Research

Working Papers

• "End of Apartheid, Not of Inequality: The Slow Transition in a Segregated Economy" with Martí Mestieri and Johanna Schauer (STEG Small Research Grant) New!

[Slides] [Township list and location] 

Despite the formal end of Apartheid in 1994, South Africa remains one of the most unequal countries in the world. This paper investigates the mechanisms behind this persistence of inequality by developing a spatial dynastic model with heterogeneous agents, incomplete markets, and endogenous choices over education, occupation, savings, and location. Drawing on newly assembled micro-geographic data, we document a shift in inequality from being primarily across races to increasingly within the Black population, with spatial segregation—especially the legacy of Townships—playing a key role. Our model, disciplined by detailed spatial and socioeconomic data, captures the slow intergenerational convergence in education and occupational outcomes observed in post-Apartheid South Africa. It shows how inherited spatial disadvantages—through high commuting costs, disparities in school quality, and limited access to credit—continue to shape households’ educational, occupational, and locational choices, reinforcing inequality long after formal legal barriers have been removed. Quantitatively, we find that removing persistent spatial distortions in Townships accelerates the transition to a race-blind equilibrium by 40% and reduces income inequality by 10%.

CEPR Discussion Paper No. DP20472 

• "Global Innovation Spillovers and Productivity: Evidence from 100 years of World Patent Data"  with Enrico Berkes and Martí Mestieri  New Version!

[Slides]   [Fields of Knowledge Correspondence]

We use a panel of patent data spanning a wide range of countries over the past century to analyze the effects of international knowledge spillovers on innovation, productivity, and income per capita. Guided by a multi-sector growth model, we leverage the historical network of patent citations to estimate the elasticity of innovation with respect to international knowledge spillovers, which we find to be approximately 0.5. Using this elasticity as the first stage of our analysis, we quantify the impact of spillover-induced innovation on productivity and aggregate income per capita. Our estimates suggest that an increase of one standard deviation in log patenting increases sectoral output per worker growth by 1.1 percentage points. We find results of similar magnitude for sectoral TFP growth and income per capita growth.

CEPR Discussion Paper No. DP17285

Coverage: VoxEU

• "Land Property Rights, Financial Frictions, and Resource Allocation in Developing Countries"  [Julia code]

[Slides] 

Traditional communal land-use systems that lack private land ownership and documentation are common in low-income countries. The absence of deeds or titles for land generates imperfections in markets for land and amplifies frictions in credit markets. This paper quantifies the aggregate and distributional impacts of these frictions, as well as the role of their interaction. I develop a dynamic general-equilibrium model that incorporates imperfections in both land and credit markets, linked via the use of collateral in the economy. Micro-level data from Tanzania discipline the model and let me show that substantial frictions in land and financial markets affect resource allocation and economic efficiency in agriculture. Using the model to simulate a reform that privatizes all land holdings, I find that such a reform increases agricultural and non-agricultural output and reduces the share of households employed in agriculture while encouraging financial inclusion. I also show, that while financial reform could deliver comparable aggregate effects, land reform is more pro-poor and reduces consumption inequality. At the same time, the presence of multiple frictions in the connected markets limits the positive impact of any reform that eliminates imperfections in only a single market.

STEG WP046

Coverage: World Bank Development Impact Blog

Published Papers

• "Gendered Impacts of Covid-19 in Developing Countries" with Titan Alon, Matthias Doepke and Michèle Tertilt

AEA Papers and Proceedings, Vol. 112, 272-276, May 2022.

Also available as IZA DP No. 15013, January 2022.

In many high-income economies, the recession caused by the Covid-19 pandemic has resulted in unprecedented declines in women's employment. We examine how the forces that underlie this observation play out in developing countries, with a specific focus on Nigeria, the most populous country in Africa. A force affecting high- and low-income countries alike are increased childcare needs during school closures; in Nigeria, mothers of school-age children experience the largest declines in employment during the pandemic, just as in high-income countries. A key difference is the role of the sectoral distribution of employment: whereas in high-income economies reduced employment in contact-intensive services had a large impact on women, this sector plays a minor role in low-income countries. Another difference is that women's employment rebounded much more quickly in low-income countries. We conjecture that large income losses without offsetting government transfers drive up labor supply in low-income countries during the recovery.

