Bachelor Thesis by Iva Venneman January 2017, Political Science Guided by dr. Sebastian Krapohl
Select, Send, Develop
A comparative case study of the effect of
education level on remittances
1 Abstract:
In this research, the relationship between the education level of a migrant and the amount of money that a migrant remits has been reevaluated. Two competing theories about the nature of this relationship were tested with a comparative case study of four cases. Hereby is controlled for the circumstances in both the migrant-sending countries and the migrant-receiving countries. The results are twofold. It turned out that not all cases suited the chosen research method. This gave interesting insides for follow-up research. The remaining two cases showed a positive relationship between the education level of a migrant and the amount of remittances that are sent to the country of origin.
2
Content
1. Introduction ... 3
2. Literature review ... 4
2.1 Education level and remittances ... 5
2.3 Migration policy of receiving country ... 7
3. Method and case selection ... 9
3.1 Case selection ... 9
3.2 Data and Indicators ... 12
4. Data analysis ... 12
4.1 Individual cases in depth ... 12
4.2 Analysis ... 14
4.3 Alternative explanations ... 15
5. Conclusion ... 17
Appendix ... 19
3
1. Introduction
Since the end of the Second World War, a significant amount of people migrated from the Global South to the Global North. Partly encouraged by decolonization, migrants left underdeveloped countries and came to what is called First World countries. Consequently, the question came up what effect South-North migration has on the development of the migrants country of origin. Do underdeveloped countries in the Global South profit from the fact that people migrate to the Global North or does it hamper their development? Put differently, the large-scale South-North migration induced interest in the relationship between migration and development within the Global South.
Nowadays, the migration-development relation is a topic of interest which has been studied by numerous people. Throughout the years, it has become clear that the perspectives on this relationship can roughly be divided into two groups. On the one hand, you have the negativists, who state the fact that the people who migrate are often the elite of a country; wealthy people who have received an education. When they migrate valuable knowledge leaves the developing country as well. In the literature, this phenomenon is revered to as brain drain. On the other hand, you have the positivists in this debate who argue that the migrant population of a country is more diverse than the elite. Not only higher educated people migrate, but also lower educated people. This reduces the unfavorable brain drain effect. Another argument that positivists give in favor of South-North migration is that migrants can bring valuable knowledge back to the country upon their return. Moreover, migrants can send money back to their families who stayed behind in the country of origin and this can stimulate the development of the country. This transferal of money is called remittances.
The flow of remittances has been growing rapidly in the last decades. Partly, due to the increase of migration to countries in the Global North but also because of the improvement of the world’s financial infrastructure (Kapur 2003: 11). In some cases, remittances make up a higher share of a country’s General Domestic Product than foreign investment does (Nyberg-Sorensen et al. 2002: 14; Gammeltoft 2002: 183). Hence, the effect of remittances has more and more become a topic of interest within both the academic- and the political world. How are remittances being used? Who receives it? And who sends these remittances?
Although all these questions have already received attention, one thing still remains unclear with regard to the last question. That is what effect the education level of a migrant has on the amount of money that is remitted by him or her to the country of origin. Some argue that this is a positive effect. Higher educated migrants earn a higher salary, which causes that the amount of money they remit is higher in comparison to lower educated migrants (Bollard et al. 2011: 11). Others argue that, unlike lower skilled migrants, higher skilled migrants come from wealthy families, are able to bring their closest family with them when they migrate, and they stay abroad for a long time period. These factors would decrease the amount of money
4 that higher skilled migrants remit (Faini 2007: 189 12; Niimi et al. 2008: 15). In short, there are different theories which suggest a different mechanism through which education level effects remittances.
Meanwhile, the reality regarding South-North migration has been changing since the start of the migration-development debate. Almost all countries in the Global North have become more selective in their migrant admission policy throughout the last decade. Primarily in terms of education level. Where it used to be only the Anglo-Saxon countries that selected incoming migrants based on education level, now all OECD countries have a focus on higher skilled migrants (OECD 2008: 103; Belot et al. 2012: 1107). By virtue of that, it is now more important than ever to have a clear view of the relationship between education level and remittances. If this is a negative relationship, it would be fair to question whether the chosen policy direction of the OECD countries contributes to the development of countries in the Global South.
The fact that there is not a clear view on this relationship yet, is in part because there is no combined data available on education level and remittances per person. Despite this, I will make another attempt to research the nature of this relation, since the relevance of this topic is so high. In order to overcome the previously mentioned constraints of lack of data, this research will be conducted with a method that has not been used before in this field; a comparative case study. This method will be of help to answer the main question of this research: What is the relation between the education level of migrants from the Global South who live in the Global North and the amount of money they remit to their country of origin? Before I will do the comparative case study, the existing literature on this topic and the chosen research method will be exemplified. The main questions of this research will be answered in the conclusion.
2. Literature review
Remittances are the international flow of income between a migrant’s country of birth – in this research further mentioned as sending country – and a migrant’s country of destination – further addressed as receiving country (Brown 2006: 55). The existence of a relationship between remittances and the development of the sending country is debatable.
The way that remittances are spent, has been used to argue that the inflow of remittances it is not stimulating development in the migrant sending-country. According to some scholars, remittances are spent on consumption goods instead of on productive investment and this would not increase the development of a country as a whole (Durand and Massey 1992:26). Others, such as Hein de Haas, argue that this is a ‘migration myth’. Remittances are also spent on education, healthcare, housing, and sanitation. Investments like these could improve people’s well-being, freedom, and human capital. (De Haas 2005: 1274).
