Altruism based on Religious Ties as a Motive to Remit?
Author: K. Popal1 Supervisor: Dr. J. Bolt MSc. Thesis University of Groningen June 2017 AbstractWith evidence that a significant share of migrants are religious and altruistic behavior is central in many religious teachings, this paper examines the question whether altruism based on religious ties is a determinant of remittances. The research question is examined through a fixed-effects panel estimation on 123 remittances’ receiving countries over the period 2000-2010. Overall, this paper finds a positive (but not significant) relationship between the share of religious people and remittance inflows in the home country. More interestingly, the results are positive and significant in the low- and lower-middle income group countries. The results suggest that the share of religious people in Eastern-Asia and Pacific and Sub-Saharan Africa countries significantly corresponds with higher remittance inflows. On the other hand, in the upper-middle income group and Europe-Central Asian countries, this paper finds an inverse relationship.
Keywords: remittances, religion, altruism, remittance decay hypothesis JEL code: J10, J15, Z12
1University of Groningen, Faculty of Economics and Business, P.O. Box 800, 9700 AV Groningen, The
2
1. Introduction
In 2015, worldwide remittance flows through formal channels have exceeded $601 billion; and developing countries received $441 billion. In some 25 developing countries remittance inflows constitute more than 10 percent of GDP.2
Consequently, policymakers and researchers have become increasingly interested in the motives and factors that affect the willingness of migrants to remit (Adams, 2009; de Haas, 2007; Agarwal and Horowitz, 2002; Hagen‐Zanker and Siegel, 2007; and Rapoport and Docquier, 2006). Remittances are the most tangible evidence that migrants do not break ties with their families and countries of origin even as they migrate (Guarnizo, 2003). Indeed, the current literature on the determinants of remittances has found consistent evidence supporting the notion that ‘’altruism based on family ties’’ motivate much of the remitting that occurs (Rapoport and Docquier, 2006; and Carling, 2008). Stronger family ties and kinship between the migrant and its family at home should raise the degree of altruism (Rachlin and Jones, 2008; and Hamilton, 1964). Consequently, a higher degree of altruism should increase the volume of remittances migrants send to family members.
However, the remittance decay hypothesis (RDH) suggests that the level of remittances sent by migrants will decline over time as migrants’ commitment and links to their relatives and home country diminishes (Stark, 1978). Stark (1978) attributes the decline in remittance behavior through time to the weakening of altruism. The decrease in remittance behavior intensifies, even more, when generational changes occur, making receivers and senders more distant relatives (Erdal, 2012). Even with continued migration, the decline in remittance rates is still likely to occur due to family reunification and better integration of the migrants in the host countries (Brown, 1998). In fact, it is family migration that accounts for most of the total migration flows in OECD countries (OECD, 2013)3.
However, why do migrants nevertheless continue sending remittances to their countries of origin? Also, if it is true that altruism weakens through time, what causes the commitment to remit to be sustained?
2 (KNOMAD & World Bank Group, 2016)
3 Global Trends in Family Migration in the OECD: Adapting to Changes in Family Migration: the Experience of OECD
3 This paper addresses the two questions by extending the boundaries of this debate through the notion of ‘’altruism based on religious ties’’ as an alternative to the current household-level approach. More specifically, this paper suggests that as time passes, remittance practices change and the significance of family ties is nuanced by other relationship dimensions. In the context of family reunification and the weakening of social- and family ties over time, altruism based on religious ties could provide alternative explanations of what motivates remittances, and why they decline - or continue through time.
The first part of the introduced concept deals with ‘’(religious) altruism’’ as a motive for migrants to remit. A significant share of migrants worldwide is religious (see Figure 1). Nearly half of the world’s migrants in 2010 are Christian, and more than a quarter are Muslim. In the case of remittances, the willingness to send money back home may have religious underpinnings. This is because many religious teachings encourage people to be more generous with their possessions towards those that are not equally endowed (Eckel and Grossman, 2004). For example, the level of remittances from migrants from Hong Kong and Malaysia to Bangladesh increases substantially during the Muslim religious festivals such as Eid-ul-Fitr and Eid-ul-Azha (Ahsan Ullah, 2010). More generally, a Gallup Poll survey conducted in 140 countries worldwide between 2006 and 2008 show that those who identify themselves as highly religious are more likely to engage in ‘’helping behaviors’’.4 The second part of the introduced concept deals with
‘’(religious) ties’’ as a means to sustain the commitment to send remittances over time. Kaplan and Gurven (2005) argue that social connectedness than kinship per se motivates altruistic behavior among humans. In mainly developing countries, common religious customs seem to give people access to social networks and opportunities to forge long-lasting relationships.5 In turn, these relationships and
social ties offer a safety net in times of crisis. Likewise, a Gallup Poll survey concerning the importance of religion finds that religiosity is of particular
4 Brett Pelham and Steve Crabtree, ‘’Worldwide, Highly Religious More Likely to Help Others’’, Gallup World Poll,
October 8, 2008, http://www.gallup.com/poll/111013/Worldwide-Highly-Religious-More-Likely-Help-Others
5 Steve Crabtree and Brett Pelham, ‘’Religion Provides Emotional Boost to World’s Poor’’, Gallup World Poll, March 6,
4 significance in the poorest countries.6 More specifically, the respondents in
countries that said religion was "important in [their] daily life’’, are typically located in the Sub-Saharan Africa and Eastern-Asia and Pacific regions (see Figure 2).
Figure 1: Religious affiliations of migrants worldwide (%) in 2010
Source: own calculation using data from the Pew Research Centre database7
Figure 2: Results of a 2008/2009 Gallup poll on whether respondents said that religion was "important in [their] daily life."
