The effect of remittances on GDP growth

Abstract

This thesis studies the relation between remittances and GDP growth. Since remittances are becoming a larger part of FDI in developing countries, it is interesting to see what the effects are on a macroeconomic scale. Therefore, this thesis answers the question if remittances have a positive impact on GDP growth. Using the Solow growth model as basis, remittances either have a negative effect because the labor force decreases, or a positive effect since remittances stimulate capital investment. The question is answered through empirical research, using panel data consisting of 42 lower and middle income countries and covering the years 1995 until 2015. Using a fixed effect OLS regression method this paper finds a positive significant effect of remittances on GDP growth. This indicates that remittances are used as capital for investment and stimulate the economy of a developing country in this way.

BSC Thesis Economics and Business Lydia van der Vegt

Specialization: Economics 10183329

Supervisor: Rutger Teulings January 31, 2017

The effect of

remittances on GDP

growth

1

Statement of Originality

This document is written by Student Lydia van der Vegt who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

2

Contents

4. Relationship between remittances and GDP growth ... 7

5. Methodology ... 10

5.1 Regression model ... 10

5.2 Data ... 12

6. Empirical findings and analysis ... 14

7. Conclusion and discussion ... 17

Bibliography ... 18

Appendix ... 20

1. List of countries included in analysis categorized by income level ... 20

2. Data sources ... 21

3. Correlation matrix ... 21

4. Results Hausman test ... 22

3

List of tables

Table 1 Descriptive Statistics ... 13 Table 2 Results regression analysis ... 14 Table 3 Results regression analysis for different income levels ... 15

4

1. Introduction

People who migrate often send money to their family that stayed behind. These flows of money are called remittances. Over the past years remittances have become a more important source of capital inflows in developing countries. The amount of remittances is growing and is one of the greatest sources of foreign direct investment in developing countries (World Bank, 2006, pp. 85-86). Because migrants often stay part of their community after migration, remittances can be seen as a part of a private welfare system, since money flows from richer parts of the community to the members that are more poor (Gupta, Pattillo, & Wagh, 2009, p. 105). So, on a microeconomic level remittances are an opportunity to increase purchasing power and a way to escape poverty, but there is still little consensus about the effect of remittances on a macroeconomic level.

Therefore, this paper attempts to answer the question if remittances have a positive effect on economic growth, more specific GDP growth.

Existing literature on this topic has not find a specific relationship between remittances and GDP growth. Some state that since the nature of remittances is a compensatory family transaction, primarily used as extra income for family consumption and not a source for investment, it may lead to economic inactivity of the family that receives the remittance. This would cause a negative relation between remittances and economic growth (Chami, Fullenkamp, & Jahjah, 2003, p. 21). On the other hand, remittances can be a source for investment, especially in countries where financial markets are less developed and therefore remittances can have a positive impact GDP growth (Guiliano & Ruiz-Arranz, 2009, p. 150). While others state that, countries with a higher level of financial development can utilize remittances better and therefore remittances can be channeled more efficiently into a source of capital causing GDP growth (Catrinescu, Leon-Ledesma, Piracha, & Quillin, 2009, p. 90).

In this paper, I investigate the relationship between remittances and GDP growth by using panel data consisting of 42 countries from the years 1995 until 2015. Based on the literature I expect that remittances influences GDP growth either positively or negatively, because it has the potential to be capital for investment or can promote economic inactivity at the receiving end. The empirical analysis finds evidence that remittances have a significant positive effect on GDP growth, indicating that remittances are used as capital for investment.

The structure in the remaining of this paper is as follows. Frist, section 2 gives an overview of literature on GDP growth and the Solow model. Section 3 discusses remittances and section 4 the relationship between remittances an GDP growth. This is followed by the methodology in section 5. Next, the results of the regression will be analyzed in section 6 . Finally, I conclude in section 7.

5

2. GDP growth

This section explains the theory on GDP growth and which macroeconomic factors determine GDP growth. In this thesis, GDP per capita growth is taken as primary unit of analysis. GDP per capita is the Gross Domestic Product of a country divided by the population of that country. This is often used as a measure of the productivity of a country and the relative performance of a country economically. Therefore, in this thesis the relationship between remittances and the productivity of a country is further elaborated.