5 Additionally, remittances can be beneficial in countries with lesser developed financial systems, because local entrepreneurs can use remittances as an alternative way of financing investment in new production activities (Giuliano and Ruiz-Arranz 2008: 145).
Multiple case studies are conducted in order to get a complete view of the way in which remittances influence the migrant sending-country. A comparative case study that looked into migration and remittances in South-America found that remittances on average reduce moderate- and extreme poverty (Fajnzylber and Lopez 2008:8). However, the poverty reduction effect of remittances is not in all regions in the Global South to the same extent as in Latin America. In Latin America, it leads to lesser income inequality, whereas in other regions this is not always the case (IMF 2005: 35). Nevertheless, other case studies that were conducted in Sub-Saharan Africa, Pakistan and Nepal also showed that remittances have a positive effect on extreme poverty (Gupta et al. 2009; Lokshin et al. 2010; Javed et al. 2017). Simultaneously, evidence was found that remittances stimulate households into using formal financial markets, which could stimulate the financial deepening of a country as a whole (Gupta et al. 2009: 111).
It is important to recognize that remittances not automatically lead to more development in a sending country. It highly dependents on the policies and social circumstances in both the sending- and the receiving countries (De Haas 2005: 1275). This is why scholars often advise governments to formulate a policy that stimulates both migration and the sending of remittances (Awan et al. 2015: 794). Nevertheless, it cannot be denied that the inflow of remittances coffers for most migrant-sending countries a higher share of the annual GDP than development aid and in some cases it is even more than foreign direct investment. This is the reason why remittances are more and more recognized to be of assistance to the development of countries in the Global South (Brown 2005:56).
2.1 Education level and remittances
In 2007 Ricardo Faini published an article that approached the migration-development debate differently. Faini argued against the widely accepted assumption that the education level of migrants has a favorable effect on the amount of money they remit. Due to this, a new discussion within the migration-development debate started.
According to Fani’s, higher skilled migrants remit less money than lower skilled migrants do, owing to several differences between the circumstances of lower- and higher skilled migrants. First of all, higher skilled migrants come from wealthier families in comparison to lower skilled migrants. These wealthier families do not need to be financially supported as much. Secondly, higher skilled migrants are able to bring their closest family with them, which also reduces the necessity of sending back money to the migrant-sending country. Finally, Faini points to the fact that higher skilled migrants are expected to stay abroad longer in comparison to lower skilled migrants. This negatively influences the amount of money that
6 is remitted, because the longer migrants are away from the sending country, the weaker the financial ties with the country of origin become (2007: 179). Although Faini obtained a positive relationship between education level and the amount of money migrants remit, the relation wasn’t statistically significant. So, the hypothesis was further tested by others. Research that looked into the relationship between tertiary education and per capita remittances found a significant negative relation (Niimi et al. 2008:15). This proves Faini’s theory that education level negatively influences remittances.
Despite this significant prove, the findings were further questioned due to the used method of cross-country analysis. It has been argued that his method only explains if migrant-sending countries with a higher educated population receive more or less remittances. It does not explain anything about the influence of skill level on the behavior of a migrant. Furthermore, it does not shed light on differences between migrant-sending countries that could influence the amount of money migrant remit. For example differences in poverty (Bollard et al 2011: ). Hence, Bollard, Morten, and Rapoport conducted new research that analyzed collected microdata from interviews of migrants living in different countries. The results show that higher educated migrants remit more than lower educated migrants because higher educated migrants earn a higher salary. When the relationship between income and remittances was tested, there was no significant relation found. Thus, income was appointed as the key mechanism through which higher education affects remittances positively (Bollard et al. 2011: 11).
It is notable that research that uses aggregate data finds a negative relationship between education level and the amount money that migrants remit, while research that uses microdata finds a positive relation between these two variables. This leads to two competing theories about the relationship between education level and remittances, which both point to different mechanisms that determine this relationship. The first theory is the negative relation theory of Faini. The second one is the positive relation theory of Bollard, Morten, and Rapoport. Naturally, this raises the question which one of the two theories is the most accurate description of reality? Below you see a graphic display of the two theories.
Table 1: Negative relation between education level and remittances
Source: Faini 2007 EdEducation levelti l l - Wealth of family - Ability to bring close family - Duration of stay Remittances
7
Table 2: Positive relation between income and remittances
Source: Bollard et al. 2011
The implications of these two theories are completely different. According to the negative relation theory, selecting migrants on education level could cause that less money is remitted which as a consequence could hamper the development of the migrant-sending country. The positive relation theory encourages selection on the education level of migrants because it suggests that more remittances will be sent back, what might positively influences the development of the migrant-sending country. These ambiguous implications make it vital to test these two competing theories again. Accordingly, I will conduct a comparative case study.
