Keywords

1 Introduction

With the acceleration of globalization and the widening income gap between high-income countries and so-called developing countries, migration from the latter to the former is inevitable. In 2019, the estimated stock of emigrants worldwide was 271.6 million, with India being the largest migrant-sending country, recording an overseas population of 17.5 million (UNDESA, 2020). Due to the current COVID-19 pandemic, as of the end of April 2021, the migration stream has almost halted; however, it will certainly resume once the situation improves.

Since international migrants, particularly labor migrants, are the source of international remittances, the large stock of migrants abroad is naturally translated into the large amount of money sent back to their home countries.Footnote 1 In 2019, the total size of remittances worldwide was US$714 billion,Footnote 2 of which US$554 billion was sent to developing countries. India and China, which not only dispatch many migrant workers overseas but have also established a vast overseas human network, are by far the largest remittance-recipient countries, followed by the Philippines and Mexico.Footnote 3 India and China received remittances of US$83.1 billion and US$68.4 billion respectively in 2019. South Asia is the largest remittance-receiving region, accounting for 25.1% of the total remittances to developing countries in 2019. The size of remittances to developing countries is equivalent to 1.55% of the combined GDP of all developing countries. This number may seem insignificant; however, in many developing countries, such remittances account for more than 10% of their GDPs. For example, in 2019, they comprised more than 30% of the GDPs of countries such as Tajikistan, Kyrgyz, Tonga, and Nepal, which constitutes a significant contribution to these nations’ economies (KNOMAD, 2020).

Remittance flows to developing countries have been steadily increasing, and the incremental rate has accelerated since the early 2000s. They have exceeded the amount of official development aid (ODA) to developing countries since the early 1990s. In 2019, the total volume of remittances was more than three times the size of ODA (US$167.8 billion in the same year)Footnote 4 and has surpassed the size of foreign direct investment (FDI) in developing countries for the first time since the beginning of 1990s (KNOMAD, 2020).

Pakistan has been recognized as one of the most active migrant-sending countries. The stock of Pakistani emigrants was estimated to be 6,303,286 as of 2019 (UNDESA 2019), the eighth largest in the world. In 2019 alone, 625,203 Pakistanis left the country to work abroad through official channels, mainly to oil-producing countries in the Gulf region, such as Saudi Arabia (ILO, 2019). Due to this large volume of overseas migrants and the continued flow of labor migration, the remittance flow to Pakistan reached US$22.5 billion in 2019, representing roughly 7.9% of the nation’s GDP. This amount of remittances was the 8th largest in the world.

Remittances from overseas migrants impact the macro and micro economic and social dimensions of recipient countries in several ways. At the macro level, remittances are an important source of foreign exchange and contribute to the stability of the balance of payments. At the micro level, remittances are also important sources of income for recipient households and ease the credit constraints they face. Remittances are spent variously according to household needs. Because of their importance, numerous studies have examined how remittances are used and their impact on the well-being of households in developing countries.Footnote 5

This chapter examines the impact of remittances on children’s school enrollment in Pakistan. Since developing the human capital of children is almost the only way for poor households to escape poverty in a sustainable fashion, it is crucial to analyze whether remittance-financed expenditures on education have increased child enrollment. Several studies have traced the effects of migration and remittances on child education in Pakistan. The evidence from these studies is mixed: some find that remittances affect child education positively, while others observe negative or no impacts.

Analyzing the 2000–2001 Pakistan Socio-Economic Survey, Arif (2004) shows that the impact of foreign remittances is positive for girls’ school enrollment but insignificant for that of boys. Mansuri (2006), using data from the Pakistan Rural Household Survey 2001–02, concludes that temporary migration has a positive and significant impact on child education. Remittances from migrants increase investments in child schooling, especially for girls. Based on a field survey in rural Chakwal, Punjab province, Oda (2007) reports a positive influence of internal remittances on school enrollment but an insignificant impact of international remittances. Ahmed et al. (2010), using data from the Household Integrated Economic Survey of Pakistan 2005–06, find that the share of expenditures on education increases with the availability of remittances for migrant households, but the increment was only 2.9%. Khan and Khan (2016) use the Pakistan Social and Living Standards Measurement (PSLM) Survey2010–11 and show a positive impact of remittances on child school enrollment, in particular, on girls’ enrollment.

