Do Remittances Encourage Investment in Education? Evidence from Eritrea

Dr. Temesgen Kifle

Abstract The purpose of this study is to estimate the impact of remittances on child education in Eritrea using sample data from remittance-receiving households that have school-age children. Key findings from the regression analysis show that households’ education ratio (the ratio of remittances spent on child education to the total amount of remittances received by households) decreases with the increase in remittances received; however, there is still a positive relationship between remittances and expenditure on child education out of remittances received. The remittance elasticity for investment in child education and the marginal remittances share to child education confirm that remittance-receiving households spend a higher proportion of remittances on child education. Unlike the conventional view that household income from remittances is used exclusively for consumption purposes, the paper argues that, in Eritrea, remittance-receiving households with school-age children spend a significant portion of the remittances on child education.

JEL Classification: F22; J24

Keywords: Remittances; Child education; Eritrea

Introduction

There is a general view that remittances are mostly spent on consumption items and are not used for productive investment that would have an impact on long-run development. This view ignores, however, that spending remittances on education and training is, in fact, investment in human capital. In addition, there are a number of expenditure items that are considered as consumption, but, directly or indirectly, have a positive effect on child education. For instance, expenditure on child health, nutrition, and clothing are part and parcel of investment in human capital. Many studies suggest that the use of remittances for education has significant social and economic effects for individuals, their families and society at large.

Eritrea’s small economy depends on remittances from Eritreans overseas. It is known that remittances from abroad are equivalent to 40-50% of GDP. In many Eritrean households, remittances are an important source of income.

It is estimated that one-quarter of the Eritrean population lives outside the country. The Eritrean Diaspora is a community of Eritrean people outside the country who generously support their families and relatives in Eritrea with recurrent transfers (Yamauchi, 2004). The available evidence indicates that Eritreans overseas have strong links with each other and with communities of origin. A survey by Koser (2002) found that the majority Eritreans (living abroad) maintained contact with their relatives and had returned to Eritrea to visit families and friends within the last two years of the survey. It is also found that most migrants send remittances on a fairly regular basis (Koser, 2003). In urban areas of Eritrea, it is estimated that assistance from households in Eritrea and abroad accounts for about 60% of all transfer income and almost one out of three households in Eritrea has income from gifts and transfers as its main source of livelihood (Arneberg and Pedersen, 2001). [1] The findings from the household consumer survey in Eritrea show that 75% of the respondents receive remittances from abroad (Tewolde, 2005).

Although there is a strong argument for the importance of remittances to recipient households in Eritrea, their use for education has not yet been analyzed econometrically. Thus, the aim of this study is to estimate the effect of remittances on child education in Eritrea. In a country where participation in education is low the use of remittances to fund education has significant socio-economic impact. In Eritrea school participation is still low, though enrolment rate has been rising continuously since independence. As a percentage of relevant age group, the Gross Enrolment Ratio (GER) at primary level is only 66%, and this is low when compared with the corresponding GER in low income countries, which is 100% (World Bank, 2006). Even if the total cost of education in Eritrea seems reasonable, it is not cheap if the average number of children per household (which is 5.1 children per woman) and the percentage of population below poverty line (which is around 50%) are taken into account. Besides, one has to consider the opportunity cost of education when calculating the total cost of education. Arneberg and Pedersen (2001) found that as much as 40% of girls and 20% of boys 15 or 16 years of age from the poorest quantile of urban households in Eritrea did not go to school and the majority dropped out without completing basic education. This implies that household poverty in Eritrea not only delays school start but also keeps a large number of children out of school. As reported in the 2002 Demographic and Health Survey (NSEO, 2003) enrolment ratios at all levels of education in Eritrea increase with the increase in the quintile of the wealth index. This indicates that household assets play a significant role in determining the demand for child education. From the above survey findings one can infer that remittance-receiving households have a greater opportunity of sending their children to school, thereby breaking the cycle of poverty and improving living conditions for future generations.

Together with this introductory part the paper is structured into five sections. Section two provides review of literature on the use of remittances. Section three focuses on data and methods of estimation. Section four offers estimated results. Finally, section five presents conclusion, implication and recommendation.

Literature Review

Although remittances flows are becoming increasingly important to the economies of the migrant-sending countries, data on the nature, volume and use of remittances is scant. The officially recorded remittances represent only the tip of the iceberg. The recorded and unrecorded remittances to Eritrea were estimated at US$400 million in 1999 (EIU, 2002). A recent survey shows that 66.7% of remittances to Eritrea are transferred informally (Tewolde, 2005).