An influential model used to explain GDP growth is the neo-classical Solow growth model (Solow, 1956). The Solow model tries to predict long-term economic growth. The model seeks to find an answer why total output of a country is growing and why different countries grow at different rates. He models total output as a function of physical capital, labor and the level of technology, so growth can only be explained by a change in one of these factors. Labor and the level of technology grow at an exogenous growth rate, respectively n and g. Physical capital growth is dependent on the savings rate in a country, since in the model the savings in a country are used for investment in new capital. At the same time, the capital stock depreciates. As long as a country invests more than the depreciation rate, the capital stock is growing, and so is total output. When a country invests less than the depreciation rate, the capital stock is declining, which results in declining total output. This results in a steady state where investments are equal to the depreciation rate. Economies that are further away from their steady state experience a higher economic growth than economies that are already close to, or in their steady state, since these economies move to their steady state.

In the steady state, an economy still grows at the exogenous rate. This rate is dependent on population growth and technological progress. One of the hypothesis of the model is convergence, in which it is assumed that all countries eventually have the same steady state and converge to that point. This explains why countries with a lower initial level of GDP grow faster than countries with a higher level of GDP, since they are further away from the steady state.

Mankiw, Romer & Weil (1990) extend the Solow model. They conducted an empirical research and find that human capital has to be endogenous in the model. So, economic growth depends on physical capital, labor, technological state and human capital. Besides this, they also find that physical capital and population growth have a larger impact on growth rates than the traditional Solow model predicts. Physical capital has a higher effect, because a higher savings rate leads to higher income, which then leads to a higher level of human capital, which leads to a higher income. The effect of population growth is also larger, because when the population grows, not only physical capital is spread more thinly over the population, but also human capital is spread more thinly. This leads to lower income. So, differences in economic growth can be explained by the Solow model when human capital is also taken into account.

6 Barro (2003) has conducted an empirical analysis of GDP growth rates based upon the Solow growth model. He conducted several regressions measuring the importance of initial human capital and initial level of capital on convergence. In his analysis he also includes policy variables and national characteristics such as international openness, the ratio of government consumption, rule of law, democracy, fertility rate, the ration of real gross domestic investment to real GDP, and the inflation rate. He finds that countries with a low initial level of real GDP per capita have a higher growth rate, but this is dependent on the initial level of human capital. When the initial level of human capital is high, countries with a low initial level of real GDP per capita tend to grow faster than countries with a low initial level of real GDP per capita with a low initial level of human capital.

Barro (2003) also finds other factors that influence GDP per capita growth. He finds that government consumption is shown to have a negative relation with economic growth. Another important factor determining GDP growth is the openness to international trade, where Barro finds a significant positive relation. Also a positive relation between terms of trade and GDP growth is found. The investment ratio has shown a positive relation with GDP growth as well. At last, the relation between inflation and GDP growth has shown to be negative.

3. Remittances

Migrants often send their earned money back to their families. Those payments are called remittances. So the total of remittances received by a country is constructed of millions of smaller payments. The amount of remittances a country receives is considered substantial and of increasing importance (Migration Policy Institute, 2006, p. 7). In developing countries remittances are growing rapidly. The increase in remittances received between 2001 and 2005 was 73% for developing countries. The amount of remittances is becoming the largest share in foreign direct investment (FDI) and becomes more important than official development assistance received by developing countries (World Bank, 2006, p. 88).

A lot of research is done on the relation between migration and remittances on a microeconomic level. Dustmann & Kirchkamp (2002) research the relation between the remittances a migrant sends to their families and the economic activity of a migrant after returning. They acknowledge that remittances can be a source of savings, but they can also be used for consumption and income for the family of the migrant. Remittances are also a source of capital for investment when capital markets are not developed and it is hard to find finance to start a business. They indeed find that when migrants return, they are more likely to start their own business, though this probability decreases with age. So, Dustmann & Kirchkamp show that remittances are used in many ways that eventually might influence GDP growth.

