Peer-reviewed publications

. Do Anti-Poverty Programs Sway Voters? Experimental Evidence from Uganda. forthcoming in Review of Economics & Statistics, 2018.

PDF

Working papers

What determines whether some asylum seekers are granted refugee status while others are rejected? I draw upon archival records from a representative sample of 4,000 asylum applications filed in France between 1976 and 2016 to provide new evidence on the determinants of asylum decisions. Comparing accepted applicants with those who were rejected, I find that Muslim applicants are 30 percent less likely to be granted asylum than otherwise similar Christian applicants. In addition, linking archival records to detailed administrative data, I show that bureaucrats at the French asylum office initially discriminate against Muslims but stop after about a year on the job. Assessing potential mechanisms of discrimination, I do not find support for the claim that discrimination is driven by bureaucrats’ preferences or ideology. Instead, I argue that discrimination occurs because bureaucrats underestimate the probability that Muslims are persecuted. This novel finding has direct implications for strategies to curtail discrimination in courtrooms and administrations. Paper available upon request.
2018

Ongoing projects

We analyze barriers to refugee integration in France using original survey data. Leveraging access to French asylum office’s administrative database, we construct a representative sample of refugees who were granted refugee protection between 1989 and 2015. After encouraging results from a pilot survey that we completed in April 2018, we will launch in September 2018 fall the first large-scale survey of refugees ever undertaken in France. This project will examine three different aspects of the integration process of refugees (waiting time, refugee housing and welcoming contracts) using causal research designs like regression discontinuity and difference in differences. For example, previous work has shown that the time refugees have to wait to receive refugee status, which varies from a few weeks to several months, can have detrimental effect on integration prospects. Preliminary results using French administrative data reveal that an increase in this waiting period significantly reduces naturalization rates. To explain this pattern, the survey will explore two potential mechanisms: Refugees could either be less likely to apply for naturalization, or they could be less likely to meet the citizenship requirements as a result of the additional waiting time. As refugees continue to arrive in Europe, this work will provide a better understanding of policies that help refugees integrate.
2018

Successful integration of immigrants has become one of the principal challenges facing Europe, North America, and Australia, where the foreign-born population averages 10 to 20 percent of the total population. In this paper, we analyze the extent to which immigrants integrate successfully in France using an extraordinary dataset – Trajectoires et Origines (2008) and a newly developed multidimensional index of integration (Harder et al. 2018). While our results confirm that immigration integration in France has been a story of success, the data reveal an integration gap between Christian and Muslim immigrants that is exacerbated over generations. Over time and across generations, immigrants to France express values on a multidimensional integration scale similar to those of the native population. But Christian immigrants and their descendants do so faster than Muslims. We explore the explanatory power of alternative mechanisms proposed in the literature ranging from the new opportunities available to Muslims for connection to an international jihadist enterprise (Tournier 2013) to studies that point to various forms of economic, social and political discrimination faced by Muslims leading to a discriminatory equilibrium (Adida et al. 2016; Dancygier 2017).
2018

Teaching

Political Methodology III - Stanford University
Model-based inference in political science. Topics covered will include likelihood-based inference, generalized linear models, discrete choice models, regularization, missing data, latent-variable models, and (a brief introduction to) optimization approaches.

Instructor: Jens Hainmueller
Quarter: Spring 2018

Political Methodology II - Stanford University
Causal inference, i.e. methods designed to address research questions that concern the impact of some potential cause (e.g. an intervention, a change in institutions,economic conditions, government policies) on some outcome (e.g. vote choice, income, election results, levels of violence). Topics include experiments, matching, regression, panel methods, difference-in-differences, synthetic control methods, instrumental variable estimation,regression discontinuity designs, quantile regressions, and sensitivity analysis.

Instructor: Jens Hainmueller
Quarter: Winter 2018

Political Economy - Sciences Po Summer school
The course will introduce advanced undergraduate students to the field of political economy and will familiarize them with a wide range of empirical methods used in this literature. The course will be organized around a series of topics(including voting, political agency and accountability,special interests politics,conflict and violence, the origins and impact of political institutions), and will be primarily based on the reading of research papers and the in-depth discussion of their theories, methods, and results

Instructor: Ruben Durante
Quarter: Summer 2017

Political Methodology III - Stanford University
Topics include maximum likelihood estimation, and then turn tomodels for discrete responses: binary (e.g., turning out to vote) ordinal (responses to Likertitems on surveys, rating schemes) and unordered or multinomial outcomes (e.g., voting in amulti-party system). We will also briefly look at models for counts (e.g., number of terroristattacks).

Instructor: Simon Jackman
Quarter: Spring 2015

Elementary Algebra - San Quentin State Prison https://prisonuniversityproject.org/
Quarter: Spring 2015

Sampled students’ evaluations

“Mathilde is really good at breaking concepts down into intuitive pieces that are easier to understand. We would often speed through topics in class and Mathilde would slow down in section and make sure we really understood what was going on.”

“Very approachable and easy to ask question to. Mathilde walked through concepts and processes very well and I did not feel like I couldn’t ask her my ‘dumb’ questions. She knows how to make students feel welcomed and comfortable in the classroom.”

“Mathilde is also really funny and nice and approachable. I never felt scared to come ask her questions, even if I knew they were dumb, and she never made me feel dumb for not understanding something.”

“Spent a lot of time preparing materials. I could tell Mathilde really cared because of the amount of work she put into making section slides and code.”

“Mathilde is really great walking through outside examples in section so that we apply what we’re learning to actual data, which helps us understand the topic a lot better.”

“Mathilde is extremely helpful, friendly and funny. Her slides are really great and I can tell that she is constantly trying to make the more effective to our learning. She is also extremely open to feedback and suggestions and always asks what material will help us more. Improvements: nothing really here.”

“Mathilde is super helpful in section and she does a really good job of breaking concepts down into component parts on her slides. This is especially useful this quarter because Jens sometimes goes through slides quickly in class, and Mathilde’s sections have been really good at helping me understand what I didn’t originally get in class. Also her slides are beautiful and I greatly appreciate how much time she spends preparing them.”