CHAPTER FOUR

DATA ANALYSIS AND INTERPRETATION OF RESULTS

4.1 Introduction

This chapter presents the results of data analysis as well as their interpretation. The
chapter is divided into five sections; the first section introduces the chapter and explains the
focus of the remaining sections. The second section presents the results of the effect of
income on health outcomes in selected African countries. The third section contains the
estimates of the effect of education on health outcomes in selected African countries using
the feasible generalised least squares (FGLS). The same estimation techniques were used to
estimate the effect of employment on health outcomes in selected African countries in
Section four. Section five presents the results of the joint effect of income, education and
employment on health outcomes in selected African countries. The results for each objective
include the descriptive statistics for the variables of interest used in the analysis, correlation,
slope heterogeneity, cross-sectional dependence test, unit root test results, as well as
cointegration results. The final section discusses the findings and provides the economic
insightful behind them.

4.2 The effect of income on Health Outcomes in selected African countries (objective 1)

This section presents the results of the effects of income on health outcomes (life
expectancy and infant mortality) in selected African countries. The analysis starts with
descriptive statistics, which describe the characteristics of the variables of interest used in this
study. The correlation analysis, unit root test, slope heterogeneity, cross-sectional
dependence, and cointegration tests results were all presented as pre-estimation procedures to
the final results of the linear regression with feasible generalised least squares (FGLS).

4.2.1 Descriptive Analysis of the Drivers of Gender Inequality in Developing Countries

The descriptive statistics presented in Table 4.1 offer a comprehensive overview of the
central tendencies and distributional characteristics of key indicators influencing gender
inequality in a panel of developing countries. The variables analyzed include the Gender
Inequality Index (GII), education gender gap (EDG), public health expenditure as a
percentage of GDP (PHE_GDP), employer gender gap (EMP_G), employment-to-population

1
gender gap (EMP_PG), the logarithm of GDP per capita (LGDPC), labour force participation
gender gap (LFP_G), population growth rate (PGR), urban population growth rate (UPR),
and wage and salaried workers gender gap (WSW_G). These indicators reflect multifaceted
dimensions of gender disparities across the social, economic, and demographic structures of
developing nations.

The Gender Inequality Index (GII) has a mean of 0.48, indicating a moderate level of gender
disparity on average across the sampled countries. With a standard deviation of 0.16 and a
negative skewness of –0.58, the distribution leans slightly towards lower inequality in a few
countries. However, the Jarque-Bera statistic (77.70, p = 0.00) rejects the null hypothesis of
normality, suggesting significant heterogeneity in gender outcomes across the sample. The
education gender gap (EDG) records an average of –3.53, reflecting persistent female
disadvantages in educational attainment. The wide range of values (–84.56 to 81.14) and a
high standard deviation (34.34) indicate large disparities across countries. The distribution is
approximately symmetric (skewness = 0.08) and relatively mesokurtic (kurtosis = 2.89),
though the Jarque-Bera test (1.96, p = 0.37) does not signal a statistically significant
departure from normality.

Public health expenditure as a share of GDP (PHE_GDP) has a mean value of 2.28%,
signifying limited investment in health sectors among developing nations. The variable is
positively skewed (0.71), suggesting that while most countries invest minimally, a few
outliers allocate relatively higher percentages. A standard deviation of 1.44 and significant
non-normality (Jarque-Bera = 112.88, p = 0.00) further emphasize the uneven commitment to
public health financing. The employer gender gap (EMP_G) has a mean of –2.71,
underscoring the underrepresentation of women among employers. The extremely negative
skewness (–2.64) and high kurtosis (10.61) reflect the concentration of data around a few
extreme values, with the Jarque-Bera statistic (4,831.88, p = 0.00) pointing to a strong
deviation from normality. This suggests that women face structural and cultural barriers to
entrepreneurship in many of these countries.