2.3 Migration policy of receiving country
The relation between education level and remittances exists within a certain context and so the migration policy of the receiving country also needs to be considered. All individual migrant-receiving countries have different policies in place when it comes to migration. Based on this, migrant-receiving countries can roughly be divided into two groups. On the one hand, there are selective countries which are highly selective on the education level of immigrants. Selective policy is often accompanied by an open policy on family reunification. An open family reunification policy means that it is easy for migrants to bring their family with them to the receiving country. The opposite of a selective policy is being open for every migrant, no matter what education one has received. Countries that are more open in terms of education are often also more restrictive when it comes to duration of stay and family reunification. This is why these countries are called restrictive migration countries (Docquir et al. 2012: 827).
The largest migrant-receiving countries can be divided into groups from restrictive to selective. The Persian Gulf countries are highly restrictive. They favor guest workers and they are highly reluctant to grant permanent status to migrant workers. The Western offshoots, the United States, Canada, and Australia, are traditionally known to be highly skill selective in terms of education but at the same time, they are open for family reunification. European Union countries have always been somewhere in between these two extremes, but they have moved more and more towards the end of the selective spectrum in the last decade (idem:24). Put graphically, it looks the following:
EdEducation levelti l l
Remittances
8
Table 3: Migration policies divided into categories from restrictive to selective
Restrictive Selective
Persian Gulf countries EU countries Western offshoots
Source: Docquir et al. 2012
Docquir, Rapoport, and Salomone conducted a research using bilateral data on remittances and found that selectiveness of a policy in a receiving country influences the amount of money migrants remit (idem:6). They found that the more selective the migrant-receiving country is, the less the effect of education level is on the amount of money that is remitted (2012: 828). It seems likely that this has something to do with the implications that a selective policy migration has, namely that it is accompanied by a generous family reunification policy.
With this in mind, it is interesting to reconsider the negative relation theory. This theory suggests that one of the mechanisms through which education level influences the amount of money migrants remit is the ability to bring your close family with you is. However, it appears that is something that is not dependent on the education level of a migrant but on the migration policy in the receiving country. For this reason, one should control for this variable if one wants to know what the effect of education level on remittances is. Considering this, the relation between migration and remittances is most likely to look like the following:
Table 4: Possible relation between education level and remittances with migration policy as the second variable
Source: Docquier et al. 2012
When looking to table 3, it could be possible that one thinks that an arrow is missing in this chart, namely between migration policy receiving country and education level. If a migration policy of a country is selective or restrictive has in a lot of cases influence on both the ability of a migrant to bring close family and on the education level of the migrants living in this receiving country. But there are also exceptions where countries indeed have a selective migration policy in place together with a generous family reunion policy, but where lower
Education level - Wealth of family - Duration of stay - Income Remittances Migration policy
9 educated migrants can also be welcome. This might sound contradictory, but this is for example the case when a country once received a lot of lower educated guest workers or asylum seekers, and this is followed up by a second stream of lower educated family members. At the same time, restrictive countries also receive higher educated migrants.
To recapitulate, based on the two theories about education level the main question could have two possible outcomes, namely a positive- or a negative relation between the education level of migrants and the amount of money they remit. In case of a positive relation, income is appointed as the key mechanism in which education level effects remittances. In case of a negative relation, the wealth the family of migrants and the duration of stay turn out to be the strongest mechanisms in the education level–remittance relation. Since the theory of Docquier et al. shows that is important to control for differences in migration policy of the receiving country, I will control for this in the comparative case study. Thereafter, I can see which one of the two competing theories on education level still holds up.
This results in two possible hypothesis (H1 and H2) that are formulated the following: H1: The higher educated a migrant from the Global South that lives in the Global North is, the more remittances one will send back to his or her country of origin.
H2: The higher educated a migrant from the Global South that lives in the Global North is, the less remittances one will send back to his or her country of origin.
3. Method and case selection
Multiple bilateral relations will be compared in order to find what effect education level has on the amount of money that migrants send remit to the country of origin. The relations that will be used as cases are the following: the remittances flows from the United Kingdom to Sri Lanka, from France to Cambodia, from Canada to Sri Lanka and from the United States to Cambodia. I will exemplify why a comparative case study is the best way to test the hypothesis and how I came to the selection of the cases.
3.1 Case selection
There is no combined data available on remittances and education level. Where one would normally statistically determine if there is a relationship between two variables, that appears to be impossible in this case. Nevertheless, with a comparative case study it is possible to vary the independent variable as much as possible and see what effect this has on the dependent variable. It is important to do this while controlling for alternative explanatory factors. Otherwise, I can’t be certain if the variance in remittances is due to the variance in education level. This is why this case study has a most similar cases design.
The education level of the migrants who are living in the receiving country is the independent variable in this research. If a migrant is higher educated or not is partially based
10 on the education in the sending country but also influenced by the policy of the receiving country. When one only varies the education level of the sending country, it remains uncertain if the migrant population in the receiving country reflects the overall education level of the sending country. It is possible that only the higher educated migrants go to a certain receiving country. For that reason, I will first select two sending countries that vary as much as possible in terms of education and control for the alternative expletory factors. After that, I select two receiving countries for both sending countries. In that way, I am certain of the largest possible variance in the independent variable.
There is not much theory on alternative explanatory factors in the migrant-sending country that could affect the amount of money that is remitted. Nevertheless, I will discuss a few factors that are possibly of influence on the remittance behavior of migrants. First of all, the existence of a domestic conflict in the sending country seems likely to influence the remittance behavior of migrants. When one’s friends or family still live in a country with a domestic conflict, it seems likely to send back more remittances regardless of all other factors. Therefore, it is preferable to select two cases where a domestic conflict is absent.
