Publications
"The Value Added of Machine Learning to Causal Inference: Evidence from Revisited Studies" with Andrea A. Naghi
Forthcoming at The Econometrics Journal
A new and rapidly growing econometric literature is making advances in the problem of using machine learning methods for causal inference questions. Yet, the empirical economics literature has not started to fully exploit the strengths of these modern methods. We revisit influential empirical studies with causal machine learning methods aiming to connect the econometric theory on these methods with empirical economics. We focus on the double machine learning, causal forest and generic machine learning methods, in the context of both average and heterogeneous treatment effects. We illustrate the implementation of these methods in a variety of settings and highlight the relevance and value added relative to traditional methods used in the original studies
"The Long-Run Effect of Migration on Firms' Trade: Evidence from China" with Christina Ammon
Forthcoming at Economic Development and Cultural Change
We empirically assess the long-term effect of migration on the exporting behavior of firms in the migrants’ origin country. We focus on ethnic migrant networks formed as a result of a mass migration wave of ethnic Cantonese people from the province of Guangdong in Southern China to the United States in the late 19th century. Using firm-level data for 2004 for Guangdong, we show that exposure to the Cantonese ethnic network has a positive effect on firms’ exports after several decades, both at the intensive and extensive margin. We also find that networks have a positive impact on other firm outcomes, such as employment, wages, high-skilled workers, domestic capital, and fixed assets.
"The Effect of Plough Agriculture on Gender Roles: A Machine Learning Approach" with Andrea A. Naghi
Accepted at the Journal of Applied Econometrics
This paper undertakes a replication in a wide sense of Alesina et al. (2013), which examines the relationship between historical plough agriculture and current gender roles. We revisit the main research question with recently developed causal machine learning methods, which allow to model the relationship of covariates with the treatment and the outcomes in a more flexible way, while also including interactions and nonlinearities that were not considered in the original analysis. Our results suggest an even larger negative effect of the historical plough adoption on female labor force participation than what the original analysis found. The paper highlights the benefits of using causal machine learning methods in applied empirical economics.
Working Papers
"Banning Women from STEM: Evidence from Iran" with Laura Hering and Julian Emami Namini (submitted)
(Older version: TI working paper 21-073/V)
We study the effect of an Iranian educational policy implemented in 2012 that restricted access to higher education for women in 30% of Iran’s public universities, mostly in the prestigious and popular field of engineering. We analyze the impact of this policy on higher education, the labor market and the marriage market, by exploiting differences in exposure to the policy across gender, cohorts and regions. We find that the resulting unexpected reduction in programs had a negative impact on university education for women, in particular for those in urban areas. We further show that the policy had a mixed impact on the labor market, with positive effects for women without university education and negative effects for women with university education. Finally, we find that the policy also had an impact on the marriage market, as it decreases the probability of young women to marry.
"The Persistent Effect of Gender Division of Labour: African American Women After Slavery"
This paper explores the role of historical gender division of labour in shaping gender norms. To answer this question, I analyse whether differences in the gender division of labour during slavery have a persistent effect on African American women’s labour market outcomes after the end of slavery. I use variation in the production of cotton and tobacco across counties during slavery as a proxy for gender division of labour: tobacco was characterized by a starker gender division of labour compared to cotton. Using data from 1870 to 2010, I show that women living in counties with lower degrees of gender division of labour (higher cotton production relative to tobacco) are more likely to participate in the labour market and have higher occupation income scores, for at least 70 years after emancipation. To disentangle gender roles from local labour demand effects, I analyse the labour force participation of migrants from counties with high historic cotton and tobacco production who relocated to urban areas. Furthermore, I explore intergenerational transmission mechanisms from mothers to daughters.
"The Value Added of Machine Learning to Causal Inference: Evidence from Revisited Studies" with Andrea A. Naghi
Forthcoming at The Econometrics Journal
A new and rapidly growing econometric literature is making advances in the problem of using machine learning methods for causal inference questions. Yet, the empirical economics literature has not started to fully exploit the strengths of these modern methods. We revisit influential empirical studies with causal machine learning methods aiming to connect the econometric theory on these methods with empirical economics. We focus on the double machine learning, causal forest and generic machine learning methods, in the context of both average and heterogeneous treatment effects. We illustrate the implementation of these methods in a variety of settings and highlight the relevance and value added relative to traditional methods used in the original studies
"The Long-Run Effect of Migration on Firms' Trade: Evidence from China" with Christina Ammon
Forthcoming at Economic Development and Cultural Change
We empirically assess the long-term effect of migration on the exporting behavior of firms in the migrants’ origin country. We focus on ethnic migrant networks formed as a result of a mass migration wave of ethnic Cantonese people from the province of Guangdong in Southern China to the United States in the late 19th century. Using firm-level data for 2004 for Guangdong, we show that exposure to the Cantonese ethnic network has a positive effect on firms’ exports after several decades, both at the intensive and extensive margin. We also find that networks have a positive impact on other firm outcomes, such as employment, wages, high-skilled workers, domestic capital, and fixed assets.
"The Effect of Plough Agriculture on Gender Roles: A Machine Learning Approach" with Andrea A. Naghi
Accepted at the Journal of Applied Econometrics
This paper undertakes a replication in a wide sense of Alesina et al. (2013), which examines the relationship between historical plough agriculture and current gender roles. We revisit the main research question with recently developed causal machine learning methods, which allow to model the relationship of covariates with the treatment and the outcomes in a more flexible way, while also including interactions and nonlinearities that were not considered in the original analysis. Our results suggest an even larger negative effect of the historical plough adoption on female labor force participation than what the original analysis found. The paper highlights the benefits of using causal machine learning methods in applied empirical economics.
Working Papers
"Banning Women from STEM: Evidence from Iran" with Laura Hering and Julian Emami Namini (submitted)
(Older version: TI working paper 21-073/V)
We study the effect of an Iranian educational policy implemented in 2012 that restricted access to higher education for women in 30% of Iran’s public universities, mostly in the prestigious and popular field of engineering. We analyze the impact of this policy on higher education, the labor market and the marriage market, by exploiting differences in exposure to the policy across gender, cohorts and regions. We find that the resulting unexpected reduction in programs had a negative impact on university education for women, in particular for those in urban areas. We further show that the policy had a mixed impact on the labor market, with positive effects for women without university education and negative effects for women with university education. Finally, we find that the policy also had an impact on the marriage market, as it decreases the probability of young women to marry.
"The Persistent Effect of Gender Division of Labour: African American Women After Slavery"
This paper explores the role of historical gender division of labour in shaping gender norms. To answer this question, I analyse whether differences in the gender division of labour during slavery have a persistent effect on African American women’s labour market outcomes after the end of slavery. I use variation in the production of cotton and tobacco across counties during slavery as a proxy for gender division of labour: tobacco was characterized by a starker gender division of labour compared to cotton. Using data from 1870 to 2010, I show that women living in counties with lower degrees of gender division of labour (higher cotton production relative to tobacco) are more likely to participate in the labour market and have higher occupation income scores, for at least 70 years after emancipation. To disentangle gender roles from local labour demand effects, I analyse the labour force participation of migrants from counties with high historic cotton and tobacco production who relocated to urban areas. Furthermore, I explore intergenerational transmission mechanisms from mothers to daughters.