Enhancing the role of remittances through social capital: Evidence from Indonesian household data

Abstract

The main purpose of this study is to determine whether remittance-receiving households in Indonesia have truly experienced a measurable increase in their welfare. It focuses on how social capital may enhance the efforts of Indonesia’s female migrant workers to improve their and their family’s welfare at home. Our findings confirm that social capital enhances the impact of remittances by increasing the welfare of the migrant workers and their immediate family members. Remittance-receiving households with strong and wide-ranging social capital tend to use the extra income from remittances more for investment. In turn, this generates future income, which boosts the impacts of their remittances on their overall welfare.

Keywords

remittances female migrant workers social capital Indonesia

Introduction

The rapid increase in the flow of international remittances has fueled interest in their impact on the economies of recipient countries, especially with regard to poverty and inequality, and particularly for migrant workers who come from poor or vulnerable families. According to the World Bank Group (2018a), migrant workers’ income increases three to six times when they move from lower-income to higher-income countries. For most Indonesian labor migrants, migration is seen as a livelihood strategy for alleviating poverty for them and their families. A study by the World Bank Group (2017) corroborates this, since the majority of migrant workers in Indonesia come from relatively poor regions.

A number of studies have found that remittances have a significant impact in reducing poverty, both at the macro and micro levels. Various studies have shown that poverty is reduced more when international remittances are included in the household income (Adams and Page, 2005; Adams, 2006; Gupta et al., 2007; Adams and Cuecuecha, 2016; Yoshino et al., 2017). However, there is a need to further investigate how remittances actually serve as a pathway out of poverty for households. As argued by Gupta et al. (2007), the impact of remittances very much depends on how they are used by the recipient households. Recipient households can use remittances for investment or as working capital for entrepreneurial activities, which Spitzer (2016) identifies as return migrant entrepreneurship. The concept refers to migrant workers who started business activities upon returning and migrant workers who have initiated, or at least had the plan to initiate, business activities while they are still working abroad and are sending remittances home for this purpose. Her research shows that there is a high likelihood for entrepreneurship to fail and when it is successful, it is only beneficial to some individuals and family members. We argue that return migrant entrepreneurship presents a viable alternative for migrant workers who intend to stop working abroad because they can have sources of income even in the face of difficult domestic economic and job market conditions. We propose that by initiating or planning the business while the migrant workers are still working abroad may lead to better results. To successfully implement this, it is important that the families of migrant workers have social capital that will enable remittance-receiving households to access information and other resources in setting up or sustaining business ventures.

Therefore, this study seeks to identify what determines an increase in investment by remittance-recipient households. In particular, this study focuses on female migrant workers (FMWs)(FMWs) and their families in Indonesia; women comprise the majority of migrant workers from Indonesia. According to the Indonesian Migrant Worker’s Protection Bureau (BP2MI) (2020), the share of Indonesia’s female migrant domestic workers in the last three years, 2017–2019, has steadily stayed around 70.2 percent (184,640 out of 262,899), 70.2 percent (198,974 out of 283,640) and 69.2 percent (191,238 out of 276,554), respectively. Most women migrant workers go abroad to support or boost the existing income-generating activities of their husbands or families. The remittances women migrants send back to their families have the potential to be used for investment.

According to Grootaert (1999), Borja (2014) and Kiboro (2017), the social capital of households is one of the determinants that can contribute to the efficient use of remittance. Kiboro’s (2017) study in Kenya shows that local-level associations, a measure of social capital, can be leveraged by households to obtain essential services, which, in turn, increase their livelihood and reduces poverty. In Indonesia, Grootaert (1999) found that membership in local-level associations raised per capita household spending, savings, assets and access to credit. Borja (2014) presents evidence that remittance-recipient countries with significant social capital are likely to leverage remittances for economic growth. However, our study differs from Borja’s (2014) in that we employed the data at the household level, thus filling the research gap on what may motivate households to use remittances for investment and generating productive assets. Furthermore, by using household level data, this study aims to determine what kind of specific social capital could induce a sustainable increase in household welfare. To date, there has not been any detailed empirical research on how the role of social capital can lead to better use of remittances by recipient households. It is assumed that FMWs, being mothers and/or wives, will stop going abroad to work at some point. Thus, the efficient use of remittances they send is a crucial step in allowing them to resume their roles at home and not have to work abroad anymore.

The rest of the article has the following sections: a discussion of the background that motivated this study; a review of the conditions of FMWs and responses to address their welfare; review of the related literature; the data set and econometric strategy; findings and analysis; and conclusion and policy implications.

FMWs in Indonesia

The temporary labor migration of women in Asia, including those from Indonesia, has often been framed in terms of economic push and pull factors. As summarized by Yazid (2013), the demand for workers in destination countries has been met by an excess in the supply of workers in origin countries. In addition, women’s migration also has cultural underpinnings. As explained by Silvey (2004a), in the concept of the ideal family in Indonesia, women and men have specific roles. For women, the constructed ideal roles are as mothers and wives, thus their identities are rooted in the family and in the domestic sphere. This culturally, and in many cases, religiously, infused concept closely links women to the family and the home, such that when they migrate, their absence from the family and, by extension, the nation, is perceived as cultural transgression. However, these cultural notions have been modified by the economic needs of families and households on the one hand, and by the demand for domestic workers in more developed economies on the other. Furthermore, the feminization of migrant workers also needs to be understood by taking into account the migration infrastructure—this includes the institutions, networks and people, such as brokers and middlemen, involved in the mobility and processing of migrant workers—which enables the mobility of women and influences the demand for FMWs (Lindquist, 2010; Lindquist et al., 2012). Brokers or middlemen tend to create more opportunities for females to play a more significant role in labor migration. Unlike men, FMWs do not have to raise funds to cover pre-migration costs. Instead, pre-migration costs are charged against their salaries. The salary deductions can extend for several months, and in some cases, migrant workers do not receive any salary for a certain period of time. The migration of women can have exploitative and empowering consequences for women migrant workers (Chan, 2014). Moreover, FMWs do not form a homogenous group with similar narratives; rather, they have diverse experiences as victims and agents acting in culturally and morally specific ways (Chan, 2018).

Almost half of the Indonesian women migrant workers are married and leave their husbands and children behind. The research is focused on married women migrant workers in line with our hypothesis that the family can be both the motivation and the support system for an investment initiative by migrant workers. From the preliminary interviews that we conducted from April to June 2018 in three origin areas, namely Indramayu, Wonosobo and Jember, we found that the main reason for married women to go overseas was to provide or contribute to the income of their nuclear family, particularly to support their children’s education.¹ Increasing welfare in the long run depends mainly on how family members use the extra funds. As Silvey (2004b) points out, due to transnational labor migration, motherhood and family relations undergo changes, the families experience rising income, and families also evince increasing consumer desires and practices.

