Publication - Working Paper
The impact on cities of a large-scale rural migration programme in Bangladesh
Seasonal hunger is perhaps the biggest challenge to the reduction of global poverty that has remained largely under the radar. About three hundred million of the world’s rural population live above the poverty line but still suffer from seasonal income insecurity, which occurs between planting and harvest when the demand for agricultural labour falls and the price of food rises.
During this two- to three-month lean season, families typically miss meals daily. Our current research builds upon a previous randomised control trial (RCT) supported by IGC, through which poor households in Northern Bangladesh were offered small travel subsidies (valued at U$9-15) to temporarily migrate for employment opportunities elsewhere in the country. Its results showed that these subsidies had not only direct positive impacts for those who used them but also benefited others within the same village. Around 60% of households offered an incentive sent a migrant to work in a nearby area, up from around one third of households not offered the incentive. Households that received the incentive had higher earnings in the lean season and took in 550-700 more calories per person per day, equivalent to an extra meal in a period when meals are regularly skipped. And this pattern raised incomes even among those who did not migrate from the same villages, by lowering the number of people competing for the scarce jobs available during the lean season. Moreover, multiple years after incentives stopped being disbursed, members of households that had received the incentive continued to seasonally migrate at higher rates.
After the initial success of this intervention, Evidence Action, an NGO committed to testing and scaling up evidence-based solutions, has been implementing the programme at gradually larger scales in Bangladesh. This year, the programme is expected to disburse about 40,000 travel subsidies (in contrast to a sample of 1,900 households in the original RCT). The move from an RCT to a scaled up program, however, requires careful analysis, as there are important effects that may only arise at scale. In this iteration of our research, we focus on the following questions: Do rural migrants compete with the existing urban poor for work or do they complement local, low-skilled labor? And might an increased influx of migrants have non-economic effects, such as a social and political backlash among urban residents?
For this year’s study, we have worked with the program implementation team at Evidence Action to ensure that the design gives us the opportunity to measure impacts at destination areas. In particular, we set up the targeting as an RCT, and have chosen treatment villages that are near each other, creating clusters of treated and untreated areas, essentially varying the influx of migrants to certain cities/neighborhoods in cities and not others. With this design, we will measure the effect in destination areas through surveys of urban residents searching for day labour opportunities; and, given that rickshaw-pulling and construction work are the two main job lines among migrants, surveys of rickshaw garage owners, other rickshaw pullers, and construction-site managers.