For the period 1990 to 2001 Eritrea was the 10^th top remittance-receiving country in Africa (Sander and Maimbo, 2003). In terms of its dependence on remittances per capita Eritrea is ranked first in the African continent (da Cruz, Fengler and Schwartzman, 2004). On the external account, remittances from Eritreans in the Diaspora are the largest source of foreign currency inflows in the country. In 2003, remittances to Eritrea were 633% of exports and 30.1% of GDP (IMF, 2003). The large trade and fiscal deficit in the country has been financed by remittances sent by Eritreans living abroad (Addison, Le Billion and Murshed, 2001; Hansson, 2001).

It is known that poor households in high rainfall zones of Eritrea depend primarily on remittances, whereas poor households residing in medium rainfall zones report high dependence on international remittances as well as non-agricultural sources of income (GSE, 2004). A survey shows that a large number of households in Central administrative zone of Eritrea receive remittances from extended families living within and outside the country (NFIS, 2005)

For the economies of the migrants’ home countries, it is not just the scale of remittances but the way in which they are used is of crucial significance. The most widely recognized argument is that remittances are exclusively spent on consumption and other unproductive or personal items. The conventional view stresses that remitted earnings are really squanderables. However, such a pessimistic view has shifted to more optimistic beliefs as expenditure on health and education is increasingly seen as investment in human capital (Carling, 2004). Besides, there is an argument that to the extent the material well-being of those remittance recipients is improved, expenditure of remittances on food items should not be considered as simply unproductive because the formation of human capital needed for development cannot be realized without such consumption (Serageldin et al., 1983; Straubhaar, 1985; Choucri, 1986).

Research findings from Africa indicate the positive impact of remittances on education. Surveys of estimates for Zimbabwe by Lachaud (1999) and de Haan (2000), for Ghana by Schoorl et al. (2000), for South Africa by Sibanda (2004), for Somalia by Lindley (2005), and for Ghana and North Africa by Adams (2006) show that remittance-receiving households spend a great share of total income on education.

In Eritrea, an in-depth analysis of the use of remittances has not been done so far. The existing evidence shows only a descriptive analysis of the dataset. The proportion of remittances spent on education can be understated, if remittance dataset aggregates all remittance-receiving households, irrespective of school-age children recipients have. One cannot expect recipients without school-age children to spend much on education. To avoid the aggregation problem this paper focuses on remittance-receiving households that have children between the age 7 and 20. Though the age bracket for primary and secondary levels of education in Eritrea is 7-17, students might need extra years to accommodate late school start, repetition and dropout and reenroll.

Data and Method of Estimation

The data is randomly drawn from 125 remittance-receiving households that have dependent children between the age 7 and 20. Data collection was carried out by the author (during 2001-2002) in two administrative zones of Eritrea, namely Central and Southern administrative zones. Since the aim of the survey is to analyze the expenditure of remittances on investment in child education, respondents were asked to give information on their yearly income from employment, yearly amount of cash remittances received from abroad, yearly amount of remittances spent on child schooling, number of family members in the age group 7-20 who attend school, number of children in the 7-20 age bracket who do not go to school, father’s occupation (dummy), father’s level of education, and the total number of years the mother completed.

While the average years of schooling of the father were 9.5, the average number of years of schooling the mother completed was 7. Excluding remittances, households got about 20,387 Nakfa per year from employment on average. [2] On average, households received around 7,144 Nakfa per year from remittances. Despite the assertion that respondents might understate the total amount of (cash) remittances received, the proportion of yearly average remittances received by households to yearly average income from employment is not wide. This shows the largeness of the size of remittances from abroad. The share of remittances that was spent on investment in child education was around 26.2% on average. The average number of children between the age 7 and 20 who were not attending school was only around 0.6.

It is difficult to assess the monetary value of in-kind remittances received by households and thus the data only captures cash remittances.