Source: GALLUP WorldView (2009)8
6Steve Crabtree, ‘’Religiosity Highest in World’s Poorest Nations’’, Gallup World Poll, August 31, 2010,
http://www.gallup.com/poll/142727/religiosity-highest-world-poorest-nations.aspx
7Pew Research Centre, The Global Religious Landscape. Available at
http://www.pewforum.org/files/2012/03/global-fact-sheet.pdf
8By Kamalthebest - CC BY-SA 4.0, https://commons.wikimedia.org/w/index.php?curid=53704107
5 This paper contributes to the development of the current economic theory by analyzing the question whether altruism based on religious ties could also be a determining factor for remittance inflows. Therefore, the hypothesis to be tested is that a higher share of religious people in the country of origin should lead to higher remittance inflows. The assumption is that migrants remain socially connected through a common spiritual bond and continue sending remittances, even if the emotional attachments with the family decays over time.
6 the home country and remittance inflows. Likewise, this holds for the relationship between the real interest rate and remittance inflows. However, as expected, the number of migrants- and income in the host countries do positively correlate to remittance inflows.
The rest of this paper is organized into five main sections. Section 2 discusses the current literature regarding the determinants of remittances. Section 3 sets out the conceptual framework of the model and discusses the data. Section 4 presents the empirical results, and implications of this paper and Section 5 summarizes the main results and concludes.
2. Theoretical Background and Literature Review
2.1 Altruism based on Family Ties
7 families and communities in response to their growing needs (Ratha, 2013). Therefore, based on Hamilton’s kin selection theory, strong family ties/kinship between migrants and remaining households would enhance the probability of migrants remitting for altruistic reasons (Hamilton, 1964). In other words, the total amount of remittances sent should increase with the migrant’s income (the migrant’s capacity to remit) and the strength of altruism (the strength of family ties) (Funkhouser, 1995). Intuitively, this would only be the case if the migrant has close family members left in the home country and if he/or she remains close ties with those remaining family members.
However, the ‘’remittance decay hypothesis’’ (RDH) suggests that the emotional attachment with the household at home vanishes over time, and social links in the countries of origin become weaker (Stark, 1978). This so-called ‘’remittance decay’’ would be more so if, in the meantime, migrants become permanent residents in the host countries (Merkle and Zimmermann, 1992). Migrants permanently abroad are under less pressure to remit as their social- and family commitments become less intense and less significant (Shankman, 1976). For example, findings from the survey in Kharian show that kinship links are relevant for remittance sending, particularly in the early stages of migration when remittances are sent back to members of one’s household (Erdal, 2012). However, over time, total remittance flows will decline unless the rate of new migration is sufficient enough to offset the declining remittance levels among the stock of ‘’first generation’’ migrants in the host countries (Brown, 1998). In fact, the global increase in the total migrant stock is not due to new immigrants, but rather, due to family reunification in the host countries (OECD, 2013).9 Migrants are often
followed by others from the same (direct) household over time and hence, have even less family pressure to remit. Carling (2008) has found that married migrants who are accompanied by their partners in the host country due to family reunification have less pressure to remit compared to those who have a spouse left in the country of origin. In other words, as long as important direct family members in the home country are in need of remittances, migrants will continue
9 Global Trends in Family Migration in the OECD: Adapting to Changes in Family Migration: the Experience of OECD
8 to send. Hence, on the one hand, the remittance decay hypothesis (RDH) suggests that the pressure to remit should decrease over time as social- and family ties diminish. On the other hand, family reunification is the single largest component of the total immigration flows and therefore should reduce the pressure to remit as well. Evidently, this challenges the mainstream assumption that altruism based on family ties is the primary determinant of the significant amount of remittances we see today.
2.2 Altruism based on Religious Ties
This paper makes the case that it is not the degree of family ties that determines altruistic behavior. It is rather altruism itself, inherent in the migrant’s and remaining household’s shared belief systems such as religion, that determines both altruistic behavior and the degree of social- and family ties. In general terms, religiosity positively affects altruistic behavior among individuals.10 Research demonstrates that religiously affiliated individuals donate
more money to charitable causes than secularists (Eckel and Grossman, 2004). Cadge and Ecklund (2007) argue that highly religious people (those who participate in religious organizations) are often concerned with improving some version of the common good. Moreover, religious identity becomes more central in the daily lives of migrants than they did before in the country of origin (Warner and Wittner, 1998). Migrants need religion to help them preserve spiritual identity and connection to a social network (Williams, 1988). This is important as there is evidence that religiosity plays a particularly significant role in the world's poorest countries.11 Those poor and developing countries are usually the
migrant-sending countries, and in turn, the main recipients of remittance flows. Therefore, it is reasonable to assume that migrants remain socially connected with their family members at home through the notion of a shared religion. In turn, this strong social bond remains an important determinant for sending remittances, potentially even offsetting the adverse effects of weakening emotional- and family
10Brett Pelham and Steve Crabtree, ‘’Worldwide, Highly Religious More Likely to Help Others’’, Gallup World Poll,
October 8, 2008, http://www.gallup.com/poll/111013/Worldwide-Highly-Religious-More-Likely-Help-Others
11 Steve Crabtree, ‘’Religiosity Highest in World’s Poorest Nations’’, Gallup World Poll, August 31, 2010,
9 attachments over time. For example, survey data in Pakistan show that often direct relatives send remittances, but when no close relatives are abroad, remittances may still come from more distant relatives or even people who are not kin (Erdal, 2012). So, although kinship links are an essential prerequisite for receiving remittances, in most cases, other factors seem to be of importance as well. A husband might be altruistic and send as much as possible to his wife and children back home, while a distant relative might feel it is a moral duty to remit occasional amounts due to norms and values inherent in its religious and cultural characteristics (Sana and Massey, 2005). Finally, migrant’s education level, marital status, gender and the number of household members have all been found to be important determinants of remittances by some studies (Lucas and Stark, 1985; and Durand et al., 1996). However, religious affiliation also matters because it has an impact on the perceived costs and benefits of various interrelated decisions that people make over the life cycle such as choosing the marital partner, marriage, and divorce (Lehrer, 2004). In other words, religion drives those important determinants of remittances in the first place, and those determinants, in turn, affect the amount of remittances sent. Hence, it is intuitively more rational to start with religion as the main unit of analysis instead of the current approach based on family ties.