Second of all, it is unknown in what way cultural factors influence remittance behavior but is assumable that this could have an effect. In some cultures, it might be more common to financially take care of your parents than it is in others. This is why it is convenient to select two sending countries with a most similar culture. Geographical closeness and religion are two factors that help to control for differences in culture. These factors often go hand in hand. After taken all regions into consideration, the focus was put on finding two South East Asian sending countries. This is because of the lack of large domestic conflicts in this region and the existence of variation in education level between countries in this region.
Sri Lanka and Cambodia were in this context selected as most similar cultures with the most variance in education level. The education level was measured with the United Nations Education Index, which gives a score between 0 and 1 based on the years of schooling and the expected years of schooling a country. Sri Lanka scored 0,738 on this index, while Cambodia scored 0,495 in the latest results of 2013 (UN 2013). Out of 15 South East-Asian countries, Sri Lanka has the highest score after South-Korea and Hong Kong. Cambodia is the 10th country on
this list. Apart from this variance in education level, Cambodia and Sri Lanka have a lot in common. In Cambodia and Sri Lanka Buddhism is the main religion and in both countries, 7 percent of the population lives abroad (IOM 2015). The countries do differ in terms of GDP per capita the in favor of Sri Lanka, but this difference has become less in the last decade (World Bank 2017a).
There are also alternative explanatory factors in the receiving country that influence the amount of money migrants remit. The migration policy of the receiving country is the most important one. More specifically, if you can easily bring your family with you to the receiving country (Docquier et al. 2012: 828). To that end, it was the aim to select receiving countries
11 that have a comparable policy in this regard. Selective countries accept family reunification, while restrictive countries have a less generous family policy. Selective countries are mostly located in the Global North and restrictive countries are often migrant receiving countries in the Global South. The topic of this research is the flow of remittances between the Global North and the Global South. Therefore, it is most convenient that the receiving countries that are selected are selective countries in the Global North.
The independent variable education level of the migrant should vary as much as possible between the cases. The first step in order to achieve this was varying the education level in the migrant-sending country, but this is not enough. It is necessary to be sure about the composition of the migrant community in every case. In order assure this, a more in-depth analysis of the receiving countries is necessary.
Unicef’s Migration Profile of Sri Lanka shows that for Sri Lankans the two most popular destination countries in the Global North are the United Kingdom and Canada. The most common destination countries for refugees from Sri Lanka are France and Canada. For students, the most popular destinations are the United Kingdom and Australia (Unicef 2013a:2). Based on this, one would expect that most of the higher educated Sri Lankan migrants are living in Great Britain, while in Canada a lot of lower educated Sri Lankan migrants live. This image is confirmed by data from the receiving countries. Sri Lankans are in Canada the migrant group that is least likely to have a university degree, was shown by Canadian statistical research (Xiu and Xu 2010: 2). The Sri Lankan community in Britain was however proven to be relatively higher educated and skilled (Siddhisena and White 1999: 532).
The largest receiving country for Cambodian migrants is by far the United States. France is both the largest receiving country for students as for asylum seekers from Cambodia (Unicef 2013b:3). A research that compared drinking behavior of Cambodian woman in the United States and France showed that from the Cambodian women who live in the United States 40 percent has been formally educated in Cambodia. Under Cambodian woman living in France, this is 77 percent (D’Avanzo and Barab: 2017: 326). Also, statistics from the United States show that only 14 percent the Cambodian Americans has a Bachelor’s degree or higher (CAP 2015:1). Based on this data it is likely that the Cambodian population in France is in general higher educated than the Cambodian population in the US.
This research thus contains four bilateral relations as cases. These cases can be summarized and put in chronological order of education level the following:
● Case 1: United States – Cambodia
Lower educated sending country + lower educated community in receiving country ● Case 2: Canada – Sri Lanka
12 ● Case 3: France – Cambodia
Lower educated sending country + higher educated community in receiving country ●Case 4: United Kingdom – Sri Lanka
Higher educated sending country + higher educated community in receiving country
3.2 Data and Indicators
For every case, the average amount of remittances per migrant will be calculated. This is further referred to as remittances per capita (RpC). In order to calculate the remittances per capita, I need two sources of information. First of all, the size of the migrant population in every case (MP). Secondly, the amount of money that is remit from the sending country to the receiving country for every individual case (R). The remittances per capita is the total amount of money that has been remitted per case divided by the total amount of migrants in every case. Written in a formula this looks the following:
𝑅𝑅𝑅𝑅𝑅𝑅 =MP R
R will be based upon data of the International Organization of Migration, the United Nations migration agency (UN 2015). Their database contains information of the migration population from and within every country. In order to establish MP, I will use the Bilateral Remittance Data 2016 of the World Banka. This data contains the sum of all recorded remittances flows between the two countries. The data is based upon all the international bank transactions in the world. Of course one could criticize this latter source because the data doesn’t give inside in the composition of the transactions. Also, it has been questioned if this database is fully complete (Kapur 2003: 3). However, this is the only existing data source on bilateral remittances flows. Therefore, I will nevertheless use this data source in this research.
With the remittances per capita of every case, it is possible to see what effect difference in education level has on the amount of money that migrants remit. Most interesting to analyze is what the difference is between each individual case and between the groups of cases: Sri Lankan migrants compared to Cambodian migrants and higher educated community in receiving country compared to the lower educated community in receiving country.