7 Lucas and Stark (1985) study the reasons for migrants to send remittances. They distinguish between pure altruistic reasons, reasons of pure self-interest, and a combination of both. They find that the amount remitted rises when the income of the sender increases. Also, migrants who are perceived as head of the family and have children in school in their home country remit more than migrants who do not feel that they have a family to take care of. So, one of the biggest reasons for people to remit is to take care for a family and pay for education of their children. They also find that when migrants stay away for longer periods the amount of remittances send decrease. This is due to “out of mind, out of sight” reasoning, but people who keep identifying as a family member keep on remitting.

There are also studies that research the relationship between remittances and poverty reduction. These are often regional and combine country panel data with household survey data. Gupta, Pattillo & Wagh (2009) research the relation between remittances and poverty reduction in Sub-Saharan Africa and find that remittances have a poverty reducing effect due to the fact that migrant transfers help people to overcome budget constraints and provide the opportunity for people to participate in the financial sector. Acosta et al. (2008) study poverty reduction in Latin America and they also found that remittances have a poverty reducing effect, but that this effect is dependent on the elasticity of poverty reduction with respect to remittances. They find that relative poorer countries have a lower elasticity and relative richer countries have a higher elasticity, which means that richer countries are better at utilizing the cash inflows coming from remittances.

So, remittances are essentially a private transfer of money. Chami et al. (2008) explain that this has certain characteristics that have to be taken into account when studying remittances on a macroeconomic level. The private nature of remittances make that they differ from other types of international cash flows in the way that they are more stable over time. Also, the literature does not find a correlation between remittances and other foreign exchange inflows. Chami et al. explain this since remittances are private transfers, and thus behave differently from other foreign exchange flows.

4. Relationship between remittances and GDP growth

In the previous sections, the focus was solely on GDP growth and remittances. This section will combine those and give an explanation of how remittances fit into the theoretical model of GDP growth.

Chami, Fullenkamp and Jahjah (2003) have constructed a framework to model the effect of remittances on economic growth. This model is based on the economics of the family, since remittances have a direct impact on families. Besides families, they also recognize the interaction

8 between family and the market, since families that have not participated in migration are still a participant in the domestic labor market. The last part they take into account, is that the remitter and the receiver are often separated from each other by a large distance. That makes this relationship subject to an asymmetrical information problem. They argue that, due to this asymmetrical information problem, there can be cases of moral hazard where the family receiving the remittance withdraws from the labor market. In their empirical research they find a negative relationship between remittances and GDP growth, from which they conclude that the effects of moral hazard predominates.

Catrinescu, Leon-Ledesma, Piracha and Quillin (2009) argue that the research conducted by Chami et al. (2003) is subject to ommited variable bias since it did not take the development of the financial market into account. Catrinescu et al. study the relationship between remittances, financial development in a country and GDP growth. They test whether countries with a highly developed financial market can utilize remittances better since they can be better channeled into investments or savings and therefore result in a higher GDP growth. The other side is that remitances can also be a source of capital used for investments when capital markets are less developed and thus result in a higher GDP growth. They find that the first is indeed the case, a highly developed financial market results in a higher GDP growth, due a better use of remittances.

This is in contrast with the findings of Giuliano and Ruiz-Arranz (2009). This can be caused due to the fact that they only look at the financial development and Catrinescu et al. (2009) also take other institutions into account. But, Giuliano and Ruiz-Arranz do not only regress the effect of remmitances on GDP growth, they also do a regression on the effect of remittances on investment and found that remittances have a positive result on the level off investment especcialy in countries with lower developed financial markets. Besides that, they found that remittance sending follows a pro-cyclical behavior, meaning that there are more remittances send in times of favourable economic circumstances with more posibility for investment. They suggest that the investment channel is most likely to explain the fact that they found a positive relationship between remittances and GDP growth. Aggarwal, Demirgüc-Kunt and Martínez Pería (2011) have also investigated the relationship between remittances and financial development. They have directly researched the effect remittances have on financial development. So, if remittances stimulate financial development. They find a significantly positive relation between remittances and financial development and they argue that this might be another channel through which remittances affect GDP growth.

