The Effect of Inward and

Outward Remittances on Welfare in Africa






Rowland Tochukwu Obiakor1, Emeka Okoro Akpa2, Andy Titus Okwu3



Abstract: We estimated the effect of inward and outward remittances on welfare using data on 17 SSA countries from 2005 to 2019, with welfare proxied as consumption. Previous studies on remittances had tended to be focused more on the effect of inward remittances on macroeconomic variables, while few that considered inward and outward remittances were single-country analysis. The study applied the system-GMM estimation approach to accommodate the dynamic nature of the empirical model. Results showed that inward remittances exerted a positive and significant effect on consumption, while the effect of outward remittances was negative and also significant. However, the coefficient of inward remittances, in absolute terms was found to be larger than the coefficient of outward remittances. The instruments used in the estimation was found to be valid by the Sargan and Hansen tests while the absence of serial correlation was confirmed by the Arrellano-Bond tests. Policy around remittances, given the greater weight of inward remittances on consumption should be such that will make remitting cheaper.

Keywords: Remittances; Welfare; GMM

JEL Classification: F24; E21; C23




1. Introduction

The welfare of economic agents is often the concern of policy makers and governments across the world, and in no less a region than sub-Sahara Africa (SSA) where welfare is disappointingly low. SSA has a high burden of diseases (World Bank, 2013), is riddled with poverty (Asongu & le Roux, 2018) and suffers from gross underdevelopment (Grell-Brisk, 2019). To drive home the level of welfare in SSA, Addison et al. (2017, p. 11) posited that “from 1960 to 1980, life expectancy increased by 8.3 years from 41.5 to 49.8 years, while it only increased by 1.6 years to 51.4 years between 1980 and 1999. Due to renewed progress after 2000, 9 years were added by 2015, by which point life expectancy in SSA had reached a level of above 60 years”. Researchers have investigated some of the determinants of welfare in Africa to include employment, level of education and dependency ratio (Biyase & Zwane, 2017). One of the factors that have influenced welfare in Africa is globalization which has enabled the movement of goods, people and ideas across the world with little friction, enabled migrants from home countries move to host countries, earn an income and remit some to their countries of origin (World Bank, 2018). In recent times, remittances have attracted the attention of scholars and policy makers because of the possible push factors from migrants’ home countries like adverse economic and political conditions, and pull factors in the migrants' receiving countries like the promise for a better economic outcome (see Mashayekhi et al., (2013)). Attention to remittances has also resulted from the amount being sent to receiving countries by remitting migrants which inched close to foreign direct investment (FDI) and has exceeded official development assistance (ODA) (World Bank, 2019). According to World Bank (2018), in 2017, remittances grew by 8.5% to reach $466 billion, after a previous, consecutive two year decline.

In SSA, remittances is often viewed as consisting only those that come from Europe, North America and the Gulf-Arab countries. Yet, we find in World Bank (2016) that SSA countries are as much top migrant destinations (especially for other SSA countries) as they are top migrant source countries, thus outward remittances from SSA should be considered in empirical studies.

Most studies on the macroeconomic effect of remittances in SSA are predominantly on the effect of inward remittances, with few studies - for example, Al Akayleh (2016) in Saudi Arabia and Ewubare & Okpoi (2018) in Nigeria - considering the macroeconomic effects of outward remittances. Therefore, the objective of this study is to investigate the dynamic welfare effect of inward and outward remittances in SSA, with welfare proxied as household consumption. The rest of the study is divided as follows: section 2 contains a review of empirical literature, section 3 provides the empirical methodology, while section 4 presents data sources and measurement, section 5 presents results and discussion while section 6 contains the conclusion.

2. Empirical Literature

Several studies have been carried out on the role of remittances in the welfare of economies in Africa, with welfare taking different forms – consumption and declining poverty.