3. Methodology and Data
3.1 Empirical Framework
The goal of the empirical analysis in this paper is to examine whether or not altruism based on religious ties is an important determinant for remittances. The empirical framework is based on country-level data and builds on a simple model, which has also been applied by others in examining the common determinants of remittances (Freund and Spatafora, 2005; and Singh et al., 2009).12 However, the innovation in this study is the religiosity term that captures
the possible altruistic motive based on a common religion as a determinant for
10 remittances. Therefore, based on the empirical strategy by Singh et al. (2009), the following double-logarithmic equation is used to answer this paper’s research question (see Eq.1):
log REMit= β0+ β1log FINDEVit+ β2log INCHOMEit−1+ β3log INCHOSTit+ β4log EXPATRit+
β5log INTit+ β6log RELit+ εit (i = 1,…,N) (t=2000,…..2010) (Eq.1)
where 𝑅𝑅𝑅𝑅𝑅𝑅𝑖𝑖𝑖𝑖 denotes remittances in current US$; 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑅𝑅𝐹𝐹𝑖𝑖𝑖𝑖 is the proxy for
financial development in the home country measured as the ratio of domestic credit to GDP; 𝐹𝐹𝐹𝐹𝐼𝐼𝐼𝐼𝐼𝐼𝑅𝑅𝑅𝑅𝑖𝑖𝑖𝑖−1 is the first lag of real GDP per capita in the home
country13; 𝐹𝐹𝐹𝐹𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑇𝑇𝑖𝑖𝑖𝑖 is the average real GDP per capita in the 17 OECD
destination countries and 𝑅𝑅𝐸𝐸𝐸𝐸𝐸𝐸𝑇𝑇𝑅𝑅𝑖𝑖𝑖𝑖 is the international migrant stock in the 17
expatriates-receiving OECD countries14; 𝐹𝐹𝐹𝐹𝑇𝑇𝑖𝑖𝑖𝑖 is the real interest rate in the home
country; and 𝑅𝑅𝑅𝑅𝑅𝑅𝑖𝑖𝑖𝑖 is the share of religious people as a percentage of the total
population in the remittances’ receiving country, for country i and year t.
In selecting the control variables, this paper follows the standard list of determinants for remittances:
(1) According to Freund and Spatafora (2005), financial development of a country is expected to stimulate the inflow of remittances through formal channels. Likewise, Singh et al. (2009) have found the same positive relationship and argue that well developed financial institutions would ease the process of money transfers.
(2) The lagged per capita GDP is taken to solve the endogeneity problem of remittances and GDP. Based on the altruism motive, the income in the home country is expected to be negatively correlated with the amount of remittances (Ratha, 2013; and Singh et al., 2009). Remittances should act countercyclically because usually if a household has higher income, they are less likely to need outside assistance (Samson, 2011). However, remittances can also be procyclical if they are motivated by investment
13 The lagged per capita GDP is taken to solve the endogeneity problem of remittances and GDP.
11 and profit considerations (Lueth and Ruiz-Arranz, 2007). Hence, the expected sign can not a priori be determined.
(3) One of the variables that are commonly used to explain remittances is the income of the remitter/migrant (Lucas and Stark, 1985; and Rapoport and Docquier, 2005). Many studies show that the level of earnings of the migrant has a positive effect on the amount of remittances (Durand et al., 1996; Funkhouser, 1995; and Hoddinott, 1994). Following Singh et al. (2009), this paper utilizes the income in the host country (the average of real GDP per capita across 17 selected OECD destination countries) as a proxy for migrant’s earnings. Therefore, higher income in the host country is expected to increase the amount of remittance flows to the countries of origin.
(4) Studies have also generally found that variables such as the total number of expatriates in the host countries consistently have a positive effect on the inflow of remittances (Freund and Spatafora, 2005; Elbadawi et al., 1992; and El-Sakka and Mcnabb, 1999). Hence, following the approach by Singh et al. (2010) and Freund and Spatafora (2005), the number of expatriates in the 17 OECD host countries, which proxies the number of international migrants, is expected to be positively related to the amount of remittances sent. A higher number of migrants in the host country should increase the amount of remittances received in the countries of origin.
(5) If investment returns at home motivate migrants, then the real interest rate should have a positive and significant impact on remittance inflows. Several studies have found that migrants tend to remit more money when interest rates at home are high and positive (El-Sakka and McNabb, 1999; Singh et al., 2010).Gupta (2005) and Chami et al. (2009), however, found in India and the developing countries respectively, that remittances and interest rates were not significantly correlated.
12 religious people in the home country, the more likely it is that the migrant shares the same religion as his family. Therefore, altruism based on religious ties is more likely to play significant role in the more religious countries. A positive sign would imply that the volume of total remittances is higher in countries with a higher proportion of religious people, supporting the hypothesis of complementarity of social connectedness through religion and altruistic behavior.
3.2 Variables and Data
This paper uses publicly available relevant data from two sources. First, data on remittances, home and host country incomes, total migrant stocks and financial development are collected from The World Bank’s World Development Indicators (WDI) database.15 The second source of data comes from The Religious
Characteristics of States Dataset (RCS). The RCS reports estimations of religious demographics, country by country. On the whole, the data series in this study covers a sample of 123 countries for the period 2000-2010 (see Table 1 below).16
Compared to previous studies, this paper includes a much larger number of countries from different regions of the world to account for most of the remittance flows in the world. The sample accounts for about 65-70 percent of global remittance flows during the analyzed period, as shown in Figure 3. During the analyzed period, it is estimated that the amount of remittances sent back to the country of origin increased to 301 billion USD in 2010 compared to 73 billion USD in 2000.
Remittances represent a significant part of international capital flows among the studied countries; see Figure 4. The amount of remittances surpassed the net official development assistance- and aid inflows (ODA) during the period 2000-2010. Moreover, compared to other international capital flows, remittances remained rather stable; the only decrease of remittances was between 2008 and
15 The WDI data represents current transfers by migrant workers and wages and salaries earned by nonresident workers. The
data is reported by countries in their balance of payments (BoP).