Table 1 Descriptions, means and standard deviation of variables

Variables	Description	Mean	S.D.
Y	Yearly household income from employment	20,387	14,001
R	Yearly cash remittances received by households	7,144	11,401
SF	Years of schooling the father completed	9.5	5.06
SM	Years of schooling the mother completed	7.0	5.36
CS	Number of children between the age 7 and 20 who are in school	3.36	1.64
CNS	Number of children 7 to 20 years of age who are not in school	0.58	0.91
RE	Yearly remittances expenditure on child education	1,873	2,111.5
FOC	Father’s occupation, coded 0 for unskilled and 1 for skilled	0.328	0.47

In this study, the equation of interest is derived from the economic theory of consumer demand and thus a modified Engel function, which states the “education ratio”, is used. The dependent variable used in this study is the “education ratio”, and this variable measures the proportion of remittances which are allocated to education goods and services (i.e. household expenditure on education out of remittances received). Since the focus in this study is on remittance-education relationships, the chosen mathematical form of the model must have a slope that is free to change with remittances, which this implies that remittances elasticities as well as marginal propensities (out of remittances) to invest in education need to be calculated. Hence, a semi-log ratio function is chosen as a basic functional form. The equation for this can be expressed as:

where RE is yearly remittances expenditure on child education, R is yearly (cash) remittances received by households, lnR is logarithm of cash remittances received by households and b is a parameter for estimating educational expenditure out of remittances.

In using this functional form to compare the expenditure behaviour of households (i.e. households’ expenditure on child schooling out of remittances) with different sizes of remittances, a set of household characteristic variables should be taken into account. Considering a priori information, the observed differences in remittances’ expenditure on education are attributed to household income from employment, number of children between the age 7 and 20, parent’ years of schooling, father’s occupation, and some unobservable characteristics. The complete model can be written as:

The independent variables R and Y are logged at base e (natural logarithm). From this equation the remittances elasticity for investment in child education (E) and the average and marginal remittances shares (ARS and MRS respectively) can be derived as follows: [3]

While E refers to the proportional change in the amount of remittances spent on child education as a ratio of the proportional change in the total amount of remittances households receive, MRS shows the amount that expenditure on child education (out of remittances) changes in response to an incremental change in remittances received by households.

Estimated Results

From the regression results presented in Table 2, it is evident that remittances receiving households spend part of the remittances on child education, but the education ratio decreases with the increase in remittances received. With remittances’ expenditure on education as the numerator of the dependent variable, the negative and significant terms for yearly remittances received by households indicate that as total remittances rise, the share of spending on child education falls. This implies that households with high remittances spend a small proportion of their remittances on education on average than those who receive low remittances. As can be seen in Table 2, a 1% increase in remittances received leads to a 0.173 percentage points decrease in education ratio. However, households that receive higher remittances spend more on child education (out of remittances) than households that receive lower remittances.

Table 2 Regression results (Dependent Variable: Household Expenditure on Child Education out of Remittances (i.e. Education Ratio))

Variable
Constant	1.485	(3.989)
ln(R)	-0.173	(-8.239)***
ln(Y)	0.03915	(1.54)*
CS	0.03209	(2.440)**
CNS	-0.00546	(-0.226)
SF	-0.0118	(-1.765)*
SM	0.004664	(0.838)
FOC	0.06022	(0.969)
R2 (adjusted)	0.417
F-value	13.672***
Sample size	125

The positive relationship between households’ earnings (excluding remittances) and the education ratio indicates that families who get higher income spend a higher proportion of remittances on education, and vice versa. Since rich families can cover the cost of living from the earnings received, they can spend a higher proportion of their remittances on productive items, such as education. However, for those who earn low salary, the share of remittances that is spent on child education is also low because it is difficult for such families to spend much of remittances on child education without having enough to consume. Therefore, an increase of 1% in household income increases the education ratio by around 0.04 percentage points. The estimated regression shows that the education ratio increases by 0.03 if a household sends one more child to school. This is expected because families who send their children to school spend more remittances on education. So long as the expenditure of remittances on education is confirmed, the existence of children attending school will definitely influence the share of remittances that is spent on education. Father’s level of education has a negative effect on education ratio, though not important to have a significant effect. Each additional year of father’s education decreases the education ratio by 0.0118. This is due to the fact that the private rate of return on an additional year of schooling is quite substantial and thus parents can pay the cost of schooling for their children out of their earnings. [4]

Table 3 Remittances’ expenditure behaviour on education for remittance-receiving households

Households ranked by remittances received (%)	Households in each group (%)	Mean of remittances received (in Nakfa)	Mean of remittances’ expenditure on education (in Nakfa)	Remittance elasticity	Marginal remittances share to education
Lowest 20%	24.0	1,100	763.3	0.75	0.52
Second 20%	24.0	2,197	1,314.7	0.71	0.42
Third 20%	12.0	3,434	1,664.7	0.64	0.31
Fourth 20%	22.4	6,885.7	2,280.7	0.48	0.16
Top 20%	17.6	24,991	3,771.4	0.15	0.02
All	100	7,139.67	3,487.73	0.64	0.31