Some studies have examined the effect of remittances on consumption. For example, Nwaru et al. (2011) estimated the effect of migrant remittances on 120 households in South Eastern Nigeria, made up of 60 each of remittance receiving and non-remittance receiving households. Results from the analysis indicated that welfare in remittance receiving households was higher than in non-remittance receiving ones. Abbas et al. (2014) examined the welfare improving capability of workers’ remittances on households in Pakistan. Drawing samples from 240 households, divided into remittance receiving and non-remittance receiving households across 10 villages in Tehsil 18 Hazari, District jhang, Punjab, Pakistan, and employing the OLS and logit estimation methods, the study concluded that remittances receipt had a positive and significant effect on household welfare. Awan et al. (2015) examined the cost of migration on households and the welfare benefits to households of migration. The study found a significant impact of migrant remittances on household consumption, savings, health and education spending. Munyegera & Matsumoto (2016) in a study on the impact of remittances on household welfare in Uganda using the OLS, Probit and Tobit models, showed that households with access to mobile money had a higher probability of receiving remittances thus improving their per capita consumption expenditure and welfare. Salman (2016) investigated the welfare enhancing effect of migrant remittances among households in Nigeria. The study, adopting the propensity score matching (PSM), endogenous switching probit (ESP), and treatment–effects models (TEM) found that remittances had a positive impact on economic welfare. Specifically, remittance receiving households had a little over 92 percent more per capita expenditure (proxy for welfare) than non-receiving households. Akanle & Adesina (2017) examined the welfare effects of remittances on households in Nigeria. Households for the study were divided up into remittance receiving households and non-remittance receiving households. A total of 1,115 household members were sampled for the study from 2015 and 2016. Using the OLS estimation method, the study found that there was a significant positive relationship between remittances and welfare. Cuong & Linh (2017) examined the welfare effect of remittances and migration on 4, 157 households in Vietnam. Adopting the fixed effects regression, the study found that international remittances had a positive effect on household income and expenditure. More specifically, remittances received improved the per capita income and expenditure of receiving households and the overall welfare of the households. Ajaero et al. (2018) examined the relationship among international migration, remittances and welfare of 2000 households in Nigeria using the World Bank Household Surveys for African migration. Adopting the analysis techniques such as quintile estimation, the ordinary least square and probit regression, findings from the study showed that welfare of household with migrants abroad who send remittances was more than that of non-remittance receiving households. Kangmennaang et al. (2018) studied the effect of migration and remittances on the welfare of 1,000 households in Northern and Central Malawi. Using the propensity score matching (PSM), with the augmented inverse probability weighting (AIPW) as the matching tool, the study found that households where migrant members existed had less chances of being food insecure, recorded increases in household assets and improvement in household welfare. Adelowokan, et al. (2019) examined the effect of remittances on private consumption in 29 SSA countries from 2002 to 2017 using the GMM estimation method and found that remittances exerted positive but insignificant effect on consumption; a similar finding was reached by Akpa (2018) in a study on Ghana. Bahadir et al. (2018) found a positive effect of remittances on welfare and consumption of receiving countries. Javed et al. (2015), studied the impact of remittances inflow on the welfare of 400 households in Toba Tek Singh, district of Punjab, Pakistan. The study adopted the propensity score matching method and found statistically significant impact of remittances on the expenditures of migrant households on food and non-food items, thus enhancing the welfare of remittance receiving households. This result was found to be significantly different for non-remittance receiving households. Also applying the propensity score matching technique, Wadood & Hossain (2017) investigated the microeconomic impact of remittances on the welfare of households in Bangladesh. The study found that remittances had a significant effect on poverty reduction while increasing consumption. This result is however more significant from external than internal remittances. Marta et al. (2020) examined the effect on household welfare of rural-urban migration in Indonesia using the difference-in-difference estimation technique on 2,581 households of the Indonesian Family Life Survey (IFLS) of 2007. Results from the study indicate that migration significantly increased household welfare through the impact of remittances investment.

Other studies have focused on the poverty effect of remittances. For instance, Koc & Onan (2004) in the study on the welfare effect of remittances on families in Turkey found that remittances receiving households had better welfare outcomes than non-remittance receiving ones. Households that received remittances spent it to meet daily expenditure. While richer households invested their remittances, poorer ones used it to meet everyday consumption needs. Quartey (2006) studied the impact on household welfare in Ghana of migrant remittances. The study adopted the random effects GLS methodology and found that shocks to household welfare was minimized by migrant remittances, with a positive relationship between migrant remittances and household welfare. Viet (2008) investigated the effects of remittances on poverty and inequality in Vietnam. In a sample of 4,008 households, and using the fixed-effects estimation technique, the study found that remittances decreased poverty for remittance receiving households, but increased inequality. Assaminew et al. (2011) examined the impact of migration and remittances on poverty in Ethiopia from 1971 to 2009 in the VAR model from 1,490 respondents, using the binary outcome model. Findings from the study showed that international remittances had a significant negative effect on poverty in urban households. Javid et al. (2012) studied the poverty reduction effect of remittances in Pakistan using the ARDL method and found a negative and significant impact of remittances on poverty in the long-run, but a positive effect in the short run. Perhaps, this reflects the time period it takes the migrants to settle in their host countries and earn enough to meet their needs and remit. The study by Azam et al. (2016) on the impact of foreign remittances on poverty alleviation in lower middle, upper middle and high income countries using the fully modified OLS estimation method revealed that remittances had a positive but insignificant effect on poverty in high and middle income countries. Imai et al. (2017) investigated the impact of remittances on growth and poverty alleviation in Asia. The study adopted the Generalized Method of Moment (GMM) method of Instrumental Variables (IV) in the study. Findings indicated that remittances significantly reduced poverty and promoted economic growth. Imran et al. (2019) studied the impact of foreign remittances on poverty in the Punjab province of Pakistan using logistic and IV regression. The study covered 36,400 households in the region. The study revealed that international remittances had a significant negative effect on the incidence and severity of poverty. Kumar (2019b) found that households that received international remittances had about 28% chance of not being poor. Furthermore, poverty in household receiving remittances was lower than in households that did not. The study was caried out on 216 households in 2018 in Bangladesh using the logistic regression model. Akhter & Islam (2019) studied the effect of migration and migrant remittances on poverty in Bangladesh. The study was conducted on 8,449 households in 2014 using the propensity score matching and logistic regression techniques. The study found that internal and international remittances exerted a positive and statistically significant effect on poverty in Bangladesh. The welfare effect of remittances received through its effect on poverty in Haiti is different based on the study of Cardozo et al. (2019). The study found that while remittances received had a negative effect on extreme poverty, it had no effect on moderate poverty. In the study, the moderately poor were considered to spend more than US $346 but lower than US $680. The extremely poor spent less than US $346. In a study on the poverty reduction potential of remittances in 10 Commonwealth of Independent States (CIS) from 2000 to 2015 using the Least Square Dummy Variable estimation technique, Abduvaliev & Bustillo (2019) found that remittances was responsible for about 2% fall in poverty. Kumar (2019a) studied the effects of remittances on poverty and welfare in Bangladesh. The study surveyed 360 households in Cumila district in Bangladesh. Adopting the one-way ANOVA, findings from the study indicated that poverty was less and welfare more in remittance receiving households than non-remittance receiving households. The study by Samaratunge et al. (2020) noted that in Sri Lanka, internal and international remittances were a major source of money to the households. In their study, they found that international remittances reduced poverty and enhanced welfare, improved human and physical capital and increased consumption. Butkus et al. (2020) evaluated the poverty effect of remittances on seven Central and Eastern European (CEE) countries from 2006 to 2015 using the POLS, fixed and random effects and 3SLS estimators. The study found that remittances had a significant negative effect on poverty headcount, poverty depth and poverty risk. In a study on the dynamic effect of remittances on poverty and inequality in Kosovo, Arapi-gjini et al. (2020) found that absolute and relative poverty in Kosovo were alleviated by remittances. The study used the propensity score matching and dose-response estimations on a cross-sectional dataset of 8,000 households in 2011.