Barajas, Chami, Fullenkamp, Gapen and Montiel (2009) argue that researching the relationship between remittances and economic growth can stay subject to endogeneity and therefore it stays hard to find a definite answer on the question if remittances have an effect on

9 economic growth. Their approach in finding an answer is constructing an instrumental variable that captures changes caused by microeconomic changes. This instrument is the “ratio of remittances to GDP of all other recipient countries” (Barajas, Chami, Fullenkamp, Gapen, & Montiel, 2009, p. 14). This instrument captures the effect of global reductions in transaction costs. This is better than lagged driven variables, since this has no direct link to other domestic macroeconomic variables. Eventually, they do not find a positive significant impact of remittances on long-term economic growth. The reason they give for this is that remittance are used to support a family in basic life needs and not as a source for investments.

Taking the Solow model into account, economic growth is a function of labor, physical capital and state of the technology. So GDP growth has to come from a positive change in either one of these variables. All articles have suggested a change either in labor or in physical capital. Chami et al. (2003) argue that a moral hazard problem results in economic inactivity of the family. This leads to a negative change in labor and therefore a negative relation between remittances and GDP growth. Catrinescu et al. (2009), Giuliano and Ruiz-Arranz (2009) and Aggarwal et al. (2011) all studied the relation between remittances and the financial market. This suggests a relation between remittances and physical capital, because through financial markets remittances finance a large part of physical capital accumulation in a country. So when remittances are used as a way to accumulate physical capital, this stimulates the GDP growth of that country.

10

5. Methodology

In this section, the methodology of this research is described. Subsection 5.1 explains the regression model. Subsection 5.2 contains the data and the definition of the variables used.

5.1 Regression model

The effect of remittances on GDP growth is estimated using a multiple OLS regression with fixed effects. The model is specified as followed:

(1) ℎ = + + + + + +

+ + + + + ,

where i refers to the receiving country and t refers to time. The independent variables are specified as follows: Rem is the amount of remittances received by a country, Pop is the population growth in a country, Infl stands for the inflation, GFCF is the gross fixed capital formation, Govcon is the government consumption, Optrade is the openness to international trade, GDPit-1 is the initial level of

GDP, and GDPgrowthit-1 is the initial level of GDP growth. measures time fixed effects and

measures country fixed effects. is the error term.

To measure percentages, some variables are transformed. GDP growth, remittances, gross fixed capital formation, government consumption, and openness to international trade are transformed into natural logarithms.

The choice of the control variables has their basis in the theory described in section 2 regarding theory on GDP growth.

Population growth is measured as the population growth rate for a year expressed as a percentage. It counts all residents in a country not taken legal status or citizenship into account.

The inflation indicator used is measured by the consumer price index and it reflects the annual percentage change in the cost of the average consumer basket. This indicator is included since previous research has shown a significant negative relationship between inflation and GDP growth.

Gross fixed capital formation is used to control for the effect of physical capital accumulation on GDP growth. It includes land improvements, plant, machinery, and equipment purchases. Also the construction of roads, railways, schools, offices, hospitals, private residential dwellings, and commercial and industrial buildings are taken into account when measuring the gross fixed capital formation.

Government consumption includes all government current expenditures for purchases of goods and services. Also, it includes most expenditure on national defense and security, but military expenditures are not taken into account, as those are a part of government capital formation.

Openness to international trade is used to control for the international trade activities of a country, since it has been shown that the openness to international trade is related to GDP growth.

11 The openness to international trade is measured as the share of total of imports and exports to GDP in current US dollars.

To decide between a random effects model and a fixed effects model I conducted a Hausman test1_{. Under the null hypothesis the random effects model is preferred. When the null hypothesis is} rejected the random effects model is preferred. The test statistic of this test is 0,0001, which means that the null hypothesis is rejected and the alternative hypothesis is accepted. Thus the fixed effects model can be applied.

A panel data regression is subject to assumptions. The first assumption is that the error term has a conditional mean of zero. This is the case when there is a constant in the model, because the constant absorbs the expected value of the error term. The model predicted in this thesis contains a constant so this assumption is met.

The next assumption is that large outliers are unlikely. Table 1 shows the descriptive statistics. There it can be seen that large values are found but can be explained by economic events. For example the highest overall inflation value of 709.35 can be explained by the hyperinflation in Belarus in 1995, which are rare cases and unlikely to happen.