The employment-to-population gender gap (EMP_PG) shows an average of –24.33,

indicating a significant disadvantage for women in employment relative to men. The wide
variation, with a standard deviation of 16.98 and values ranging from –65.91 to 6.73,
highlights stark inequalities. The distribution is slightly negatively skewed (–0.51), with
moderate kurtosis (2.11), but again, the Jarque-Bera statistic (103.79, p = 0.00) confirms a

2
significant departure from normality. Economic welfare, captured by the log of GDP per
capita (LGDPC), has a mean of 7,129.88 with an exceptionally high standard deviation
(10,766.43), showing substantial disparities in income levels across developing countries.
The positive skewness (2.28) and high kurtosis (7.78) reflect that a few relatively wealthier
economies drive the upper end of the distribution. The Jarque-Bera test (2,457.51, p = 0.00)
further supports the presence of non-normality.

The labour force participation gender gap (LFP_G) averages –24.79, reaffirming the
substantial under-participation of women in the labour market. The variable is moderately
left-skewed (–0.63) with a standard deviation of 18.77. The Jarque-Bera statistic (123.50, p =
0.00) again shows a significant deviation from normality, suggesting that structural and
institutional barriers are common across many countries. The population growth rate (PGR)
has a mean of 2.03%, characteristic of the youthful demographic profile typical of developing
nations. However, the positive skewness (1.30) and high kurtosis (9.39) suggest outliers and
asymmetry. The Jarque-Bera test (2,676.96, p = 0.00) indicates non-normality, highlighting
that demographic pressures vary widely, with implications for social infrastructure,
particularly education and health services.

Urban population growth rate (UPR) presents an average of 3.36%, reflecting rapid
urbanization trends in developing countries. The data is highly right-skewed (5.58) with an
extremely leptokurtic distribution (kurtosis = 42.97), and the Jarque-Bera test (96,894.66, p =
0.00) points to severe non-normality. This suggests that while some countries are urbanizing
gradually, others are experiencing explosive urban growth that could strain public
infrastructure. The wage and salaried workers gender gap (WSW_G) has a mean of –5.29 and
a standard deviation of 13.02, indicating significant disparities in access to formal
employment for women. The distribution is negatively skewed (–1.21) and leptokurtic
(kurtosis = 8.31), with the Jarque-Bera test (1,913.04, p = 0.00) confirming non-normality.
These findings reinforce the persistent inequality in labour market outcomes, especially in
formal wage sectors.

In summary, the descriptive statistics reveal a landscape of deep-rooted gender disparities and
structural inequalities across developing countries. Most variables demonstrate non-normal
distributions, high skewness, and elevated kurtosis, emphasizing the asymmetry and
heterogeneity in gender outcomes. These statistical properties suggest that policy responses
must be context-specific and multidimensional, addressing disparities in education, labour

3
market access, healthcare investment, and demographic pressures. Moreover, given the
presence of non-normality in most variables, it is crucial to employ appropriate data
transformations and robust econometric techniques to ensure reliable and unbiased empirical
results.

Table 4.1,
Summary Statistics for Drivers of Gender Inequality
GII IMR GDPC CPI REX CHE_GDP PGR

Mean 0.48 50.53 2565.96 131.92 99.37 4.79 2.23

Median 59.95 47.50 1828.47 110.00 100.00 4.34 2.26

Maximum 71.37 109.60 8646.06 524.91 141.61 8.58 3.15

Minimum 47.19 15.50 402.94 29.60 69.83 2.49 0.65

Std. Dev. 6.69 24.27 2008.20 89.11 15.97 1.63 0.64

Skewness 0.18 0.56 1.36 2.01 0.43 1.00 -0.58

Kurtosis 2.00 2.20 3.78 7.81 3.29 2.87 2.35

Jarque-Bera 5.63 9.33 40.11 196.21 4.10 20.11 8.81

Probability 0.06 0.01 0.00 0.00 0.13 0.00 0.01

Observation
120 119 120 120 120 120 120
s

Source: Computed by the Author based on the World Bank WDI (2025)