Other studies have found that when entrepreneurs received remittances, they increased investment, output and wages which enabled wage earners increase their consumption. However, when remittances went to wage earners, it contracted welfare as a result of decline in investment and output. In essence, remittances tended to be expansionary in the hands of entrepreneurs – increasing output and wage, but contractionary in the hands of wage earners – reducing output and investment Bahadir et al. (2018); remittances increased inequality (Wouterse, 2010), improved economic growth (Evans & Kelikume, 2018), reducing remittances reduced economic growth, investment, consumption in households, with the attendant increase in poverty and reduction in welfare (Ahmed et al., 2010). Furthermore, Sulemana et al. (2018) found that there was a direct relationship between remittances and food security in SSA, while (Sulemana et al., 2019) found that remittances had a significant effect on subjective wellbeing in SSA.

Few studies have considered the welfare effects of both inward and outward remittances. For example in Al Akayleh (2016), outward remittances was found to have a negative effect on consumption and investment in Saudi Arabia; Ewubare & Okpoi (2018) estimated the effect of inward and outward remittances on poverty reduction in Nigeria using the ARDL estimation technique. Findings from the study indicated that in the long-run, inward remittances had a negative effect on poverty while outward remittances had the opposite effect.

From the literature reviewed, it is observed that most of the studies focused on the effect of inward remittances on welfare, with welfare proxied as private consumption, economic growth or poverty; and few on the effects of both inwards and outwards remittances. This study contributes to literature by considering the potential effect on welfare - proxied as consumption - of inward and outward remittances. While inward remittances is more common in literature, we have introduced outward remittances because the sampled countries in SSA migrate as much to countries outside SSA as they migrate to countries within SSA as seen in Table (2).

3. Empirical Methodology

In the Two-Gap model of Chenery & Strout (1966), developing countries like those in SSA face a savings gap needed for take-off due to the discrepancy between savings and investment and a foreign currency gap needed for development due to differences in exports revenue and imports spending. This gap is closed when developing countries earn additional income. This additional income improves economic welfare through its effect on consumption and investment expenditure. One of such avenues for additional income is through remittances. Following this theoretical background, this study specifies an empirical model by adapting that of Iheonu & Nwachukwu (2020) where remittances was a significant contributor to household consumption in Burkina Faso and Senegal, but not so in the full sample. The model of Iheonu & Nwachukwu (2020) is presented in the following form:

(1)

However, this current study will differ from that of Iheonu & Nwachukwu (2020) by including outward remittances as potentially affecting household consumption. Other variables affecting consumption to be included in an extension of the model in equation (1) are those identified by Biyase & Zwane (2017) which include employment, level of education and dependency ratio, and exchange rate which Okwu et al. (2020) posited had a significant effect on consumption. Thus, the model in equation (1) is augmented and presented in equation (2):

(2)

Equation (2) is a dynamic regression equation. In equation (2), is household consumption expenditure in country i at time t which is used to proxy economic welfare expressed in its natural log form and is the lagged value of the dependent variable; is inward remittances in country i at time t expressed in its natural log form; is outward remittances in country i at time t expressed in its natural log form. The control variables are as follows: is inflation in country i at time t; is per capita income in country i at time t expressed in its natural log form; denotes exchange rate in country i at time t expressed in its natural log form; is credit to the private sector in country i at time t; represents unemployment in country i at time t; is primary school enrollment rate which is used to proxy human capital in country i at time t; and denotes dependency ratio in country i at time t.

In equation (2), is the error term and is defined thus:

(3)

In equation (3), is the country fixed effects while is the idiosyncratic error term. The dynamic panel model specified in equation (2) suffers from two sources of persistence over time:

  1. The inclusion of a lagged dependent variable as a regressor may result in the problem of autocorrelation.

  2. The unobserved heterogeneity among the units (Olubusoye, Salisu and Olofin, 2016)

Resolving these two sources of persistence requires the General Method of Moment (GMM) estimator. The GMM uses some sets of instruments to deal with the issue of potential correlation between lagged dependent variable and the error term. The difference GMM developed by Arellano & Bond (1991) can be used to deal with this persistence. However, Arellano & Bover (1995) and Blundell & Bond (1998) proposed the system GMM approach which is more efficient than the difference GMM approach because it uses both the levels and first differences of instruments in estimation.

Following a similar study by Adelowokan, et al. (2019), this study will use the system GMM estimator for this study.



4. Data Sources and Measurement

This study examined the effect of inward and outward remittances on welfare in SSA. Given that SSA countries are as much a host of migrants as they are source of migrants, data has been sourced from 17 SSA countries for which data on inward and outward remittances are available. The data scope for the study is from 2005 to 2019, which aims to capture the period when SSA remittances (inward and outward) grew rapidly (see Akpa et al., 2020). The variables names, measurement, a priori expectation about δ and sources are presented in Table (1)



Table 1. Characteristics of Data

Variable Name

Variable Description

Variable Measurement

A priori expectation of δ

Source

Household Consumption Expenditure (HCE)

Market value of goods (including durables) and services purchased by households

In current US$ million dollar

Nil

World Development Indicators (WDI, 2020)

Inward remittances (IREM)

Personal transfers and compensation of employees received by households in home countries

In current US$ million dollar

Positive (+)

World Development Indicators (WDI, 2020)

Outward remittances (OREM)

Personal transfers and compensation of employees sent from host countries.