Another assumption is that there is no perfect multicollinearity. Appendix 2 shows a correlation matrix. It shows that there is no high correlation between the variables, so there is most likely no multicollinearity.

There is also an assumption about the variance of the error term. If the error term is heteroskedastic robust standard errors must be used. As Stock and Watson put it: “Economic theory rarely gives any reason to believe that the errors are homoscedastic. It therefore is prudent to assume that the errors might be heteroskedastic unless you have compelling reasons to believe otherwise.” (Stock & Watson, 2015, p. 209)Since the data gives no reason to assume homoscedastic error terms, I correct for heteroscedasticity.

The last assumption is no autocorrelation. This is tested by a Woolridge test, where the null hypothesis is that there is no first order autocorrelation. The test statistic is 0,000, which indicates that the null hypothesis has to be rejected and the data is auto-correlated. Therefore, the regression is also corrected for serial correlation by using clustered robust standard errors.

12

5.2 Data

In this paper, I use panel data consisting of 42 countries2 _{covering the years 1995 until 2015. So, there} are a total of 882 observations. The countries are all lower and middle income countries, which is also shown in appendix 1. The dataset is balanced for all variables. The data for all variables is from the World Bank Development Indicators, which consists of officially recognized international sources. The data presents the most current and accurate global development data available3_.

Even though data on remittances is available for many countries and several years, the quality is often not optimal. There is no consensus on what should be measured, only workers remittances or also compensation of employees and migrant transfers. There are certain countries where data collection is of low quality, which causes remittances to be unrecorded. Also, when only a small amount of money is remitted, it is not mandatory to report that amount. There are cases where remittances are sent through an informal channel and therefore are not recorded. There are also cased where remittances are sometimes recorded as regular FDI, which has the consequence that the amount of remittances that are reported is lower than the actual amount remitted (World Bank, 2006, p. 86). The World Bank Development Indicators measures remittances as the sum of workers remittances, compensation of employees and migrant transfers (World Bank, 2006, p. 87).

Table 1 shows the descriptive statistics. It shows the mean for the total dataset, the standard deviation and the minimal and maximal value for the total dataset, the values between the countries and between the time range.

To get more insight if the effect on remittances changes per income group the data is separated by the classification the World bank gives to countries regarding income. They distinguish low-income, lower-middle-income, upper-middle-income and high-income countries. For my data this is also shown in appendix 1. Since there are only three countries in the low-income category, I clustered these with the lower-middle-income countries. Therefore the analysis on the low income group is more reliable, since it has a larger sample.

2 _{A list of countries used in the analysis is available in Appendix 1} 3 _{A table of the data sources for all variables used is given in Appendix 2}

13 Table 1 Descriptive Statistics

Variable Mean Std. Dev. min max

GDPgpc overall 3.14 3.94 -14.35 33.03 between 1.72 0.95 8.75 within 3.55 -17.71 29.98 Rem overall 4.57 5.49 0.00 31.75 between 4.88 0.23 18.85 within 2.61 -8.62 22.16 Pop overall 1.43 0.97 -2.05 4.26 between 0.92 -0.65 2.91 within 0.33 -0.29 2.82 Infl overall 12.00 33.46 -8.53 709.35 between 13.52 1.78 82.62 within 30.68 -63.59 638.72 GFCF overall 22.38 6.71 5.46 57.71 between 5.34 10.04 39.14 within 4.14 9.20 51.26 Govcon overall 13.23 4.52 4.58 63.94 between 3.84 5.09 21.91 within 2.45 4.91 60.56 Optrade overall 0.54 0.28 0.12 1.87 between 0.26 0.18 1.53 within 0.10 0.10 0.94 N=882, n=42, t=21 De scriptive statistics

14

6. Empirical findings and analysis

Table 2 shows the results of the OLS regressions. From the table, it can be concluded that when controlled for population growth, inflation, gross fixed capital formation, government consumption, and openness to international trade remittances have a positive significant effect on GDP growth, indicating that remittances cause GDP growth. This can be explained by the fact that remittances are used as a source of capital for investments and these investments result in a higher GDP growth. Table 2 Results regression analysis

Initial level of GDP is not significant, but the sign implies a negative effect on GDP growth .This means that a higher initial level of GDP results in a lower growth in GDP. This is in line with the predictions of the Solow model, since economies with a higher level of GDP are closer to their steady state, which results in a lower GDP growth. The initial level of GDP growth is significant and positive. This is also in line with the Solow model since a higher initial level of growth suggests a higher growth the next year compared to countries with a lower initial GDP growth.

