Projects for the Ph.D. Thesis:
Charitable Behavior and Public Intervention: a Survey Experiment (joint work with Francesca Leombroni)
In this paper we measure the extent of charitable behavior crowding out public intervention, and how this phenomenon affects the welfare of the poor. To achieve this objective we collect novel survey data on a representative sample of the U.S. adult population. In the survey, respondents are asked to go through several hypothetical scenarios which are built starting from a simple model of public good contribution, in order to learn about their preferences and expectations regarding donations and taxation. We find that when donations are available, government expenditure on the poor is lower in equilibrium. Yet, households in need are better off due to disproportionately higher donations. This means that, in our setting, private charity crowds out public intervention only to a limited extent, affecting equilibrium-level taxes only slightly. Moreover, as equilibrium tax rates without donations are not high enough to compensate for the lack of private charity, our results suggest that people are less driven by inequity aversion than by the direct utility of donating (warm glow). As an implication, we conclude that in the United States the widespread availability of private charity plays a pivotal role in alleviating poverty, which government intervention cannot substitute for due to the structure of voters' preferences.
I build a dynamic equilibrium model of household behavior with unobserved heterogeneity in the desired number of children to examine how policies targeting the housing market affect choices of fertility, location, and house size of young households. I estimate the model's structural parameters using data from Hungary to evaluate the dynamic effects of the Family Housing Allowance policy, which provided a sizeable lump-sum subsidy for house purchases, with built-in commitment regarding the number of children the family would have. The model suggests that the combination of lower interest rates and the allowance increases house prices substantially compared to the baseline, which for poorer households counteract some of the positive welfare effects of the policy. While according to the model, completed fertility increases due to the policy by around 5-10\% on average, mainly driven by poorer households, their housing conditions worsen in the long run due to the elevated house prices. Richer households experience no adverse effects of the policy, however, their completed fertility remains unaffected.
The Impact of Childcare on Maternal Employment (joint work with Judit Berei, Márton Csillag, Hanna Erős, Judit Krekó, Ágota Scharle, 2022)
This paper examines the effect of childcare availability on maternal employment in Hungary based on 2016 Microcensus data. We exploit the exogenous variation in access to childcare due to informal admission practices based on the date of birth, to identify the effect of childcare availability on maternal employment and the children’s enrolment. We find that on average, expanding the coverage of nurseries to the same level as kindergartens would lead to around 7.3 percentage points higher maternal employment, an around 25% higher employment rate compared to the baseline of mothers with a child aged 2-2.5 years. At the same time, the decomposition of the link between childcare availability and employment shows that enrolment would increase by 17.7 percentage points due to the higher coverage, close to 40% compared to the baseline. Enrolment in childcare would increase maternal employment probability by around 41 percentage points, around two-thirds of the employment rate of mothers. We also investigate the heterogeneities of the effect along demographic characteristics using causal forests, and the economic cycle by expanding the analysis to the 2011 Census. We find that in 2016 the childcare availability effect is higher for mothers with 3 children, living in villages, or municipalities without nurseries. The employment effect is lower in the 2011 Census, while the effect on enrolment in formal childcare remains similar, suggesting the importance of weaker labour demand in 2011.
Measuring the Influence of Coaches on NBA Players’ Development, with Spillovers (Data Science Term Project, 2017)
In this paper I construct a quantitative measure to evaluate the influence of NBA coaches on long-term player development, based on the players’ productivity measured in win shares. Using scraping methods I collect data of 17 seasons of NBA statistics (1999-2016) on all players, coaches and assistant coaches, recording their performance and occupation, to create a network dataset with two types of nodes: mentors and mentees. Mentees are young players who are potentially influenced by coaches, assistant coaches, and older players (the mentors). Controlling for their initial skill level I quantify their prime production and link it to the previous network of mentors. The mentors are allowed to have direct and indirect effects, the latter resulting from spillovers through their mentees becoming mentors later in their careers.