In current US$ million dollar

Negative (-)

World Development Indicators (WDI, 2020)

Inflation (INF)

Persistent rise in the general prices of goods and services.

Annual percentage change in consumer price index

Negative (-)

World Development Indicators (WDI, 2020)

Per capita income (PCI)

Output per person within a country, measured in monetary terms

GDP/population in constant US dollar

Positive (+)

World Development Indicators (WDI, 2020)

Exchange rate (EXR)

The value of a currency against another. In this case, the US dollar

Monthly averages of annual local currency to the US dollar exchange rate

Positive (+) or negative (-)

World Development Indicators (WDI, 2020)

Credit to private sector (CPS)

The financial resources in forms of loans and trade credits extended to firms in the private sector by financial institutions

% of GDP

Positive (+)

World Development Indicators (WDI, 2020)

Unemployment rate (UNE)

The share of labour force that is out of work, but ready and willing to work when work is available

% of total labour force

Negative (-)

World Development Indicators (WDI, 2020)

Primary school enrollment rate (PSE)

Primary school provides basic reading, writing and numeracy skills.

% of gross enrollment

Positive (+)

World Development Indicators (WDI, 2020)

Dependency ratio (DPR)

People younger than 15 and older than 64 who depend on those in the working age (aged 15 to 64)

% of working-age population

Negative (-)

World Development Indicators (WDI, 2020)

Source: Authors’ compilation 2021

Table 2. Top Migrant Destinations of Selected Countries

S/N

Country

Top migrant destination country

1

Liberia

Guinea

2

Comoros

France

3

The Gambia

Spain

4

Cabo Verde

Portugal

5

Togo

Ghana

6

Guinea-Bissau

Portugal

7

Senegal

France

8

Mali

Cote d’Ivoire

9

Ghana

Nigeria

10

Nigeria

United States

11

Congo, Democratic Republic

The Republic of Congo

12

Uganda

Rwanda

13

Burkina Faso

Cote d’Ivoire

14

Kenya

United Kingdom

15

Malawi

Zimbabwe

16

Eswatini

South Africa

17

Rwanda

Congo, Democratic Republic

Note: South Sudan is the top migrant destination for Uganda but its next top migrant destination, Rwanda, has more data than South Sudan, thus it is used in place of South Sudan.

Source: Compiled by authors from World Bank (2016)



5. Results and Discussion

5.1. Pre-estimation

5.1.1. Descriptive Statistics

Table (3) and (4) contain the statistical features of the variables employed in the study. This descriptive statistics contains measures of central tendency, measures of dispersion, minimum values, maximum values, peakedness and normality tests. From Table (3 and 4), the average values of HCE, IREM, OREM, PCI, EXR, INF, CPS, UNE, PSE, DPR are 2.49E+10, 1.69E+09, 137.6770, 1336.130, 335.7514, 7.053439, 18.73065, 7.151086, 104.1723, 85.40300 respectively. All variables are positively skewed except DPR. From the table, PSE is platykurtic since its kurtosis is below three, while the rest of the variables are leptokurtic since their kurtosis is above three. The variables in the study are individually not normally distributed as the probability value of the Jarque-Bera statistics are statistically significant.

Table 3. Descriptive Statistics of Study Variable


HCE

IREM

OREM

PCI

EXR

Mean

2.49E+10

1.69E+09

137.6770

1336.130

335.7514

Median

6.33E+09

2.14E+08

52.24447

789.4973

156.9138

Maximum

3.92E+11

2.10E+10

2607.106

4773.917

2586.890

Minimum

5.25E+08

746883.9

0.000000

315.2639

0.905209

Std. Dev.

6.24E+10

4.60E+09

286.8328

1178.160

397.0278

Skewness

4.139319

3.498474

5.149149

1.585844

3.010395

Kurtosis

20.36223

13.71516

37.52494

4.318897

16.95255

Jarque-Bera

2682.376

1187.346

9410.690

85.54342

1674.196

Probability

0.000000

0.000000

0.000000

0.000000

0.000000

Source: Authors’ Computation (2021)



Table 4. Descriptive Statistics of Study Variables (Continued)


INF

CPS

UNE

PSE

DPR

Mean

7.053439

18.73065

7.151086

104.1723

85.40300

Median

5.776217

15.61786

5.819000

103.4257

87.46531

Maximum

44.80416

65.74181

28.24000

148.2303

103.2776

Minimum

-4.294873

1.596296

1.017000

58.32719

50.12106

Std. Dev.

6.902635

13.63315

5.972676

20.83669

11.78542

Skewness

1.717385

1.916831

2.097026

0.364735

-0.859316

Kurtosis

8.015326

6.802944

7.252442

2.340648

3.581887

Jarque-Bera

267.8958

211.4053

258.6316

7.009815

23.86911

Probability

0.000000

0.000000

0.000000

0.030050

0.000007

Source: Authors’ computation (2021)

5.1.2. Correlation Analysis

In Dormann et al. (2012), when correlation between two variables is such that |r| is ≤ 0.7, then there is no possible multicollinearity problem, otherwise, it exists. Following this rule, the correlation matrix presented in Table (5) shows that the independent variables do not suffer from multicollinearity problem.