13 2009. This was caused by the global economic crisis, which affected most of the major remittance sending countries.
Table 1: Countries per income group and region in the research sample
Region of the World
Number of Countries out of Total in Region
Part of the Sample, %
Income Group
Number of Countries out of Total in Income Group
Part of the Sample,% East Asia &
Pacific
16 out of 37 13% Upper middle income
48 out of 56 39%
Europe & Central Asia
21 out of 58 17% Lower middle income
48 out of 52 39%
Latin America & Caribbean
22 out of 42 18% Low income 27 out of 31 22%
The Middle East & North Africa
13 out of 21 11%
North America 0 out of 3 0% South Asia 8 out of 8 7% Sub-Saharan
Africa
43 out of 48 35%
Total 123 out of 217 100% Total 123 out of 139 100% Source: Own calculations using World Development Indicators
Figure 3: Personal remittances received (current US$) and the sample-to-world ratio (%).
Source: own calculations using World Development Indicators (WDI)
Moreover, while this was also the case for the net inflow of foreign direct investment flows (FDI), the decrease in the inflow of remittances remained rather modest. Remittances as a share of GDP are relatively small, but the average share for all the analyzed regions has increased during the analyzed period (2000-1.6 percent of GDP in all the regions; 2010-2.05%); see Figure 5.
55% 60% 65% 70% 75% 0 100 200 300 400 500 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Samp le -to -w orl d ra ti o (% ) Pe rs on al r em it tan ce s (c ur re nt US $) Billion s Year
14
Figure 4: Comparison of remittances- and external capital inflows (FDI and ODA).
Source: own calculations using World Development Indicators (WDI)
Figure 5: Personal remittances received (current US$) and as a share of GDP (%) in different regions of the sample.
Source: own calculations using World Development Indicators (WDI)
The growth in remittance flows in the period 2000-2008 could be explained due to the economic growth in the main remittance sending countries. Assuming migrants living in those countries have benefited from the improved economic conditions in their host country, their capacity (earnings) to remit was improved as well. Another explanation could be that the improving economic conditions in the major destination countries are an important part of the reason some immigrant workers come to these countries and eventually acquire permanent resident status. Indeed, during the analyzed period, the percentage of world population that lived outside their country of birth, increased from 2,8 percent in 2000 to 3,2 percent in 2010; see Figure 6. More specifically, in 17 selected OECD
0 100 200 300 400 500 600 700 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 B illio ns (U S$ )
Sample: Personal remittances (current US$) Sample: Net ODA (current US$)
Sample: FDI (BoP, current US$)
World: Personal remittances (current US$)
0,0% 1,0% 2,0% 3,0% 4,0% 5,0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Re m itta nc es to G D P r ati o (% ) Year
15 host destination countries17, the total migrant stock increased with the GDP per
capita (US$) during the analyzed sample period. This suggests the complementarity of improved economic conditions in the 17 OECD countries and the inflow of new immigrants. However, as mentioned earlier, the increase in the total number of immigrants is mainly due to family reunification than new migrants (OECD, 2013). When we look at the annual number of new asylum applications in these 17 selected OECD countries for the period 1980-2010, the number of asylum claims reached a peak during 1992 and gradually decreased after that period (see Figure 7). While the number of international migrants increased during the analyzed period, the number of new asylum applications diminished in the period 2000-2010. This is because usually there is no need to apply for asylum in the case of family reunification in the host country. There are no annual data on the stock of expatriates are available during 2000-2010. To measure the total number of expatriates as a proxy for the total number of migrants, this variable is estimated by the average international migrant stock in the above OECD countries for the period 2000-2010. Using World Development Indicators for the years 2000, 2005 and 2010, we inferred the international migrant stock in the 17 OECD countries as a proxy for the stock of expatriates.
Figure 6: Total migrant stock and GDP per capita (US$) for 17 OECD countries.
Source: OECD, WDI
17 The data for these 17 OECD countries are collected from the 2009 International Migration Outlook: Australia, Austria,
Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Netherlands, Norway, Spain, Sweden, United Kingdom, United States
0 10 20 30 40 50 60 0 50 100 150 200 250 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 G D P pe r ca pi ta (U S$ ) Tho us ands M ig ra nt sto ck , to ta l M ill io ns Year
16
Figure 7: Total number of new asylum applications for 17 selected OECD countries for the period 1980-2010.
Source: OECD (Paris: OECD, 2009)
Unlike previous efforts to estimate the determinants of international remittances, this paper considers the role that the migrants’ religious motives to remit play in conditioning this empirical relationship. Religious influence is often neglected in economic studies on remittances because it is difficult to quantify. There are no datasets available that take religious altruism into account. Since this paper uses macro determinants for remittances, the key is to find a variable that is both time variant, related to religion and differs across the countries of the sample. Consequently, this paper uses the religious landscape of the country of origin as a proxy for altruism based on religious ties. The assumption here is that a common religion between the migrant and its household indicates the social connectedness that motivates altruistic behavior. Hence, countries with more religious people will have, ceteris paribus, a higher inflow of remittances. The Religious Characteristics of States Dataset (RCS) is the only available data that allows us to include the different religious affiliations of the countries in our model. It estimates the share of religious adherents among the population in 202 countries. Figure 8 below shows the average share of religious people in the three different income group countries of the sample. The low-income group countries in the sample have on average the highest share of religious people. Looking at the individual regions separately, we can see that on average the South-Asia, Sub-Saharan Africa and the Middle-East and North African regions have the highest share of religious population; see Figure 9.
17
Figure 8: The share of religious people as a percentage of total population in the three different income groups (%).
Source: own calculation using the Religious Characteristics of States Dataset (RCS)
Figure 9: The share of religious people as a percentage of total population in the six different regions (%).
