
Across the world, women do more housework (including childcare) than men, which can negatively affect both fertility and women’s career, and could be a reason why in many countries fertility and female employment are now positively correlated. I argue that rebalancing the housework load increases female employment and men’s opportunity cost of children, so it increases demand to outsource childcare which, if met, can also raise fertility. Using detailed data from Italy, I estimate a life-cycle model of housework, employment and fertility, with limited nursery school availability. To rationalize gender differences, men’s weight in housework production should be 30% as women’s, even accounting for labor market gaps, perceived housework skills, and preferences. Housework imbalance has a similar effect on employment to an additional 50% gender wage gap, and accounts for 70% of the child penalty. Closing all gender gaps and providing full nursery school coverage increases fertility to 2.25 children per woman and female employment to 0.96 (two thirds of the employment effect driven by housework). The model generates negative correlation between fertility and female employment in Southern Italy, and positive correlation in the North. It also replicates historical trends in both areas from 1980 onwards.
"Committing While We Still Can: Marital Regimes, Labor Supply and Household Intertemporal Behavior"
(with Alessandra Voena)
[Previously circulated as: "Prenuptial Contracts, Labor Supply and Household Investments" by Denrick Bayot and Alessandra Voena.]
This paper studies how couples use premarital agreements to commit to how marital assets will be divided in case of divorce. We examine a setting in which the cost of such agreements is minimal: by marking their choice on the marriage license application, Italian couples can choose at no financial cost between community property, in which assets are split equally between husband and wife, and separation of property, in which each spouse retains the assets in their name. Using multiple administrative and survey datasets, we document that a sizable minority of Italian couples chooses community property, and that their selection is systematic: couples in which the wife undertakes substantial household-specific investments are especially likely to elect community property, and the take-up of the regime declines as female labor market attachment increases. We show that these patterns are not consistent with a standard limited-commitment model of household behavior with divorce as the only outside option to marital cooperation. Instead, we develop a dynamic model of marriage, female labor supply, savings, and divorce in which spouses can engage in non-cooperative behavior prior to divorce, and pre-commit to an equal split of assets to promote the wife's household-specic investments. The estimated model implies that, as female labor force participation rises and the gender wage gap narrows, the value of this commitment falls. Lowering the cost of prenuptial contracting generates substantial welfare gains for couples and raises the value of marriage.
"The Distribution of Value of Time: An Analysis from Traffic Congestion and Express Lanes”
(Honorable mention, 2021 UEA European Meeting)
The value of time (VOT) determines the allocation of non-labor time to tasks and is crucial in travel demand and infrastructure, where congestion is a major source of loss. A large literature has estimated the mean VOT over a variety of subpopulations, but commuting choices and welfare effects of congestion policies depend on the individual VOT of the policy-relevant population. In this paper, I estimate the full VOT distribution for a population of drivers, using a rich new dataset in the unique context of highway Express Lanes (ELs), which offer time savings in exchange for a toll. A continuous function of traffic density sets the toll and rounds it to the nearest $0.25, creating 32 separate discontinuities which provide identifying variation. The analysis is divided in three parts. First, using an RDD, I show that EL drivers have a mean VOT of $66.56 per hour saved, substantially exceeding estimates from the literature. Second, the full VOT distribution for all drivers, which rationalizes EL aggregate traffic shares and RD results, shows wide heterogeneity: the median is $17.42 per hour and the 95th percentile is $166.05. Third, I build a structural model that endogenizes departure time (a key form of adjustment) to assess the welfare consequences of a range of counterfactual policies. I find that the EL is welfare-reducing because the value of the increase in travel times for non-users outweighs the benefits for users by $25.68 per year, more than what half of drivers spend on the EL in a year.
"The Dynamics of the Earnings Gap between Spouses in the US and Europe"
When husbands in heterosexual couples prefer to be the primary earners, women's outcomes both during and before marriage can be negatively affected. Some papers claimed that the distribution of the wife's share of household earnings has missing mass to the right of 0.5 and attributed this regularity to a “male breadwinner” norm. If this norm exists, economic theory predicts that spouses would respond to violations of the norm ex-post, or prevent ex-ante violations. These responses do not emerge from the static distribution of the earnings gap between spouses, but from its dynamics. To test the theory, I provide the first dynamic analysis of the earnings gap, characterize its Markov transition matrix and the persistence of its shocks. The methodology is easily replicable, because it only requires a two-year panel of both spouses’ earnings. The gap converges to a stationary distribution that reproduces extremely well the static distribution of the wife's share of earnings. Shocks to the earnings gap are more persistent when they are close to the quantile of the initial gap. Finally, the local transition of the earnings gap around 0 is driven by couples bunched exactly at 0 earnings gap. The behavior of spouses with similar earnings is not consistent with them holding a male breadwinner norm. When the wife earns slightly more than the husband, in the following period it is more likely that the husband earns more without any spouses' response. All the results hold true across the US and 20 European countries.
Work in progress:
"Women's Control over Reproductive Life, Education and Career”
(runner-up, NBER 2022 SI Gender in the Economy)
I study the link between women’s control over reproductive life and their education and career choices. In particular, I consider policies like the Affordable Care Act (ACA, which mandated no-cost coverage of contraception), access to reproductive healthcare and access to abortion clinics. In preliminary reduced-from results I show that the ACA significantly reduced average out-of-pocket cost of prescribed contraception but it had little to no effect on demand for contraception and for more effective methods, such as the IUD. Women currently in school are significantly more likely to use contraception but I find no differential effect post-ACA. These results might imply, for example, that women’s contraception choice would react to price changes but is constrained on other margins like access to care, or that women’s long-term family planning choice is not significantly affected by price. I am currently assembling a dataset of access to reproductive care, with the aim of using it together with the ACA price variation to provide the first estimate of willingness to pay for reducing distance to contraception and abortion providers. I complement the reduced-form with a structural life-cycle model contraception method choice, education and labor supply. In the model, contraception provides insurance against pregnancy risk, which is salient during long-term human capital investments. Each method provides a different insurance level, and women’s method choice is endogenized. Expectations of low future expectations decrease the value of averting pregnancy, which endogenizes the event of a teenage pregnancy. These features of the model are novel.