The Effects of Fathers’ and Siblings’ Migration on Children's Pace of Schooling in Rural Bangladesh

Abstract

This paper studies the impact of fathers’ and siblings’ migration on children's pace of schooling in Matlab, an area of rural Bangladesh with high rates of rural–urban and international out-migration. A large sample of children aged seven to 14 participated in the 1996 Matlab Health and Socio-economic Survey. The overseas migration of their fathers and both the overseas and internal migration of their brothers were associated with improvements in the pace of the children's schooling. The migration of sisters had no effect. Migration of mothers in the location of this study remained too rare to be addressed in this study. Research relating migration to improved schooling outcomes among the left-behind raises concern that migration effects merely proxy for greater investment capital or higher aspirations. By integrating MHSS data with a unique Health and Demographic Surveillance System, the effects of wealth measured at the time of the survey and 14 years before the survey were controlled for. Migration effects, though reduced, remained significant even after controlling for current and prior socio-economic status.

Keywords:

• pace of schooling
• internal migration
• overseas migration
• migration effects

This work was graciously supported by the US National Science Foundation Grant #BCS0422976, ‘Migration and Marginalization: Social Networks, Inequality and Health in Indonesia and Bangladesh’; National Institute on Aging Grant R01AG016308-02, ‘Social Networks and Adult Survival in Rural Bangladesh’; and NIA grant #5P30AG017248-02, ‘University of Colorado Population Aging Center’. The author also thanks Ali Protik for data management assistance, Linda Mamoun for editorial assistance, and Rachel Silvey for her advice. Most of all, the author thanks the staff of the Matlab Health and Demographic Surveillance System of ICDDR,B and the residents of Matlab for their patient participation in the research endeavour.

Notes

1. For instance, migration could alter the incentives for investing in particular social goods such as education.

2. Roughly, purdah can be defined as a system in which the status of a family or lineage is tied to the maintenance of women's modesty and immobility.

3. Bari refers to a cluster of traditional dwelling (huts) occupied by an extended family.

4. The within-bari sample consists of one randomly chosen household, and one based on a purposive selection process that gives preference to close kin. The analysis is weighted according to the likelihood of inclusion in the sample, where households from large baris are under-represented.

5. Age 16 is chosen as the cut-off because it is the age at which large numbers of daughters begin to marry out of the household, and sons begin to migrate to the city. Without relying on non-household child reports in constructing the sample, 16 is the last reasonable cut-off.

6. The slight preponderance of girls here is explained by the removal of a small number of girls (57 in total) who had married into this household. This presumably reflects a minor selection among the older female respondents towards those who have yet to marry, who make up 96 per cent of 16-year-old girls.

7. The individual survey was administered to the household head and spouse of head, all household members age 50 and over and their spouses, and two additional randomly selected adults (ages 15–50).

8. Mothers of young children, who were young themselves, had a lower probability of Book III selection.

9. A simple linear regression predicted MHSS total household assets (logged) in terms of asset measures also available in the 1982 census. Using these relationships (which explained 56 per cent of the variance in total household assets), I imputed household assets for 1982 census households.