Presentation Slides: [PDF], Codes and Data: [Link to Open ICPSR]

Coverage: WBBM Newsradio

Selected Work in Progress

• "The Origins of Structural Transformation" with Julieta Caunedo and Mayara Felix

Description: Labor reallocaton out of agriculture is a key feature of economic development. While income effects and changes in relative sectorial forces guide this process, is this reallocation mostly a consequence of "push" effects that release workers out of agriculture, or of "pull" effects that increase the demand for workers in sectors other than agriculture? We propose a novel general equilibrium model of structural change featuring a non-trivial size distribution of firms and farms in different sectors, and employment reallocation across sectors and locations. In this framework, we can study pull effects (i.e., firm entry in non-agriculture, that shift labor demand towards those sectors). Using rich data from Brazil spanning 1986 to 2017, we document the impact of "pull" effects from the entry of large non-agricultural firms in a municipality; into labor, plot sizes, productivity and capital intensification in the agricultural sector of surrounding municipalities. We find that non-agricultural firm entry reallocaties labor away from agriculture in surrounding municipalities, increases tractor adoption, drives small farms out of business and increases the number of large farms. We use this evidence to discipline out model and to assess the relative role of pull and push effects in the process of structural transformation thorugh counterfactual exercises. Overall, our paper contributes to a classic and rich literature on structural transformation in developing countries, emphasizing the importance to pull factors in non-agricultural sectors as a trigger of capital deepening and land consolidation in agriculture, mostly through shifts in relative prices.  

• "What is the Aggregate Impact of Pandemic Education Disruptions in Low-Income Countries" with Titan Alon, Matthias Doepke and Michèle Tertilt

Description: The Covid-19 pandemic has led to prolonged school closures in most countries around the world. This paper aims to quantify the potential impact of pandemic learning losses in developing countries, with a specific focus on sub-Saharan Africa. We argue that both micro and macro channels imply that the repercussions of pandemic education disruptions are more severe in poorer countries than richer economies. First, the evidence suggests that children in low-income countries suffer more significant learning losses. This obtains partly because of the lower availability and efficiency of alternative learning channels such as virtual instruction and partly because of a higher impact on dropout rates, which are amplified by income losses during the pandemic. Second, a given learning loss has a more significant medium-run impact on the economy because recent school graduates make up a larger fraction of the total labor force in low-income economies and because older cohorts have relatively little formal education. We quantify these channels using a macro-development model that matches household-level and aggregate data from Nigeria.

• "The Impacts of Mobile Money on Village Economies: A General Equilibrium Approach"

Description: This paper aims to assess the aggregate effects and the distributional consequences of the adoption of mobile money technology for the rural economy. I propose a two-sector heterogeneous agent model that incorporates occupational choice, endogenous wages, and forward-looking saving decisions. The more developed (urban) sector is motivated to make transfers to a less developed (rural) sector. Moreover, due to limited access to financial services in rural areas, saving technology is subject to potential losses. The introduction of mobile money technology reduces transfer friction between the urban and rural parts of the household and improves rural saving technology. I use existing micro evidence on the effects of mobile money to discipline the model.

Discussions

• Brooks, W. &  Donovan, K. “Optimal Fertilizer Policy After a Global Shock” July, 2025 (NBER SI Economic Growth 2025) Slides

• Dai, P., Müller, K. &  Verner, E. “Credit Allocation, Collateral, and Economic Development” January, 2025 (ASSA 2025, AFA) Slides

• Mitra, D., Pham, H. &  Ural-Marchand, B. “Exports and Intergenerational Mobility” September, 2024 (Trade and Uneven Development Conference, WB & JIE) Slides