Research

Job Market Paper

Weather Shocks, Labor Reallocation, and Structural Transformation: Micro Evidence from Africa

Job Market Paper

Structural transformation, which is the movement of labor from agriculture to more productive non-agricultural activities, is a key driver of economic development. Yet labor released from agriculture does not always transition into productive non-agricultural employment. This paper develops a structural framework that decomposes structural transformation into three sequential stages: labor release, labor reallocation, and labor absorption, allowing the roles of subsistence constraints, mobility barriers, land transferability, and labor absorption capacity to be separately identified. Using household panel data from the World Bank Living Standards Measurement Study (LSMS) for Niger, Malawi, and Ethiopia, combined with high-resolution climate and geospatial data, I first provide reduced-form evidence on how weather-induced agricultural productivity shocks interact with these constraints to shape household labor allocation. I then estimate the structural model using the simulated method of moments and conduct counterfactual policy analysis. The results reveal substantial heterogeneity in the mechanisms governing structural transformation. Subsistence constraints are the primary barrier in Niger and Malawi, whereas labor reallocation and labor absorption constrain structural transformation in Ethiopia. By identifying the binding constraints at each stage of structural transformation, the proposed framework provides a unified approach for understanding labor allocation and evaluating development policies.

Paper: Draft (PDF)

Working Papers

Can the Service Sector Lead Structural Transformation in Africa? Evidence from Côte d'Ivoire

with Jeremy Foltz

Paper: SSRN

Standard models of structural transformation view manufacturing as the primary engine of economic development because it combines high productivity with the capacity to absorb labor released from agriculture. Yet many African economies have largely bypassed manufacturing, raising concerns about whether service-led development can sustain long-run growth. Can the service sector drive structural transformation by simultaneously generating productivity growth and labor absorption? We address this question using firm panel data from C\^ote d’Ivoire, one of Africa’s fastest-growing ``leopard” economies. Building on the structural-change decomposition, we show that services can drive structural transformation when they combine high productivity with labor absorption. Using proxy-variable estimates of firm-level total factor productivity, we find that services are more productive than both agriculture and manufacturing and account for the largest expansion in formal employment. We then examine the firm-level mechanisms behind this expansion. More productive service incumbents grow faster, more productive entrants expand after entry, and more productive firms are less likely to exit, with employment growth concentrated particularly among unskilled workers. Together, these findings show that productive service firms can generate sustained labor absorption and contribute to structural transformation. The results suggest that development policies should promote productive services alongside manufacturing in Africa.

Book Chapters

Perspectives on The Agricultural Wealth of Nations

with Héctor Paredes

Forthcoming, Proceedings of the British Academy

Work in Progress

Gender Constraints and Labor Allocation during Structural Transformation

Structural transformation requires workers to move from agriculture into more productive sectors, yet women’s participation in non-agricultural employment remains low across much of Sub-Saharan Africa. This paper asks whether gender differences in labor allocation reflect preferences or binding constraints. Using LSMS panel data from Niger, Nigeria, Malawi, and Ethiopia combined with satellite-based measures of weather-driven agricultural productivity, we examine gender differences in sectoral transitions. We find that women respond more strongly than men to productivity improvements by moving into non-agricultural work, suggesting that gender-specific entry constraints bind more strongly in low-productivity states and are partially relaxed as agricultural productivity rises. The results highlight how barriers to women’s mobility can slow structural transformation.

Climate Change and Conflict: Causal Evidence using New Machine Learning Methods

with Jeremy Foltz

Climate shocks may affect conflict not only where they occur, but also through exposure in neighboring areas. Using a panel of 9,691 grid cells across Africa, we combine geocoded conflict events with rainfall, land-cover, and other geospatial data and estimate causal effects using double/debiased machine learning. We find no significant effect of own-cell rainfall shocks on next-year conflict incidence when neighborhood exposure is ignored. Once spatial exposure is considered, however, high rainfall in neighboring cells increases conflict risk, with effects varying across land systems. These findings highlight the importance of spatial spillovers and ecological heterogeneity in understanding the climate–conflict relationship.