Table 3 shows the remittances’ expenditure behaviour on education for the 125 remittance-receiving households. Remittance-receiving households are first ranked into five quintile groups on the basis of remittances received. The average in each quintile is evaluated on the basis of the size of remittances received by households. The above findings are consistent with Engel’s Law. [5] Households in the poorest quintile group spend 52% of their increments to expenditure on education, but those in subsequent groups spend smaller proportions. At the mean level of remittances, households devote 31% of their increments (to remittances) to education. This implies that remittance-receiving households spend 31 cents on child education, if they get one extra Nakfa of remittances. [6] Similarly, remittances elasticities for expenditure on education are less than one for all quintiles, implying that the amount of remittances spent on child education is inelastic with respect to the total amount of remittances received by households. On average, a 10% increase in the total amount of remittances received would increase the amount of remittances spent on child education by 6.4%.

Conclusion, Implication and Recommendation

In Eritrea, worker remittances from abroad play a significant role in the country’s economy. Domestic output (GDP) is substantially augmented by remittances and many families in Eritrea depend on remittances for their livelihood. There are widespread criticisms of dependency on migrant remittances. The main argument is that remittances are spent on consumption and other unproductive items. However, such a view is no longer acceptable, as expenditures on education and training represent investment in human capital. Besides, those consumption goods that have a positive impact on education and health outcomes should not be considered as unproductive.

To assess the expenditure of remittances on child schooling in Eritrea data was collected from remittance-receiving households that have children between the age of 7 and 20. From the analysis it can be concluded that households with higher remittances tended to spend more on child education than households with lower remittances, but that the share of expenditure on child education (out of remittances) in the total remittances received tended to vary inversely. As a whole, the proportion of each additional Nakfa of household income from remittances that is used for expenditures on child education is less than one (but still positive), implying that an increase in remittances leads to a smaller increase in expenditure on child education. The positive relationship between household’s income (excluding remittances) and education ratio indicates that the opportunity to spend more remittances on child education is relatively higher for families who earn higher income.

Since the amount of remittances that is spent on child education increases with the increase in remittances received, a strategy that motivates emigrants to remit more is essential. For instance, the introduction of attractive exchange rate system is one among different strategies for boosting the flow of remittances. In addition, improvement in transfer system, together with quick, easy and reasonable charge, has a positive impact on the flow of remittances. Strategies for directing remittances to business investment have also an indirect positive effect on education ratio because investment usually brings higher income and higher income, in turn, results in higher education ratio. Finally, development of community awareness of the importance of child education and expansion of adult education are believed to increase the share of expenditure on child schooling out of remittances. Above all a sustainable return to peacetime is fundamental to development in general and to efficient and productive use of remittances in particular.

The paper acknowledges the importance of further research on the use of remittances other than for education. It is crucial to collect the necessary data and run a regression analysis by spending category.

Acknowledgements

I would like to thank Richard Brown for helpful comments. I also wish to thank an anonymous referee for reviewing the manuscript and Afeworki Paulos for expediting the review process

About the Author

Dr. Temesgen Kifle is a postdoctoral research fellow at the School of Economics, University of Queensland, Australia. E-mail: [email protected]

Notes

1. The term transfer income in this survey is composed of housing rent subsidy, pensions, annuities, assistance from government and non-government organization, and assistance from households in Eritrea and abroad.

2. Nakfa is the standard unit of money in Eritrea.

3. For more information about the model and its functional form refer to Adams (1991).

4. For more information on the private rate of return to schooling in Eritrea refer to Kifle (2004).

5. According to Engel, consumers will tend to spend an increasing proportion of any additional income upon luxury goods and a smaller proportion on stable goods; so that a rise in income will lower the overall share of consumer expenditure spent on stable goods (such as foodstuffs) and increase the share of consumer expenditure on luxury goods (Pass, Lowes, and Davies, 1993).

6. Since we don’t have any evidence for the overall uses of remittances, it is difficult to know whether the remaining part of remittances received is used exclusively for consumption or for investment purpose. Therefore, a further study on this area is essential.

References

Adams, R. (1991), “The economic uses and impact of international remittances in rural Egypt”, Economic Development and Cultural Change,39(4): 695-722.

Adama, R. (2006), “Migration, Remittances and Development: The Critical Nexus in the Middle East and North Africa”, UN/POP/EGM/2006/01, United Nations Expert Group Meeting on International Migration and Development in the Arab Region.

