Ongoing research projects

“Does education make Africa work?”

Authors: Taryn Dinkelman

Abstract: Sub-Saharan Africa (SSA) has the world's fastest-growing labor force (Lam et al 2019).  Expanded access to education has increased the pool of skilled workers in SSA countries over time.  While increasing human capital is an important precondition for economic development (Case 2000), schooling gains may not always lead to prosperity at the national or individual level. Specifically, many SSA countries are struggling to create enough good jobs to absorb their growing, and more educated, labor force. One reason is that the structure of production in African economies seems ill-fitted to both the top and bottom ends of the human capital distribution: large fractions of low-skilled workers are underemployed for at least some parts of the year, while many of the highly skilled are unemployed and waiting for jobs (Filmer and Fox 2014). Over half of Africans surveyed in public opinion surveys identify unemployment as their top issue of concern (Afrobarometer 2016). This labor market dysfunction may have political consequences, as large gaps between labor market aspirations and actual economic opportunities can be a recipe for political uprisings (Campante and Chor 2012). Against this backdrop, it is essential to understand what types of jobs are currently being generated in sub-Saharan Africa, how those jobs match the skills workers possess, and how political attitudes and participation among Africans have changed in response to increasing levels of education. This project has two parts. I aim to characterize the nature of work and of workers in Africa, and causally estimate how labor market opportunities and political attitudes have changed in response to population-level schooling gains over the last several decades. In each part, I exploit publicly available individual-level Census and Afrobarometer data, together with policy-driven variation in access to education generated by a wave of free primary education (FPE) laws introduced in the 1990s and early 2000s. I apply quasi-experimental methods (difference-in-differences and regression discontinuity-in-differences techniques) to generate causal evidence of the impact of human capital on employment outcomes, and on political attitudes/aspirations, across eight African countries that account for almost half of Africa’s population. 

“From Conflict to Commercialization: Promoting Agribusiness Growth through Market Access in Rural Colombia”

Authors: Joe Kaboski, Viva Bartkus, Wyatt Brooks, and Carolyn Pelnik

Abstract: We study an intervention designed to assist former coca-growing farmers to substitute toward licit income generating activities by create local market makers to collaborate in distributing and marketing their products. The program is conducted by our NGO partner, PASO Colombia and works with former coca-growing farmers in remote areas of Colombia which were previously controlled by FARC and cut off from outside markets. PASO assigns extensionistas to farmers in these areas. Extensionistas help farmers develop skills needed for modern farming and connect communities to national markets. The broad goal is to move farmers out of coca production and into other financially sustainable farming activities with specific intentions to increase yields, establish market relationships needed to produce cash crops, improve agricultural productivity,and earn competitive prices on non-coca output in a post-conflict economy.

Partner: PASO Colombia

“Gender Gaps in Entry into Competitive Careers: Evidence from Indian Civil Services”

Authors: Niharika Singh, Kunal Mangal

Abstract: We explore the causes of under-representation of women in the Indian civil services, where placement depends on succeeding in a highly competitive examination process. Using a dataset from a large Indian state that covers the universe of applicants and their placement outcomes to state-level civil service jobs between 2012-2019, we uncover where gaps arise in the recruitment process. We find that test re-taking is a key constraint for women: successful placement typically requires multiple attempts, but women—particularly those that score well on initial attempts and are most likely to be successful—are less likely to retake the exam than men. This dynamic selection of applicants across exam attempts contributes to a gender gap in placement outcomes. We provide suggestive evidence that marriage and household formation pressure may constrain high-ability women from making more attempts.

“How Information on Labor Market Conditions Shapes Search and Employment: Evidence from India”

Authors: Niharika Singh

Abstract: I provide over 6,000 job seekers in India with customized information about labor market conditions on a job portal for their preferred city and occupation. I find that treated respondents are 10% more likely to be employed than control respondents six months after treatment. Preliminary analysis suggests that these positive treatment effects are a function of prior beliefs about finding new employment as well as baseline employment status. This experiment shows how information about local labor market conditions can reduce frictional unemployment.

“Improving Women’s Access to Public Employment Opportunities: Experimental Evidence from India”

Authors: Niharika Singh, M.R. Sharan

Abstract: Access to paid work opportunities can change women’s lives. In India, the flagship workfare program is a demand-driven paid employment option available to rural households. Yet, disadvantaged low-caste women face several barriers to accessing this employment. Can collective action among low-caste women improve their access to NREGA employment and their overall economic and social wellbeing? We use a randomized evaluation to study the impacts of introducing women’s ‘worker groups’ to 127 village councils in Bihar, India. We will combine administrative program data and household surveys to study the impacts of these worker groups on women’s public and private employment and social networks. In addition, we also induce random variation in group participation to understand selection into groups and women’s demand for collective action.