Variable

(1)

(2)

lnRem

0.09**

0.07**

(0.04)

(0.03)

L_GDPcur

-3.59e-14

(3.54e-14)

L_GDPgpc

0.03*

(0.02)

Pop

-0.29**

(0.11)

Infl

-2.08e-4

(1.68E-3)

lnGFCF

0.39*

(0.19)

lnGovcon

-0.36

(0.30)

lnOptrade

-8.30e-3

(0.17)

Constant

1.04***

0.89

0.02

(0.82)

Observations

761

761

R-Squared

0

0.21

*significant at 10% level **significant at 5% level ***significant at 1%level

Re gre ssion analysis

GDP Growth

Table notes: The regression results are corrected for year and time fixed effects, using a dummy for time fixed effects but those are not included in the table. Robust standard errors are adjusted for 42 clusters.

15 Also in line with the Solow model is the population growth, which is both negative and significant. The Solow model predicts that population has a negative impact on GDP per capita growth, since the total income (GDP) has to be spread out over more people. Inflation has a small negative effect on GDP growth. However, this effect is insignificant at all significance levels. Though, it is expected that the relation is negative, other literature finds higher and significant coefficients for this relation (Barro, 2003, p. 256). Gross fixed capital formation does have a significant effect on GDP growth. As expected by the Solow model this effect is positive, since gross fixed capital formation measures the physical capital accumulation. The openness to international trade has a very small negative effect on GDP growth, and is also not significant.

Table 3 shows the regression results when countries are divided by income level. The World Bank classifies low-income, lower-middle-income and upper-middle-income countries. Since there are only three low-income countries in my sample I clustered those with the lower-middle-income group. Though neither coefficient is significant, it is remarkable that the coefficient is much larger for low-income and lower-middle-low-income countries then for upper-middle-low-income countries. This might be since in lower-middle-income countries financial markets are less established then in upper-middle-income countries. Therefore, the remittances are used as capital for investment which results in a higher economic growth due to remittances. In the upper-middle-income countries, financial markets are more established and therefore the effect of remittances on economic growth is smaller. Because, financing to start a business is also provided by the capital markets and does not have to come from remittances. Which results in a lower impact of remittances on GDP growth.

16 Table 3 Regression analysis per income group

Variable (1) (2)

lnRem 0.04 1.98e-3

(0.06) (0.04)

L_GDPcur -9.34e-14 5.93e-15

(1.79e-13) (4.85e-14) L_GDPgpc 0.02 0.04** (0.02) (0.15) Pop -0.28* -0.32* (0.13) (0.19) Infl -0.01 -1.09e-3 (0.01) (1.56e-3) lnGFCF 0.56*** 3.15e-3 (0.16) (0.33) lnGovcon -0.50 -0.61 (0.31) (0.49) lnOptrade -0.05* 0.33 (0.19) (0.31) Constant 0.99 2.75* (1.00) 1.59 Observations 367 359 Groups 20 22 R-Squared 0.25 0.27

*significant at 10% level **significant at 5% level ***significant at 1%level

Re gre ssion analysis pe r income group

GDP Growth

Table notes: regresion 1 is the lower and lower-middle-income countries, regression 2 is the higher-middle income countries. Standard errors are robust.

17

7. Conclusion and discussion

This thesis has addressed the question if remittances have an effect on GDP growth. This question is answered through empirical research using panel data on 42 countries and 21 years. Through an OLS regression there is shown that remittances have a significant positive effect on GDP growth. This can be explained by the fact that remittances are an extra source of capital, which can be used as an investment when financial markets are less developed. Also, remittances are an extra income that can be used for private consumption and which can also affect GDP.

The results per income group are not significant, though the effect in the lower-income and lower-middle-income countries is substantially larger than the effect in the upper-middle-income group. This might be because financial markets are less established in lower and lower-middle-income countries and therefore remittances have a greater use as capital for investment. But, the effect of remittances per income group needs further research.