Data from the World Bank (2018a) show that 70 percent of Indonesian migrant workers have been able to increase the welfare of their families. As elaborated by Rahman (2009) through his study on Indonesian domestic workers, remittances are important in fulfilling daily consumption, education, house renovation and savings. Remittances become one of the sources of income, if not the only one, for the family and thereby a strategy to increase their living standard. Nationally, remittances generated by migrant workers constitute around 1 percent of Indonesian gross domestic product and 10 percent of its tax revenue. In 2019, Bank Indonesia (2020) recorded that Indonesian migrant workers’ remittances reached USD11.4 billion.

Data compiled by Bank Indonesia (2020) suggest that most remittances that Indonesia receive come from countries that are the main destinations for Indonesian FMWs employed as domestic workers, such as Malaysia, Hong Kong, Taiwan and Saudi Arabia. Although it has been suggested that those who migrate are not necessarily the poorest of the society, according to the World Bank Group (2017), almost two-thirds of the origin areas of migrants are relatively poor regions with poverty rates higher than the national average. The women, most of them with low educational background, work abroad to support their family. Therefore, since most migrant workers come from poor or vulnerable households, remittances have a disproportionately positive impact on helping families cover their daily needs. Given their important role in reducing poverty, households are expected to prioritize using remittances on more long-term goals, such as savings, education or starting a business (World Bank Group, 2017).

Discussions on how to turn remittances into either businesses or investments that can generate enough sustainable income to replace the need to work abroad are important. The risks attendant to migration, such as the risks to marriages and the difficulties in bringing up their children from afar, make it most logical to render working abroad for women a temporary solution. Migrant workers—especially women—remain extremely vulnerable to abuse in the absence of a strong legislative or administrative framework to protect their basic human rights (Palmer, 2016). They are thus vulnerable to abusive practices, such as: unpaid wages; sexual harassment; verbal, physical, sexual and psychological abuse by employers; unfair termination of employment; and inability to depart from the country of origin for reasons beyond their control, including bureaucratic problems with government (Human Rights Watch, 2016; Killias, 2018). Therefore, it is important to ask the question: What can be done to realize the potential of turning remittances into economic investment? In other words, what can be done to incentivize people to use remittances for purposes other than consumption?

Randolph (2015), puts worker migration in the context of inclusive growth, in which growth is not based merely on economic output, but also on broad-based economic opportunities. We agree with his statement that building a connection between the discourses on inclusive growth and impacts of worker migration still represents a challenge for the government. For a long time, worker migration has been perceived as a way to release some of the tension caused by unemployment. After almost four decades of regulated worker migration from Indonesia, the point where new jobs are created in the origin countries due to the sustainable economic activity brought about by the flow of remittances into the country has yet to be reached. Randolph (2015) strongly argues that remittances are able “to sustain job-creating local economies.” Furthermore, he argues that returning migrants are perceived more as consumers rather than as producers who have the potential to establish job-creating economic activities.

These arguments show how the perception of worker migration has shifted significantly to a more enabling one, which allows migrant workers to become more active players in the efforts to turn remittances into more sustained and job-creating economic activities in the origin areas. With this new perspective, there needs to be a preliminary assessment that looks into current conditions and practices in the utilization of remittances by identifying the determinants of decisions to use remittances for investment and recording the narrative of how investments have been made in three major areas of origin in Indonesia.

Literature review

As explained in the introduction, the role of remittances in improving the livelihood of migrant worker families is still inconclusive. Several studies, such as those by Adams (2006), Gupta et al. (2007), Adams and Page (2005), Adams and Cuecuecha (2016) and Yoshino et al. (2017), show a significant and positive relationship between remittances and household welfare. There are, however, studies that also imply this is not the case for some countries—Randolph (2015) and Chami et al. (2008) obtained results which could not provide robust evidence that remittances affect income growth in a positive way. Their studies implied that the effectiveness of remittances in improving household welfare is conditional on certain factors. In the case of Indonesia, Nahar and Arshad (2017) provide evidence that increases in remittances have a small effect on poverty reduction, possibly due to the low educational background of migrants, low-wage jobs, expensive remittance costs and a lack of information on how to remit their money through formal institutions.

We aim to fill the gap by advancing that the investment channel is one of the determinants that may improve the welfare of the families of migrant workers significantly and sustainably and, in turn, it could play a role in poverty eradication. The fact that the use of remittances for investment and productive assets has a multiplier effect on the consumption channel is at the core of our argument. Past studies also corroborate our hypothesis. The study of Cuecuecha and Adams (2016) in Indonesia presents evidence that households that have a higher proportion of investment in their total budget expenditure are more likely to be not poor. They also found that the remittance-receiving households spend more money on food consumption and investment in education. This study focuses on what it is that determines whether the FMWs’ households spend their extra income on investment.

Based on their research, Anwar and Chan (2016) highlighted the notion of a positive link between financial remittances and development. They compared two migrant-origin villages in Java—one is characterized by fierce economic competition while the other one has a more inclusive environment. In the former village, return migrant enterprises tended to build on higher levels of distrust, exclusion and competition, whereas in the latter village, there was more cooperation through locally initiated systems, thereby increasing a sense of belonging, communal pride and equality. In line with their arguments, we posit social capital as the main determinant of how recipient households use remittances. Social capital refers to “the ability of actors to secure or benefit by virtue of membership in social networks or other social structures” (Portes, 1998: 6). According to Portes, social capital includes factors such as government, values, the rule of law, civil and political liberties at the macro level, and networks and norms that govern interactions among individuals, households and communities at the micro and meso levels (Portes, 1998). We hypothesize that the remittance-receiving households, which have strong social capital will have a greater probability of using the extra income from remittances on productive assets or investments rather than on consumption, mainly because of greater access to knowledge and networks.

To our knowledge, there has not been any detailed empirical research on how social capital affects the use of extra income from remittances, especially at the micro and meso levels. At the macro level, however, Borja’s (2014) panel study of 88 countries from 1990 to 2009 showed that remittance-recipient countries with developed social capital in place are more likely to leverage remittances for economic growth. In the study, social capital was broadly defined to include institutions, attitudes and values at the macro/country level. The effect of social capital infrastructure, however, is rather small. At the micro level, the study found that remittances are likely to be used for consumption. The present study addresses one of the gaps of Borja’s (2014) study, which is to explain the ways in which remittances are used and channeled.