Part A: The model of Developing Countries

4.2.2a Correlation Analysis for Drivers of Gender Inequality

The correlation analysis presented in Table 4.2a shows that GDP per capita (LGDPC)
has a positive but weak association with life expectancy (LLEB), with a correlation
coefficient of 0.310. This suggests that increase income is modestly linked to longer life
expectancy across the selected African countries. Similarly, real exchange rate (REX) and
health expenditure as a share of GDP (CHE_GDP) also show positive relationships with life
expectancy, with coefficients of 0.070 and 0.349 respectively, implying that favourable
exchange conditions and increased public health spending may support better health
outcomes. Conversely, inflation rate (LCPI) and population growth rate (PGR) exhibit

4
negative and weak correlations with life expectancy, at -0.306 and -0.369 respectively. This
indicates that rising prices and rapid population growth may place pressure on health systems,
thereby limiting improvements in longevity. Importantly, all correlation coefficients fall well
below the conventional multicollinearity threshold of 0.8 (Gujarati & Porter, 2008; Hair et
al., 2009). This implies that the independent variables are not highly collinear, supporting the
validity and robustness of further regression analysis.

Table 4.2a,
Correlation Matrix for Drivers
PHE_G EMP_ EMP_P LGDP LFP_ WSW_
GII EDG DP G G C G PGR UPR G
GII 1.000
-
EDG 0.054 1.000
PHE_G - -
DP 0.289 0.225 1.000
EMP_G 0.192 0.165 -0.178 1.000
EMP_P -
G 0.185 0.188 0.056 0.118 1.000
- -
LGDPC 0.682 0.133 0.465 0.181 0.044 1.000
-
LFP_G 0.242 0.240 -0.017 0.134 0.739 0.026 1.000
- - -
PGR 0.409 0.121 0.279 0.201 -0.076 0.750 0.127 1.000
- - -
UPR 0.386 0.135 0.328 0.192 -0.140 0.767 0.186 0.762 1.000
WSW_ - - - -
G 0.217 0.230 0.139 0.642 0.028 0.044 0.024 0.156 0.104 1.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.2.3a Test for multicollinearity (Variance Inflation and Tolerance Factor Test)

Based on the findings from the previous section, the Variance Inflation Factor (VIF)
and tolerance values offer a robust assessment of potential multicollinearity among the
independent variables in the regression model. The VIF identifies how much the variance of a
regression coefficient is inflated due to multicollinearity, while tolerance, the inverse of VIF
measures the proportion of variance in a predictor not explained by other predictors.
Tolerance values close to 1 indicate low multicollinearity, whereas values near 0 suggest high

5
multicollinearity. As a rule of thumb, a tolerance value above 0.1 and a VIF below 10 are
acceptable thresholds (Miles, 2014). As shown in Table 4.3a, all VIF values are below 10 and
tolerance values exceed 0.1, indicating that multicollinearity is not a concern in this study.

Table 4.3a,
Variance Inflation Factor (VIF) and Tolerance level for Income and Life Expectancy
Variable VIF 1/VIF
CHE_GDP 5.06 0.197652
LGDPC 4.92 0.203196
PGR 3.52 0.283805
LCPI 1.99 0.503041
LREX 1.63 0.615246
MeanVIF 3.42

4.2.4a Slope Heterogeneity Test for Income and Life Expectancy

Table 4.4a presents the results of the Pesaran and Yamagata (2008) slope
heterogeneity test examining the relationship between income and life expectancy. At the 1%
significance level, both the Delta and adjusted Delta statistics yield p-values of 0.000,
indicating strong statistical significance. Consequently, the null hypothesis of slope
homogeneity is rejected. This confirms the presence of slope heterogeneity across the panel
data, implying that the relationship between income and life expectancy varies across the
selected African countries.