Table 5. Correlation Matrix


lnIREM

lnOREM

INF

lnPCI

lnEXR

CPS

UNE

PSE

DPR

lnIREM

1









lnOREM

0.6105

1








INF

-0.2155

-0.2306

1







lnPCI

0.3480

-0.0647

-0.1943

1






lnEXR

0.0575

0.3355

-0.2164

-0.5277

1





CPS

0.1429

0.0187

-0.3483

0.5647

-0.0720

1




UNE

-0.1725

-0.2024

-0.1397

0.6656

-0.5440

0.2369

1



PSE

-0.4287

-0.2801

0.3584

-0.0422

-0.0753

0.0074

0.0142

1


DPR

-0.1073

0.1309

0.1461

-0.6831

0.4802

-0.6973

-0.2765

-0.2760

1

Source: Authors’ computation (2021)



5.2. Model Results

The results of the estimated model are presented in Table (6). The instruments used in this analysis was set at a limit of (1 3) for the level equation. The results showed that the one-period lag of household consumption was significant at the one percent level. This shows that consumption in SSA is persistent through time. It is also revealed that inward remittances exerted a positive and statistically significant effect on consumption at the five percent level. The coefficient value indicates that a 1 percent rise in inward remittances would lead to about 0.09 percent rise in consumption. Outward remittances is seen to exert a negative and significant effect on consumption. The coefficient indicates that for a 1 percent increase in outward remittances, consumption falls by about 0.05 percent. This is in consonance with the study by Al Akayleh (2016). The results for inward and outward remittances meet our a priori expectations. Judging by the value of their coefficients, it would be observed that the negative effects on consumption of outward remittances is compensated for by the positive effects of inward remittances otherwise the gains from inward remittances would be wiped by the trend in outward remittances from the region.

The control variables have shown varied results. From the estimated results, it is shown that inflation exerted a positive and statistically significant effect on consumption, against the negative relationship that was expected. From the result, a 1 percent rise in inflation would lead to about 0.021 percent rise in consumption. This phenomenon may not be entirely unexpected in SSA where the pressure of inflation often force households to spend more on the same goods, especially those goods that are non-household durables such as food and water. The result also showed that the coefficient of per capita income is negative and statistically not significant, against the a priori expectation. This indicates that households in SSA consume less the wealthier they become, even if this pattern is not significant. Furthermore, exchange rate is shown to exert a positive and significant effect on consumption. The coefficient shows that consumption rises as exchange rate depreciates. This finding is similar to the long-run positive effect of positive change in exchange rate on consumption found in Okwu et al. (2020). Credit to private sector exerted a positive but not significant effect on consumption; the sign of the coefficient of credit to private consumption meets the a priori expectations. From the results, if credit to private sector increases by 1 percent of GDP, consumption increases by about 0.2 percent. Additionally, unemployment rate is seen to exert a positive, but not statistically significant effect on consumption, against the a priori expectation of a negative relationship. From the results, if unemployment rises by about 1 percent of labour force, consumption is raised by about 0.8 percent. Rising unemployment in SSA increases the dependency on those working, in a region not known for having benefits for unemployed people. Additionally, primary school enrollment rate coefficient is in agreement with our a priori expectations. It exerted a positive effect on consumption even though the effect is not statistically significant. From the results, an increase in primary school enrollment rate by 1 percent of gross enrollment induces about 0.3 percent rise in consumption. Finally, dependency ratio is shown to have a positive effect on consumption, even though it is insignificant; this did not meet our a priori expectations. From the results, a 1 percent rise in dependents as a proportion of working-age population induces an increase in consumption by about 0.5 percent.

The statistically significant F-statistics showed that the model is significant overall.

Table 6. Two-Step System-GMM Estimation Results

Dependent Variable: lnHCE

Variables

Coefficient

z-statistics

p-value

Lagged dependent variable (L1.lnHCE)

0.937

(0. 041)

22.67***

0.000

lnIREM

0.094

(0.036)

2.58**

0.010

lnOREM

-0.035

(0.017)

-2.01**

0.045

INF

0.022

(0.006)

3.40***

0.001

lnPCI

-0.022

(0.117)

-0.19

0.851

lnEXR

0.030

(0.013)

2.34**

0.019

CPS

0.002

(0.003)

0.74

0.461

UNE

0.008

(0.008)

0.98

0.329

PSE

0.003

(0.003)

0.88

0.380

DPR

0.005

(0.005)

0.86

0.387

Constant

-1.128

(1.222)

-0.92

0.356

Note: The standard errors in this estimation are Windmeijer robust standard errors, and they are presented in brackets while the statistical significance is defined as *** p<0.01, ** p<0.05, & *p<0.1. All variables in column 1 are as defined in Table (1).

F-Stat = 12550.52 Prob-value (F-Stat): 0.000

Source: Authors’ computation (2021)



5.3. Post-estimation Analyses

5.3.1. Serial Correlation Test

Due to the presence of the lagged dependent variable as a regressor in the model, the GMM model tested for serial correlation/autocorrelation. These tests are Arellano-Bond test of autocorrelation: AR(1) and AR(2). The hypothesis to be tested here is that there is no autocorrelation. In Table (7) we present the serial correlation test result which showed that given the probability values of AR(1) and AR(2) at 0.004 and 0.097, at the 5 percent level of significance, we reject the presence of autocorrelation found in AR(1) using the result of AR(2). Thus, the system GMM model adopted for this study is valid.

Table 7. Serial Correlation


Z-statistics

Probability value

AR(1)

-2.88

0.004

AR(2)

-1.66

0.097

Source: Authors’ computation (2021)

Sargan and Hansen Tests

The Sagran and Hansen tests validate the instruments used in the model, testing for over-identification restrictions. The null of the test is that all instruments are valid. In Table (8), we present the result of the Sargan and Hansen tests which show that the instruments employed in the model are valid given that the p-value of the Sargan and Hansen tests are 0.641 and 0.992 respectively.

