Habibu Mohammed Umar1, Ibrahim Muhammed Dahiru2, Ibrahim Idris3
1&2
Department of Economics and Development Studies Faculty of Management and Social Sciences Federal University Gashua, Yobe State, Nigeria
3Department of Economics, School of Arts and Social Sciences Aminu Saleh College of Education, Azare
Bauchi State, Nigeria
Correspondence email: [email protected]
Abstract
This paper investigates determinants of welfare and sources of regional welfare disparity in Nigeria, focusing on the role of educational attainment and its distribution among households.
The paper used data from the „Living Standards Measurement Survey‟ (LSMS, 2013) on Nigeria and applied a decomposition method. The study found that the household characteristics associated with better demographics such as a low number of dependants, urbanisation, the type of occupation, educational attainment and its distribution are associated with a higher standard of living in Nigeria. It is also found that the difference in the distribution of educational attainment among households is the leading factor to account for observed welfare gap across the regions. The findings emphasize the role of household characteristics in explaining regional welfare disparity and substantial improvements of those characteristics in the northern region will ameliorate the welfare gap. Therefore, investing in programs that ensure equal access to education and the support for poor people to improve their human capital endowments may be a very helpful strategy.
Keywords: Welfare, Educational Distribution, Regional Disparity, Decomposition Analysis.
Introduction
Addressing the problem of inequality among individuals and regions has been an enduring development challenge everywhere in the world. In many countries of the world, it is common to observe that the indicators of living standard are higher in some areas than the national average while is the opposite in some other areas. Similarly, it is a common trend, especially in developing countries, to find the concentration of the major portion of wealth in the hands of a few people while the vast majority of the population in that country or region is living in poverty. Such disparities between and within regions may stem from the concentration of people with better demographic and human capital characteristics (concentration hypothesis) or due to the variations in the level of distribution of these characteristics as well as their associated returns ( distribution and geography hypotheses).
The concentration of people with more favourable characteristics such as education is possible due to migration or historical experience, while differences in returns and productivity stem from specific geographic factors, such as the quality of institutions, access
112
to infrastructure, distance, and size of the markets (Skoufias & Katayama 2011).
Understanding whether people‟s characteristics or geography explain regional economic disparities has important policy implications calling for public policies and programs to invest in people in the particular lagging areas.
The Nigerian gross domestic product (GDP) has grown remarkably for over a decade, averaging 7% per annum in real terms (National Bureau of Statistics (NBS), 2013). , Despite the rapid economic growth, there is little or no progress in level of poverty reduction in the country. While other growing economies such as Brazil, China and India have successfully lifted substantial number of their people out from poverty over the last decade. Nigeria, on the other hand, is experiencing the reverse. In Nigeria, poverty incidence based on the international poverty line (living on less than US$1 per day) has been on the increase as shown on figure 1. Regional disparities in Nigeria offers an interesting case study because of the pronounced and long existing economic differences between the Northern and Southern regions and rural and urban areas of the country despite the recorded economic growth and poverty reduction programs (UNDP, 2012). The recent Nigeria survey on households‟
standards of living by the country‟s Bureau of statistics shows a variation in average income levels of the households as well as GDP per capita level across regions in the country (NBS, 2013). The northern region has the lowest average household income and the per capita GDP level and also shows that relative poverty is most apparent in the region of the country. The North-West and North-Eastern regions of the country have poverty rates as high as 77.7%
and 76.3% respectively. The South-West has only 59.1% poverty rate and the trend continuous with almost all socio-economic indicators. This scenario, if not checked can cause distrust between the regions and plant a state of hopelessness in the minds of people in the lagging regions which could be counterproductive to the nation‟s economy as a whole.
Recognizing the adverse social, economic, and political consequences of the above trends, as exemplified in the Arab Spring, it is imperative to pursue policy guides aggressively to ameliorate the disparities.
This paper is aimed at analysing regional welfare disparities in Nigeria by focusing on the role of educational attainment and its distribution. The paper is aimed at quantifying and separating the sources of regional welfare gap emerging from the differences in observable regional characteristics of households and associated differences in marginal returns to these characteristics. Specifically, the study first investigate how differences in average educational attainment, educational inequality and family size result in different living standards.
The study is important due to the following reasons. First, it verifies the relative role of educational attainment, its distribution and other factors in determining regional welfare status. Second. It evaluates how these factors shape the regional disparity in Nigeria.
