Migration, development and welfare: findings from a household survey in two selected villages in Bangladesh

Abstract

This paper contributes to the debate about the developmental impact of overseas migration using household level data from two selected villages in the southeastern region of Bangladesh. It argues in turn that remittances do contribute to household wealth accumulation but longer term investment in agriculture that can potentially lead to a change in agrarian structure is largely missing. An asset index is also employed to determine the welfare effects of both overseas and domestic migration. This paper begins with a discussion on theories of migration and how these have evolved to encompass new empirical features of the migration phenomenon. The survey results are then presented with an analysis of what this means for the nexus between migration and development within the specific context of Bangladesh.

Keywords:

• migration
• development
• Bangladesh

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. These household assets included livestock/poultry, almirah (wardrobe), refridgerator, car/truck, rickshaw/van, mobile phone, radio, television, motorcycle/scooter, bicycle, tin roof.

2. Hasanpur village is located in the Anandapur Union of Fulgazi Thana in the Feni district. As this village is located in the basin of two rivers, it is generally prone to flash floods. In contradistinction, Purbalach village which is located in the Raipur Union of Raipur thana in the Lakshmipur district is also part of the flood plain but has become generally flood-free due to the Chandpur Irrigation Project. Both villages have a very high density of population and thus exert great pressure on land that is already prone to floods though in varying intensity. Both Feni and Noakhali districts are known to have a high incidence of migration in comparison to other regions of Bangladesh. Migration from these regions, in fact, dates as far back as the eighteenth century during the British period. The southeastern region of Bangladesh where the greater Noakhali lies with its proximity to sea ports allured many migrants abroad through the British navy (Siddiqui, 2003). Another reason for the high incidence of migration from regions such as Noakhali could be due to the recurrence of flooding thus making agriculture a more precarious source of year round employment (Siddiqui, 2003). Even during the Pakistan colonial period, there were many households from Feni district and Hasanpur, in particular, who migrated to then West Pakistan (Interview, Ullah, 2011).

3. In Hasanpur, a total of 61 households were enumerated. Out of this sample, a the incidence of households with migrants overseas currently was 36%. The incidence of households with domestic migrants was also approximately 36%. One household had a migrant overseas at some point in time though not currently. The remaining households (26%) had no migrants either overseas or domestic.

4. In Purbalach, the incidence of migration was as follows: 45%, current overseas migrants; 5%, households with migrants overseas at some point in time; 8.3%, domestic migrants; 25%, no migrants.

5. In Hasanpur, 12 households and in Purbalach 20 households thus representing approximately 20 and 33.3% of the village samples respectively.

6. The World Bank has also cited some of these geographic differences and in particular a possible East West divide in incidence of foreign remittances for Bangladesh in its Poverty Assessment for Bangladesh, World Bank (2008).

7. Note that the actual value of the remittance was not used as these values are prone to error. Rather, in order to explore the separate effects, the asset indices of households with solely migrants overseas with that of households with only domestic migrants were compared. In so doing, the results produced varied sharply across the two sample villages as shown below.

8. When dependency ratios of households were correlated with asset indices, the Pearson correlation coefficient was negative but was not statistically significant.