At the micro and meso levels, several studies have shown that there is a close relationship between social capital and household income and/or poverty. Kiboro (2017) in his study in Kenya uses membership in local level associations as a measure of social capital and explores the relationship of this membership with the livelihood outcomes of internally displaced persons. His findings demonstrate that poor households that have better social networks and are members of associations have the advantage of receiving support to achieve a better livelihood. Furthermore, in Grootaert’s (1999) study in three provinces in Indonesia (Jambi, Central Java and East Nusa Tenggara), she found that the active participation of households in local associations correlates closely with higher expenditure, but participation at the community (meso) level does not significantly contribute to household expenditure. This explains how community participation, measured by the percentage of people in their respective provinces who have membership in community-initiated organizations, affects household welfare in an indirect way. A study by Rustiadi and Nasution (2017) in rural Indonesia also supported Grootaert’s (1999) conclusion. By using nationally representative datasets and a logistic regression model, they found that the probability of households being poor is lower when households have a higher social capital index. The index consists of six dimensions: groups and networks; trust and solidarity; collective action and cooperation; information and communication; social cohesion and inclusion; and empowerment and political activities.

In the context of Indonesia, Randolph (2015) states that there is no evidence yet to suggest that remittances have induced economic development and job opportunities. One of the reasons as to why remittances have not contributed to sustainable development is that the Indonesian government sees the returning migrants as consumers rather than producers. Therefore, remittances are not yet seen as an effective way out of poverty since they function only to cover consumption rather than to increase welfare. Therefore, this study could have strong policy implications regarding the effectiveness of remittances in increasing household welfare—especially for low-income families—through the mechanism of social capital.

Methodology and data

Econometric model and method

The main motivation for workers to migrate is to improve livelihood since most likely, they come from low-income families (Adams and Page, 2005). In this study, we define low-income families as the poor and the vulnerable. Therefore, the first stage of this study investigates whether the hypothesis is true in the case of Indonesia, particular for the families of FMWs, which is represented by equation (1):

D_F M W_{i} = π + λ D_P V_{i} + Y_{i} ϑ + μ_{i}

(1)

As can be seen from the model specification, this is a cross-sectional model where i represents individual households. D_FMW is the dummy for households of FMWs, where a value of 1 indicates that one of the family members is a woman migrant worker, and a value of 0 otherwise. D_PV represents the income status (dummy) of individual households (whether they are poor or vulnerable). A value of 1 indicates that they are the poor or vulnerable families, and 0 otherwise. In equation (1), we included the matrix of Y as the control variable. First, we included the dummy of university degree, where 1 indicates that the head of household held a university degree. Therefore, it was assumed that this would have a negative correlation with the D_FMW since, if the head of household had a university degree, he would have a high probability of having a good job and would be able to provide for his family; thus, the wife would not to need work overseas. Second, we also found in our fieldwork that most of the FMWs from Indonesia are Muslim and go to Saudi Arabia. One of their main motivations while working there is to perform the Hajj pilgrimage. Therefore, we included the dummy of religious value, where a value of 1 indicates that they were religious, and 0 otherwise; we also assumed that it would have a positive correlation with D_FMW.

The main framework of this study is that the presence of social capital can enhance the social welfare of remittance-recipient households. In this study, the investigation focused on the families of FMWs, and the main thrust of our research is that the role of remittances in increasing the social welfare of a family depends on the social capital of that family which, we assume, positively affects consumption and investment.

Thus,

D_P V_{i} = γ + η (Z_{i}^{n}) + K_{i} θ + ε_{i}

(2)

Z_{i}^{n} = α + β (D_F M W_{i}^{*} D_S C_{i}^{p}) + δ Credi t_{i} + X_{i} φ + e_{i}

(3)

Equation (2) asserts that the higher the level of consumption and/or investment, the lower the probability that a household will be poor. The K variable is a matrix of control variables, which includes whether there is an entrepreneur in the household and if there is cigarette consumption. The presence of an entrepreneur family member (measured by a dummy of 1, and 0 otherwise) is assumed to negatively affect the D_PV, which stems from the argument that entrepreneurs most likely have higher consumption and investment, and are thus not “poor or vulnerable.” Equation (2) includes cigarette consumption since, in the case of Indonesia, cigarette consumption contributes significantly to the consumption basket of the poor and has a tendency to sacrifice essential goods for cigarettes (Nasrudin et al., 2013). Therefore, cigarette consumption may well have a negative impact on the purchasing power of other essential goods, pushing their income status into the region of the poor or vulnerable group.

In equation (3), Z is defined as an endogenous variable, which is consumption and investment expenditure, both in real terms. D_SC is the dummy for social capital, where 1 indicates having certain social interactions with the community, and 0, otherwise. The superscript p of D_SC shows there are several indicators of social capital. In this study, six different indicators are used, based on a household’s participation in community meetings, cooperatives, religious activities, PNPM (Program Nasional Pemberdayaan Masyarakat, National Program for Public Funding)² and village savings and loans. It was expected that those who participated in these community activities would be better off, not only in terms of consumption, but also in terms of investment. This assumption is based on the logic that a better social network gives households more opportunities and knowledge of how to use and multiply their income, thus making their total welfare higher than the welfare of those who did not participate or had limited social networks. The subsequent variables in the model are control variables. These variables include access to credit (Credit) and household demographic characteristics (X), such as home ownership status and education level of the household members. As indicated by Quach et al. (2005), household credit has a positive effect on the economic welfare of poor households in rural Vietnam, since it may provide a source of funding for households to carry out entrepreneurial activities. The status of home ownership is also assumed to have a positive correlation with investment and consumption, since those who own a house do not need to use their income for rent but use it for consumption and/or investment activities. The variable of education, in this regard, is a dummy variable where 1 indicates that the head of household holds a university degree and 0 otherwise; and holding a university degree would have a positive effect on consumption and investment, since a higher degree would correlate with a higher income.

Furthermore, we assume that the exogenous variables in equation (3)—especially the social capital indicators—are strictly exogenous toward the level of consumption and investment. In making this assumption, it seems safe to say that the choice of participation in community activities is not affected by consumption and investment. Based on the results of our fieldwork in six regions in Indonesia, FMWs’ families participated in local community activities based on their respective village environment and characteristics. For example, the high participation in religious activities in most of the villages in the province of Nusa Tenggara Barat is influenced by local values and beliefs.