Source: own calculation using the Religious Characteristics of States Dataset (RCS)
See Table 2 in Appendix for an extended list of the definitions and sources of the variables employed in this paper and Table 3 for the descriptive statistics for all the variables utilized in the regression.
3.3 Estimation Strategy
The data presented in Table 3, shows the correlation matrix and reveals that there is a weak correlation between the variables and therefore multicollinearity is not a problem. Looking at the variables individually, we can see that the religiosity term negatively correlates with remittances (at 5 percent significance level). This correlation is not reassuring as this is not in accordance with this paper’s hypothesis. On the other hand, the variables financial development, the number of expatriates and income in the host country have the
88,00% 90,00% 92,00% 94,00% 96,00% 98,00% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sh ar e of r el ig iou s peop le (% ) Year
Low-income Lower-middle income Upper-middle Income
80,00% 85,00% 90,00% 95,00% 100,00% 105,00% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sh ar e of r el ig ou s peop le (% ) Year
18 expected positive relationships with the remittances variable. Finally, the variables measuring the initial income in the home country and the real interest rate do not seem to be correlated according to the a priori expectations.
Table 3: Correlation between the variables
(1) (2) (3) (4) (5) (6) (7) (1) Remittances 1.000
(2) Financial development 0.317* 1.000
(3) Initial home income 0.270* 0.368* 1.000
(4) Income in the host country 0.188* 0.079* 0.266* 1.000
(5) Migrant stock 0.169* 0.087* 0.263* 0.082 1.000
(6) Religiosity -0.104* -0.136* -0.091* 0.017 0.017 1.000
(7) Real interest rate -0.189* -0.172* -0.167* -0.232* -0.199* -0.064 1.000 Notes: 1) * is significance at 5 percent level
2) All variables are expressed in logarithms 3) STATA 14 output
If we assume the “country specific” effects to be constant across countries and the “time specific” effects are not present, then the model is estimated by the Ordinary Least Squares (OLS) method or restricted OLS method. Earlier studies have relied on OLS in modeling the determinants of migrant remittances (Lucas & Stark, 1985). However, by pooling all the observations together, we neglect the cross-sectional and time-series nature of the data. If the country effects are not constants but rather are disturbances, the model is estimated by the generalized least square (GLS) method which yields the random-effects model. However, if the country specific effects are constant but not equal, this yields a one-way fixed effects model. To decide between a random effects and fixed effects model, researchers often rely on the Hausman (1978) specification test. Performing a Hausman-test for choosing between a random and fixed effects model gives a Chi-squared test statistic of 3.18 and a p-value of 0.045. The null hypothesis is rejected in favor of the fixed-effects model, and this model will be used for further analysis.
19 heteroscedasticity could be problematic. Performing the Breusch-Pagan test for heteroscedasticity yields the test statistic NR2=4.59 and a p-value of 0.03. The test
statistic is distributed as a Chi-square, and we reject the null H0 of
homoscedasticity at 5% significance level and conclude that heteroscedasticity is present. As there is heteroskedasticity found in the model, the routinely computed standard errors and corresponding tests are not appropriate. Therefore, the model is estimated with heteroscedasticity-robust standard errors.
4. Results and Interpretation
Table 4 reports the fixed-effects estimation results for the determinants of remittances per income group and region. All the coefficients represent elasticities since the estimated equation is a double-logarithmic model.
4.1 Financial Development
20
Table 4: Determinants of remittances per income group and region
Variable (all in logs) FE
Dependent variable: Remittances (L) (L-M) (U) (SA) (EUCA) (MENA) (EAPA) (SUB) (LAT)
Financial development -.049 (.077) -.299 (.221) .143 (.116) -.261*** (.094) .164 (.231) -.339 (.209) -.423*** (.146) .927*** (.288) -.196 (.143) -.076 (.134) Initial home income .713***
(.113) 1.841*** (.484) .796*** (.177) .561*** (.106) .778** (.347) .745*** (.265) .474* (.280) .465 (.362) .836** (.337) .137 (.124) Income in the host country 1.016***
(.173) .464 (.549) 1.240*** (.268) 1.000*** (.175) .660* (.327) 1.661*** (.542) .839** (.312) .863* (.485) .855* (.442) 1.518*** (.169) Migrant stock .603* (.355) .599 (1.068) .483 (.567) .127 (.362) .789 (.737) .901 (.797) .145 (.712) 1.120 (1.069) 1.084 (.834) .422 (.372) Religiosity .408 (1.354) 15.191* (7.943) 4.996** (2.471) -2.378** (1.102) -41.109 (88.165) -3.080* (1.717) 46.809 (30.629) 7.695* (4.236) 11.515* (6.377) .898 (1.669) Real interest rate -.097***
(.030) -.268** (.104) -.014 (.044) -.113*** (.030) -.081 (.066) -.143** (.071) -.034 (.034) .092 (.084) -.262*** (.075) -.063 (.041) Constant -7.142 (5.529) -6.402 (17.045) -7.937 (8.840) 3.321 (5.590) -7.263 (11.648) -18.214 (13.961) 8.489 (10.547) -16.449 (16.629) -14.900* (13.366) -3.925* (5.579) Observations 766 158 310 298 52 124 64 117 237 172 Countries 104 23 40 41 7 17 10 15 35 20 AIC 1209.251 346.233 487.194 205.851 6.098 177.305 22.167 215.368 498.074 37.137 BIC 1241.739 367.671 513.350 231.730 19.757 197.047 37.279 234.704 522.351 59.169 Notes: 1) Standard errors in parentheses
2) Legend: *p<.1; ** p<.05; *** p<.01 3) All variables are expressed in logarithms
4) L=Low income group; L-M=Lower middle-income group; and U=Upper middle-income group
21 According to Ratha (2006), informal channels account for about 50% of remittances. Therefore, it is plausible that the domestic credit provided in the home country, as a proxy for financial development, fails to capture the impact on remittance receipts.
