Migrants and their remittances: Patterns and trends from Ethiopia

This blog was written by Yuki Kimura, a Research Assistant at the University of Sussex. 
(Email address: yukikimura96@gmail.com)

Given the increasing prominence of migration especially due to climate change and globalisation, deepening our knowledge of who and why they migrate and remit is necessary since migration may be a key way in which individuals and households would be able to escape poverty.

One of the central discussions in the migration literature is over remittances, which is a crucial mechanism via which migration can lift households out of poverty. Theoretical debates over why migrants remit has explored altruism, exchange or self-interest, or the mixture of both “impure altruism” or “enlightened self-interest”[1], as the possible reasons for why migrants may remit.[2]

Another debate focuses on whether there is a trend in how much migrants remit as their time away increases, or more commonly known as the “remittance decay hypothesis”, that is, the longer a migrant is away from home, the lower the probability of sending remittances and also the lower the amount. In this realm, empirical evidence has been mixed, with some studies finding that time has a negative effect on the amount and probability of remittances, whilst others have found a null effect of time.

There has also been a finding of an inverted U-curve relationship between time and both the amount of and the probability of remittances.[3] The theoretical reasoning for such a relationship is the following: due to set-up costs (such as migrating, finding a job, finding a place to live) and low or uncertain income when migrants have just migrated, the probability of remitting and the amount of remittances are both likely to be low at the beginning of their migration, meaning that as they establish themselves over time, both the probability and the amount of remittances will rise.

However, there is also a counteracting effect as time goes on: over time, migrants’ social connections with the community and their families are likely to diminish or weaken, leading to a decrease in remittances. For example, Liu and Reilly (2006) have found that the peak of remittance sending for internal migrants in Jinan, China is between 25 to 48 months.

Migrating out of Poverty in Ethiopia [4] collected information in 2018 on over 7000 individuals, consisting of non-migrants, internal, and international migrants. Just over 30% of the migrants remitted cash back to the household in the last 12 months. There are significant differences between international and internal migrants over whether they remit, however: 21.3% of internal migrants remitted cash back compared with 63% of international migrants.

Table 1. Descriptive Statistics

International migrants comprised 22.36% of the total migrants, most of whom a. migrated to the Middle East and b. were female, reflecting the recent surge of female Ethiopians migrating to the Middle East to work as domestic workers.[5] As for internal migrants, most (54.5%) migrated within the zone of their household and there is an even distribution of males and females.

Exploring the correlations:  

The correlations between characteristics of migrants and their likelihood of sending remittances home/how much they send were explored by regressing whether they remit or not/how much they remit on their age, months away, whether they are married, their education level, whether they have a child at home, gender, whether they are  an international or an internal migrant, and their relationship with the household head.

Preliminary analysis of the data suggests some interesting patterns. We explored the correlations between various characteristics of migrants and their likelihood of sending remittances home, and how much they send.[6] Firstly, not only does the data support the existence of an inverted-U relationship between time away and both the likelihood of migrants remitting and the amount of remittances sent home to their families, it also suggests that these peak at around 5 years.

For example, Figure 1, which shows the predicted probabilities that migrants remit back home against time, illustrates that although the probability of sending remittances home gradually increases as their months away increases to roughly 40% when their time away reaches around 5 years, the probability of remitting then declines as time away from home increases.

Similarly, from Figure 2, which shows the predicted amount that migrants remit back home on average per month against time, we can see that the predicted remittance amount increases gradually to about 5000 Birrs (roughly US$168.92), again at around 5 years before declining as time away increases.

This is longer than that typically found by the literature (2 to 3 years). This suggest that Ethiopian migrants take longer to settle down at their migration destination than we see in other contexts. Alternatively, it may imply that the amount of time it takes for social ties to erode between an Ethiopian migrant and their family is longer: given that kinship and familial ties play an important role in the Ethiopian community, remittances as informal familial arrangements may be more strongly institutionalised in their community compared with elsewhere.[7]

Breaking down the data by gender, we observe some differences between males and females, as illustrated by Figures 3 and 4. Although men and women have similar probabilities of remitting on average, this pattern over time is different.

Firstly, at any given time, a. women have a higher probability of remitting than men (Figure 3) and b. women are predicted to remit a larger amount back home (Figure 4).

Secondly, both the likelihood of remitting and the amount of remittances sent continues to increase for a longer period of time for women than for men: for women (men), the former continues to increase until 65 (53) months and the latter peaks at 65 (60) months for women (men). This may imply that women take longer to set up and establish themselves to be able to remit back home, or that the familial ties they have with their family are stronger than men.

Lastly, marital status matters for women but not for men: married women decrease their likelihood of remitting by 12.8 percentage points compared to unmarried women and they also, on average, remit nearly 2000 Birrs (US$68) less.