Table 8. Sargan and Hansen TEST


Chi2 statistics

Probability value

Sargan test

11.57

0.641

Hansen test

4.44

0.992

Source: Authors’ computation (2021)

5.4. Robustness Check of System-GMM Model

In Bond (2002), the system GMM result is valid only if the coefficient of the lagged value of the dependent variable in the GMM estimate lies between the pooled OLS estimates and Within-group estimates (Fixed effects estimates). From the results presented in Table (9), it is observed that the lag of the system-GMM estimate lies between the OLS and fixed effects estimates. That is 0.9947702 > 0.9371555 < 0.6567627.



Table 9. Pooled OLS/Fixed Effects Estimation Results

Dependent Variable: lnHCE

Variables

Coefficient

t-statistics

p-value

Lagged dependent variable (L1.lnHCE)

0.995/0.657

(0.015/(0.061)

65.45***/10.84***

0.000/0.000

lnIREM

0.013/0.050

(0.014)/(0.018)

0.96/2.70***

0.339/0.008

lnOREM

-0.012/0.006

(0.009/(0.010)

-1.33/0.64

0.187/0.524

INF

-0.002/-0.002

(0.002/(0.002)

-0.72/-0.74

0.473/0.461

lnPCI

0.002/0.634

(0.046)/(0.216)

0.04/2.94***

0.972/0.004

lnEXR

-0.011/-0.155

(0.012)/(0.052)

-0.93/-2.95***

0.353/0.004

CPS

0.004/-0.003

(0.001/(0.003)

0.00/-0.74

0.997/0.459

UNE

-0.003/-0.009

(0.004)/(0.009)

-0.67/-1.00

0.502/0.318

PSE

0.0001/0.001

(0.001)/(0.002)

0.94/0.51

0.349/0.614

DPR

0.002/-0.002

(0.002)/(0.006)

1.03/-0.37

0.304/0.711

Constant

-0.244/3.462

(0.346)/(1.395)

-0.71/2.48**

0.480/0.014

Note: The statistical significance is defined as *** p<0.01, ** p<0.05, & *p<0.1. All variables in column 1 are as defined in Table (1).

Source: Authors’ Computation (2021)

6. Conclusion

Remittances into SSA has been growing in recent years, so also is outward remittances from SSA. While literature on the macroeconomic effects of inward remittances into SSA are abundant; especially its effects on household consumption, there is scarce literature on how both inward and outward remittances affect welfare in SSA. The few studies on the phenomenon have reported varying results. For example, while Al Akayleh (2016) reported a negative effect of outward remittances on consumption and investment in Saudi Arabia; Ewubare & Okpoi (2018) reported a long-run positive effect in the long-run in Nigeria. We have adopted the System GMM estimation model in our study because the number of cross-sections were more than the number of years and it is a useful model in analyzing short dynamic panels such as the one used in this study. Findings from the study showed that both inward and outward remittances contributed significantly to consumption (proxy for welfare) in SSA. While the contribution of inward remittances was positive, the effect of outward remittances was negative. However, the coefficient of inward remittances was seen to be larger than that of outward remittances, meaning that the negative effects of outward remittances is countered, almost twice by inward remittances. The implication of this is that while outward remittances cannot be stopped, measures should be taken to make inward remittances easier and cheaper, so that more can be attracted and the negative effect of outward remittances significantly compensated for.



References

Abbas, K.; Muhammad Sabir, H.; Shehzadi, A. & Abbas, Q. (2014). Impact of Workers’ Remittances on Household Welfare in District Jhang A Case Study of Tehsil 18 Hazari. Journal of Finance and Economics, 2(4), pp. 131–135. https://doi.org/10.12691/jfe-2-4-5.

Abduvaliev, M. & Bustillo, R. (2019). Impact of remittances on economic growth and poverty reduction amongst CIS countries. Post-Communist Economies, pp. 1–22. https://doi.org/10.1080/14631377.2019.1678094.

Addison, T.; Pikkarainen, V.; Rönkkö, R. & Tarp, F. (2017). Development and Poverty in sub-Saharan Africa, 2017 / 169. WIDER Working Paper.

Adelowokan, O. A.; Adesola, B.; Akpa, E. O. & Maku, O. E. (2019). Remittances , Foreign Aid and Private Consumption in Sub-Saharan Africa ( SSA ): A System GMM Estimation. KIU Journal of Social Science, 5(4), pp. 67–76. https://doi.org/10.2139/ssrn.3526808

Ahmed, V.; Sugiyarto, G. & Jha, S. (2010). Remittances and household welfare: A case study of Pakistan. ADB Economics Working Paper Series (194). https://doi.org/10.2139/ssrn.1632200

Ajaero, C. K.; Nzeadibe, C. T.; Obisie-Nmehielle, N. & Ike, C. C. (2018). The linkages between international migration, remittances and household welfare in Nigeria. Migration and Development, 7(1), pp. 40–54. https://doi.org/10.1080/21632324.2017.1322171.

Akanle, O. & Adesina, J. O. (2017). Remittances and Household Welfare in Nigeria. African Population Studies, 31(1), pp. 3194–3211.

Akhter, N. & Islam, M. K. (2019). The Impact of Migration and Migrant Remittances on Household Poverty in Bangladesh. Asian Development Perspectives, 10(1), pp. 43–59.

Akpa, E. O.; Awode, S. S.; Okwu, A. T. & Oseni, I. O. (2020). The Global Financial Crisis (GFC) and Remittances Received In Africa : Any Lessons For Covid-19? South-Eastern Europe Journal of Economics, 18(2), pp. 217–239.

