Essays on India's Employment Guarantee

Abstract

This thesis studies three issues related to the labor market and distributional impacts of a large rural workfare program in India. The first chapter is co-authored with a graduate student at the Paris School of Economics, Clement Imbert. The second and third chapters are single-authored.

Chapter One uses the gradual roll-out of India's National Rural Employment Guarantee Act (NREGA), a large rural workfare program, to estimate the program's impact on wages and aggregate employment. We use a difference-in-differences strategy, which compares changes in districts that received the program earlier to districts that received it later. Our estimates reveal that following the introduction of the program, public employment increased by 0.3 days per prime-aged person per month (1.3% of private sector employment) more in early districts than in the rest of India. Casual wages increased by roughly 4.5%, and private sector work for low-skill workers fell by 1.7%. These effects are concentrated in the dry season, during which the majority of public works employment is provided. Our results suggest that public sector hiring crowds out private sector work and increases private sector wages. We use these estimates to compute the implied welfare gains of the program by consumption quintile. Our calculations show that the welfare gains to the poor from the equilibrium increase in private sector wages are large in absolute terms and large relative to the gains received solely by program participants. We conclude that the equilibrium labor market impacts are a first order concern when comparing workfare programs with other anti-poverty programs such as a cash transfer.

Chapter Two describes the design and implementation of a survey consisting of in- terviews with 2,224 adults from 705 households from 70 villages in three states in India (Rajasthan, Gujarat, and Madhya Pradesh). Data collection emphasized migration and participation in the NREGA. Over half of the adults in the sample, belonging to 80 percent of households, left the village for work at least once from summer 2009 through summer 2010. Migration was short-term and seasonal with the majority of migrants leaving during the agricultural off-season. Most migrants, even those spending six to 11 months outside the village per year, engaged in economic activities when they returned to the village suggesting important economic ties to the village. Participation in the NREGA is seasonal and common with 72% of surveyed households reporting at least one member working for the program within the past year. We document substantial unmet demand for work under the program and sharp differences in provision of work under the program across states.

Chapter Three uses the data from the survey described in Chapter Two to present evidence that the NREGA decreased migration. Thirty-five percent of the respondents report leaving the village for work during summer 2009. Of these migrants, 35% report working for the NREGA, and 88% report that they would work more for the NREGA if more work were available. Eight percent of adults report that they would have migrated had they not worked for the NREGA. We document that adults living in a state that provides more days of public works under the program, even conditional on demand for NREGA work, spend less time outside the village for work compared with other states. This cross- state difference in days outside the village is statistically significant only during the summer season during which most of the NREGA work is provided. Demand for NREGA work is high among migrants despite the fact that wages for work outside the village are roughly double the earnings for NREGA participants. Using wage differentials and survey measures of demand for NREGA work, we estimate that the average migrant incurs monetary and non-monetary costs of migration equivalent to 75% of the average earnings per day spent outside the village.