In contrast, Nasir et al. (2011) have analyzed household data from four cities in Khyber Pakhtunkhwa and report a non-significant effect of migration and remittances on the school enrollment of children belonging to migrant households. They argue that the positive impact of remittances on school enrollment is countered by the negative impact of the absence of fathers or other close relatives due to overseas migration. A similar result is also observed by Hassan et al. (2013), while Arif and Chaudhry (2015) find otherwise. Focusing on a set of Punjab household data, Arif and Chaudhry (2015) show a significant positive influence of overseas migration on the school enrollment of younger children and explain that the positive income effect of remittances outweighs the negative effect of parental absence as a consequence of migration.

The current study aims to add new insights and knowledge to existing studies by investigating the issue from different perspectives. The study differs from existing studies on the relationship between remittances and school enrollment in several dimensions. First, school enrollment in this study considers attendance at middle and secondary (matriculation-level) schools but not primary schools. Middle and secondary schools in Pakistan correspond to an age range of about 11 to 15 years old (middle school: 11–13 years old; secondary school: 14–15 years old). Previous studies have typically examined the impact of remittances on enrollment at the primary school level or on enrollment regardless of the grade level. However, given the differences in the nature of primary and post-primary schooling, this study focuses on middle and secondary school enrollment. Post-primary education is essential for skill development and higher income generation, whereas primary education is limited to providing students with basic knowledge and skills. Secondary education is also effective for narrowing the income gap in society (Abdullah et al., 2015). Middle and secondary school enrollments are considerably lower than primary school enrollments in Pakistan. The net enrollment rate at the primary level (ages 6–10) was 66%, while the rates at the middle level (ages 11–13) and the secondary level (ages 14–15) were 38% and 27% respectively in 2018–19 (Government of Pakistan, 2019). Analyzing this trend could potentially increase our understanding of it and provide policy implications.

In addition, this study considers the school enrollment of relevant age children. That is, only children in middle or secondary schools within the relevant age range were counted as enrolled. It is clear from the gap between gross enrollment rates and net enrollment rates that overage schoolchildren are common in Pakistan. One extreme example in our dataset was a 15-year-old child who was enrolled in Grade 1 (the grade for six-year-olds). These were counted as enrolled children in many previous studies. Lastly, remittances are categorized as either internal (received from a source in Pakistan) or international (sent by migrants abroad). The nature of these two kinds of remittances is the same, as they comprise money, but the characteristics of senders and receiving households tend to differ, and the volume of overseas remittances per household is usually higher. Thus, it would be inappropriate to conflate these two types of remittances.

In this study, rural household data on Khyber Pakhtunkhwa province, taken from the PSLM Survey 2014–15, are examined. As Khyber Pakhtunkhwa is underdeveloped and rural areas have limited job opportunities, remittances are considered important sources of additional income for rural households in this province.

The organization of this chapter is as follows: The next section briefly presents the situation of overseas migration from Pakistan and the trends of international remittance flows to Pakistan. The following section analyzes household data and estimates the impact of two types of remittances on middle and secondary school enrollments of relevant age children. Then estimated results and findings are presented. The last section concludes.

2 Migration and Remittances in Pakistan

Pakistan is the 8th largest emigration country in the world. According to UN data, the number of Pakistani out-migrants in 2019 was around 6.3 million (UNDESA 2019).Footnote 6 In regional terms, the largest concentration of Pakistanis is found in the six countries of the Gulf Cooperation Council (GCC): Saudi Arabia, the United Arab Emirates (UAE), Kuwait, Oman, Qatar, and Bahrain, which host 3.3 million Pakistanis. Out of these, 1.4 million are in Saudi Arabia, and nearly 1 million are in the UAE. As of 2019, other major destinations are the United Kingdom (with a Pakistani population of 605,016), the United States (406,509), and Canada (215,409). The majority of Pakistani migrants in the GCC states are labor and temporary migrants, which means that they return to Pakistan when their working permits expire. Conversely, the majority of migrants in Western countries, such as the UK and the US, are mostly permanent and naturalized in those respective countries.