Moreover, third, it recommends policy changes that will help reduce poverty incidence in the country. The remaining part of the paper is structured as follows. Section 2 presents the background information describing welfare disparities and structural differences across the regions. Section 3 briefly describes the data used for the analysis. It also explains the methodology and presents the variables used in the analysis. The results of the decomposition are presented and discussed in Section 4 while Section 5 concludes the paper by summarizing the findings and proposing how this work can be carried forward as future study.
113
Background information
Poverty rate has not been declining in Nigeria despite the steady economic growth the country recorded over the last two decades. The Sustainable Development Goals (SDGs) requires that the percentage of the population living in poverty to be below 22% in 2015.
However, the data shows that poverty rates in terms of the minimal requirements necessary to afford minimal standards of food, clothing, healthcare and shelter (absolute poverty) is far above the 22% threshold. The recent statistics on poverty about Nigeria shows that 40.1% of the country‟s population are poor (NBS, 2020)
The differences in regional poverty rates still remain a pressing issue in the Nigeria due to a wide gap between the southern and northern regions. Disparities in poverty rates were rising in Nigeria mainly because of the sharp variation in the economic performance among the regions. The northern region had a poverty rate of 59.5% in 2019, while the Southern region has a poverty rate of 24.3 %. The level of poverty is lower in the southern region, so also the rate at which it changes over the time. This disparity is highlighted on Table 1.
Table 1: Regional absolute poverty rate in Nigeria %
Year National Northern region Southern region
2004 54.4 58.0 50.2
2019 40.1 59.5 24.3
Source: NBS Press briefing on Nigeria Poverty Profile 2020 Report
There are also variations in the level of income between the regions in Nigeria. The recent Nigeria survey on households‟ standards of living by the country‟s Bureau of statistics (NBS, 2019), shows a variation in average income level of the households as well as GDP per capita level across regions in the country. The northern region is having the lowest average household income and the per capita GDP level. Thus, it implies that lower standard of living is most apparent in the Northern region of the country. Figure 1 depicts the regional variation in income in the country.
Figure 1: Average per Capita Income Distributions across Regions in Nigeria
Source: Author‟s calculations from the living standards measurement survey on Nigeria (2019)
Data and Methodology
The data utilized in this paper is from the Nigeria household living standards measurement survey (LSMS) for 2013, which was collected by the country‟s national bureau of statistics with technical support from the World Bank and made available on the World Bank database.
The survey covered 5000 households nationwide. The sample of the survey is drawn using
114
stratified two-stage random sampling procedure ensuring representativeness of the regional dimensions, as well as the urban and rural divide of the country.
Following Wodon, 1999 and Wan & Zhou, (2005) the determinants of welfare can be evaluated based on the following specification:
(1)
Where log (Y) is a log of real household expenditure per capita; and X is a vector of variables representing the characteristics of households which likely affect the expenditure per capita The D is a regional dummy (1 for the southern region, zero otherwise); is a vector of coefficients; and is the error term. Subscript stands for individual household head (i = 1...
N). the variables explaining the welfare in the above specification (Eq. 1) include demographics, employment, and education categories. The category of demographic variables includes, gender, age and household size (hhsize; number of dependents of less than 25 years excluding the spouse). The employment variables includes dummy for the primary occupation or industry of the head of household (Inddmy; 1 for agriculture and 0 otherwise) and for the head of household living in the urban or rural sector (sector; 1 for urban, 0 otherwise). The education category consists of the level of educational attainment of the household head measured by average years of schooling and the measure of educational distribution (inequality) using Theil index.
.
Per capita expenditure is chosen rather than income as a proxy for household welfare. This is because of the suggestions in the literature that the expenditure is a good proxy for permanent income and also for long-term average well-being of an individual or group (Li & Xu, 2008;
Balisacan, Pernia & Asra, 2003). For example, a low-income family unit can withdraw its savings or obtain a loan to spend in order to keep up its relative living standard. In contrast, a high-income but indebted family unit need to cut part of its income to pay off the obligation.
Additionally, information on consumption is less hard to assemble than those on income, particularly in the developing nations where self-employed people are hesitant to reveal their income unequivocally. Along these lines, in this study, as outstandingly used in the previous studies, per capita household expenditure is utilized as an approximation for household welfare.