However, based on the model specifications of equation (2), it is likely that a reverse causality exists between the level of consumption and/or investment and household income status (poor or vulnerable), which suggests an endogeneity problem. An increase in consumption and/or investment implies a rise in income, potentially raising the household out of poverty. It is logical, then, to assume that the amount a household can spend on consumption and/or investment is also affected by the level of income. In our case, families living under the poverty line and those whose income is close to the poverty line may have a low level of investment relative to those in upper-class and middle-class families.

Bellemare et al. (2017) stated that lagged explanatory variables are commonly used in response to endogeneity concerns in observational data. However, this approach could not be used in this study, since using the time variable in the Indonesian Family Life Survey (IFLS) at the household level data that we employed in this study was not possible. Another way to solve the endogeneity problem is by using an instrumental variable (IV) and doing a two-stage least squares (2SLS) regression on the exogenous variable that is suspected of having the property of endogeneity (Murray, 2006). In employing this method, Levitt (1997, 2002) and Murray (2006) suggest that the inclusion of an IV should be robust since including a weak and invalid instrument means that the cure could make the disease worse. The inclusion of an IV can be achieved by testing the equation using several instruments. To this end, the model in equations (2) and (3) was constructed as a simultaneous equation model, where the social capital (six indicators) was selected as the instrument variable in predicting whether the increase in consumption and/or investment could help the FMWs’ families out of poverty. These social capital indicators include participation in community meetings, cooperatives, youth activity groups, religious activities, village savings and loans, and the PNPM. Consequently, whichever social capital significantly affected the consumption and/or investment was chosen as the IV—in our case, the principal social capital (PSC).

The standard 2SLS method could not be applied directly to our model specifications, however, since the endogenous variable in equation (2) is not continuous but categorical. Therefore, the Instrumental Variable Probit (IV Probit) developed by Buzas and Stefanski (1996) was applied. The IV Probit was able to solve the problem in our model specification since, in principle, it is similar to the 2SLS and allows the model to be both categorical and instrumented. Based on the model construction, Figure 1 presents the conceptual framework of this study.

Figure 1.

Conceptual framework.

Data

In this study, data from the 2014 IFLS (5th wave) and Indonesian Central Statistical Bureau (BPS) were used. Note that the IFLS was not intended as a survey of remittances and migrant workers. Because of that, data have had to be recalculated to represent the variables in equation (1). First, we truncated the full sample of IFLS 2014 by identifying the remittance-recipient households, both domestic and international. These families were identified by adding the dummy variable, with a value of 1 if they receive remittances and 0 if otherwise. Using this identification, there were 10,909 observations in all. Second, in order to determine whether a household was the recipient of international remittances from a female migrant worker, we added dummy (1) if the household received remittances from female spouses or mothers living in other countries.³ Using this method, we identified 185 households that were recipients of international remittances from FMWs. Third, we defined consumption as the average monthly expenditure on all items of consumption in the IFLS, and investment as the amount of productive assets used for business purposes. Furthermore, as the consumption and investment data in the IFLS are in nominal terms, we deflated both variables using the Consumer Price Index on the district level, and it was assumed that every household living in the same district faced the same price index.

The sample identification of FMWs’ families is representative of the actual condition in Indonesia. Around 80 percent of FMWs’ families reside in Java and West Nusa Tenggara (NTB), as can be seen in Table 1. According to the World Bank, the provinces of NTB contribute the largest number of migrant workers, and in 2016, approximately 65 percent of those migrant workers were female (World Bank Group, 2017). Based on this fact, our sample was limited to households in Java and NTB, since our interests focused mainly on FMWs’ families. Using this limitation, there were a total of 4,338 observations, where 124 of them (2.86 percent) represented the families of FMWs.

Table 1.

Percentage of FMWs’ households to the total sample of migrant worker households in the Indonesian Family Life Survey (IFLS).

Province	Number of samples	Percentage
West Java	36	19.4
East Java	37	20.0
West Nusa Tenggara	41	22.2
Central Java	20	10.8
Banten	10	5.4
Others	41	22.2
Total	185	100.0

Source: Authors’ calculations from the IFLS 2007 and 2014.

The participation in local level associations has been used as a proxy for social capital, which consists of household members participating in community meetings, cooperatives, youth group activities, religious activities, village savings and loans and the PNPM. Note that other local level association types are not included. The reason is that some activities are more likely participated in by women (arisan, women’s association activities and the community weighing post), while the subjects of this study were the remittance-recipient households, most often the FMWs’ spouses.

To determine whether a household was poor or vulnerable, the question on the IFLS regarding whether or not the household is a recipient of the “rice for the poor” (beras miskin, henceforth Raskin⁴) was used where those who answered yes (recipient of Raskin) were identified as poor or vulnerable families. The “yes” in the IFLS includes those who actually buy the rice at a lower price using the coupons given to them and those who receive money to replace the coupons. Out of a total of 4,338 observations, 2,623 (60.5 percent) of them were identified as Raskin recipient households (either poor or vulnerable). Access to credit was gained by identifying whether or not a household had knowledge of borrowing from formal financial institutions (including the government). Table 2 presents a summary of the statistics of variables used in this study.

Table 2.

Summary of statistics.

Variable	Mean	Minimum	Maximum	Standard deviation
Dummy
Poor or vulnerable households	0.605	0.000	1.000	0.489
Female migrant worker households	0.029	0.000	1.000	0.167
Community meetings	0.311	0.000	1.000	0.463
Cooperatives	0.068	0.000	1.000	0.252
Youth activity groups	0.160	0.000	1.000	0.367
Religious activities	0.708	0.000	1.000	0.455
Village savings and loans	0.046	0.000	1.000	0.210
National Program for Public Funding	0.103	0.000	1.000	0.304
Access to credit	0.826	0.000	1.000	0.379
Home ownership status	0.902	0.000	1.000	0.399
University education	0.173	0.000	1.000	0.378
Entrepreneurial activities	0.401	0.000	1.000	0.490
Ln Real investments	5.880	0.000	20.934	8.008
Ln Real consumption	16.813	11.341	20.761	0.842
Ln Cigarette consumption	0.810	0.000	4.395	1.175

Source: Authors’ calculations.

Findings and analysis

As was explained in the introduction, it was assumed that most international remittances in Indonesia are received by low-income families, and the FMWs tend to choose to work abroad, as the expected income from working abroad is often higher than their comparable domestic income. As shown in Table 3, calculations using logistic regression for equation (1) show that the families of FMWs tend to receive Raskin, which indicates that they are from poor or vulnerable families. This result is robust since when we added the control variables in (3.2), the coefficient only marginally decreased, and statistically, the dummy “poor or vulnerable” is significant in both (3.1) and (3.2). This implies that the workers migrated in order to increase the family income, and this result is in line with the argument of Adams and Page (2005), who found that poverty was the main driver of migration for most migrants.