4. Data analysis
4.1 Individual cases in depth
It is useful to first fully understand the individual cases before analyzing the remittances per capita. Although the cases have been selected on similarity, it is possible that there are still differences between the cases that can influence the remittance behavior of the migrants. It is
13 important to be aware of these differences so that they can be taken into account while
analyzing the date. In the table below you see some important features about the migrant population in every case.
Table 5: Features of migrant population in cases
US-CAM CAN-SRI FRA-CAM UK-SRI
Migrant group in total (MP) Migrant group of total migrants in receiving country (in percentage) Migrant group of total migrants from sending country (in percentage) 163,039 0,3 13,7 146,998 1,9 9,0 63,343 0,8 5,3 138,752 1,6 8,5 Source: UN 2015
The information in table 5 shows that the former colonial powers, Great Britain and France, are not the largest receiving countries for migrants of their former colonies. Furthermore, Sri Lankan migrants seem to be more rooted in their two top destination countries in comparison to Cambodian migrants. This observation is based on the fact that Sri Lankan migrants make up a larger part of the total migrant population in the receiving country than Cambodian migrants do.
The spread of migrants across the country can also tell us something about the rootedness of the migrant population. The more spread migrants are across the country, the harder it might be to establish a strong migrant community in the receiving country. In Case 1 the Cambodian are spread across the United States (CAP 2015:1). In Case 2 it is clear that Sri Lankan’s in Canada are mostly living in the Toronto area. This is also the largest Sri Lankan community outside of Sri Lanka (Singam 2014). Cambodian’s in France are mostly living in Paris, but there are also concentrations in other regions in France (L’Echo 2014:6). In Case 4 82 percent of the Sri Lankans in the United Kingdom are living in London, so this migrant population highly concentrated (Siddhisena and White 1999: 521).
All in all, the Cambodian migrant population tends to be more spread across the receiving country than the Sri Lankan community. In combination with the fact that they make up a lower share of the migrant population in the receiving countries, one can conclude that the migrant communities in Case 1 and Case 3 are less rooted than the communities in Case 2 and Case 4.
The economic situation in both the receiving country and the sending country could also be of influence on the remittance behavior of the migrant. It is therefore that additional information on this is presented in the following table.
14
Table 6: Economic situations of the countries involved
US-CAM CAN-SRI FRA-CAM UK-SRI
GDP per capita sending country (in Us dollars in 2016) GDP per capita receiving country (in US dollars in 2016) 1,269.91 57,466.76 3,835.39 42,157,93 1,269.91 36,854.97 3835,39 39,899.39
Source: World Bank (2016a)
First of all, General Domestic Product (GDP) per capita is an instrument to measure the average wealth of a country (OECD 2017). Sending country Cambodia has a lower GDP per capita than sending country Sri Lanka, which is an indicator that the average income per person in Sri Lanka is higher than in Cambodia. As for the receiving countries, the two European countries where the higher skilled migrants live, have a lower GDP per capita than the North-American receiving counties. Based on this information it is not possible to say what the migrant populations earn, but the setting in which the migrants and their families live has now been set. To sum up, Cambodian migrants are less rooted in the receiving countries than Sri Lankan migrants, Cambodia has a lower GDP per capita than Sri Lanka and the GDP in the North-American receiving countries is higher than in the European receiving countries.
4.2 Analysis
Table 7 shows the total remittances per case and the remittances per capita per individual case. Based upon this the education level – remittances relationship can be analyzed.
Table 7: Remittances per capita
USA-CAM CAN-SRI FRA-CAM UK-SRI
Total remittances flow (R) (in million US dollars, 2016) Remittances per capita (RpC)1
(in US dollars per year, in 2016) 77 472.27 493 3,353.79 25 394,67 530 3,819.76
Source: World Bank (2016b)
The remittances per capita show a remarkable result. Multiple, contradictory conclusion can be drawn. First of all, the remittance flows to Sri Lanka is in general much bigger than the remittance flows to Cambodia. Both in Canada as in the United Kingdom, the remittances per capita is much higher than in France and the United States. Another finding is that the differences between the higher- and lower educated population from the same sending country show a different pattern. When the education level of Cambodians increases, they tend to send back less remittances back to their country of origin. When the lower educated Sri
15 Lankan in Canada are compared with the higher educated community in the United Kingdom, one sees a reverse, positive trend. As a consequence, the order of education level among the cases does not give a clear pattern of increase or decline of remittances per capita. Therefore, it is not possible to say if education level positively or negatively influences the amount of money that migrants remit.
However, when you compare the difference in remittances per capita between Case 1 and Case 3 for Cambodian migrants and Case 2 and Case 4 for Sri Lankan migrants you see that the difference between education level is only substantial among Sri Lankan migrants. Sri Lankan migrants in the United Kingdom sand back 466 dollars per year more on average than Sri Lankans in Canada. The difference among Cambodians is with 80 dollars negligible. This raises the questions if there are differences between the cases of Cambodia and Sri Lanka that might have been overlooked.
4.3 Alternative explanations
The difference between Cambodian- and Sri Lankan migrants are interesting. How is it possible that two sending countries that were selected based on similarities show such a big difference in remittance behavior? In order to find out, I will look into alternative explanations.