4.2 Income in the Home Country
Contrary to Ratha (2013), remittances do not seem to play a shock-absorbing role. The first-lagged coefficient of real per capita GDP in the home country is positive. In this case, it seems that migrants send remittances when the home country experiences a higher income. More specifically, ceteris paribus, a 1% increase in the initial real GDP per capita in the country of origin (at t-1) is associated with an increase of 0.713% in remittance inflows (at 1% significance level). This pro-cyclical relationship seems to hold in all the regions and income groups. The largest impact of initial home income on the amount of remittance inflows is found in the Sub-Saharan Africa region (+0.84% at 5% significance level), while the smallest effect is found in the Latin-American region (+0.14% but not significant). There is, thus, some indication that migrants tend to send remittances when the economic situation in the country of origin is favorable, possibly in search of investment opportunities consistent with the findings of Lueth and Ruiz-Arranz (2007). This pro-cyclical behavior is in contrast with the general assumption of altruism based on family ties as the primary determinant for remittances. In other words, migrants do not send remittances so that those left behind can maintain a certain quality of life.
4.3 Income in the Host Country
As expected and consistent with previous studies (see Durand et al., 1996; Rapoport and Docquier, 2005; and Singh et al., 2009), the coefficient of the host country income is positive (at 1% significance level). This result implies that the wealthier the country where expatriates live, the higher the remittances they send back home. More specifically, holding all other variables constant, a 1% increase in the host country income18, is associated with a 1.02% increase in the amount of
22 remittance inflows in the country of origin (column FE). This same positive result holds for all the three income groups separately. However, the relationship in the low-income group countries is not significant. In all six regions, earnings in the host country have a significant and positive effect on the amount of remittance inflows in the country of origin. The largest impact is in the Europe-Central Asia region (+1.66% at 1% significance level), while the smallest effect is found in South-Asia (+0.66% at 10% significance level).
4.4 The Number of Expatriates
With respect to the total number of migrants in the 17 OECD destination countries, the results are consistent with previous studies (Singh et al., 2009). Overall, the fixed-effects model indicates a positive relationship with the number of migrants abroad and the amount of remittances inflows. Ceteris paribus, a 1% increase in the total number of migrants in the 17 OECD destination countries, is associated with a 0.60% increase in the amount of remittances inflows (significant at 10% level). When we run the regression for separate income groups and regions, the relationship remains positive. However, the effect is no longer significant for the various income groups and regions separately. The ambiguous findings could be related to the limitation of reliable data and the fact that the number of migrants is proxied by observations of the migrant stocks in the 17 OECD countries in only three years (2000, 2005 and 2010).
4.5 Real Interest Rate
23 with the findings of Gupta (2005) and Chami et al. (2009) where remittances and interest rates have been found to be not significantly correlated. Assuming that households are rational and forward-looking, an increase in the domestic real interest rate should lead households to save more. By smoothing consumption, households aim to maintain a relatively steady consumption pattern over time. The higher the interest rate relative to the rate of time preferences, the more it pays to depress current consumption levels (save), to have more consumption in the future (Gruber, 2013). In other words, an increase in the real interest rate decreases households’ (cash) spending. Since remittances (particularly, those sent for altruistic motives) are used primarily for household consumption purposes, the reduced household consumption signals the migrant that his/or her family is less likely in need of assistance.
4.6 Religious Ties
The results of our variable of interest are somewhat striking. Overall, the marginal effect of a 1% increase in the number of religious people as a percentage of total population in the home country, is associated with a 0.41% increase in the amount of remittances. However, this result is not significant. Religion provides a motive to engage in more altruistic behavior (Eckel and Grossman, 2004). Therefore, this altruistic motive should play a more significant role in the poorest countries as religiosity typically tends to be more of importance to people living in the world’s poorest countries.19 The results of this paper seem to suggest this as
well when we look at the different income groups separately. More specifically, in the low-income group countries, a 1% increase in the percentage of religious people in the country of origin, is associated with a 15.19% increase in remittance inflows (significant at 10% level). This positive and significant relationship also holds for the middle-income group countries (+4.99% at 5% significance level). However, in the upper-middle income group countries, the inverse relationship suggests that religiosity in the countries of origin is negatively correlated with the amount of remittances received. Ceteris paribus, a 1% increase in the percentage of religious
19Steve Crabtree, ‘’Religiosity Highest in World’s Poorest Nations’’, Gallup World Poll, August 31, 2010,
24 people in the home country, is associated with a 2.38% reduction in the remittance receipts (significant at 5% level). If we look at the individual regions separately, the results are only significant in the Europe-Central Asia, Eastern-Asia and Pacific, and Sub-Saharan Africa regions (at 10% significance level). In countries belonging to the Europe-Central Asian region, a 1% increase in the percentage of religious people corresponds with a 3.1% decrease in remittances received. However, religiosity and remittance inflows are positively related in the Eastern-Asia and Pacific countries (+7.70%), with the largest effect found in the Sub-Saharan Africa region. More specifically, a 1% increase in the share of religious people in countries belonging to the Sub-Saharan Africa region, is associated with an 11.52% increase in remittance inflows. These results are exciting as the most religious countries tend to be located in the Eastern-Asia and Pacific and Sub-Saharan Africa regions as well (as is shown before in Figure 2). However, it could be argued that remittance inflows should always be higher in developing countries, regardless of the share of religious people. Migrants, who usually migrate from developing countries themselves, send remittances to family members to help them cope with the adverse economic conditions in the home country. Therefore, developing countries receive remittances primarily because of their poor economic situation, and not because religion seems to be important for poor countries. However, as mentioned before, this paper’s results suggest that remittances are not compensatory in nature. In fact, remittance inflows increase when the initial income in the home country increases. This holds for all regions and income group countries.