Table 3.

Logistic regression of the poor or vulnerable on FMWs’ families.

Dependent: FMW family	(3.1)	(3.2)
Dummy poor or vulnerable	0.522**	0.482**
	(0.202)	(0.203)
University degree		−0.393
		(0.281)
Religious value		2.008**
		(1.007)
Constant		−5.746***
		(1.012)
Prob > Chi²	0.008	0.000
Observations	4,338	4,338

Source: Authors’ calculations.

Nevertheless, it should be emphasized that in the case of labor migration, workers still need a certain amount of capital to pay for the cost of migration, so that it is most likely that they do not come from the low-income families, which is one of the central arguments of De Haas (2017). He argues that migration requires significant resources and extreme poverty immobilizes people since they cannot afford to leave their homelands. However, in the case of Indonesia, low-income families finance the cost of migration through debt, either from formal or informal channels (Platt et al., 2013; Yuniarto, 2015; World Bank Group, 2017). Therefore, the fact that they are financially disadvantaged implies that they come from low-income families yet might be not the poorest of the poor. Although our study sample has its limitations in identifying which families are indeed the poorest of the poor, the identification procedure would also be in line with the argument. In this matter, low-income families who have FMWs in their families are not the poorest of the poor. However, most of those families should still be categorized as “poor or vulnerable” since there are still many cases where the migration process is financed through debt.

As explained in the methodology section, we argue that the appropriate way to estimate equations (2) and (3) are least squares regression and IV Probit, and this was based on two factors. First of all, the least squares method is used to answer whether having social capital can significantly influence the increase in investment in productive assets. This also tests whether social capital can serve as a valid instrument in predicting differences in levels of household consumption and investment. Second, IV Probit is employed to determine whether increases in investments as a result of remittances and having social capital help FMWs’ families out of poverty. Tables 4 and 5 present the results of the least squares regression of equation (3), where the endogenous variables are “consumption” and “investments,” respectively. There are six different regressions each in Tables 4 and 5, which differ based on the social capital indicator. Regression (4.1) and (5.1) show participation in community meetings, while (4.2) and (5.2), (4.3) and (5.3), (4.4) and (5.4), (4.5) and (5.5) and (4.6) and (5.6) represent cooperatives, youth group activities, religious activities, village savings and loans and the PNPM, respectively.

Table 4.

Regression results on the impact of social capital ownership on consumption.

	(4.1)	(4.2)	(4.3)	(4.4)	(4.5)	(4.6)
FMW*social capital	0.200*	0.220	0.109	0.110	−0.575	−0.035
FMW*social capital	(0.105)	(0.229)	(0.154)	(0.087)	(0.376)	(0.231)
Access to credit	0.363***	0.385***	0.385***	0.382***	0.390***	0.395***
	(0.035)	(0.035)	(0.035)	(0.035)	(0.035)	(0.035)
Home ownership status	0.056*	0.084***	0.072**	0.076**	0.085***	0.086***
	(0.031)	(0.030)	(0.031)	(0.031)	(0.030)	(0.031)
Highest education	0.306***	0.325***	0.313***	0.319***	0.329***	0.329***
	(0.034)	(0.034)	(0.034)	(0.034)	(0.034)	(0.034)
Constant	16.34***	16.36***	16.35***	16.19***	16.35***	16.36***
	(0.041)	(0.041)	(0.041)	(0.074)	(0.041)	(0.041)
R-squared	0.072	0.057	0.063	0.058	0.056	0.055
F-stat	55.38***	40.56***	47.12***	41.78***	40.85***	39.93***
Observations	4,338	4,338	4,338	4,338	4,338	4,338

Note: Numbers in parentheses are standard error; significant at α= *10%, **5%, ***1%.

Source: Authors’ calculations.

There are several interesting results from these regressions. Firstly, the majority of these six social capital indicators have a positive correlation to the level of real consumption and investments of FMWs’ families, though the participation in village savings and loans and PNPM do not contribute to the consumption level. Secondly, however, there are only three indicators (community meetings (5.1), youth group activities (5.3) and religious activities (5.4)) that have a significant and positive impact on the level of investment, and only one, community meetings (4.1), weakly affected the level of consumption. These results are somewhat surprising since we had expected that membership in cooperatives, village savings and PNPM, would be significantly associated with higher investment level.

These insignificant results imply that the three local-level associations did not function ideally, at least for the FMWs’ families. As also stated by the World Bank (2018b), the PNPM has not been very effective in helping marginalized people improve their welfare, as the local elite still dominate the decision-making process. Kirwati et al. (2018) also support this conclusion, and their study on the effectiveness of Rural PNPM in the North Denpasar district gives an indication that the program selection and impact is not yet pro-poor. The results also imply that the development of the program’s effectiveness needs strong social capital, especially in the context of bridging the social capital between the poor and the elite groups in the community. This also seems to be the case for cooperatives and village savings and loans, where members are very seldom involved in decision-making, and only a few elite groups in the cooperatives make policy. The regulatory framework is disengaged from the local values (adat), resulting in a lack of trust from the people on the cooperatives’ function (Sugarda, 2016). These arguments are also supported by the data sample, where only 6.8 percent and 4.6 percent of households participated in cooperatives and village savings and loans, respectively. These numbers are even lower when considering only the families of FMWs, where only 5.6 percent and 2.42 percent participated in these same two activities.

Nevertheless, the results confirmed the initial hypothesis, in which having social capital contributed strongly to an increase in investments in productive assets for the families of FMWs, rather than on consumption. The results also show that the highest coefficient occurs from the interaction between the dummy of the families of FMWs and the dummy of youth activity groups (Karang Taruna). The Karang Taruna itself was formed by the government under the Ministry of Social Affairs to help young people develop and to empower them through productive economic activities (Goodwin and Martam, 2014). Therefore, when one of the household members is active in the Karang Taruna, the probability that a household will invest in productive assets is relatively high. This result is corroborated by Jumirah and Wahyuni (2018), who also found that the youth group activity contributes significantly to household welfare.

The results shown in Tables 4 and 5 also demonstrate that the role of community meetings in inducing investment activities by the remittance-receiving households is essential, and its parameters are only behind those of the youth group activity. Not only would community meetings widen the household local network—essential in building an enterprise—but those who are very active in the community have a high probability of having community knowledge. In this context, people are then well informed about what is happening in their communities and may have the upper hand in detecting demand in the community, relative to those who are not very active.