Large-scale labor migration from Pakistan started with the economic boom in the Middle East after the quadrupling of oil prices by OPEC countries in 1973.Footnote 7 The number of Pakistani workers who went overseas through formal channels was 3534 in 1971, which increased to 140,445 in 1977.Footnote 8 With increasing oil revenues, the Gulf countries embarked on development projects that created a huge demand for labor. In particular, manual workers were in high demand on construction sites. Because of surplus labor and the prevalence of poverty in rural areas, particularly in rain-fed farming areas in the northern parts of the country, Pakistani workers flew to the Gulf countries to grab these new job opportunities. Geographical and religious proximities were also factors that influenced their decision to migrate.

This trend continues today. In 2019, 625,203 Pakistanis left the country to work overseas.Footnote 9 The number of outflows of Pakistani workers increased from 382,439 in 2018 but declined from the peak of 946,571 in 2014. In 2019, 53.2% migrated to Saudi Arabia and 33.8% to UAE. If labor migrants to Oman and Qatar are included, these four countries received 95% of Pakistani migrant workers (ILO, 2019). Among the four provinces in Pakistan, most labor migrants came from Punjab province, followed by Khyber Pakhtunkhwa. In 2019, migrants from Punjab constituted 50% of all labor migrants, and migrants from Khyber Pakhtunkhwa constituted 29.8% (ILO, 2019). In terms of migrants per million inhabitants, the figure is 5925 migrants for Khyber Pakhtunkhwa and 2860 for Punjab.Footnote 10 As for the origin of labor migrants, Swat and Lower Dir of Khyber Pakhtunkhwa are recorded as the most and the second-most migrant-sending districts among all districts in Pakistan. Thus, such evidence verifies the significant migration activity in Khyber Pakhtunkhwa that this chapter examines.

Since international migrants, particularly labor migrants, are the source of remittances from abroad, the large stock of Pakistani migrants is naturally translated into the large number of remittances sent back to Pakistan. The remittance flow to Pakistan has been constantly increasing since the beginning of the 2000s (Fig. 9.1). It was US$22.5 billion in 2019, representing roughly 7.9% of the nation’s GDP, the eighth largest in the world. Remittances from Saudi Arabia were largest in 2017, amounting to US$5.78 billion, followed by the UAE with US$5.67 billion and the UK with US$1.689 billion. Remittances from the six GCC countries constituted around 70% of all remittances to Pakistan in the same year.

Fig. 9.1
A line graph of remittance inflow in million dollars versus years from 1980 to 2018. It begins at 2500 in 1980 and continues to remain low at 2000 in 2000 and increases gradually to 23000 in 2018.

Trend of remittance flows to Pakistan. (Source: World Bank Remittance and Migration data https://www.worldbank.org/en/topic/migrationremittancesdiasporaissues/brief/migration-remittances-data)

Full size image

Internal labor migration (i.e., migration within Pakistan) also forms an important dimension of mobility in the country. Data regarding internal migration are not easily accessed, but it is possible to trace internal remittances from household surveys. Here, the PSLM Survey 2014–15, conducted by the Bureau of Statistics, Government of Pakistan, is employed to see how many households receive internal remittances by province and urban–rural classification, together with information on international remittances (Table 9.1).

Table 9.1 Internal and international remittances by province and region from the PSLM survey 2014–15
Full size table

Several characteristics of internal and international remittances in Pakistan are observed in Table 9.1. First, more households receive internal remittances than international remittance. As shown in the Table 9.1, 10.6% of the households are recipients of internal remittances, while 5.9% are recipients of international remittances. Second, rural households receive more internal and international remittances than urban households. In rural Pakistan, 11.5% of the households receive internal remittances, and 6.2% receive international remittances. Meanwhile, in urban Pakistan, the corresponding percentages are 6.2% and 4.9% respectively. Third, most of the remittance flows in Pakistan are concentrated in Khyber Pakhtunkhwa and Punjab provinces. It would be worthwhile to investigate such concentration of remittance flows as rural Sindh and Balochistan are as underdeveloped as rural Khyber Pakhtunkhwa. Fourth, Khyber Pakhtunkhwa has the highest percentage of households that receive remittances in all categories, showing their high dependence on remittances.

3 Data and Estimation

This study utilizes the Pakistan Social and Living Standard Measurement (PSLM) Survey 2014–2015 to investigate the effect of remittances on children’s middle and secondary school enrollment. The PSLM contains data on 78,635 households and 513,945 individuals from four provinces and from Islamabad. Of these, the rural household data of Khyber Pakhtunkhwa province are used for estimation. Located in the western part of Pakistan and adjacent to Afghanistan, Khyber Pakhtunkhwa is known as the home base for internal and international migrants. Because of the lack of industries other than agriculture, migration has been an important source of income for families in this province. Migration from rural areas is common where employment opportunities are limited.