“Leveraging Savings Groups with Cash Injections”

Authors: Joe Kaboski, Wyatt J. Brooks and Danice Guzman

Abstract: Rural Ugandan households have very little access to formal credit markets, which limits their ability to respond to negative shocks or undertake profitable investments. In Uganda, local community savings groups, promoted by Catholic Relief Services, attempt to mitigate this problem by accumulating savings from members that then finance loans to other members, but the quantity of financing that these groups can provide is typically limited. We conduct a randomized controlled trial intervention that injects outside capital into savings groups to increase loan supply. We collect baseline and endline household-level surveys of members on borrowing, savings, consumption, income and other household outcomes, together with a baseline exercise measuring female empowerment within the household in order to assess the relationship between borrowing behavior and family dynamics. Empirically, the exogenous increase in loan supply does not cause an increase in defaults and does not affect savings by group members but leads to a reduction in missed meals and an increase in agricultural investment. All of these results are concentrated among households where the exercise conducted at baseline was consistent with relatively empowered female decision-making within the household.

Partner: Catholic Relief Services

“Making connections: exploring the impact of social and professional network assortativity on the labour market for South African youth”

Authors: Patrizio Piraino, N. Herath

Abstract: Social networks are an important source of information about the labour market, but in South Africa they are deeply unequal in the quality of information they provide due to network assortativity – the clustering of labour market participants into groups with similar levels and types of professional experience. We conduct a pilot study of an intervention designed to expand and/or strengthen the social and professional networks of youth participating in a large scale 12-month work experience program. The program is run by the Youth Employment Service (YES), in partnership with South African government and industry. Our full study aims to create exogenous variation in network composition by experimentally enhancing the program with a training element that bolsters the professional networking activities of treated participants. This allows us to explore the impact of alleviating network constraints on learning and performance in the workplace and on post-programme search behaviours and employment outcomes.

Partner: Youth Employment Service

“Monopsony Power in Mexico”

Authors: Joe Kaboski, Alejandro Estefan-Davila, Illenin Kondo, and Wei Qian

Abstract: In contrast to advanced economies, firms in developing countries can evade labor income taxation by hiring workers informally, or operate in the formal sector yet avoid the taxation by employing outsourced workers. We document that both margins of informality and outsourcing have increased in Mexico over the last 30 years. We also show empirical evidence that larger firms in the informal sector pay a lower share of their production to workers relative to materials. We then estimate labor markdowns as a measure of a firm's monopsony power in its local labor market. We find that markdowns are correlated with the firm's degree of informality, and markdowns are higher among firms that outsource. Motivated by these facts, we are developing a general equilibrium model with rich firm heterogeneity to study the impacts of monopsony power on employment and wages in the presence of informality and outsourcing in labor markets. We will use the model to evaluate the labor market and distributional effects of counterfactual enforcement policies aimed at reducing the degree of informality.

“School Management and Re-Enrolling Dropouts: Experimental Evidence from Schools in Rural Uganda”

Authors: Joe Kaboski, Viva Bartkus, Wyatt Brooks, Carolyn Pelnik, and Maurice Sikenyi

Abstract: In low-income countries, two interrelated causes of low educational achievement are students falling behind grade level and dropouts, problems that have been exacerbated by the COVID pandemic and associated lockdowns. To address this problem, in collaboration with a Ugandan NGO named Building Tomorrow, we consider whether a management intervention targeting school leaders can offer a cost-effective means of improving public service delivery in schools in Uganda. Specifically, we ask whether a fellows program to support school principals by engaging community leaders can increase student enrollment and learning. We will measure the effect of this intervention on a school’s outputs, namely student enrollment and test scores, across 300 rural schools.

Partner: Building Tomorrow

“The Natural Rate of Structural Change”

Authors: Joe Kaboski, Francisco J. Buera, Marti Mestieri

Abstract: The literature on structural transformation has made important recent advances in explaining the data, but little progress on developing normative lessons. This project develops diagnostic methods data on structural transformation to identify potential distortions to growth, capital accumulation, and structural change. Starting with the standard efficient model, we are applying the latest advances in explaining structural change patterns to cross-country panel data on growth, investment, and sectoral allocations to measure black box wedges or distortions. We will then project these distortions on potential explanatory factors to identify policy distortions and/or avenues for refinements to the standard model.