As stated earlier, the data on remittances is often not complete. People who send remittance might do this through informal channels, or remittances might not be recorded as remittances. Therefore, the number used is likely to not display the total amount of remittances received by a country accurately. Future research also has to show whether remittances influences GDP growth through investment or consumption, such that there is a better view on the causal relationship between remittances and GDP growth.

Since this thesis is about GDP growth, I want to add a critical note about GDP and what it does and does not measure. GDP is often used as one of the primary measures of the economic health of a country, since it measures the output level of a country measured in dollars in a specific country over a specific period. Many researchers also see a high GDP per capita as a measure of progress and that a country has a high social welfare. Though there is critique of using GDP as a measure for social welfare and progress. Since GDP per capita is an average of the production level it does not take into account the income distribution in a country and the ability of people to have access to basic needs (van den Bergh, 2009, p. 119). While these factors may be more important for people who experience poverty every day. Therefore, it should be noted that while this paper covers GDP growth, it does not suggests that a high GDP growth is necessarily an improvement for the people living in those countries. This thesis does however give an insight in the contribution of remittances on the production level of a country.

18

Bibliography

Acosta, P., Calderón, C., Fajnzylber, P., & Lopez, H. (2008). What is the Impact of International Remittances on Poverty and Inequality in Latin America? World Development, 36(1), 89-114. Aggarwal, R., Demirgüc-Kunt, A., & Martínez Pería, M. S. (2011). Do remittances promote financial

development? Journal of Development Economics, 96(2), 255-264.

Barajas, A., Chami, R., Fullenkamp, C., Gapen, M., & Montiel, P. (2009). Do Workers' Remittances Promote Economic Growth. IMF Working paper, 1-22.

Barro, R. J. (2003). Determinants of Economic Growth in a Panel of Countries. Annals of Economics

and Finance, 4, 231-274.

Barro, R. J., & Lee, J.-W. (2001). International data on educational attainment: updates and implications. Oxford Economic Papers, 53(3), 541-563.

Catrinescu, N., Leon-Ledesma, M., Piracha, M., & Quillin, B. (2009). Remittances, Institutions and Economic Growth. World Development, 37(1), 81-92.

Chami, R., Barajas, A., Cosimano, T., Fullencamp, C., Gapen, M., & Montiel, P. (2008).

Macroeconomic Consequences of Remittances. Washington DC: International Monetary

Fund.

Chami, R., Fullenkamp, C., & Jahjah, S. (2003). Are Immigrant Remittance Flows a Source of Capital for Development. IMF Working Paper, 189(3), 1-47.

Dustmann, C., & Kirchkamp, O. (2002). The optimal migration duration and activity choice after re-migration. Journal of Development Economics, 67(2), 351-372.

Fischer, S. (1993). The role of macroeconomic factors in growth. Journal of Monetary Economics, 485-512.

Guiliano, P., & Ruiz-Arranz, M. (2009). Remittances, financial development, and growth. Journal of

Development Economics, 90(1), 144-152.

Gupta, S., Pattillo, C. A., & Wagh, S. (2009). Effect of Remittances on Poverty and Financial Development in Sub-Saharan Africa. World Development, 37(1), 104-115.

Kormendi, R. C., & Meguire, P. G. (1985). Macroeconomic determinants of growth: Cross-country evidence. Journal of Monetary Economics, 16(2), 141-163.

Landau, D. (1983). Government Expenditure and Economic Growth: A Cross-Country Study. Southern

Economic Journal, 49(3), 783-792.

Lucas, R. E., & Stark, O. (1985). Motivations to Remit: Evidence from Botswana. Journal of Political

Economy, 93(5), 901-918.

19 Mankiw, N. G., Romer, D., & Weil, D. N. (1990). A contribution to the empirics of economic growth.

NBER Working Paper Series, 1-45.

Migration Policy Institute. (2006). Remittances and Development: Trends, Impacts and Policy

Options. Washington DC: Migration Policy Institute.

Solow, R. M. (1956). A Contribution to the Theory of Economic Growth. The Quarterly Journal of

Economics, 70(1), 65-94.