Test for structural difference (Chow test)
A Chow test is used to test whether the welfare function is different for the two regions. In other words, this is to show that one subsample has different intercept and slopes than another. The test can be used to detect structural breaks in time-series models or to determine whether, in this case, the South‟s welfare is determined differently from that in the northern part of the country (Adkins, 2012). The gretl's built-in chow command is used to test for a change in the regression. It must follow an executed regression of equation one (Eq.1)
Oaxaca-Blinder Decomposition
To examine how educational attainment and its distribution (inequality) explain the regional disparity in Nigeria, a decomposition method suggested by Oaxaca and Blinder, (1973) is used. This method can be used to analyse differences across groups in any continuous and unbounded outcome variable as demonstrated in Jann, (2008). Similarly, the method was used empirically by O‟Donnell, van Doorslaer, Wagstaff and Lindelow, (2008) in their work on regional health inequalities.
115
Having two regions south (1) and north (2) for the study; it is assumed that the natural logarithm of welfare can be summarized by linear regressions:
(2)
In the above specification (Eq. 2), is the vector of the coefficients associated with educational attainment, educational distribution and household size included in the vector X and is a random disturbance term. The coefficients in equation 2 summarize the marginal welfare gains in terms of consumption for each particular explanatory variable. The marginal welfare gains to the family unit attributes are additionally influenced by a number of other geographic factors of a region, for example, markets access, infrastructure, traditions and other cultural factors which are excluded in the vector of explanatory variables of equation 2.
This will form the unexplained portion of the decomposition analysis.
Given the two regions (i.e. South and North) with the outcome variable Y (expenditure) and a set of predictors comprising educational attainment and inequality, the disparity between the two regions with the respect to the outcome variable is set to be accounted for by regional differences in the predictors. This can be expressed as:
Where R is the difference in the outcome variable between the regions, and E(Y) denotes the expected value of the outcome variable based on the following model:
Where X is a vector containing the independent variables, β contains the slope parameters and the intercept, and ∈ is the error term. Following Jann, (2008) the mean outcome difference can be expressed as the difference in the linear prediction at the regions -specific means of the predictors. Hence;
This is a „three-fold‟ decomposition in which the outcome difference is decomposed into three parts: compactly expressed as:
The first summand in the right hand side of equation X, is the
„endowment effects‟ i.e. the part of the differential due to regional differences in the predictors. The second summand which is measures the contribution of the differences in the coefficients including intercept. The third summand;
is the interaction term accounting for the simultaneous existence of the differences in endowments and coefficients (returns) between the two regions.
The above decomposition (eq.5) is formulated from the viewpoint of the North. this implies that, the regional disparities in the indicators are weighted by the coefficients of the Northern region to determine the endowment effect (E). As it were, the E part measures the normal change in one region's mean outcome (say Northern region), if the region had the other region's (i.e. the Southern region) levels of the predictors. Likewise, for the second part (C), the differences in coefficients are weighted by the Northern region's indicator levels. That is;
the second segment measures the expected change in Northern region's mean outcome, if the region had the Southern region's coefficients.
116
Results
This section presents the estimated results. The first step involved the regression of the welfare function as specified in equation one (Eq.1). This is followed by the Chow test and finally the decomposition analysis. The results are presented in tables 4 and 5.
The null hypothesis for the Chow test is that the coefficients of the two subsets (i.e. the regions) are equal and the alternative is that they are not. In this case, the p-value associated with the test is less than 1% level of significance (p-value = 0.000), thus providing sufficient evidence to believe that the welfare levels are different with the respect to regions in Nigeria.
However, as shown in table 4, the output of the Chow test contains the coefficient of a regional dummy (Region); which is coded „1‟ for Northern region and „0‟ for the Southern region. The coefficient of the regional indicator is significant with a negative sign (β= -0.175). This implies that the Northern region is differently lower than the southern region in terms of welfare with about 17%. This result paves the way for the decomposition analysis in an effort to analyse the role of the educational attainment level and its distribution in explaining the regional welfare disparity in Nigeria.