Akpa, E. O. (2018). Private Remittances Received and Household Consumption in Ghana (1980- 2016): An ARDL Analysis with Structural Breaks. International Journal of Management and Economics Invention, 04(05), pp. 1771–1777. https://doi.org/10.31142/ijmei/v4i5.10.

Al Akayleh, F. (2016). Do Foreign Workers’ Positive Contributions to GDP Outweigh the Negative Effect of Their Outward Remittances on GDP? A Case Study of Saudi Arabia. Advances in Economics and Business, 4(12), pp. 657–666. https://doi.org/10.13189/aeb.2016.041204.

Arapi-gjini, A.; Möllers, J. & Herzfeld, T. (2020). Measuring Dynamic Effects of Remittances on Poverty and Inequality with Evidence from Kosovo. Eastern European Economics, pp. 1–26. https://doi.org/10.1080/00128775.2020.1720517

Arellano, M. & Bond, S. (1991). Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations. Review of Economic Studies, 58(2), pp. 277–297. https://doi.org/10.2307/2297968.

Arellano, M. & Bover, O. (1995). Another Look at the Instrumental Variable Estimation of Error-Components Models. Journal of Econometrics, 68, pp. 29–51. https://doi.org/10.1016/0304-4076(94)01642-D.

Asongu, S. A. & le Roux, S. (2018). Understanding Sub-Saharan Africa’s Extreme Poverty Tragedy WP/18/012; A G D I Working Paper, Issue WP/18/012.

Assaminew, E.; Ahmed, G.; Aberra, K. & Makonnen, T. (2011). International Migration, Remittances and Poverty Alleviation in Ethiopia. Ethiopian Journal of Development Research, 33(1). https://doi.org/10.4314/ejdr.v33i1.68616.

Awan, M. S.; Javed, M. & Waqas, M. (2015). Migration, Remittances, and Household Welfare: Evidence from Pakistan. The Lahore Journal of Economics, 20(1), pp. 47–69. https://doi.org/10.35536/lje.2015.v20.i1.a2.

Azam, M.; Haseeb, M. & Samsudin, S. (2016). The impact of foreign remittances on poverty alleviation: Global evidence. Economics and Sociology, 9(1), pp. 264–281. https://doi.org/10.14254/2071-789X.2016/9-1/18.

Bahadir, B.; Chatterjee, S. & Lebesmuehlbacher, T. (2018). The macroeconomic consequences of remittances. Journal of International Economics. https://doi.org/10.1016/j.jinteco.2018.01.010.

Biyase, M. & Zwane, T. (2017). An Empirical Analysis of the Determinants of Poverty and Household Welfare in South Africa. MPRA Paper, 77085, pp. 1–19. https://doi.org/10.1353/jda.2018.0008.

Blundell, R. & Bond, S. (1998). Initial Conditions and Moment Restrictions in Dynamic Panel Data Models. Journal of Econometrics, 87, pp. 115–143. https://doi.org/10.1016/S0304-4076(98)00009-8.

Bond, S. R. (2002). Dynamic Panel Data Models: a Guide to Micro Data Methods and Practice. Portuguese Economic Journal, 1, pp. 141–162. https://doi.org/10.1007/s10258-002-0009-9.

Butkus, M.; Matuzevičiūtė, K. & Raupytė, K. (2020). Effects of Remittances on Poverty : Evidence in CEE Countries. Organizations and Markets in Emerging Economies, 11(1), pp. 69–82.

Cardozo, A. R.; Nowak-Lehmann, F. & Djiofack, C. Z. (2019). Migration and remittances in Haiti: Their welfare impact on poor and non-poor households. Georg-August-Universität Göttingen, Ibero-America Institute for Economic Research Discussion Papers, 244, pp. 1–28.

Chenery, H. B., & Strout, A. M. (1966). Foreign Assistance and Economic Development. The American Economic Review, 56(4), pp. 679–733.

Cuong, N., & Linh, V. (2017). The Impact of Migration and Remittance on Household Welfare: Evidence from Vietnam. Munich Personal RePEc Archive, pp. 1–30.

Evans, O., & Kelikume, I. (2018). The Effects of Foreign Direct Investment, Trade, Aid, Remittances and Tourism on Welfare under Terrorism and Militancy. International Journal of Management, Economics and Social Sciences, 7(3), pp. 206–232. https://doi.org/10.32327/ijmess.7.3.2018.14.

Ewubare, D., & Okpoi, G. E. (2018). International Remittances and Poverty Reduction in Nigeria. Journal of Applied Economics and Business, 6(2), pp. 5–24.

Grell-Brisk, M. (2019). Arrested development? Sub-Saharan Africa in the stratified world economy 1965–2015. Journal of Contemporary African Studies. https://doi.org/10.1080/02589001.2019.1585525

Iheonu, C. O. & Nwachukwu, T. (2020). Macroeconomic determinants of Household Consumption in selected West African countries. WP/20/031; AGDI Working Paper.

Imai, K. S.; Malaeb, B. & Bresciani, F. (2017). Remittances, growth and poverty reduction in Asia. Ifad Research Series 15, pp. 1–47.

Imran, K.; Devadason, E. S., & Cheok, C. K. (2019). Foreign Remittances and Regional Poverty : Evidence from Household Data. International Migration, pp. 1–17. https://doi.org/10.1111/imig.12681.

Javed, M.; Awan, M. S. & Waqas, M. (2015). International Migration, Remittances Inflow and Household Welfare: An Intra Village Comparison from Pakistan. Social Indicators Research. https://doi.org/10.1007/s11205-015-1199-8

Javid, M.; Arif, U. & Qayyum, A. (2012). Impact of remittances on economic growth and poverty. Post-Communist Economies, 2(1), pp. 433–447. https://doi.org/10.1080/14631377.2019.1678094.