Table 4.4a,
Pesaran and Yamagata Slope Heterogeneity Test for Income and Life Expectancy

Pesaran and Yamagata Test statistics p-value

Delta 6.205 0.000
Delta adj 7.373 0.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.2.5a Cross-Sectional Dependency Test for Income and Life Expectancy

In addition to slope heterogeneity, testing for cross-sectional dependence (CSD) is a

critical step in panel data analysis. Cross-sectional dependence arises when individual units
within a panel, such as countries are not independent of each other due to shared shocks,
spillover effects, or regional integration. The presence of CSD can lead to biased and

6
inconsistent parameter estimates if not properly addressed. The cross-sectional dependence
(CD) test remains valid under both homogeneous and heterogeneous dynamic models, as well
as for non-stationary data structures. A p-value below 5% leads to the rejection of the null
hypothesis of no cross-sectional dependence. Table 4.5a indicates that the null hypothesis of
cross-sectional independence is rejected, confirming the presence of cross-sectional
dependence among panel units. This finding necessitates the use of second-generation panel
unit root and cointegration tests that account for cross-sectional dependence. Accordingly, the
study employs the Pesaran (2007) Cross-Sectionally Augmented IPS (CIPS) unit root test,
which adjusts for common factors across countries. It is noteworthy that all variables
demonstrate cross-sectional independence, except for the real exchange rate (LREX), which
exhibits significant interdependence across the selected African countries.

Table 4.5a,
Pesaran CD Cross-Sectional Dependency Test for Income and Life Expectancy
Variable CD-test p-value corr Abs(corr)
LLEB 14.06 0.000 0.908 0.908
LGDPC 12.82 0.000 0.827 0.827
LCPI 15.29 0.000 0.987 0.987
LREX -1.57 0.117 -0.101 0.336
CHE_GDP -2.64 0.008 -0.171 0.327
PGR 2.86 0.004 0.184 0.445
Source: Computed by the Author based on the World Bank WDI (2025)

4.2.6a Unit Root Test for Income and Life Expectancy

The cross-sectionally augmented IPS (Im, Pesaran, and Shin) unit root test of Pesaran
(2007) was performed, and the data ranged from 1995 to 2023, because the null hypothesis of
cross-sectional dependence was not denied. Table 4.6a presents the results of unit root tests
for all variables using panel data. The null hypothesis of the presence of unit roots is not

7
rejected for all the variables when utilising the level form of the variables. Using the variables
in their first-differenced form, however, the null hypothesis is rejected. This means that the
variables have been integrated to the first order, indicating that each variable is a
nonstationary variable I(1). This result led to the use of panel data cointegration techniques.
Table 4.6a,
CIPS root tests (Income and Life Expectancy)
Cross-sectionally Augmented Im, Pesaran, and Shin (CIPS) test
Level Stat (critical First Difference Stat Remark
Variables level@5% (critical level@5%
LLEB -2.022(-2.33) -4.097(-2.33) I(1)
LCPI -1.270(-2.33) -2.655 (-2.33) I(1)
LGDPC -2.320(-2.33) -3.816 (-2.33) I(1)
PGR -1.386(-2.33) -2.552 (-2.33) I(1)
LREX -1.997(-2.33) -4.426 (-2.33) I(1)
CHE_GDP -1.477(-2.33) -3.876 (-2.33) I(1)
Source: Computed by the Author based on the World Bank WDI (2025)

4.2.7a Panel data Cointegration Test for Income and Life Expectancy

The Westerlund Panel Cointegration test, which accounts for cross-sectional

dependence, was employed to examine the existence of a long-run relationship among the
variables. In line with Levin, Lin, and Chu (2002), the test was adapted to reflect cross-
sectional dependence. The test evaluates the null hypothesis of no cointegration against two
alternatives: that at least some panels are cointegrated, or that all panels are cointegrated. As
shown in Table 4.7a, the null hypothesis is rejected when testing for cointegration across all
panels, indicating a common long-term relationship among the variables. However, the null is
not rejected under the assumption that only some panels are cointegrated. Overall, the
rejection of the null for all panels confirms the presence of cointegration, suggesting that a
stable long-run relationship exists among the variables.
Table 4.7a,
Modified Westerlund Panel Cointegration Test for Income and Life Expectancy
Some panels are cointegrated All panels are cointegrated

