UC Berkeley

Rising youth unemployment is a key concern for many poor countries, where school dropout rates at the primary or secondary level are very high. The relationship between skills training and youth employment is thus a critical question of interest to economists. In the first chapter of this dissertation, I examine this relationship in the context of Bangladesh, a lower-middle income country. The relationship between credit and agricultural development is also an important question of interest to economists, as agricultural development is directly linked with food security and poverty. The theoretical literature shows that credit can positively affect agricultural development through, among others, decreasing sharecropping relative to fixed rental contracts. The second chapter of this dissertation empirically tests this proposition. The third chapter of this dissertation examines whether providing potential migrants with information along with administrative and community support reduces the risk of international migration.

In the first chapter, I estimate the effects of a youth training program in Bangladesh on labor market outcomes. The program provides on-the-job and classroom training to disadvantaged and unemployed youths. On-the-job training is provided through apprenticeship under a local master crafts person. Classroom training curriculum includes theoretical training on specific trades as well as soft-skills training. The program is implemented by BRAC, the largest NGO in the world. Using the data generated by BRAC’s internal research unit, I show that six months after the intervention, on-the-job training increases labor market participation of program participants by 22.6 percentage points, total time devoted to earning activities by 59%, and earnings by 44%. It increases both self- and wage employment. The effect on employment is found to be larger for females. Additional effects of classroom training over on-the-job training on overall employment and earnings are small in magnitude. Results, however, indicate that if classroom training is added to on-the-job training, the effects shift from self- to wage employment. Results also show that employment in firms where the apprenticeship took place is a channel for the effect on wage employment. The benefit-cost ratio for on-the-job training is estimated to be 6.34, demonstrating high returns of the investment made under this initiative. I also show that, at the scale at which the program was implemented, employment effects for beneficiaries were not achieved through displacement of non-beneficiaries.

Choice of a land rental contract has figured prominently in the theory of industrial organization. While a share contract is inefficient in a first-best world, it may be the preferred option under second-best conditions. Theory has thus predicted the existence of sharecropping as the potentially preferred contract under liquidity constraint, insurance market failure, and market failure for non-contractible inputs not owned by the tenant. Rigorous empirical evidence is, however, still lacking on this basic tenet of theory. In the second chapter, co-authored with Alain de Janvry and Elisabeth Sadoulet, we study a randomized experiment in a credit program for landless workers and marginal farmers organized by BRAC in Bangladesh to show that access to credit has a large positive effect on the choice of fixed rent over share rent contract, both in terms of number of contracts and area contracted. As predicted by theory, the magnitude of this shift is enhanced when the tenant is less exposed to risk. However, we do not find any conclusive evidence that the shift is increased by differential possession of non-contractible farm management experience. Our results suggest that development programs that give access to credit to potential tenants can help them move away from inefficient land rental contracts.

In the third chapter, coauthored with Alain de Janvry and Elisabeth Sadoulet, we investigate whether providing information along with administrative and community support to aspiring migrants reduces the risk of migration. The risks we focus on include: (i) the risk of failure to depart the home country after having spent resources during the migration process; and (ii) the risk of not finding a job abroad with the expected salary. BRAC implemented the intervention in Bangladesh. It recruited community volunteers to implement the intervention. We randomized the intervention at the union level - the lowest administrative unit in the country. Results show that the intervention has no statistically significant effect on the overall migration success or on migration failure. Similarly, there is no significant effect on the salary received abroad. However, in areas where poverty rates are relatively high, the intervention reduces migration failure. The program also reduces the cost of failure, as well as migration costs among individuals who attempt to migrate through informal channels. Additional results suggest that the program is more likely to decrease migration failure in areas where the levels of education of volunteers are higher. We also find that treatment effect on migration success is larger in areas where the proportion of volunteers who are returning migrants or migrant family members is higher. These results suggest that the program might be more effective if it engages more educated individuals and/or more returnee migrants/migrant family members as volunteers.