Showing posts with label data. Show all posts
Showing posts with label data. Show all posts

Tuesday, 8 March 2016

Migration as a route out of poverty in Zimbabwe: remittances and gender

by Julie Litchfield


This year on International Women’s Day, 8 March 2016, I will be speaking at a conference on migration at the National Gallery of Zimbabwe in Harare, presenting some of the Migrating out of Poverty research undertaken in collaboration with the Centre for Applied Social Sciences (CASS) at the University of Zimbabwe.[1]

Colleagues at CASS have carried out extensive qualitative fieldwork in migrant-sending communities across the country and one of the observations they have made is that families often describe women migrants as the “saviour of the family”, less likely to forget their families at home and more likely to send a range of cash and goods than male migrants.

Just over a year ago in April 2015 we carried out a fairly large survey of households to probe deeper into migration and remittances. We have been able to collect data on the type of remittances sent home by individual migrants, the frequency and value of remittances as well as data on each household’s living standards, and I have been using this data to investigate the belief that women are the saviours of the family.

First I wanted to explore if families regard migration as a positive or a negative development. We asked households several questions about their perception of migration as a route out of poverty and were surprised that only around half felt that households with migrants were generally better off than those without. Digging deeper, I found that migration is seen as a good thing by those households who receive remittances, especially among female headed households.

Looking at who remits, I discovered that what migrants remit differs. Male migrants are slightly more likely to remit cash than female migrants, a large proportion of whom send goods, such as food, clothing and school items. Interestingly however, proportionately more women than men migrants send both cash and goods, and while the cash they send is lower in value, the value of goods is higher. This resonates strongly with findings of our CASS colleagues in the qualitative research. Perhaps this mixture of remittances suits households’ needs better and that is why women migrants are described as saviours.

To explore the idea deeper, I estimated a simple econometric model of remittance receipt. This helps me to explore what factors affect the probability that a household receives remittances. As my independent factors, I included variables on the household size and composition, whether the migrant has left behind children at home and also variables that capture the household’s assets and current living conditions. The theory behind this model is that remittances might be motivated by altruistic reasons (concern over aging parents, younger siblings for example) or for exchange reasons (for example the need to continue to contribute to the household and protect inheritable assets, or to maintain a certain social standing within the home community).

My results suggest that the independent factors that affect whether or not a household receives remittances differ between male and female migrants. Remittances to households from male migrants depend heavily on the assets of the household: households with bigger houses or a higher standard of living are more likely to receive remittances than those households we might regard as poorer. In contrast, remittances from female migrants are unaffected by these factors. Instead, female remittances depend almost entirely on the demographic make-up of the household: older heads of households, larger households and households where the migrant has left children behind are the only factors that appear to play any role in the remittance model. These factors also affect remittances receipt from male migrants[2] but what is striking is the absence of any relationship between female remittances and household wealth and living standards: female migrants remit regardless of whether the household is rich or poor. 

What does this tell us? We need to be careful about jumping too quickly to the conclusion that women care more about their families than men. One interpretation is that what underlies these differences is a complex set of rules and norms that are highly gendered and work at numerous levels within the household and the wider community. Male migrants may need to send remittances home to maintain their social standing not just within their own household but also in the community. Perhaps more is at stake the better off their household, both in terms of what they might inherit or the role they might be able to fulfil when they return.  A preference for sending cash might help with this aim. Women on the other hand may be seeking only the approval of their household as a caring, dutiful daughter for example, and thus send goods which they believe the family needs and demonstrates thoughtfulness and concern.

Understanding this complex set of rules and norms is difficult and warrants further research, especially when designing policy. Policies to reduce the transaction costs of sending formal cash remittances for example may have little effect on remittances from women if they continue to place value on sending remittances in the form of goods. A so-called diaspora bond may only be attractive to male migrants. What is clear is that policy on migration in general and on remittances in particular needs to be informed by a deeper understanding of these gendered patterns of behaviour.

Julie's presentation slides can be accessed here.



[1] I would like to thank the National Gallery of Zimbabwe for their kind invitation to speak at their conference and the British Council (Zimbabwe) for generously funding my visit.
[2] Their remittances are very responsive to whether they have a child at home, much more so than female migrants. 


Dr Julie Litchfield is the Theme Leader for Quantitative Research for the Migrating out of Poverty Research Programme Consortium (MOOP) and Senior Lecturer in the Department of Economics at the University of Sussex. The Working Paper discussing aspects of the findings of the Zimbabwe survey is now available. See also Eva-Maria Egger's presentation of the preliminary findings of MOOP's household survey conducted in Zimbabwe  and her related blog discussing the context:  Migration in Southern Africa: A Visit to the City of Migrants.