8
 statistics p-value statistics p-value
Variance ratio -2.6754 0.0037 -1.7544 0. 0397
Source: Computed based on the data from the World Bank WDI (2025)

4.2.8a Heteroskedasticity Test for Income and Life Expectancy

In this study, heteroscedasticity was examined using a modified Wald test for group-
wise heteroscedasticity. The null hypothesis that the model does not contain
heteroskedasticity is to be rejected when the P-value is less than 5% or 0.05. Table 4.8a
demonstrates that the null hypothesis cannot be rejected because the probability value is less
than 5%. This suggests that the model contains heteroscedasticity.
Table 4.8a,
Modified Wald test for groupwise heteroskedasticity result for Income and Life Expectancy
Modified Wald test Statistics and P-value
chi2 (5) 44272.52

Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.2.9a Testing for Serial Correlation for Income and Life Expectancy
Table 4.9a shows the result of the Breusch-Pagan test, which shows that the
probability value of the Chi-statistics is less than 1%. The null hypothesis that there is no
serial correlation should be rejected if the p-value is less than 5% or 0.05. From Table 4.9a,
the null hypothesis that there is no serial correlation is rejected because the p-value (0.000) is
less than 0.05. This suggests that the residuals of the panel data model suffer from serial
correlation. Hence, the model is best estimated using the feasible generalised least squares
(FGLS).
Table 4.9a,
Breusch-Pagan LM test of independence for Income and Life Expectancy
Breusch-Pagan test of independence Statistics and P-value
chi2(10) 65.210
Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.2.10a Testing for the Endogeneity for Income and Life Expectancy

9
 When a regression model’s error term and the predictor variables are correlated,
endogeneity problems occur. If the endogeneity problem is disregarded, the results could be
skewed, with the true effect of a variable on the intended outcome being over- or under-
estimated depending on how accurate the estimations were. The endogeneity test in this study
was done using the Durbin-Wu-Hausman Test of Endogeneity. The null hypothesis is based
on the fact that endogeneity does not exist. Table 4.10a shows that both the Durbin and Wu-
Hausman tests have p-values that are more than 0.05, which suggests that the null hypothesis
is not rejected. This implies that the model is free of endogeneity issues
Table 4.10a,
Durbin-Wu-Hausman Test of Endogeneity
Durbin (score) chi2(1) Wu-Hausman F(1,115)
Statistics 1.93248 1.88227
P-value 0.1645 0.1727
Source: Computed based on the data from the World Bank WDI (2025)

4.2.11a The Effects of Income on Life Expectancy

The regression results based on the feasible generalized least squares (FGLS)
estimation offer important insights into the socioeconomic determinants of life expectancy
(LLEB) across five selected African countries Egypt, Nigeria, Cameroon, Kenya, and South
Africa. Table 4.11a indicates that the overall model is statistically significant, as evidenced
by the Wald chi-square statistic (Wald chi2(5) = 146.88, p < 0.01), suggesting that the
explanatory variables collectively explain substantial variation in life expectancy across the
countries in the study.

Among the key predictors, GDP per capita (LGDPC) exhibits a positive and
statistically significant relationship with life expectancy, with a coefficient of 0.0120 (p =
0.010). This implies that a 1% increase in per capita income is associated with a 0.012%
improvement in life expectancy. This finding aligns with economic theory and previous
empirical evidence, which suggest that higher income levels improve access to healthcare,
education, nutrition, and other determinants of well-being that ultimately lead to longer lives.
For countries like Nigeria and Cameroon with lower GDP per capita relative to South Africa
or Egypt, policies that promote inclusive economic growth may be particularly impactful for
public health.

10
 Health expenditure as a percentage of GDP (CHE_GDP) is also positively and
significantly associated with life expectancy, with a coefficient of 0.0049 (p < 0.01). This
result indicates that increased public investment in healthcare contributes positively to health
outcomes. Specifically, a one-unit increase in CHE_GDP results in a 0.49% increase in life
expectancy. This underscores the critical role of sustained healthcare financing, especially in
countries like Kenya and Cameroon where healthcare infrastructure may still be
underdeveloped. Governments aiming to enhance life expectancy should prioritize health
sector budgetary allocations.