Before was found that Cambodian migrants are less rooted in the two receiving countries than Sri Lankan migrants are. Could it be possible that rootedness influences the amount of remittances migrants send back? This has never been questioned in the existing literature on remittances. Besides, it seems unlikely that large immigration countries like the United States and France have an undeveloped remittance transferal system for certain countries. Hence, I will look into other aspects of the analysis in order to find a reason for the difference.
The composition of the data sources that were used to calculate the remittance per capita might have caused the differences in outcome. As previously discussed, the data of the World Bank has been criticized. The data of the World Bank consists of all national transactions over the world. It could be that Cambodian migrants don’t send their remittances back via bank transactions, while Sri Lankan migrants do. This could explain the big difference in remittances per capita. If this is the case, the reason for it could be found in a difference in trustworthy, financial infrastructure between the migrant-sending countries. Freund and Spatafora argued that migrants from migrant-sending countries with a weak developed financial system tend to send back their remittances through informal remittance channels instead of through formal channels (2005: 21).
Formal channels are all money transfers that involve formal contracts. For example, money transferals via banks or post office banks. We speak of an informal channel when a migrant sends back money through a personal relationship with a businessman, a friend or family member or via unregistered intermediaries who deliver the money upon return to the
16 sending country (Kosse and Vermeulen 2014: 214). The data of the World Bank that has been used to calculate the remittances per capita only contains data on formal remittances. There is no available data on informal remittance transfers because it is a form of unregistered money transfer. In other words, the previous analysis of remittances per capita missed out on the remittances that were sent through informal channels and this possibly explains the difference between the cases that involved Cambodia and Sri Lanka.
Informal remittances flows were estimated to make up 35 to 75 percent of the recorded remittances flows. This was estimated based on survey data in multiple migrant-sending countries (Freund and Spatafora 2005: 3). The choice of a migrant to send remittances back through formal- or informal channels dependents on a few factors. First, the functioning of the financial system in both the sending and the receiving country. Second, on the transaction cost and the amount of money that is being sent back. In 2005 transaction cost for formal transactions made on average up to 10 to 20 percent of the total amount sent, while transactions through informal channels had transaction cost between 2 and 5 percent. At the same time transactions through informal channels come with a higher risk than transactions to formal channels. It is therefore likely that dependent on the amount of the remittance a migrant wants to send back, a decision is being made whether to use a formal or an informal channel (idem: 8).
In the last years, a lot of countries across the world made an effort to reduce the high transaction costs for formal remittances. This paid off because in 2017 the average transaction costs were only 7,09 percent of the amount send (The World Bank 2017b: 2). Data of the World Bank gives inside in the transaction costs between some countries. This data is available in 2 of the 4 cases of this research. In case 2 sending 200 US dollars from Canada to Sri Lanka is involved with an average of 6.41 dollar transaction fee, so 3.2 percent of the amount send (The World Bank Data 2017c). In case 4 sending 200 US dollar remittances from the United Kingdom to Sri Lanka one has to pay 5.04 US dollars fee, which is 2.5 percent of the total amount. Unfortunately, there is no data available on the transaction cost of case 1 and case 3. Nonetheless, it is known what the transaction cost from Thailand to Cambodia are and this is 15.06 US dollar when sending 200 US dollar remittances. That is 7.5 percent of the total amount send2. Although this doesn’t tell anything about the cases, it does give an indication
that sending remittances through formal remittances channels to Cambodia is more expensive than to Sri Lanka.
When one compares Cambodia and Sri Lanka on their financial circumstances it becomes clear that these countries differ substantially. In Sri Lanka, 82.7% of the people above 15 have a bank account, while this is only 12.6% in Cambodia (The World Bank 2014). This information makes it assumable that sending country Cambodia has a weak functioning
17 financial infrastructure. This combined with the higher transaction cost in comparison to Sri Lanka makes it likely that Cambodian migrants don’t send their remittances through formal channels but through informal channels. Therefore, the method that was chosen in this research doesn’t suit the Cambodian cases. It only contains information on formal remittance channels, but these are not used in the Cambodian cases. In order to research the difference in remittance behavior of Cambodians information about the informal remittance flows is needed. A way to collect this is through surveys. Therefore, it might have been more useful to compare survey data of Cambodian migrants in France and Cambodian migrants in the United States. Unfortunately, this data is unavailable.
On the other hand, Sri Lanka has a functioning financial system and low transaction cost. As a result of this, Sri Lankan migrants do use formal remittance channels. Both in Case 2 as in Case 4 a substantial amount of remittance per capita was found. Also, the transaction costs in both cases are below average. When one makes a comparison between the lower educated Sri Lankan migrants in Canada and the higher educated migrants in the United Kingdom, it becomes clear higher educated Sri Lankan migrants remit a substantially higher amount of money per year in comparison to lower educated Sri Lankan migrants. Thereupon, it can be argued that education level has a positive influence on the remittances that are being sent through formal channels. In order to draw a conclusion on the remittance behavior of migrants based upon both formal and informal remittances channels, it is necessary to have more information.