Interestingly, active participation in religious activities also has a positive impact on investment activity, albeit a weak one. This is also borne out in our fieldwork results, where in one of the focus group discussions with spouses of FMWs in Juntinyuat on 26 August 2018, the role of local clerics (ulama or ustadz) emerged as quite significant in developing a small enterprise. The cleric explained that in Muslim religious values, the husband should be the sole (or at least the main) provider for the household. Following the cleric’s advice, the husband used the remittance sent by his wife to build small enterprises, which in turn increased his income, and he became the principal provider for his household.

The results, however, lead to one quite surprising conclusion. Whereas the access to credit does not result in a significant increase in investment, it shows significant statistical results and positively affects consumption (4.1–4.6). This implies that the majority of households that have access to credit do not take out loans for the purpose of investment. This can be explained by our sample selection, where 50.8 percent of the sample consists of poor or vulnerable households. Therefore, when these households have the opportunity to take out loans, they tend to use the funds to meet their consumption needs.

Additionally, house ownership status has the most significant impact on increases in investment. One explanation could be that those who legally own a house can optimize that ownership status by collateralizing the house and taking out loans for investment purposes. Owning their houses legally also indicates that the household is not poor and most likely has a relatively high income such that they would have additional funds in productive assets. Furthermore, the higher the education the head of household has, the higher both consumption and investment are, which is in line with the hypothesis of our model.

The results of the equation (3) regressions in Table 5 imply three principal areas of social capital that could induce household expenditure on productive assets, namely community meetings, youth groups and religious activities. Therefore, we have constructed PSC, which is the amount of participation of households in these three activities. It is assumed that households that participate in all three PSC activities will have significantly higher investment participation than households that only participate in one of the activities. The PSC then acts as the instrument variable of household investment to predict equation (2). Following Levitt (1997, 2002), we tested whether the PSC can be a valid and robust instrument of household investment. We present the regression results of the impact of participation in PSC on household investment in the Appendix. The results imply that there is a strong and significant correlation, which points to a valid and robust independent variable. The results also imply that participation in more than one activity would have a higher impact on the level of investment. Using this approach, the regression results of IV Probit based on equations (2) and (3) are presented, along with robustness tests in Table 6.

Table 5.

Regression results for the impact of social capital ownership on investment.

	(5.1)	(5.2)	(5.3)	(5.4)	(5.5)	(5.6)
FMW*social capital	2.824**	1.495	6.082***	1.538*	0.831	1.569
FMW*social capital	(1.369)	(3.168)	(1.680)	(0.845)	(4.671)	(2.444)
Access to credit	−0.053	−0.014	−0.079	−0.075	−0.062	−0.052
	(0.309)	(0.308)	(0.307)	(0.309)	(0.308)	(0.308)
Home ownership status	4.477***	4.515***	4.393***	4.462***	4.494***	4.487***
	(0.240)	(0.238)	(0.238)	(0.239)	(0.238)	(0.238)
University education	0.744**	0.765***	0.647**	0.710**	0.744**	0.752**
	(0.317)	(0.327)	(0.326)	(0.328)	(0.327)	(0.327)
Constant	2.139***	2.150***	2.063***	1.438***	2.120***	2.129***
	(0.315)	(0.316)	(0.315)	(0.316)	(0.315)	(0.315)
R-squared	0.053	0.052	0.052	0.052	0.053	0.052
F-stat	62.02***	61.65***	66.99***	62.85***	62.36***	61.67***
Observations	4,338	4,338	4,338	4,338	4,338	4,338

Note: Numbers in parentheses are standard error; significant at α = *10%, **5%, ***1%.

Source: Authors’ calculations.

Table 6.

Regression results of Instrumental Variable (IV).

Dependent: poor or vulnerable	(6.1)	(6.2)
Investments	−0.124***	−0.124***
	(0.002)	(0.002)
Entrepreneurs		−0.113***
		(0.031)
Cigarettes		0.027**
		(0.013)
Constant	0.740***	0.765***
	(0.018)	(0.024)
Instruments:	Yes	Yes
Principal social capital	Yes	Yes
R-squared	–	–
F-stat/Wald (IV)	4811***	78.34***
Observations	4,338	4,338

Note: Numbers in parentheses are standard error; significant at α = *10%, **5%, ***1%.

Source: Authors’ calculations.

After instrumenting the investment by using the PSC, presented in column (6.1), the results are rational and consistent, and it was found that the higher the level of household investment in productive assets, the lower the probability that the household is poor or vulnerable. Even after entrepreneur membership and cigarette consumption were controlled for, the results in (6.2) were robust and consistent. Note that we did not include the consumption variable in the Table 6 regression, since the results in Table 4 indicate that variables of social capital are not a valid instrument in predicting changes in household consumption level.

Based on our study, three types of social capital—community meetings, youth groups and religious activities—serve as the drivers for spending on productive assets by the households of FMWs. Participation in these activities by the FMW’s spouse (most likely in community meetings and religious activities) and children (youth groups) creates social networks that can open up economic or entrepreneurial activity. Therefore, remittances sent by the FMWs would have a useful and efficient impact and have a multiplier effect on their welfare. As a result, the probability of FMWs not continuing their work abroad in the future is higher. This is very important, since the presence of the mother has a significant effect on the development of children, especially on the development of cognitive skills (Adda et al., 2011; Brilli, 2015).

Conclusion

There is doubt as to whether remittances sent by FMWs can actually have a sustainable impact on the total welfare of their families. The effective use of remittances is essential for the FMWs to resume their roles as the mothers and/or wives, and past studies indicate that the role of women in the development of families, especially regarding any offspring, is an important one. We argue that the role of social capital in FMWs’ families is important in inducing the effective use of remittances from FMWs.

Based on IFLS data in the largest source provinces of FMWs and employing the instrumental variable approach, some interesting results were obtained. First, the majority of FMWs’ families are poor, and there is strong empirical evidence that the wife or mother chooses to work abroad to alleviate her family’s poverty status. Second, the results of this study also present strong evidence that social capital, measured by active participation in local-level associations, can indeed encourage the use of remittances sent by the FMW for productive assets, which subsequently has a relatively high probability of raising the FMWs’ welfare and alleviating poverty in general. However, in three local-level associations, active participation is not enough to induce household expenditure on productive assets. These are the cooperatives, village savings and loans and the PNPM. This finding suggests that the Indonesian government needs to increase the effectiveness of these associations at the local level.