Stock, J. H., & Watson, M. W. (2015). Introduction to Econometrics. London: Pearson.

van den Bergh, J. C. (2009). The GDP paradox. Journal of Economic Psychology, 30(2), 117-135. World Bank. (2006). Global Economic Prostpects: Economic Implications of Remittances and

20

Appendix

1. List of countries included in analysis categorized by income level

Low-income countries Lower-middle-income countries Upper-middle-income countries

Malawi Armenia Albania

Nepal Bangladesh Azerbaijan

Tanzania Bolivia Belarus

Egypt, Arab Republic Belize

El Salvador Botswana

Ghana Brazil

Guatemala China

Honduras Colombia

India Costa Rica

Indonesia Dominican Republic

Morocco Ecuador

Nigeria Jordan

Pakistan Malaysia

Philippines Mauritius

Sri Lanka Mexico

Sudan Paraguay Tunisia Peru Romania Russian Federation South Africa Thailand Turkey

21 2. Data sources

Indicator Name Source

GDP (current US$) World Bank national accounts data, and OECD National Accounts data files.

GDP per capita growth (annual %)

World Bank national accounts data, and OECD National Accounts data files.

Personal remittances, received (% of GDP)

World Bank staff estimates based on IMF balance of payments data, and World Bank and OECD GDP estimates.

Goods imports (BoP, current US$)

International Monetary Fund, Balance of Payments Statistics Yearbook and data files.

Goods exports (BoP, current US$)

International Monetary Fund, Balance of Payments Statistics Yearbook and data files.

Population growth (annual %)

Derived from total population. Population source: (1) United Nations Population Division. World Population Prospects, (2) Census reports and other statistical publications from national statistical offices, (3) Eurostat: Demographic Statistics, (4) United Nations Statistical Division. Population and Vital Statistics Reprot (various years), (5) U.S. Census Bureau: International Database, and (6) Secretariat of the Pacific Community: Statistics and Demography Programme.

Inflation, consumer prices (annual %)

International Monetary Fund, International Financial Statistics and data files.

Gross fixed capital formation (% of GDP)

World Bank national accounts data, and OECD National Accounts data files.

General government final consumption expenditure (% of GDP)

World Bank national accounts data, and OECD National Accounts data files. 3. Correlation matrix Optrade 0.0387 -0.1424 0.0391 -0.0659 -0.0208 0.0352 0.1678 0.1925 1.0000 Govcon -0.0865 0.0627 -0.0657 -0.1231 -0.0618 0.0034 0.1564 1.0000 GFCF 0.3150 0.3369 0.3320 0.0128 -0.3065 -0.0910 1.0000 Infl -0.0574 -0.0642 -0.0111 -0.1271 -0.0775 1.0000 Pop -0.2579 -0.1523 -0.2441 -0.0632 1.0000 Rem 0.0167 -0.1878 0.0329 1.0000 L_GDPgpc 0.4246 0.1331 1.0000 L_GDPcur 0.1035 1.0000 GDPgpc 1.0000 GDPgpc L_GDPcur L_GDPgpc Rem Pop Infl GFCF Govcon Optrade

22 4. Results Hausman test

5. Results Wooldridge test for autocorrelation Prob>chi2 = 0.0001

= 30.50

chi2(7) = (b-B)'[(V_b-V_B)^(-1)](b-B) Test: Ho: difference in coefficients not systematic

B = inconsistent under Ha, efficient under Ho; obtained from xtreg b = consistent under Ho and Ha; obtained from xtreg lnOptrade .1902458 .1056172 .0846286 .1476912 lnGovcon -.612206 -.3842195 -.2279865 .1840031 lnGFCF .4882369 .6550972 -.1668603 .150323 Infl -.0012135 .0003696 -.0015831 .0011525 Pop -.338899 -.1871112 -.1517877 .1012857 L_GDPgpc .0332432 .0482635 -.0150203 .0030067 L_GDPcur -1.01e-13 6.75e-15 -1.07e-13 4.06e-14 lnRem .0800996 .001907 .0781925 .0384215 fe re Difference S.E. (b) (B) (b-B) sqrt(diag(V_b-V_B)) Coefficients Prob > F = 0.0000 F( 1, 41) = 21.757 H0: no first-order autocorrelation