Kangmennaang, J.; Bezner-Kerr, R. & Luginaah, I. (2018). Impact of migration and remittances on household welfare among rural households in Northern and Central Malawi. Migration and Development, 7(1), pp. 55–71. https://doi.org/10.1080/21632324.2017.1325551

Koc, I., & Onan, I. (2004). International Mi’ants’ Remittances and WeGfare StatuS of the Lefi-Behind Families in Tbkey. International Migrants’ Remittances, 38(1), pp. 78–112. https://doi.org/10.1017/CBO9781107415324.004

Kumar, B. (2019a). Remittances, Poverty and Welfare: Evidence from Cumilla, Bangladesh. American Journal of Data Mining and Knowledge Discovery, 4(1), pp. 46–52. https://doi.org/10.11648/j.ajdmkd.20190401.17

Kumar, B. (2019b). The Impact of International Remittances on Poverty Alleviation. Remittances Review, 4(1), pp. 67–86.

Marta, J.; Fauzi, A.; Juanda, B. & Rustiadi, E. (2020). Regional Studies , Regional Science Understanding migration motives and its impact on household welfare: evidence from rural – urban migration in Indonesia. Regional Studies, Regional Science, 7(1), pp. 118–132. https://doi.org/10.1080/21681376.2020.1746194

Mashayekhi, N.; Zhang, L. & Eugui, D. V. (2013). Maxamizing the development impact of remittance. United Nations Conference on Trade and Development, 181.

Munyegera, G. K. & Matsumoto, T. (2016). Mobile Money, Remittances, and Household Welfare: Panel Evidence from Rural Uganda. World Development, 79, pp. 127–137. https://doi.org/10.1016/j.worlddev.2015.11.006

Nwaru, J. C.; Iheke, O. R. & Onyenweaku, C. E. (2011). Impact of Migrant Remittances on the Welfare of Arable Crop Farm Households in South Eastern Nigeria. Human Ecology Review, 18(2), pp. 159–166. https://doi.org/10.9734/ajea/2014/6502

Okwu, A. T.; Akpa, E. O.; Oseni, I. O. & Obiakor, R. T. (2020). Oil Export Revenue and Exchange Rate: An Investigation of Asymmetric Effects on Households’ Consumption Expenditure in Nigeria. Babcock Journal OfEconomics (BJOE), 7(E/June), pp. 13–28.

Olubusoye, E. O.; Salisu, A. A. & Olofin, S. O. (2016). Applied Panel Data Analysis, Ibadan University Press, Ibadan.

Quartey, P. (2006). The impact of migrant remittances on household welfare in Ghana. African Economic Research Consortium (AERC) Paper 158.

Salman, K. K. (2016). Have Migrant Remittances Influenced Self-Employment and Welfare among Recipient Households in Nigeria ? No. 0030; AGRODEP Working Paper, Issue July.

Samaratunge, R.; Kumara, A. S. & Abeysekera, L. (2020). Where do Remittances Go in Household Consumption ? Empirical Evidence from Sri Lanka-Wide Micro-data. International Migration, pp. 1–26. https://doi.org/10.1111/imig.12697.

Sulemana, I.; Anarfo, E. B. & Quartey, P. (2018). International remittances and household food security in Sub-Saharan Africa International remittances and household food security in Sub-Saharan Africa. Migration and Development, 00(00), pp. 1–17. https://doi.org/10.1080/21632324.2018.1560926.

Sulemana, I.; Louis, D. & Anarfo, E. B. (2019). International Remittances and Subjective Wellbeing in Sub-Saharan Africa : A Micro-level Study. Journal of Family and Economic Issues. https://doi.org/10.1007/s10834-019-09615-5.

Viet, C. N. (2008). Do foreign remittances matter to poverty and inequality? Evidence from Vietnam. Economics Bulletin, 15(1), pp. 1–11.

Wadood, S. N. & Hossain, A. (2017). Microeconomic impact of remittances on household welfare: Evidences from Bangladesh. Business and Economic Horizons, 13(1), pp. 10–29. https://doi.org/10.15208/beh.2017.02.

World Bank (2013). The Global Burden of Disease: main Findings for Sub-Saharan Africa. https://www.worldbank.org/en/region/afr/publication/global-burden-of-disease-findings-for-sub-saharan-africa.

World Bank (2018). Migration and Remittances: Recent Developments and Outlook. Migration and Policy Brief, 29.

World Bank (2019). Migration and Remittances: Recent Developments and Outlook. Migration and Policy Brief, 31.

World Development Indicators (WDI). (2020). https://databank.worldbank.org/source/world-development-indicators

Wouterse, F. (2010). Remittances, poverty, inequality and welfare: Evidence from the central plateau of Burkina Faso. Journal of Development Studies, 46(4), pp. 771–789. https://doi.org/10.1080/00220380903019461





1Associate Professor, Economics Department, Babcock University, Ilishan-Remo, Nigeria, Address: P.M.B 4003, Ilishan Remo, Ogun State, Nigeria.

2 PhD Candidate, Department of Economics, Olabisi Onabanjo University, Address: P.M.B 2002, Ago-Iwoye, Ogun State, Nigeria, corresponding author: akpaemeka@gmail.com.

3Associate Professor, Economics Department, Babcock University, Ilishan-Remo, Nigeria, Address: P.M.B 4003, Ilishan Remo, Ogun State, Nigeria.

AUDOE, Vol. 17, No. 2/2021, pp. 84-104