Addison, T., Le Billion, P, and Murshed, S. (2001), “Finance in Conflict and Reconstruction”, Discussion Paper No. 2001/44, United Nations University, WIDER.

Arneberg, M. and Pedersen, J. (2001), “Urban Household and Urban Economy in Eritrea: Analytical Report from the Urban Eritrea Household Income and Expenditure Survey 1996/97, Fafo Institute for Applied Social Science, Oslo.

Carling, J. (2004), “Policy options for increasing the benefits of remittances”, Paper presented at Workshop B9: Remittances and Development: beyond increasing incomes at the Ninth international Metropolis Conference, Geneva, 27/09 – 01/10/2004.

Choucri, N. (1986), “The hidden economy: A new view of remittances in the Arab world”, World Development,14(6): 697-712.

da Cruz, V., Fenger, W., and Schwartzman, A. (2004), “Remittances to Comoros: Volume, Trends, Impact and Implications”, Africa Region Working Paper No. 75, The World Bank Group.

de Haan, A. (2000), “Migrants, Livelihood and Rights: The Relevance of Migration in Development Policies”, Social Development Working Paper No. 4, Department for International Development, London.

EIU (Economist Intelligence Unit) (2002), “Country Profile 2002”, Economist Intelligence Unit, The Economist, London.

GSE (Government of the State of Eritrea), (2004), “Eritrea Food Security Strategy”, GSE.

Hansson, G, (2001), “Building New States: Lessons from Eritrea”, Discussion Paper No. 2001/66, United Nations University, WIDER.

IMF (International Monetary Fund) (2003), “Eritrea: 2003 Article IV Consultation Staff Report; Public Information Notice on the Executive Director for Eritrea, IMF Country Report No. 03/165 (online), http://www.imf.org/external/pubs/ft/scr/2003/cr03165.pdf, Accessed 4 July 2003.

Kifle, T. (2004), “Return to and Demand for Education in Eritrea and the Role of International Remittances”, Aachen, Shaker Verlag.

Koser, K. (2002), “From refugees to transnational communities?”, in N. Al-Ali and K. Koser (ed.) New Approach to Migration, pp. 138-152, London: Routledge.

Koser, K. (2003), “Mobilizing New African Diaspora: An Eritrean Case Study”, in K. Koser (ed.) New African Diaspora, pp. 111-123, London: Routledge.

Lachaud, J. (1999), “Envoi de fonds, inegalite et pauvreta au Brkina Faso”, Documents de travail 40, Centre d’Economie du Developpmente de l’Universite Montesquieu Bordeaux IV.

Lindley, A. (2005), “Influence of Remittances and Diaspora Donations on Education”, Paper for Conference on Somali Remittances, Washington D.C., December 1-2, 2005, Hosted by World Bank and United Nations Development Program.

NFIS (National Food Information System of Eritrea) (2005), “Baseline Study of Livelihood Systems in Eritrea: Phase 1 Preliminary Identification of Livelihood Systems Through Research on Secondary Sources (Draft), Zoba (Region) Maekel.

NSEO (National Statistics and Evaluation Office) (2003), “Eritrea: Demographic and Health Survey 2002”,Asmara, Eritrea: National Statistics and Evaluation Office.

Pass, C., Lowes, B., and Davies, L. (1993), Collins dictionary of economics, 2^nd edition, Great Britain: Harper Collins Publishers.

Sander, C., and Maimbo, S. (2003), “Migrant Labor Remittances in Africa: Reducing Obstacles to Developmental Contributions”, Africa Region Working Paper Series No. 64, The World Bank Group.

Schoorl, J. et al. (2000), “Push and Pull Factors of International Migration: Comparative Report”, Eurostat, Luxembourg.

Serageldin et al. (1983), “Manpower and International Labour Migration in the Mid East and North Africa”, published for the World Bank, OxfordUniversity Press.

Sibanda, A. (2004), “Who gets to drop out of school in South Africa? The role of individual and household attributes”, African Population Studies,19(1): 99-117.

Straubhaar, T. (1985), “Migrants’ remittances and economic activity”, Intereconomics,20: 87-92.

Tewolde, B. (2005), “Remittances as a tool for development and reconstruction in Eritrea: An economic analysis”, Journal of Middle Eastern Geopolitics, 1(2): 21-32.

World Bank (2006), “06 World Development Indicators”, Washington D.C.: The World Bank.

Ymauchi, A. (2004), “Fiscal Sustainability – The Case of Eritrea”, IMF Working Paper, African Department, WP/04/7.