5. Conclusion
In this research the following question has been evaluated: What is the relation between the education level of migrants from the Global South who live in the Global North and the amount of money they remit to their country of origin? Due to the literature study and the comparative case study, it is now possible to draw several conclusions. First of all, the amount of money that is sent is likely to be influenced by the absence or presence of a migrant’s closest family in the migrant-receiving country. Faini’s theory suggests that if a migrant can bring his or her family along to the migrant-receiving country is influenced by education level, but this is not true. Whether or not a migrant is able to bring his or her family along to the new destination country is not always given as a choice. This is something that is dependent on the migration policy of the receiving country. Therefore, I chose to control for the migration policy in the receiving country in this research. The cases that were selected, had receiving countries with a similar policy in this regard.
The comparative case study resulted in two findings. Firstly, it became clear that it is important to be aware of the financial infrastructure of countries when measuring remittance behavior. While selecting the cases there has not been controlled for the financial structure in
18 the sending countries and it might also have been impossible to control for it within this research design. However, it appeared that one of the selected sending countries had a completely different financial development and that this influences the way in which remittances are sent back to the country of origin. As a consequence, the field of informal remittances had to be introduced into this research. This is a field that is understudied for a reason because informal remittances are hard to trace.
It became clear that bilateral data could be a useful measurement instrument to research remittance behavior to migrant-sending countries with a well-developed financial system. However, for migrant-sending countries with a lesser developed financial system, this is not a good indicator because migrants do not use formal transaction channels. Survey data in the receiving- or sending country could give more clarity on the remittance behavior of migrants from these countries. Unfortunately, this data doesn’t exist yet for the selected cases in this research. Due to time- and resources constraints it has not been possible to collect this additional data and include it in this research.
The second finding of the comparative case study is based on the results of the two cases that involved remittances to Sri Lanka. Luckily, the case study was designed in such a way that it remains possible to say something about the relationship between education level and remittances. Sri Lanka is a country in the Global South with a relatively high average education level. In addition, this country has a well-developed financial system. Consequently, Sri Lankan migrants use formal remittance channels to transfer money to their country of origin. The results showed that from the United Kingdom, a country home to the higher educated share of Sri Lankan migrants, per year on average 466 euros more is remitted to Sri Lanka than from Canada, where mostly lower educated Sri Lankan migrants live.
Based on this information can be concluded that when formal remittance routes are used, the higher educated the migrant from the Global South is, the more remittances he or she sends back to the country of origin. Therefore, it is likely that the theory of Bollard, Morten, and Rapaport, that suggests that income is the key mechanism through which education level influences remittances, is the most accurate description of reality. In addition, this research does not give reason to suggests that the chosen policy direction of the OECD countries to select for education level, hampers the development in the Global South.
To conclude, this research showed once more that the behavior of migrants in regard to remitting money is influenced by several factors. One should be aware of all these factors while looking into a certain case. In this research, a light was shed on the following factors that could be of influence on the remittance behavior of a migrant: the education level of a migrant, a migrants income, the selectivity of the migrant destination country and the financial development of the migrant-sending country. When the latter factor is absent, informal remittance networks are more likely to be used than formal remittance networks. In order to get a complete picture of the remittance behavior of migrants, it is necessary to research the
19 informal remittances flow to countries with an underdeveloped financial system. With the findings of this research in mind, it can thus be recommended to collect survey data specifically on migrants from countries with an underdeveloped financial system.
Appendix
1. Remittances per capita
The Remittances per capita are calculated the following. Case 1: 77,000,000163,039 = 472.27
Case 2: 493,000,000146,998 = 3,353.79
Case 3: 25,000,00063,343 = 394,67
Case 4: 530,000,000138,752 = 3,819.76
2. Percentage transaction costs
Transaction costs of countries are calculated the following: A. Transaction costs Canada – Sri Lanka: 6.41 ×100200 = 3.2
B. Transaction costs United Kingdom – Sri Lanka: 5.04 ×100200 = 2.5 C. Transaction costs Thailand – Cambodia: 15.06 ×100200 = 7.5
References
Awan, M.S., Javed, M. and Waqas, M. (2015). ‘’Migration, Remittances, and Household Welfare: Evidence from Pakistan’’, The Lahore Journal of Economics, 20 (1): 47-69.
Belot, M. and Hattor, T. (2012) ‘’Immigrant Selection in the OECD’’, Journal of Economics, 114 (4): 1105-1128.
Bollard, A., Morten, M. and Rapoport, H. (2011). ‘’Remittances and the brain drain revisited: The Microdata Show That More Educated Migrants Remit More’’, The World Bank Economic Review, 25 (1): 132-156.
Brown, S.S. (2006). ‘’Can Remittances Spur Development? A Critical Survey’’, International Studies Review, 8 (1): 55-75.
CAP (Center for American Progress) (2015). Who are Cambodian Americans?, Washington D.C. : Center for American Progress.
20 D’Avanzo, C.E. and Barab, S.A. (2000). ‘’Drinking during pregnancy: practices of Cambodian
refugees in France and the United States’’, Health Care for Woman International, 21(4): 319-334.
De Haas, H. (2005). Migration, Remittances and Development: Myths and Facts, Thirds World Quarterly, 26 (8): 1269-1284.
Durand, J. and Massey, D.S. (1992). ‘’Mexican Migration to the United States. A Critical Review’’, Latin American Research Review, 27 (2): 3-42.