5. Conclusion
25 important role in the daily lives of many religious migrants in the host countries (Eckel and Grossman, 2004; Warner and Wittner, 1998; and Williams, 1988). Also, religiosity seems to play a vital role in mainly the world’s poorest countries, who are also likely to be the recipients of remittances.20 Therefore, over time, even if
the social- and family attachments with the home country diminishes, the strong (social) bond based on religious ties could ensure the sustained commitment of the migrant to remit. The findings of this paper suggest that overall, in remittances’ receiving countries with a higher share of religious people, the inflow of remittances is slightly higher but not significant. However, if we look at the different income-groups and regions separately, the results become consistent with this paper’s hypothesis. Altruism based on religious ties seems to play a more significant role in the low- and lower middle-income group countries rather than in the upper middle-income group countries. An increase in the share of religious people in those countries is associated with more remittance inflows. This positive and significant relationship holds particularly for the countries located in the Sub-Saharan Africa and Eastern-Asia and Pacific regions. Interestingly, Gallup Poll surveys show that it is in those regions that religion seems to play an important role.21 While the results in the low-income group countries, Eastern-Asia and
Pacific and Sub-Saharan Africa regions indicate a positive relationship, a 10% significance level is however not that accurate. Another caveat of this analysis is the fact that the number of observations is significantly reduced when the entire sample is separated into the three different regions and income groups. Nevertheless, these results are thought-provoking and extend the boundaries of the current economic theory on the determinants of remittances. Future research with more reliable micro-data on the role of religion in the migrant’s decision to remit could provide new insights on what motivates remittances, and why they continue through time.
Moreover, the message from this paper is that remittances are not compensatory in nature (i.e. if they are sent for altruistic reasons to help the family
20Steve Crabtree and Brett Pelham, ‘’Religion Provides Emotional Boost to World’s Poor’’, Gallup World Poll, March 6,
2009, http://www.gallup.com/poll/116449/Religion-Provides-Emotional-Boost-World-Poor.aspx
21 Steve Crabtree, ‘’Religiosity Highest in World’s Poorest Nations’’, Gallup World Poll, August 31, 2010,
26 in the home country). Consistent with the findings of Lueth and Ruiz-Arranz (2007), the positive relationship between the initial GDP per capita in the home country and the inflow of remittances suggests that remittances are in fact profit-driven and pro-cyclical. Also, consistent with the findings of Singh et al. (2009), the number of migrants- and income in the host country positively affects remittance inflows. However, the results suggest that financial development in the home country does not have explanatory power for remittances. Finally, contrary to the findings of Singh et al. (2009), the real interest rate in the home country is positively associated with remittance inflows.
27
Bibliography
Adams, R. H., 2009. The determinants of international remittances in developing countries. World Development, 37(1), 93-103.
Agarwal, R., & Horowitz, A. W., 2002. Are international remittances altruism or insurance? Evidence from Guyana using multiple-migrant households. World development, 30(11), 2033-2044.
Ahsan Ullah, U.A., 2010. Rationalizing Migration Decisions: Labour Migrants in South and South-East Asia.
Brown, R.P., 1998. Do migrants' remittances decline over time? Evidence from Tongans and Western Samoans in Australia. The Contemporary Pacific, pp.107-151.
Cadge, W., & Howard Ecklund, E., 2007. Immigration and religion. Annu. Rev. Sociol., 33, 359-379.
Carling, J., 2008. The determinants of migrant remittances. Oxford Review of Economic Policy, 24(3), pp.581-598.
Chami, R., Fullenkamp, C. & Jahjah, S., 2005. Are Immigrant Remittance Flows a Source of Capital for Development?, Washington, DC: International Monetary Fund (IMF), Working Paper 52(1)
Chami, R., Barajas, A., Fullenkamp, C., Gapen, M., & Montiel, P. J., 2009. Do workers'
remittances promote economic growth? IMF Working Paper WP/09/153, International Monetary Fund: Middle Eastern and Central Asia Department
Durand, J., Kandel, W., Parrado, E. & Massey, D., 1996. International Migration and Development in Mexican Communities. Demography, 33(2), pp. 249-264.
Eckel, C. C., & Grossman, P. J., 2004. Giving to secular causes by the religious and nonreligious: An experimental test of the responsiveness of giving to subsidies. Nonprofit and Voluntary Sector Quarterly, 33(2), 271-289.
Elbadawi, I. A., de Rezende Rocha, R., & Mundial, B., 1992. Determinants of expatriate workers' remittances in North Africa and Europe (No. 1038). Country Economics Department, World Bank.
El-Sakka, M. I., & McNabb, R., 1999. The macroeconomic determinants of emigrant remittances.
World Development, 27(8), 1493-1502.
Erdal, M.B., 2012. Who is the Money for? Remittances within and beyond the Household in Pakistan. Asian and Pacific Migration Journal, 21(4), pp.437-457.
Freund, C., & Spatafora, N., 2005. Remittances: Transaction Costs, Determinants, and Informal Flows. World Bank Policy Research, Working Paper 3704, (Sept)
Funkhouser, E., 1995. Remittances from International Migration: A Comparison of El Salvador and Nicaragua. The Review of Economics and Statistics, Vol. 77, No.1, (February), p.137-146 Gruber, J., 2013. Public Finance and Public Policy. New York, NY: Worth, 2013. Print. 304-305. Guarnizo, L. E., 2003. The Economics of Transnational Living. International Migration
28
Gupta, P., 2005. Macroeconomic Determinants Of Remittances: Evidence From India (No. 2005-2224). International Monetary Fund.
De Haas, H., 2009. Remittances and social development. Financing Social Policy: Mobilizing Resources for Social Development, 293-318.
Hagen‐Zanker, J., & Siegel, M., 2007. The determinants of remittances: A review of the literature. Working Paper MGSoG/2007/WP003 (June), Maastricht University
Hamilton, W. D., 1964. "The Genetical Evolution of Social Behaviour". Journal of Theoretical Biology. 7 (1): 1–16.
Hamilton, W.D., 1964. "The Genetical Evolution of Social Behaviour. II". Journal of Theoretical Biology. 7 (1): 17–52.
Hoddinott, J., 1994. A Model of Migration and Remittances Applied to Western Kenya. Oxford Economic Papers, 46(3), pp. 459-479.