On the other hand, inflation (LCPI) shows a strong and statistically significant
negative effect on life expectancy, with a coefficient of -0.0393 (p < 0.001). This suggests
that rising price levels erode purchasing power and make essential goods and healthcare
services less affordable, ultimately reducing life expectancy. A 1% increase in inflation is
associated with a 0.039% decline in life expectancy. In volatile economic environments such
as Nigeria and Egypt, inflation control remains essential not only for economic stability but
also for public health resilience.

The real exchange rate (LREX) has a positive but statistically insignificant coefficient
(0.0071, p = 0.344), indicating no clear evidence of its direct influence on life expectancy in
the selected African countries. Although exchange rate movements may affect the cost of
imported medicines and medical equipment, the insignificant result suggests that other
mediating factors likely diminish this effect in the current analysis. Population growth rate
(PGR) also shows a negative but statistically insignificant relationship with life expectancy
(coefficient = -0.0069, p = 0.147). This suggests that while high population growth may
strain public services and healthcare delivery, the effect is not robust enough in this model to
establish a definitive link. Nonetheless, the negative sign is consistent with expectations and
may warrant further investigation, especially in countries like Nigeria and Kenya where
population growth is rapid.

Overall, income and health expenditure have positive and significant effect on life
expectancy, while inflation rate has negative and significant effect on life expectancy in the
long run. In addition, real exchange rate has positive and insignificant effect on life
expectancy, while population growth has negative and insignificant effect. These findings
underscore the importance of economic growth and increased public health expenditure in
improving life expectancy in African countries. Governments in the region should invest

11
more in the health sector while ensuring macroeconomic stability, particularly through
inflation control, to foster positive health outcomes. The insignificant effects of the real
exchange rate and population growth rate also suggest the need for broader structural reforms
and targeted health interventions to manage demographic challenges and economic volatility.

Table 4.11a,
Results on the Effect of Income on Life Expectancy
Lower Upper
LLEB Coef. Std. Err. z P>|z| limit Limit
LGDPC 0.0120 0.0047 2.56 0.010** 0.0028 0.0212

LCPI -0.0393 0.0058 -6.84 0.000*** -0.0506 0.0281

LREX 0.0071 0.0075 0.95 0.344 -0.0076 0.0219
CHE_GDP 0.0049 0.0010 4.81 0.000*** 0.0029 0.0069
PGR -0.0069 0.0047 -1.45 0.147 -0.0161 0.0024
Constant 3.7761 0.0411 91.85 0.000*** 3.6955 3.8567
Wald chi2(5) 146.88
Prob > chi2 0.0000
No of Obs. 120
Source: Computed by the Author based on the World Bank WDI (2025)
Note: P-value < 1%, <5%, <10% indicates significance at 1%, 5%, and 10% respectively

Part B: The model of Income and Infant Mortality

4.2.2b Correlation Analysis for Income and Infant Mortality

The correlation analysis presented in Table 4.2b shows that GDP per capita (LGDPC)
has a negative but weak association with Infant Mortality (LIMR), with a correlation
coefficient of -0.436. This suggests that increase in income is modestly linked to decrease in
infant mortality rate across the selected African countries. Similarly, real exchange rate
(REX) and health expenditure as a share of GDP (CHE_GDP) also show negative
relationships with infant mortality, with coefficients of -0.016 and -0.468 respectively,
implying that favourable exchange conditions and increased public health spending may

12
support reduction in infant death. Conversely, inflation rate (LCPI) and population growth
rate (PGR) exhibit negative and weak correlations with infant mortality, at 0.379 and 0.514
respectively. This indicates that rising prices and rapid population growth may place burden
on health systems, thereby reducing available resources and increase financial burden on
health care consumers . Importantly, all correlation coefficients fall well below the
conventional multicollinearity threshold of 0.8 (Gujarati & Porter, 2008; Hair et al., 2009).
This implies that the independent variables are not highly collinear, supporting the validity
and robustness of further regression analysis.

