Working Papers
Her Mother’s Footsteps: The Effects of Family Norms and Social Norms on Women’s Labor Force Participation (link to draft)
Family and societal norms play a crucial role in shaping female labor force participation (FLFP), particularly in the Middle East and North Africa (MENA) region, which consistently reports some of the lowest FLFP rates globally. This study investigates how these norms affect FLFP by utilizing data from the Iran Labor Force Survey. Building on existing literature that identifies three pathways through which norms influence FLFP—(1) observing a working mother teaches daughters how to balance household and professional roles, making employment seem more attainable; (2) family norms shape attitudes toward women’s work; and (3) societal norms influence perceptions of female employment—I develop a structural model that integrates these mechanisms and allows for the decomposition of their individual effects. Moreover, empirically, I assess two of these mechanisms. First, I find that a mother’s work history has a significant positive impact on her daughter’s decision to work, with no comparable effect observed for sons. Second, using a subsample of women who migrated for family-related reasons, I demonstrate that the FLFP level in their region of origin continues to significantly influence their labor force participation, even after controlling for local labor market conditions. By focusing on movers and isolating local labor market effects, these results underscore the enduring influence of cultural norms on women’s employment decisions.
Variable Partitioning in Reduced Form Models Using Classification Trees: Case of Studying Labor Force Participation (link to draft)
Understanding the non-linear effects of continuous variables, such as age, on labor force participation (LFP) is crucial for accurate economic modeling and effective policy formulation. Current approaches often rely on intuitive age categories, which may not capture the nuanced shifts in participation behavior across different life stages. Although the reduced-form model should be grounded in theory, the variable partitions can be refined based on data-driven insights. For instance, is age 60 an appropriate threshold for defining the retired group, or would 62 be more suitable? This study proposes a novel methodology that employs Classification Trees to partition the continuous variables, such as age, into categorical groups, thereby enhancing the precision of the analysis. By integrating machine learning techniques with econometric models, this approach determines age thresholds based on data-driven insights rather than subjective judgment. I implement and compare three Probit models: (a) treating age as a continuous variable, (b) using pre-determined intuitive age categories, and (c) applying Classification Tree-derived age categories. The results demonstrate that the Classification Tree method significantly improves model performance, evidenced by higher R-squared values and increased significance levels of key predictors, which can be attributed to more precise categorization that aligns better with the underlying patterns in the data. Furthermore, this refined partitioning enhances the identification of target groups for policy interventions, ensuring a more effective allocation of resources. This study contributes to the literature by providing a robust framework for handling continuous variables with non-linear effects, thereby offering a more credible and data-driven approach to demographic analysis and policy evaluation.