Skip to main content

Advertisement

Log in

Extreme weather and migration: evidence from Bangladesh

  • Original Paper
  • Published:
Population and Environment Aims and scope Submit manuscript

Abstract

Using retrospective migration history data collected in southwestern Bangladesh, we examine the relationship between extreme weather conditions—warm spells, dry spells, wet spells, and intense precipitation—and the likelihood that male household heads make a first internal or international migrant trip. We also investigate whether and how agricultural livelihoods and having a close migrant relative are associated with migration in response to extreme weather. Findings reveal that dry spells are most consistently associated with increased migration, although we see some evidence that first trips also increase after warm spells and above average rainfall. Associations between dry spells and warm spells are more pronounced among household heads with agricultural livelihoods. We find little evidence that having migrant relatives is associated with migration after extreme weather events, but some evidence that the presence of migrant networks at the community level is associated with international migration after an environmental shock.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Notes

  1. However, women make up an increasing share of labor migrants, largely due to emergence of the ready-made garment industry in Bangladesh in the 1980s (Ward et al. 2004).

  2. Given the high prevalence of migration to India, we collected migration histories for this destination separately from other international destinations.

  3. The BEMS was part of a larger project that investigated coastal dynamics, environmental change, and migration in the low-lying southwestern region of Bangladesh, which was heavily affected by tidal processes (Auerbach et al. 2015). The BEMS is closely modeled on the sampling design and ethnosurvey used in the Mexican Migration Project (see http://mmp.opr.princeton.edu/).

  4. Of the 286 male household heads who were excluded, most (72.0%) were born outside of the study region, two (0.7%) assumed the role of household head after 2012, and another 78 (27.3%) completed a first trip before 1973.

  5. The average duration of trips for domestic moves was 3.3 years and, for international moves, 6.4 years.

  6. Although we wanted to construct a continuous measure of the number of family members who had migrated, we were unable to do so because it was rare to have more than one migrant in the household (1.9% of all person years). Thus, we use a dichotomous variable.

  7. The zero-order correlations range from −0.02 to 0.19.

  8. We also considered a seasonally splitted missing value imputation algorithm (SEASPLIT), which produced nearly identical regression results as those reported below. We decided to use the SEADEC algorithm because it generated an imputed dataset with fewer outliers and best matched the seasonal and temporal trends within the raw data.

  9. We also estimated multinomial models in which domestic and international trips are modeled as competing risks. These findings are presented in Table 5 of the Appendix.

  10. The distribution of migrant social ties is roughly equivalent across the four communities; mean person-years with social ties are 0.018, 0.012, 0.019, and 0.010 for Morrelganj, Tala, Satkhira Sadar, and Kalia, respectively. Mean person years with a first international trip for the same communities are 0.015, 0.002, 0.004, and 0.011.

  11. We included these interactions in a preliminary model; however, the standard errors associated with the interactions are very large making interpretation difficult.

  12. Socioecological System Dynamics Related to Livelihood, Human Migration, and Landscape Evolution (Grant # CNH-1716909)

  13. Available from the authors on request.

References

Download references

Acknowledgments

We presented an earlier version of this paper at the 2018 meeting of the Population Association of America (PAA). We are grateful for the generous support from the Office of Naval Research, Georgetown University, University of Colorado Boulder, and Vanderbilt University. We also thank Jonathan Gilligan for helpful comments and Bhumika Piya, Blake Sisk, and Mitra and Associates for their assistance in carrying out this research. Finally, we thank colleagues at the PAA session and the Population Research Center at the University of Colorado Boulder for their helpful comments.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Amanda R. Carrico.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

BEMS site selection

To select the nine BEMS research sites, we constructed a sampling frame of all mouzas in the five southwestern-most districts (Jessore, Narail, Satkhira, Khulna, and Bagerhat). We stratified the sampling frame by district to ensure variability in proximity to the Indian border (to the east) and to the coast (south), where there is greater exposure to salinity, storm surges, and cyclones. We also stratified by economic development using an index that sums the adult literacy rate and share of households with access to electricity in each mouza, using data from the Bangladesh census (Bangladesh Bureau of Statistics 2011). We then divided the sampling frame of mouzas into high and low economic development strata using a median split. Within each district, we selected one mouza from each stratum, resulting in two study sites in each of the five districts. We omitted one study site in Khulna District from this analysis because it was our pilot data collection site.

Table 5 Multinomial logistic regression models predicting first domestic or international trips by male household heads
Table 6 Logistic regression models predicting a first international trip by male household heads in the full sample

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Carrico, A.R., Donato, K. Extreme weather and migration: evidence from Bangladesh. Popul Environ 41, 1–31 (2019). https://doi.org/10.1007/s11111-019-00322-9

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11111-019-00322-9

Keywords

Navigation