Docquier, F., Rapoport, H. and Salomone, S. (2012). ‘’Remittances, migrants’ education and immigration policy: Theory and evidence from bilateral data’’, Regional Science and Urban Economics, 42 (5): 761 -904.
Faini, R. (2007). ‘’Remittances and the Brain Drain: Do More Skilled Migrants Remit More?’’ , The World Bank Economic Review, 21 (2):177-191.
Fajnzylber, P. and López, J.H. (2008). Remittances and Development: Lessons from Latin America. Washington D.C: The World Bank.
Freund, C. and Spatafora, N. (2005). ‘’Remittances: Transaction Costs, Determinants and Informal Flows’’, World Bank Policy Research: Working Paper 3704, Error! Hyperlink
reference not valid..worldbank.org/bitstream/handle/10986/8293/wps3704.pdf?sequence.
Gammeltoft, P. (2002). ‘’Remittances and Other Financial Flows to Developing Countries’’, International Migration, 40 (5): 181-211.
Giuliano, P. and Ruiz-Arranz, M. (2008). ‘’Remittances, financial development, and growth’’, Journal of Development Economics, 90 (1): 144-152.
Gupta, S., Pattillo, C.A. and Wagh, S. (2009). ‘’Effect of Remittances on Poverty and Financial Development in Sub Saharan Africa’’, World Development, 37(1): 104-115.
IMF (International Monetary Fund )(2005). Economic Prospects and Policy Issues. Washington D.C.: International Monetary Fund.
IOM (International Organisation for Migration) (2013). ‘’Global Migration Flows’’, Error!
Hyperlink reference not valid., Consulted on 28 December 2017.
Javed, M., Awan, M.S. and Waqas, M. (2017). ‘’International Migration, Remittances Inflow and Household Welfare: An Intra Village Comparison from Pakistan’’, Social Indicators Research, 130(1):779-797.
Kapur, D. (2003). ‘’Remittances: the new development mantra?’’, Harvard University and Center for Global Development: 1-38, https://www.g24.org/wp-content/uploads/2016/ 01/REMITTANCES-THE-NEW-DEVELOPMENT-MANTRA.pdf.
Kosse, A. and Vermeulen, R. (2014). ‘’Migrants’ choice of remittance channel. Do General payment Habits play a role?’’, World Development, 62: 213-217.
21 L’Echo du Cambodge (2014). ‘’Cambodians Living in France,’’ Error! Hyperlink reference not
valid.france-page-06.pdf, Consulted on 28 December.
Lokshin, M., Osmolovski, M.B. and Glingskaya. E. (2010). Work related migration and poverty reduction in Nepal, Review of Development Economics, 14(2):323-332.
Mansuri, G. (2006). ‘’Migration, school attainment, and child labor: evidence from rural Pakistan’’, World Bank Policy Research: Working Paper No. 3945, https://openknow ledge.worldbank.org/bitstream/handle/10986/8422/wps3945.pdf?sequence=1. Niimi, Y., Ozden, C. and Schiff, M. (2008). ‚‘Remittances and the Brain Drain: Skilled Migrants
Do Remit Less!’’, ABD Economics Working Paper Series, 126(1): 1-21.
Nyberg-Sorensen, N. and Van Hear, N. and Engberg-Pedersen, P. (2002). ‘’The Migration-Development Nexus Evidence and Policy Options State-of-the-Art Overview’’, International Migration, 40(5): 4-47.
OECD (2008). International Migration Outlook 2008. Paris: OECD Publishing.
OECD (2017). ‘’Better Life Index’’, http://stats.oecd.org/Index.aspx?DataSetCode=BLI, Consulted on 28 December 2017.
Siddhisena, K.A.P. and White, P. (1999). ‘’The Sri Lankan Population of Great Britain: Migration and Settlement’’, Asian and Pacific Migration Journal, 8(4): 511-536.
Singam, A.A. (2014). ‘’The Tamil Community in Canada: A Brief Overview’’ of Tamil Culture, https://tamilculture.com/the-tamil-community-in-canada-a-brief-overview/, Consulted on 16 January 2018.
Unicef (2013a). Migration Profiles Sri Lanka, New York City: UNICEF. Unicef (2013b). Migration Profiles Cambodia, New York City: UNICEF.
UN (United Nations) (2013). ‘’Education Index, Human Development Reports’’,
http://hdr.undp.org/en/content/education-index, Consulted on 29 December 2017. UN (United Nations, Department of Economic and Social Affairs) (2015). ‘’Trends in
International Migrant Stock: Migrants by Destination and Origin’’, Error! Hyperlink
reference not valid., Consulted on 28 January 2018.
World Bank (2014). ‘’Global Financial Institutions, Worldbank Data’’, http://databank.world bank.org /data/reports.aspx?source=1228, Consulted on 2 January 2018.
World Bank (2017a). ‘’GDP per Capita’’, https://data.worldbank.org/indicator/ NY.GDP.PCAP. CD, Consulted on 29 December 2017.
22 World Bank (2017c) ‘’Country corridors’’, https://remittanceprices.worldbank.org/en/
countrycorridors, Consulted on 2 January 2017.
Xue, L. and Xu, L. (2010). ‘’An educational portrait of postsecondary educated immigrants, 2006 Census’’ of Citizenship and Immigration Canada, https://www.canada.ca/content /dam/ ircc/migration/ircc/english/pdf/research-stats/education-summary.pdf.