Table 4: Chow test of structural difference with respect to regions
Variables ln(income)
Coefficients Std. Error
gender -0.00463 (0.0154)
age 0.000544 (0.000414)
inddmy -0.0518*** (0.0150)
sector 0.0736*** (0.0128)
hhsize -0.0439*** (0.00254)
Edu 0.0232*** (0.00136)
Theil -1.584*** (0.170)
gdpp 7.77e-05*** (1.25e-05)
D -0.175** (0.0703)
r_gender 0.00254 (0.0382)
r_age -0.000175 (0.000675)
r_inddmy -0.0224 (0.0223)
r_sector 0.108*** (0.0204)
r_hhsize 0.0196*** (0.00341)
r_Edu -0.00683*** (0.00187)
r_gdpp -1.69e-05 (1.77e-05)
r_theil 1.356*** (0.177)
Constant 4.977*** (0.0471)
Observations 4,979
R-squared 0.314
Chi-square(8) 46.23***
F-form:F(8, 4963) 5.779***
Note: Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Source: Authors‟ computation
Usually, it is expected that people involved agriculture have a lower standard of living compared with those engaged in other white collar jobs (Tamura, 2002; Trendle, 2004). The findings of Nigerian households‟ data support this expectation. As shown in Table 5, the coefficient for industry is negative and significant at 1%. This result implies that an individual living in households with a head whose main occupation is agriculture had a lower expenditure per capita compared to those living in households with a head working in other job categories. This goes in line with the findings of Anyanwu, (2013) that shows a higher incidence of poverty for those in self-employed farming than those who worked in other sectors. Similarly, the literature suggests that household size and educational inequality can affect wellbeing negatively (Berthoud & Iakōvou, 2004; Lopez, Thomas, & Wang, 1998).
117
The results of this paper support the above prediction as the level of welfare is affected negatively by those factors as shown on table 5. This implies that large households with low educational attainment are likely to have a low standard of living. On the other hand, the coefficients of education attainment and GDP per capita are all significant with positive signs, meaning that higher level of these variables is associated with a higher level of household welfare. Additionally, the households living in urban centres enjoy a higher level of welfare than households living in rural areas as shown by the coefficient on ‗sector.' Oaxaca-Blinder Decomposition
The Oaxaca-Blinder decomposition is used to explain the welfare disparities between the southern region and northern region of Nigeria. The disparity is assumed to be explained, at least partially, by the differences in the regional educational attainment, educational inequality and average household size as specified in equation 2. As shown in table 5, approximately 17 percent of the difference in the log welfare between the southern region and northern region can be explained by less favourable household endowments in the northern regions such as lower educational attainment level and its distribution as well as larger household size. The difference in the level of educational inequality is the main factor in the explained part of the gap. The results show that welfare disparity among the regions is mainly because of the differences in the regional observable characteristics (the explained part), rather than the differences associated with the returns to the characteristics. Meaning that, the role of marginal returns to the characteristics (the unexplained part) is less important.
Table 5: Blinder-Oaxaca decomposition
(1) (2) (3)
VARIABLES overall explained unexplained
group_1 4.883***
(0.00713)
group_2 4.712***
(0.00749)
difference 0.171***
(0.0103)
explained 0.206***
(0.0137) unexplained -0.0356**
(0.0152)
schooling 0.0643*** 0.0332***
(0.00434) (0.00971)
theil 0.0824*** -0.195***
(0.0130) (0.0284)
hhsize 0.0596*** -0.110***
(0.00418) (0.0178)
Constant 0.237***
(0.0393)
Observations 4,979 4,979 4,979
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
In the first column of the decomposition output shown in table 5 is the mean predictions by Regions and their difference. In the sample, the mean of the log income is 4.883 for the Southern region and 4.712 for the Northern region, yielding a regional income difference of about 17 percentage points (0.17). It shows that, almost all of the observed regional difference is explained by the differences in the level of the observed characteristics (i.e.
118
household size, educational attainment and its distribution). This implies that, adjusting the level of these factors in the North to the levels obtainable in the Southern region would clear off the observed regional welfare gap.
Conclusion
This paper analysed welfare disparities between the southern and northern regions of Nigeria by decomposing the logarithm of the welfare components associated with household characteristics (educational attainment, educational distribution and household size) and their associated returns. This helps to understand whether regional disparities in welfare are related to the concentration of people with favourable characteristics or higher returns to them in a particular region. The Chow test conducted reveals a significant difference in terms of welfare between the southern region and northern regions of Nigeria. Moreover, the decomposition analysis reveals an interesting picture with regards to the leading determinants of welfare gap among the regions in Nigeria.
The household characteristics associated with better demographics such as a low number of dependants, urbanisation, the type of occupation, educational attainment and its distribution are associated with a higher standard of living in Nigeria. Specifically, educational attainment, educational distribution and household size are found to play more important role in explaining the welfare gap between southern and northern regions of Nigeria. It is found that the northern region lag behind the south due to the concentration of households with better characteristics in the southern region in terms of education. The concentration of people with better endowments in the southern region could be a result of internal migration when people move into the region where their skills are most rewarded and because of the inherently different economic structure.
The findings of this study emphasize the role of household characteristics in explaining regional welfare disparity between Southern and Northern regions. Therefore, investing in programs that ensure equal access to education and support poor people to improve their human capital endowments may be a very helpful strategy. This will give the poor households a chance to search for better opportunities regardless of their area of residence.