Friday, 18 December 2015

Minding the Migration Data Gap: new data from the Migrating out of Poverty Consortium

By Julie Litchfield




Research on migration and development has seen a dramatic resurgence in recent years. As Michael Clemens, Çağlar Özden and Hillel Rapoport Michael A. Clemens, Çağlar Özden, Hillel RapoportMichael A. Clemens, Çağlar Özden, Hillel Rapoportoutline in their introduction to the special issue on migration and development of World Development in 2014[1], part of this renewed interest and increase in published research is due to long overdue[2] improvements in the availability and quality of data. Estimates of international migration and remittances are now published by the World Bank and the UN Population Division compiles census data to give estimates of migrants stocks.

This improvement in data quality and data availability allows us to make tentative statements about the extent of internal and international migration. One of the most serious attempts to estimate internal migration is underway by researchers at the IMAGE project who use census data to estimate that globally in 2005 there were 229 million people living within the same country but in a different part of that country compared to five years before. Estimates for lifetime internal migration are much higher, with 763 million people in 2005 living outside their region of birth.[3]  Combining these with UN estimates of international migrants of 232 million people living outside their country of birth, suggest that nearly a billion people live away from their region of birth. 

Census data is useful for providing insights into how many people are migrants and their demographic profile but is less useful for understanding the why and the how of migration. Understanding why people migrate and for how long, how that contributes to, or even changes, their and their household’s livelihoods and well-being are just some of the questions of great interest to migration researchers and policy makers. These questions can be answered with qualitative research and there are some impressive examples to draw on which provide rich and nuanced stories of migrant lives. Deirdre McKay’s ethnographic work provides new insights into the aspirations and experiences of Filipino temporary labour migrants;[4] Trond Waage uses visual anthropological tools to document the lives of young migrants in west Africa; and Migrating out of Poverty (MOOP) partners at the Asia Research Institute, National University of Singapore, have used qualitative methods to shed light on the recruitment of Indonesian domestic workers.

Complementing this qualitative work is a growing body of evidence emerging from quantitative research using household surveys. These offer the opportunity to include more people in the research sample than qualitative research typically allows, anything from a few hundred to a few thousand people is pretty normal, and to use more detailed questions that capture a wider range of data than is feasible to collect in a population census. There are a growing number of household surveys for developing countries which capture information on migrants. For example, the Mexican Migrant Project collects and publishes data on migration between Mexico and the United States, and a number of household and labour force surveys now contain supplementary modules on migration.[5]  These are encouraging signs that the migration data gap is closing but there is still some way to go.  

The Migrating out of Poverty (MOOP) consortium is contributing towards this by publishing open access micro data from a set of five comparable household surveys collected between 2013 and 2015, in five developing countries: Bangladesh, Ethiopia, Ghana, Indonesia, and Zimbabwe.   The precise sampling strategy differs across countries, and we can’t claim that our samples are nationally representative. However because we adopt a purposive approach, selecting regions which are known to be migrant-sending and sampling quotas of households with and without migrants, we generate large enough sub-samples of households and individuals across different categories of migrants and non-migrants to make us confident that our findings are robust.

Our sample sizes range between 1200 and 1400 households, with data available on every member of those households. We adopt a near identical survey instrument in each country, which facilitates comparisons to be drawn across countries. Our household questionnaire includes a complete household roster collecting social, economic and demographic data on both migrant and non-migrant members of the household, and a specially designed module that captures interactions between migrants and their households in the form of remittances and social contacts. Our survey also explores perceptions of migration as a way of improving the living standards of households.

One of the important contributions the MOOP consortium hopes to make by collecting and publishing this data is to support more research into internal and intra-regional migration. As the figures of migration estimates quoted above suggest, three out of four migrants remain within their country of birth and much international migration is within the global South. Our data will help to shed more light on these movements and help to inform policies that respond appropriately to those affected by migration.

The full data sets from the first three of our surveys undertaken in Bangladesh, Indonesia and Ghana, are now available to download for free from www.migratingoutofpoverty.dfid.gov.uk. Data is available in both STATA and SPSS formats and users can access the questionnaire and a short user guide for each survey. Data for Ethiopia and Zimbabwe will be made available in 2016. 

We want students, researchers and teachers to access the data, and policy makers to use it. Feed back to us and let us know what you do with it.


Bangladesh internal and international migration destination maps
Remittances help fund improvements in housing



[4] McKay D. 2012. Global Filipinos. Indiana University Press.
[5] See Santo Tomas et al (2009) for a useful audit of migration data.



Julie Litchfield is the Theme Leader for Quantitative Research for the Migrating out of Poverty Research Programme Consortium (MOOP) and Senior Lecturer in the Department of Economics at the University of Sussex. Working Papers discussing aspects of the findings of the Bangladesh, Ghana, Indonesia and Zimbabwe surveys are also available. A summary of key data from the Indonesia survey has also been published. See also Eva-Maria Egger's presentation of the preliminary findings of MOOP's household survey conducted in Zimbabwe  and her related blog discussing the context:  Migration in Southern Africa: A Visit to the City of Migrants.