We show that the 20th-century human capital increase contributed to structural transformation. We document that almost half of the decline in aggregate agricultural employment was driven by new birth-cohorts entering the labor market. We use data on educational attainment and compile a comprehensive list of policy reforms to interpret the differences in agricultural employment across cohorts. We find that the increase in schooling led to a sharp reduction in the agricultural labor supply by equipping younger cohorts with skills more valued out of agriculture. Interpreted through a model of frictional labor reallocation, these facts imply that human capital growth accounts for about 20% of the observed decline in agricultural employment.
How do changes in agricultural productivity affect firms? Using the predictions of a simple multi-sector general equilibrium model of the local economy and exploiting weather-induced agricultural volatility across India, I estimate the response of manufacturing firms to changes in agricultural productivity. I show that negative agricultural productivity shocks lower the cost of labor but that this does not cause firms to hire more. Firms’ production and employment in fact decrease because the shocks also reduce local income and hence the demand that firms face. My estimates provide evidence for a significant local demand effect. I then use my framework to show that this has key policy implications. I examine the introduction of a rural workfare program and assess how it affects firms. I show that the program attenuates the impact of negative agricultural shocks on firms because of its counter-cyclical effects on local wage and demand for manufacturing goods. The results highlight how policies that target households and increase their income can affect local market size and therefore the industrial sector through their general equilibrium effects.
Work in Progress
Buy from Your Kin: Microenterprise Customer Base, Productivity and Market Power — Pilot completed
This project studies whether social and kinship ties influence the productivity and market power of small businesses via their influence on business’ access to customers. In contexts where kinship identity matters, consumers may value social proximity to the business owner and be willing to trade off such proximity for other product attributes such as price and quality. This guarantees businesses from large social networks access to a large set of loyal customers, and thus puts them at an advantage relative to equally or more productive competitors that belong to smaller networks. A model of product differentiation à la Hotelling – in which physical distance to firms is replaced by social distance to firms’ owners – captures the mechanisms proposed and delivers testable predictions on the behavior of firms and consumers, as well as on market outcomes. I provide preliminary evidence in support of customer kinship loyalty using microenterprise and social network data from India. Existing data, however, lack information on customers and transactions, and hence limit the possibility to address questions related to market structure. A full test of my hypothesis will rely on a unique matched business-customer dataset that I collect for a sample of large Indian villages (currently in progress). The new data will also provide the basis for potential future interventions that vary the productivity of a subset of businesses and examine whether market shares reallocate in a pattern consistent with customer kinship loyalty.
Does Agricultural Volatility Change the Selection into Entrepreneurship?
A large fraction of households in low-income countries owns enterprises that are informal and very small. In this paper, I shed light on the motivations that drive households to participate in enterprise and study the role of agricultural volatility. I examine whether households resort to microenterprises as a response to agricultural income shocks, and assess the implications that this has on selection into entrepreneurship and business type. Using employment and firm data from India, and exploiting variation in agricultural incomes induced by seasonality and weather, I provide evidence for the following selection dynamics. First, I show that a decrease in agricultural productivity significantly raises the probability of running a microenterprise (extensive margin) as well as the time spent working in the enterprise (intensive margin). That is, engagement in enterprise behaves counter-cyclically. Second, and consistently with a selection mechanism based on entrepreneurial ability, I show that average ability among active entrepreneurs decreases when agricultural productivity is low, and that microenterprises that operate cyclically are different from their permanent counterparts across a number of dimensions, including size and profitability. Third, I find that most of the cyclical entry-exit takes place in services rather than manufacturing. This result suggests that households that operate businesses intermittently do so through activities that require lower start-up costs and fewer specific skills.
Medical procedures vary considerably across places, hospitals and physicians, and clinical and demographic factors alone are unable to fully explain this variation. In this study, we examine the role of a non-medical factor, patient-physician trust, in the choice of medical procedures. We focus on childbirth and the choice between natural deliveries and Cesarean sections. First, using GSS data, we document that individuals who trust others more (i.e., higher generalized trust) tend to have higher trust in physicians. Second, we show a strong negative relation between generalized trust and Cesarean Section rates across U.S. states and across countries. Third, we estimate the causal effect of trust by studying the delivery methods of second-generation American women. Using generalized trust in the woman’s country of origin as a source of variation for her level of trust, we show that low-trust mothers are more likely to deliver through a Cesarean section. Our results are robust to the inclusion of an extensive set of pregnancy and delivery risk factors, as well as controls that capture differences across women in socio-economic status, health care access and health care quality. We provide further evidence that trust is not capturing omitted health risk or disadvantaged background by performing a placebo exercise which shows that trust does not predict poor health outcomes such as low birth weight and pre-term birth. Our findings are consistent with a framework where asymmetric information on physician’s effort is coupled with differential returns to effort across procedures.