Post-doctoral researcher
Sciences Po
What economic sacrifices are people willing to make to transmit their culture? Using data on religious affiliation in France, I study the intergenerational transmission of religion and how it interacts with children's educational outcomes. A reduced-form analysis suggests that mothers contribute to religious transmission more than fathers; religious minorities more than majorities; and lower-educated parents more than higher-educated ones. A mechanism that can explain these patterns is that higher-educated parents have a higher opportunity cost of transmitting their religion to their children. I investigate this mechanism through a structural model, in which parents endogenously decide their time investments in their child's culture on the one hand, and in their formal education on the other hand. The analysis suggests that heterogeneities in transmission patterns are driven primarily by heterogeneities in preferences for religious transmission across genders and religious groups, rather than by differences in parents' education. Furthermore, religious minorities pay a higher price for religious transmission in terms of their children's educational outcomes. For instance, by measuring this cost in terms of the probability that the child will obtain a college education, Muslim parents pay a cost between 8 and 13 times greater than that for Christians.
This paper provides the first empirical evidence on the economic costs of wearing the Islamic veil and on motives for veiling in a secular Western country. Using French observational data rather than small-scale interviews, we demonstrate a significant negative correlation between veiling and economic participation, even conditional on the respondent's religious environment. This newly-documented fact is not consistent with the existing economic theory of veiling in Muslim-majority countries, which has focused on the socio-religious signalling effect of veiling. We then show that a model which also accounts for reduced economic opportunities for veiled women is consistent with our findings in the Muslim-minority context. Using a structural interpretation of the model, we then disentangle the various motivations behind the joint decision to veil and to be economically active. Our findings indicate that veiled women are less economically active not due to religious preferences, but rather because veiling is costly on the labor market. Additionally, our results emphasize the significance of personal religious motives in the decision to veil, rather than signalling piety to others.
Matching problems with linearly transferable utility (LTU) generalize the well-studied transferable utility (TU) case by relaxing the assumption that utility is transferred one-for-one within matched pairs. We show that LTU matching problems can be reframed as nonzero-sum games between two players, thus generalizing a result from von Neumann. The underlying linear programming structure of TU matching problems, however, is lost when moving to LTU. These results draw a new bridge between non-TU matching problems and the theory of bimatrix games, with consequences notably regarding the computation of stable outcomes.
How do rent control policies affect equilibrium in housing markets and shape investment decisions? This paper delves into these questions by proposing a novel structural model to empirically study the influence of rent ceilings on the general equilibrium in housing markets, in which some landlords are able to freely adjust their rents (thus falling under the standard transferable utility assumption), while other landlords cannot (non-transferable utility).
We propose two fast algorithms to compute aggregate equilibrium outcomes in two-sided matching problems with non-transferable utility (NTU) and logit heterogeneity. Our first algorithm is an application of Jacobi's method. We show that this algorithm converges to the equilibrium, and also that the Jacobi iterates can be written in closed form, allowing for quick computation. Our second algorithm, which alternates between Jacobi steps and damped Newton steps, is guaranteed to converge in finite time. Specifically, if X and Y are the sets of individual types on each side of the market, then our Jacobi–Newton algorithm converges in |X||Y| steps at most.
In this paper, we view matching with nontransferable utility (NTU) as the limit of matching with imperfectly transferable utility (ITU) when utility is very costly to transfer. In the corresponding limit model, utility is not only nontransferable but also disposable. We show that stable outcomes in ITU models with very costly utility transfers are close to stable outcomes in this NTU model and, under a familiar regularity condition, that every stable outcome in the NTU model with disposable utility in which identical individuals are treated equally is close to a stable outcome in some ITU model with very costly utility transfers. We conclude that assuming utility to be nontransferable and disposable is an appropriate simplification to model situations in which transferring utility is difficult but not impossible.
This paper investigates the role of market segmentation in marital assortativeness, a feature traditionally attributed to variations in the surplus of potential matches within transferable utility models. I propose a modification to the Choo–Siow model, allowing individuals to be assigned to submarkets according to their gender and other relevant traits. This segmentation introduces a new explanation for spousal assortativeness, which is accompanied by a redistribution of surplus among partners compared to the original model. The significance of market segmentation is empirically examined by focusing on the termination of the mandatory military service in France in 1996, a quasi-natural experiment that arguably altered the structure of the marriage market. Preliminary event study analysis reveals an observable shift in educational homogamy post-termination, emphasizing the influence of market segmentation. Finally, I discuss the possibility of structurally estimating this extended model.
This note investigates population scale effects in several bipartite matching models commonly used in empirical applications. Such models are typically characterized by matching functions, which predict the number of matches of any pair of types based on the number of singles of each of these types. In models for which the matching functions are homogeneous of degree 1, the equilibrium matching distribution scales linearly with the population size. In the case of models with matching functions which are not homogeneous of degree 1, however, this simple scaling property does not hold in general. We thus analyze the asymptotic behavior of the equilibrium matching distribution for such models as the population size grows to infinity.