Middle and secondary school enrollment is used as the dependent variable for this study. In Pakistan, the relevant age range for middle and secondary school students is 11 to 15 years old. For this study, the upper age limit is extended to age 16, allowing a one-year delay of schooling. A total of 13,752 children within this age range of 11 to 16 years old (male: 7616; female: 6136) and their households (7300 households) in rural Khyber Pakhtunkhwa are chosen for estimation. Children whose ages range from 11 to 16 but are enrolled in primary school are not counted as school-enrolled children; 2729 children fall into this category. Eliminating these children and 4756 children who never attended school or received less than one year of education in that age range results in 6267 school-enrolled children (45.6%) for this study, of which 4439 are males and 1828 are females. The number of households with school-enrolled child/children is 4211.

A total of 1834 households received internal remittances. The mean value in Pakistani rupees (PKRs.) was PKRs. 160,166 per recipient household. As for international remittances, 1346 households received them and its mean value was PKRs. 224,000 per recipient household. Among the recipient households, 153 received both internal and international remittances.

A probit estimation technique is applied to estimate the impacts of two types of remittances on the middle and secondary school enrollments of school-age children. The dependent variable is a binary variable that indicates a child’s school enrollment; it is assigned a value of 1 if the child is enrolled in a middle or a secondary school and a value of 0 otherwise. Remittances are classified into two categories: internal and international. These are included as dummy variables. Thus, the households are categorized as either households that receive internal remittances, households that receive international remittances, or households with no remittances. If a household receives both internal and international remittances, it is categorized as a household receiving international remittances. The reference category for remittance dummies is households that receive no remittances.

Other explanatory variables used for estimation consist of the gender of a child, the number of household members, the age of the household head, the gender of the household head, and the education level of the household head. Age, the number of household members, and the education level of the head are continuous variables, whereas the others are all dummy variables. The reference category for the gender of a child is female, and for the gender of the head, it is male. Table 9.2 provides the descriptive statistics and definitions of the dependent and independent variables.

Table 9.2 Summary statistics of the dependent and independent variables
Full size table

4 Estimated Results and Findings

The estimated results are reported in Table 9.3. Marginal effects and corresponding standard errors (in parenthesis) are shown. Three models are estimated. Model 1 uses the enrollments of both male and female children. Model 2 considers male children only, and model 3 utilizes female children only.

Table 9.3 Probit estimation of the impact of remittances on school enrollment: marginal effects
Full size table

In models 1 and 2, the marginal effects of receiving international remittances are positive as expected. The marginal effect of model 1 is 0.027 when both male and female enrollments are used, meaning that the probability of school enrollment increases by 2.7% if the household receives international remittances. The school enrollment increases by 4.3% when the household receives international remittances for model 2. The results indicate that international remittances help households finance schooling for their children, particularly for male children, and increase the probability of these children’s school enrollment. In fact, education is free at government schools for primary up to the secondary level of education under Article 25-A of the Pakistani Constitution. However, schooling is not free. It costs expenditures related to schooling, and the opportunity cost of a child going to school should also be considered a cost of schooling. International remittances ease the burden of such costs by increasing the income of the recipient household.

On the other hand, the positive impact of international remittances disappears for model 3, in which only female enrollments are considered. Once gender is considered, the effect of remittances vanishes for female children. This implies that international remittances increase the gender gap in school education instead of reducing it. This result contradicts the findings of Arif (2004), and Khan and Khan (2016), who found a positive influence of remittances on female enrollment. There are several interpretations for this difference. First, the dependent variables differ. The current study uses the middle and secondary school enrollment of children in a specific age range, whereas, for example, Arif (2004) uses primary school enrollment. Second, the difference can be explained by the parental preference for male children over female ones. In Khyber Pakhtunkhwa, the cultural norm of male domination still exists in society. Given budget constraints, parents prioritize male education. This tendency would be stronger for middle and secondary levels of education or higher. Third, as argued by Nasir et al. (2011), and Hassan et al. (2013), the impact of remittances and migration is the balance between a positive remittance effect and a negative migration effect. Female children may be affected more negatively by the absence of their parents because of overseas migration. As a result, the positive impact of international remittances might be canceled out by the negative impact of migration.