In general, the results imply that the government should place more emphasis on how to use social capital to improve the life of the families of FMWs. Encouraging participation in entrepreneurship activities for the husband and children of FMWs and promoting the usefulness of local-level associations to widen social networks could have a multiplier effect on a family’s welfare. This is important since the absence of FMWs in the long term may have negative implications for family development, especially where children are concerned.

Data gathered from fieldwork and other documents show that FMWs migrate mostly for consumption and their children’s education. However, as intrinsic assets such as childhood education are not yet regarded as investments in our research, this could be on the next agenda. Future research could focus on exploring whether or not migration increases the children’s level of education, since one of the main motivations for FMWs to migrate is to finance their children’s education. Nevertheless, this study’s results also show the importance of investment activities because if FMWs’ families successfully manage the extra funds from remittances, this will contribute to the early return of FMWs.

Footnotes

Acknowledgment

The authors thank Mr. Windhiarso Ponco Adi Putranto from Statistics Indonesia for his assistance in data analysis.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this paper.

Funding

This research was funded by the Directorate Research and Community Service, Directorate General of Research Empowerment and Development, Ministry of Research, Technology and Higher Education of the Republic of Indonesia.

ORCID iDs

Rulyusa Pratikto

Sylvia Yazid

1

In comparison, the interviews revealed that unmarried women may have non-economic reasons to work abroad, such as wanting to be independent. This observation was then strengthened by the results of the follow-up interviews conducted in five origin areas, namely Indramayu, Wonosobo, Jember, Lombok and Kupang, in 2019.

2

Program Nasional Pemberdayaan Masyarakat (National Program for Community Empowerment and Self-Reliance) is a community-driven development program established in 2007 by the Government of Indonesia. It covers both villages and urban areas. It is funded by the World Bank and the Islamic Development Bank.

3

The measurement of identification means that we cannot include female migrant workers (FMW) who are not married (or have ever been married) since the Indonesian Family Life Survey (IFLS) does not provide such data. However, as has been explained earlier, the majority of FMWs' are married or have children; therefore, the sample from IFLS data using this method of identification represents the majority of FMWs' households..

4

As explained in the website of the National Team for the Acceleration of Poverty Reduction (2020), Beras Untuk Rumah Tangga Miskin or Rice for Poor Households is a dedicated food subsidy program for poor and vulnerable households. Launched in 2002, it is part of government efforts to boost food security and provide social protection to these households. Eligible households are given a Raskin Card that allows them to buy rice at a much lower price, approximately 10–15 percent of the market price ().

References

Adams

RHJ

(2006) Remittances and poverty in Ghana. World Bank Policy Research Working Paper 3838. Washington, DC: World Bank, February 2006. Available at: https://openknowledge.worldbank.org/handle/10986/8781

Adams

RHJ

Cuecuecha

(2016) Remittances, household expenditure, and investment in Guatemala. World Development 38(11): 1626–1641.

Adams

RHJ

Page

(2005) Do international migration and remittances reduce poverty in developing countries? World Development 33(10): 1645–1669.

Adda

Bjorklund

, and Holmlund

(2011) The role of mothers and fathers in providing skills: Evidence from parental deaths. Discussion Paper Series No. 5425, Bonn: IZA Institute of Worker Economics, January 2011. Available at http://ftp.iza.org/dp5425.pdf

Anwar

Chan

(2016) Contrasting return migrant entrepreneurship experiences in Javanese villages. International Migration 54(4): 1–14.

Bank Indonesia (2020) Table V.31 Remitansi Tenaga Kerja Indonesia (TKI) Menurut Negara Penempatan in Statistik Ekonomi dan Keuangan Indonesia (SEKI) [Remittance of Indonesian migrant workers based on placement countries in economic and financial statistics]. Available at: https://www.bi.go.id/seki/tabel/TABEL5_31.pdf

Bellemare

Masaki

Pepinsky

(2017) Lagged explanatory variables and the estimation of causal effect. The Journal of Politics 79(3): 949–963.

Borja

(2014) Social capital, remittances and growth. European Journal of Development Research 26(5): 574–596.

BP2MI (2020) Laporan Pengolahan Data BNP2TKI 2019 [Report of BNP2TKI’s 2019 Data Processing]. Available at: http://www.bnp2tki.go.id/uploads/statistik/images/data_19-02-2020_Laporan_Pengolahan_Data_BNP2TKI____2019(2).pdf

10.

Brilli

(2015) Mother’s time allocation, child care, and child cognitive development. EUI working paper MWP 2015/03. Fiesole, Italy: Max Weber Programme, European University Institute, 8 January 2015. Available at: http://conference.iza.org/conference_files/transatlantic_2014/brilli_y7225.pdf

11.

Buzas

Stefanski

(1996) Instrumental variable estimation in a probit measurement error model. Journal of Statistical Planning and Inference 55(1): 47–62.

12.

Chami

Adolfo

Thomas

, et al. (2008) Macroeconomic Consequences of Remittances. Washington, DC: International Monetary Fund.

13.

Chan

(2014) Gendered morality and development narratives: The case of female labor migration from Indonesia. Sustainability 6(10): 6949–6972.

14.

Chan

(2018) In Sickness and in Wealth: Migration, Gendered Morality, and Central Java. Bloomington: Indiana University Press.

15.

Cuecuecha

Adams

RHJ

(2016) Remittances, household investment and poverty in Indonesia. Journal of Finance and Economics 4(3): 12–31.

16.

Goodwin

Martam

(2014) Indonesian Youth in the 21st Century. Jakarta: United Nations Population Fund, Indonesia, July 2014. Available at: https://indonesia.unfpa.org/sites/default/files/pub-pdf/Indonesian_Youth_in_the_21st_Century_%28Youth_Mapping%29.pdf

17.

Grootaert

(1999) Social capital, household welfare, and poverty in Indonesia. World Bank Policy Research Working Paper No.WPS 2148. Washington, DC: World Bank, 31 July 1999. Available at: http://documents.worldbank.org/curated/en/738431468751146917/Social-capital-houshold-welfare-and-poverty-in-Indonesia

18.

Gupta

SPC

Pattillo

Wagh

(2007) Impact of remittances on poverty and financial development in Sub-Saharan Africa. IMF Working Paper WP/07/38. Washington, DC: International Monetary Fund, 1 February 2007.

19.

Haas

(2017) Myths of migration much of what we think we know is wrong. Spiegel Online. 21 March 2017. Available at: https://www.spiegel.de/international/world/eight-myths-about-migration-and-refugees-explained-a-1138053.html

20.

Human Rights Watch (2016) Migrant domestic workers: Overworked and underprotected. Available at: https://www.hrw.org/news/2016/06/15/migrant-domestic-workers-overworked-and-underprotected

21.