International migrants (which over 70% are female) are 38.5 percentage points more likely to remit on average and remit over 10,000 Birr more compared with internal migrants when controlling for some of their key characteristics. International migrants’ likelihood of remitting back home increases until 58 months into their migration before decreasing, 4 months later than when the likelihood of sending remittances peaks for internal migrants.

This difference possibly highlights the larger set up costs imposed on international migrants compared to internal migrants. Another disparity between the international and internal migrants’ remittance behaviour is that contrary to international migrants, there is no pattern between how long they have been away and the amount of remittances sent back home for internal migrants once we control for some key characteristics. In other words, for internal migrants, although the probability of remitting increases and then gradually falls over time, the amount remitted does not vary according to their time away.

This may reflect their limited ability in sending remittances home: there may not be scope for internal migrants to increase their remittances over time due to minimal income compared with international migrants.

Preliminary analysis using the survey dataset on Ethiopian migrants have revealed interesting patterns and trends in their remittance patterns and in particular that Ethiopia migrants increase their likelihood and the amount of remittances for a longer period than suggested by the literature.

It has revealed that women have a higher probability of remitting throughout their time away and the results have also suggested the possibility that familial ties for female migrants (whether it be due to altruism or self-interest) take longer to break down than for male migrants.

Further, international migrants remit considerably more both in terms of likelihood of remitting and the amount remitted, highlighting the differences between internal and international migrants’ capability of sending remittances home. Further investigation into migrant behaviours and their remittances are required to understand how and when migration can play a key role in aiding individuals and households to break out of poverty.

References

• Adugna, G., 2019. Migration patterns and emigrants’ transnational activities: comparative findings from two migrant origin areas in Ethiopia. Comparative Migration Studies, 7(1), p.5.
• Brown, R.P., 1997. Estimating remittance functions for Pacific Island migrants. World development, 25(4), pp.613-626.
• Brown, R.P., 1998. Do migrants' remittances decline over time? Evidence from Tongans and Western Samoans in Australia. The Contemporary Pacific, pp.107-151.
• Czaika, M. and Spray, J., 2013. Drivers and dynamics of internal and international remittances. The Journal of Development Studies, 49(10), pp.1299-1315.
• De Regt, M., 2007. Ethiopian women in the Middle East: The case of migrant domestic workers in Yemen. Africa Studies Centre, Leiden, 15.
• Dercon, S., De Weerdt, J., Bold, T. and Pankhurst, A., 2006. Group-based funeral insurance in Ethiopia and Tanzania. World development, 34(4), pp.685-703.
• Hunte, C.K., 2004. Workers' remittances, remittance decay and financial deepening in developing countries. The American Economist, 48(2), pp.82-94.
• Johnson, G.E. and Whitelaw, W.E., 1974. Urban-rural income transfers in Kenya: an estimated-remittances function. Economic Development and Cultural Change, 22(3), pp.473-479.
• Krishnan, P., & Sciubba, E. (2009). Links and architecture in village networks. Economic Journal, 119(537), 917-949.
• Liu, Q. and Reilly*, B., 2004. Income transfers of Chinese rural migrants: some empirical evidence from Jinan. Applied Economics, 36(12), pp.1295-1313.
• Lucas, R.E. and Stark, O., 1985. Motivations to remit: Evidence from Botswana. Journal of political Economy, 93(5), pp.901-918.
• Mitra, A., 2004. What motivates transfer of resources: altruism or principle of exchange?. The Journal of Developing Areas, pp.31-48.
• Siegel, M. and Kuschminder, K., 2012. A Who’s Who in Ethiopian Migration?.
• Zewdu, G.A., 2018. Ethiopian female domestic labour migration to the Middle East: patterns, trends, and drivers. African and Black Diaspora: An International Journal, 11(1), pp.6-19.

Endnotes

[1] A term coined by Lucas and Stark (1985). 

[2] Discussion on motivations to remit available in Lucas and Stark (1985) and in Czaika & Spray (2013).

[3] Stark (1978) and Liu and Reilly (2006).

[4] The Ethiopia survey was collected as part of the Income and Remittances research theme, with the Ethiopia team led by Mr Asmelash Haile Tsegay and the Organization for Social Science Research in Eastern and Southern Africa, Addis Ababa. Data for the 2018 survey and an earlier baseline survey in 2014 are available from the Migrating out of Poverty website. 

[5] Further discussion on female Ethiopians migrating to the Middle East as domestic workers can be found in G.A. Zewdu (2018).

[6] Correlations are explored (rather than causation given selection effects) via estimating regressions using ordinary least squares. 

[7] Extended families and kinship networks provide crucial assistance in times of need in Ethiopia, such as providing help during a health crisis discussed by Krishnan and Sciubba (2009) and covering funeral costs by setting up funeral societies as found by Dercon et al (2006).