The marginal effect of internal remittances is negative for all three models. Receiving internal remittances reduces the probability of school enrollment by 6.2% in model 1, by 4.7% in model 2, and by almost 7.0% in model 3. The probability of school enrollment for female children declines more than that of male children. There are several possible explanations for this. One is that the size of internal remittances is about 70% of that of international remittances, on average. Internal remittances may not be sufficient for some recipient households to send their children to school. For households having difficulty in making both ends meet in rural areas, migration is a strategy to get out of poverty. Overseas migration is preferred because of higher remittances. However, as (Oda, 2007) argues, overseas migration opportunities are not available to every household, as they involve high direct and indirect costs. Based on his survey in Chakwal, Punjab province of Pakistan, he also finds a relatively high incidence of poverty among internal migrant households and concludes that migration within a country does not necessarily improve the economic conditions of migrant households. This is also the case for our results. Internal remittance-receiving households just spend on their daily needs, such as food items, leaving little for education. They are struggling for survival and do not have the luxury of sending their kids to middle or secondary school. As there is a tendency for parents to prioritize male children over female ones, female children are disadvantaged, and the probability of their middle and secondary school enrollment is lower than that of male children. Again, this contributes to widening the gender gap in education. This finding is consistent with Arif (2004), which notes that the gender gap persists within the migrant households.

All other explanatory variables significantly influence the probability of children’s school enrollment. Among those, gender significantly affects school enrollment. First, as the results indicate, the marginal effect of the gender of a child is 0.3. As the reference category is female, the probability of a child going to school increases by 30% for male children. A large disparity in education between males and females exists in Khyber Pakhtunkhwa and in Pakistan, overall. The gross enrollment rates in middle and secondary schools for males in Khyber Pakhtunkhwa are 82% and 68%, respectively, and the same rates for females are 46% and 31%, respectively (Government of Pakistan, 2019). In our sample, 1828 out of 6136 female children aged 11 to 16 years old are enrolled in a middle or secondary school, whereas 4339 out of 7616 male children of the same age range are enrolled. The former figure corresponds to 29.8%, and the latter corresponds to 57.0%. This huge gap indicates parental preference for the education of sons over daughters and that females’ access to education is still limited compared with males in rural Khyber Pakhtunkhwa.

Second, interestingly, the marginal effect of the gender of the head is 0.174. As the reference category is male, this means that a household headed by a female has an increased probability of her child/children going to middle or secondary school by 17.4%. Generally speaking, mothers tend to care more about their children. Given autonomy in household decisions as the heads, they tend to spend more on the education and health of their children (Durrant & Sathar, 2000; Eswaran, 2002; Maitra, 2004; Chakraborty & De Prabal, 2015). The result is supportive of the findings of Mahmood et al. (2017). Mahmood et al. (2017) finds a positive impact of female headships on child school enrollment in Pakistan.Footnote 11

5 Conclusion

This study has examined the impact of internal and international remittances on middle and secondary school enrollment in the rural Pakhtunkhwa province of Pakistan. The estimated results have indicated that (1) international remittances influence male school enrollment positively but do not affect female enrollment, and (2) internal remittances reduce the probabilities of school enrollment for both male and female children; in particular, females’ probability declines much. While international remittances increase the income of recipient households and help finance the education of their children, such positive impact is not visible for female children’s enrollment. These findings imply that remittances exacerbate the gender gap in middle and secondary school education that already exists.

Educating females is important in every aspect of the economic and social development of the country. For example, education allows females to work outside and earn income for their families. In turn, females are empowered and become more autonomous, which increases their intra-household bargaining power so that they can participate in household decision making (Dyson & Moore, 1983). As is well known from the findings of existing studies, these positive impacts of female education eventually lead to lowered mortality and fertility rates, reduced child labor, more educated children, and so on (Caldwell (1979), Dyson and Moore (1983), Luz and Agadjanian (2015)). Educated mothers are taking more care of their children. Unfortunately, as the current study has argued, females are not given priority for middle and secondary education in rural Pakistan. There is much space for Pakistan to improve this situation so that it can further develop as a country.