Further investigation is required to identify the geographical factors that are responsible for low productivity in the lagging region and how these factors affect the welfare level of the region.
References
Adkins, L. C. (2012). Using Gretl for Principles of Econometrics, 4th Edition, Version 1.03.
Agbiboa, D. (2014). Under-development in practice: Nigeria and the enduring problem of corruption. Development in Practice, 24(3), 390-404.
Aigbokhan B.E. (2000), Poverty, Growth and Inequality in Nigeria: a Case Study, African Economic consortium, Nairobi.
Akanbi, F., (2011, May 15). Nigeria: Income Disparity between North and South Widens.
Thisday. Retrieved from http://www.thisdaylive.com on June 30th, 2011
Anyanwu, J. C. (2013). Determining the Correlates of Poverty for Inclusive Growth in Africa. African Development Bank.
Balisacan, A. M., Pernia, E. M., & Asra, A. (2003). Revisiting growth and poverty reduction in Indonesia: what do subnational data show? Bulletin of Indonesian Economic studies, 39(3), 329-351.
119
Berthoud, R., & Iakōvou, M. (Eds.). (2004). Social Europe: living standards and welfare states. Edward Elgar Publishing: Cheltenham Glos.
Fan, S., Kanbur, R., & Zhang, X. (2011). China's regional disparities: Experience and policy.
Review of Development Finance, 1(1), 47-56.
Fleisher, B., Li, H., & Zhao, M. Q. (2010). Human capital, economic growth, and regional inequality in China. Journal of Development Economics, 92(2), 215-231.
Jann, B. (2008). A Stata implementation of the Blinder-Oaxaca decomposition. Stata Journal, 8(4), 453-479.
Jensen, A. R. (1998). The g factor: The science of mental ability (pp. 270-305). Westport, CT: Praeger.
JICA, J. I. C. A. (2011). Federal republic of Nigeria study for poverty profile (Africa).
Tokyo: Japan International Cooperation Agency.
Li, S., & Xu, Z. (2008). The trend of regional income disparity in the People's Republic of China (No. 85). ADB Institute Discussion Papers.
Lopez, R., Thomas, V., & Wang, Y. (1998). Addressing the education puzzle: the distribution of education and economic reform: The World Bank.
Madu, I. A. (2006). Spatial inequality in Nigeria: The imperative of geographic perspectives in the development process. Journal of Social and Economic Development, 8(2), 105.
National Bureau of Statistics NBS (2020), Poverty Profile for Nigeria, National Bureau of Statistics, Abuja- Nigeria.
Nguyen, B. T., Albrecht, J. W., Vroman, S. B., & Westbrook, M. D. (2007). A quantile regression decomposition of urban–rural inequality in Vietnam. Journal of Development Economics, 83(2), 466-490.
O‟Donnell, O., E. van Doorslaer, A. Wagstaff, & M. Lindelow. 2008. Analysing Health Equity Using Household Survey Data. A Guide to Techniques and Their Implementation. Washington, DC: The World Bank.
Obayelu, O. A., & Awoyemi, T. T. (2012). Spatial decomposition of income inequality in rural Nigeria. Journal of Rural Development, 31(3), 271-286.
Skoufias, E. & Lopez-Acevedo, G. (2009). “Determinants of Regional Welfare Disparities within Latin American Countries.‖ 2 vols. The International Bank for Reconstruction and Development / The World Bank.
Skoufias, E., & Katayama, R. S. (2011). Sources of welfare disparities between and within regions of Brazil: evidence from the 2002–2003 household budget survey (POF).
Journal of Economic Geography, 11(5), 897-918.
Tamura, R. (2002). Human capital and the switch from agriculture to industry. Journal of Economic Dynamics and Control, 27(2), 207-242.
Trendle, B., & Pears, W. (2004). The role of education in regional income determination-a cross sectional study of small areas in Queensland. Australasian Journal of Regional Studies, 10(3), 383.
Wan, G., & Zhou, Z. (2005). Income Inequality in Rural China: Regression‐based Decomposition Using Household Data. Review of Development Economics, 9(1), 107-120.
Wodon, Q. (1999). Growth, poverty, and inequality: a regional panel for Bangladesh (Vol.
2072). World Bank Publications.
World Bank. (2012). Living standard measurement survey data table. Retrieved from http://go.worldbank.org/PDHZFQZ6L0
World Bank. (2013). Living standard measurement survey data table. Retrieved from http://go.worldbank.org/PDHZFQZ6L0
120