The effect of plough agriculture on gender roles - A machine learning approach

Aug 3, 2026

Upload by Maeve

This replication study revisits research linking historical plough agriculture to contemporary gender roles using recently developed causal machine learning methods. By applying double machine learning, the analysis models relationships between covariates, treatment, and outcomes more flexibly, incorporating interactions and nonlinearities not addressed in the original study. Results indicate a more pronounced negative effect of historical plough adoption on women’s labor force participation. The paper also emphasizes the practical value of causal machine learning in applied empirical economics.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Introduction

Historical origins of gender norms

Replication strategy with causal machine learning

Empirical approach: double machine learning

Machine learning nuisance estimation

Received: 9 September 2023 Revised: 20 April 2024 Accepted: 21 April 2024 DOI: 10.1002/jae.3083 REPLICATION The effect of plough agriculture on gender roles: A machine learning approach Anna Baiardi1 Andrea A. Naghi2 1 Erasmus School of Economics, Erasmus University and Tinbergen Institute, Rotterdam, Netherlands 2 Department of Business Analytics and Applied Economics, Queen Mary University of London, London, UK Correspondence Andrea A. Naghi, Queen Mary University of London, London, UK. Email: [a.naghi@qmul.ac.uk](a.naghi@qmul.ac.uk) Summary This paper undertakes a replication in a wide sense of a recent study that 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 researchers to model the relationship ofcovariates with the treatment and the outcomes ina 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. KEYWORDS average treatment effects, causal inference, double machine learning, machine learning 1 INTRODUCTION Beliefs about the appropriate role of women in society are very different across regions. These disparities can be observed by analyzing, for instance, differences in labor force participation for women across societies. Choices about female labor supply have been shown to be partially explained by culture and norms (Fernandez, 2007; Fernández & Fogli, 2009). Thus, investigating the origin of gender norms is very important to understand the reasons behind these differences and assess whether appropriate policies can be implemented to reduce them (Hansen et al., 2015) . The paper by Alesina et al. (2013) is a seminal contribution addressing the question of the historical origins of gender roles. The authors test the hypothesis, originally developed by Boserup (1970), that today's gender norms have their roots in the agricultural practices that prevailed in pre-industrial times. The hypothesis compares the roles of shifting and plough cultivation. Since operating the plough requires considerable physical strength, men have an advantage in plough cultivation compared to women; in contrast, women could more easily participate in shifting cultivation, in which the use of the hoe and the digging stick is prevalent, and there is higher need for weeding, which was traditionally performed by women and children. Thus, where plough agriculture was prevalent, gender division of labor was more common. This division of labor would persist over time until the present day. In this paper, we perform a replication in a wide sense by revisiting the main research question in Alesina et al. (2013) with new causal inference tools, namely, causal machine learning (CML) methods. To this end, we connect the econometric theory on CML with empirical economics, serving as an illustration for applied researchers on the gains of implementing these newly available methods in observational studies. In our replication study, we employ the dou- This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2024 The Author(s) . Journal of Applied Econometrics published by John Wiley & Sons, Ltd. 2 BAIARDI and NAGHI | | | | | --- | --- | --- | | ble/debiased machine learning method (DML) introduced in Chernozhukov (2017, 2018), which provides consistent estimation and valid inference on the average treatment effect(ATE), in settingswhere high-dimensional nuisance parameters are estimated

FAQs

What does the study replicate and why?

It replicates, in a wide sense, a prior investigation into how historical plough agriculture relates to current gender roles, revisiting the central question with newer causal machine learning tools.

Which causal machine learning method is used?

The analysis uses double machine learning (DML), which supports consistent estimation and valid inference for the average treatment effect when nuisance parameters are estimated with machine learning.

What is the main finding about women’s labor force participation?

The study finds an even larger negative effect of historical plough adoption on female labor force participation compared with the original analysis.

Aug 3, 2026

Upload by Maeve