Kaplan, H. & Gurven, M., 2005. The natural history of human food sharing and cooperation: A review and a new multi-individual approach to the negotiation of norms, in:
Gintis, H., Bowles, S., Boyd, R., Fehr, E., (Eds.), Moral sentiments and material interests: The foundations of cooperation in economic life, Cambridge, MA: MIT Press, pp. 75–113 Lueth, E., & Ruiz-Arranz, M., 2007. Are workers' remittances a hedge against macroeconomic shocks? The case of Sri Lanka (No. 2007-2022). International Monetary Fund.
Lucas, R. E. & Stark, O., 1985. Motivations to Remit: Evidence from Botswana. The Journal of Political Economy, Volume 93, pp. 901-918.
Merkle, L. and Zimmermann, K.F., 1992. Savings, remittances, and return migration. Economics Letters, 38(1), pp.77-81.
Rachlin, H., & Jones, B. A., 2008. Altruism among relatives and non-relatives. Behavioural processes, 79(2), 120-123.
Rapoport, H., & Docquier, F., 2006. The Economics of Migrants' Remittances. The Institute for the Study of Labour Discussion Papers, Issue 1531
Ratha, D., 2013. The Impact of Remittances on Economic Growth and Poverty Reduction. Migration Policy Institute (MPI) Policy Brief, September, Issue 8.
Ratha, D. & Mohapatra, S., 2007. Increasing the Macroeconomic Impact of Remittances on Development, Washington, DC: The World Bank, Development Prospects Group.
Samson, M.S., 2011 "An Analysis of Remittance Tendencies of Philippine Migrant Workers". CMC Senior Theses. Paper 195.
Sana, M., & Massey, D. S. (2005). Household composition, family migration, and community context: Migrant remittances in four countries. Social Science Quarterly, 86(2), 509-528. Shankman, P., 1976. Migration and underdevelopment: The case of Western Samoa. Westview Press.
29
Vargas-Silva, C. & Huang, P., 2006. Macroeconomic Determinants of Workers' Remittances: Host versus Home Country's Economic Conditions. Journal of International Trade and Economic Development, 15(1), pp. 81-99.
Warner, S.R. & Wittner, J. G., 1998 eds. Gatherings in Diaspora: Religious Communities and the New Immigration, (Philadelphia: Temple University Press)
30
Appendix
Table 1: Countries in the research sample (total:123). Source: World Development Indicators
Afghanistan Georgia Nigeria
Albania Ghana Pakistan
Algeria Guatemala Panama
Angola Guinea Papua New Guinea
Argentina Guinea-Bissau Paraguay
Armenia Guyana Peru
Azerbaijan Haiti Philippines
Bangladesh Honduras Romania
Belarus India Russian Federation
Belize Indonesia Rwanda
Benin Iran, Islamic Rep. Samoa
Bhutan Iraq Sao Tome and Principe
Bolivia Jamaica Senegal
Bosnia and Herzegovina Jordan Serbia Botswana Kazakhstan Sierra Leone
Brazil Kenya Solomon Islands
Bulgaria Kosovo South Africa
Burkina Faso Kyrgyz Republic Sri Lanka
Burundi Lao PDR Sudan
Cabo Verde Lebanon Suriname
Cambodia Lesotho Syrian Arab Republic
Cameroon Liberia Tajikistan
Chad Libya Tanzania
China Macedonia, FYR Thailand Colombia Madagascar Timor-Leste
Comoros Malawi Togo
Congo, Dem. Rep. Malaysia Tunisia
Congo, Rep. Maldives Turkey
Costa Rica Mali Turkmenistan
Cote d'Ivoire Mauritania Uganda
Cuba Mauritius Ukraine
Djibouti Mexico Uzbekistan
Dominican Republic Moldova Vanuatu
Ecuador Mongolia Venezuela, RB
Egypt, Arab Rep. Montenegro Vietnam
El Salvador Morocco West Bank and Gaza Equatorial Guinea Mozambique Yemen, Rep.
Eritrea Myanmar Zambia
Ethiopia Namibia Zimbabwe
Fiji Nepal
Gabon Nicaragua
31
Table 2: Variable sources and definitions
Variable Description Details and Sources
REM Personal remittances, received (current US$)
Personal remittances comprise personal transfers and compensation of employees. Data are the sum of two items defined in the sixth edition of the IMF's Balance of Payments Manual: personal transfers and compensation of employees. Data are in current U.S. dollars.
Source: World Bank staff estimates based on IMF balance of
payments data. INCHOME GDP per capita (current
US$)
GDP per capita is gross domestic product divided by midyear population. Data are in current U.S. dollars.
Source: World Bank national accounts data, and OECD National
Accounts data files. FINDEV Domestic credit provided
by financial sector (% of GDP)
Domestic credit provided by the financial sector includes all credit to various sectors on a gross basis, with the exception of credit to the central government, which is net.
Source: International Monetary Fund, International Financial
Statistics and data files, and World Bank and OECD GDP estimates. INCHOST Average GDP per capita
(current US$) in 17 OECD countries
Source: World Bank national accounts data, and OECD National
Accounts data files. EXPATR Total of international
migrant stock in 17 OECD countries
International migrant stock is the number of people born in a country other than that in which they live. It also includes refugees. The data used to estimate the international migrant stock at a particular time are obtained mainly from population censuses. The estimates are derived from the data on foreign-born population--people who have residence in one country but were born in another country.
Source: United Nations Population Division, Trends in Total
Migrant Stock: 2012 Revision. REL Share of population in the
home country with a religious affiliation (%)
The RCS reports estimates of religious demographics, country by country. RCS was created to fulfil the unmet need for a dataset on the religious dimensions of countries of the world, with the state-year as the unit of observation. It covers 202 states plus 22 selected non-state political entities, for every year from 2010 back to 1900. It estimates populations and percentages of adherents of 87 religious denominations.
Source: RCS
Table 3: Summary statistics for variables