Jumirah

Wahyuni

(2018) The effect of social capital on welfare in Indonesia. Journal of Indonesian Economy and Business 33(1): 65–76.

22.

Kiboro

(2017) Influence of social capital on livelihood outcomes for the internally displaced persons in Kenya: A social capital approach. European Scientific Journal 13(26): 266–280.

23.

Killias

(2018) Follow the Maid: Domestic Worker Migration in and from Indonesia. Copenhagen: Nias Press.

24.

Kirwati

Setiawina

Yasa

IGWM

(2018) Efektivitas dan Dampak Program Nasional Pemberdayaan Masyarakat-Mandiri Perdesaan (PNPM-MPd) di Kecamatan Denpasar Utara. [The effectiveness and impacts of the National Program on Community Empowerment of Rural Independency in the sub-district of North Denpasar]. E-Journal Ekonomi dan Bisnis Universitas Udayana 7(2): 367–396.

25.

Levitt

(1997) Using electoral cycles in police hiring to estimate the effect of police on crime. The American Economic Review 87(3): 270–290.

26.

Levitt

(2002) Using electoral cycles in police hiring to estimate the effect of police on crime: Reply. The American Economic Review 92(4): 1244–1250.

27.

Lindquist

(2010) Labour recruitment, circuits of capital and gendered mobility: Reconceptualizing the Indonesian migration industry. Pacific Affairs 83 (1): 115–132.

28.

Lindquist

Xiang

Yeoh

BSA

(2012) Opening the black box of migration: Brokers, the organization of transnational mobility and the changing political economy in Asia. Pacific Affairs 85(1): 7–19.

29.

Murray

(2006) Avoiding invalid instruments and coping with weak instruments. The Journal of Economic Perspectives 20(4): 111–132.

30.

Nahar

Arshad

MNM

(2017) Effects of remittances on poverty reduction: The case of Indonesia. Journal of Indonesian Economy and Business 32(3): 163–177.

31.

Nasrudin

Trialdi

Hartono

Ahsan

(2013) Tobacco economic of Indonesia: Poor households’ spending pattern, tax regressivity and economic wide impact of tax simplification. Working Paper in Economics and Business No. 201302. Jawa Barat: Faculty of Economics and Business, University of Indonesia.

32.

Palmer

(2016) Indonesia’s Overseas Labour Migration Programme, 1969–2010. Leiden: Brill.

33.

Platt

Yeoh

BSA

Baey

, et al. (2013) Financing migration, generating remittances and the building of livelihood strategies: A case study of Indonesian migrant women as domestic workers in Singapore. Migrating Out of Poverty RPC Working Paper 10. Migrating Out of Poverty Consortium. Brighton, UK: University of Sussex, 1 January 2013. Available at: http://www.migratingoutofpoverty.org/files/file.php?name=wp10-platt-et-al-final.pdf&site=354

34.

Portes

(1998) Social capital: Its origins and applications in modern sociology. Annual Review of Sociology 24(1998): 1–24.

35.

Quach

Mullineux AW Murinde

(2005) Access to Credit and Household Poverty Reduction in Rural Vietnam: A Cross-sectional Study. Birmingham: The University of Birmingham. Available at: http://www.grips.ac.jp/vietnam/VDFTokyo/Doc/1stConf18Jun05/OPP01QuachPPR.pdf

36.

Rahman

(2009) Gender dimension of remittances: A study of Indonesian domestic workers in East and Southeast Asia. Bangkok: United Nations Development Fund for Women (UNIFEM) East and Southeast Asia Regional Office. Available at: http://ssrn.com/abstract=1938306–

37.

Randolph

(2015) Worker Migration and Inclusive Growth: Toward Creating Employment in Origin Communities. The Transformation of Work: Challenges and Strategies Research Series. Washington, DC: Solidarity Center. Available at: https://www.solidaritycenter.org/wp-content/uploads/2015/05/Migration.Toward-Creating-Employment-in-Origin-Communities.5.15.pdf

38.

Rustiadi

Nasution

(2017) Can social capital investment reduce poverty in rural Indonesia? International Journal of Economics and Financial Issues 7(2): 109–117.

39.

Silvey

(2004a) Consuming the transnational family: Indonesian migrant domestic workers to Saudi Arabia. Global Networks 6(1):23–40.

40.

Silvey

(2004b) Transnational domestication: State power and Indonesian migrant women in Saudi Arabia. Political Geography 23(3): 245–264.

41.

Spitzer

(2016) Return migrant entrepreneurship and the migration and development agenda: A focus on Filipino and Indonesian migrant workers. Migration, Mobility & Displacement 2(2): 24–39.

42.

Sugarda

(2016) Cooperatives: Indonesia’s sleeping giant or a total failure? A legal perspective. The Social Sciences 11(18): 4338–4392.

43.

The National Team for the Acceleration of Poverty Reduction (2020) At a glance. Available at: http://tnp2k.go.id/program/using-the-kps-for-the-raskin-programme

44.

World Bank Group (2017) Indonesia’s Global Workers: Juggling Opportunities and Risks. A World Bank Indonesia Report, November 2017. Washington, DC: The World Bank. Available at: http://pubdocs.worldbank.org/en/357131511778676366/Indonesias-Global-Workers-Juggling-Opportunities-Risks.pdf

45.

World Bank Group (2018a) Moving for Prosperity: Global Migration and Labor Markets, Policy Research Reports, 14 June 2018. Washington, DC: The World Bank. Available at: https://openknowledge.worldbank.org/bitstream/handle/10986/29806/211281ov.pdf

46.

World Bank Group (2018b) After five years, PNPM Mandiri becomes an integral part of the development of communities across Indonesia. Available at: http://www.worldbank.org/en/news/feature/2012/08/07/after-five-years-PNPM-Mandiri-becomes-an-integral-part-for-the-development-of-communities-across-indonesia0

47.

Yazid

(2013) Indonesia’s Civil Society in the Age of Democratization: NGO Responses on the Issue of Labour Migration. Baden-Baden: Nomos Publisher.

48.

Yuniarto

(2015) Indonesian migration industry in Taiwan: Some socio-economic implications and improvement challenges. Jurnal Kajian Wilayah [Journal for Area Studies] 6 (1): 17–33.

49.

Yoshino

Taghizadeh-Hesary

Otsuka

(2017) International remittances and poverty reduction: Evidence from Asian developing countries. ADBI Working Paper Series No.759, July 2017. Tokyo: Asian Development Bank Institute. Available at: https://www.adb.org/publications/international-remittances-and-poverty-reduction