Table 4.2b,
Correlation Matrix for Income and Infant Mortality
LIMR LGDPC LCPI LREX CHE_GDP PGR
LIMR 1.000
LGDPC -0.436 1.000
LCPI 0.379 0.486 1.000
LREX -0.016 -0.057 0.397 1.000
CHE_GDP -0.468 0.762 0.067 -0.394 1.000
PGR 0.514 -0.786 -0.272 0.092 -0.784
Source: Computed by the Author based on the World Bank WDI (2025)

4.2.3b Test for multicollinearity (Variance Inflation and Tolerance Factor Test)

Based on the findings from the previous section, the Variance Inflation Factor (VIF)
and tolerance values offer a robust assessment of potential multicollinearity among the
independent variables in the regression model. The VIF identifies how much the variance of a
regression coefficient is inflated due to multicollinearity, while tolerance, the inverse of VIF
measures the proportion of variance in a predictor not explained by other predictors.
Tolerance values close to 1 indicate low multicollinearity, whereas values near 0 suggest high
multicollinearity. As a rule of thumb, a tolerance value above 0.1 and a VIF below 10 are
acceptable thresholds (Miles, 2014). As shown in Table 4.3a, all VIF values are below 10 and
tolerance values exceed 0.1, indicating that multicollinearity is not a concern in this study.

Table 4.3b,
Variance Inflation Factor (VIF) and Tolerance level for primary school attainment and
longevity Model
Variable VIF 1/VIF
CHE_GDP 5.06 0.197652

13
 LGDPC 4.92 0.203196
PGR 3.52 0.283805
LCPI 1.99 0.503041
LREX 1.63 0.615246
MeanVIF 3.42

4.2.4b Slope Heterogeneity Test for Income and Life Expectancy

Table 4.4b presents the results of the Pesaran and Yamagata (2008) slope
heterogeneity test examining the relationship between income and life expectancy. At the 1%
significance level, both the Delta and adjusted Delta statistics yield p-values of 0.000,
indicating strong statistical significance. Consequently, the null hypothesis of slope
homogeneity is rejected. This confirms the presence of slope heterogeneity across the panel
data, implying that the relationship between income and Infant Mortality varies across the
selected African countries.

Table 4.4b
Pesaran and Yamagata Slope Heterogeneity Test for Income and Infant Mortality

Pesaran and Yamagata Test statistics p-value

Delta 9.051 0.000
Delta adj 10.754 0.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.2.5b Cross-Sectional Dependency Test for Income and Infant Mortality

In addition to slope heterogeneity, testing for cross-sectional dependence (CSD) is a

critical step in panel data analysis. Cross-sectional dependence arises when individual units
within a panel, such as countries are not independent of each other due to shared shocks,
spillover effects, or regional integration. The presence of CSD can lead to biased and
inconsistent parameter estimates if not properly addressed. The cross-sectional dependence
(CD) test remains valid under both homogeneous and heterogeneous dynamic models, as well
as for non-stationary data structures. A p-value below 5% leads to the rejection of the null
hypothesis of no cross-sectional dependence. Table 4.5b indicates that the null hypothesis of
cross-sectional independence is rejected, confirming the presence of cross-sectional
dependence among panel units. This finding necessitates the use of second-generation panel
unit root and cointegration tests that account for cross-sectional dependence. Accordingly, the

14
study employs the Pesaran (2007) Cross-Sectionally Augmented IPS (CIPS) unit root test,
which adjusts for common factors across countries. It is noteworthy that all variables
demonstrate cross-sectional independence, except for the real exchange rate (LREX), which
exhibits significant interdependence across the selected African countries.
Table 4.5a,
Pesaran CD Cross-Sectional Dependency for Income and Infant Mortality
Variable CD-test p-value corr Abs(corr)
LIMR 14.86 0.000 0.959 0.959
LGDPC 12.82 0.000 0.827 0.827
LCPI 15.29 0.000 0.987 0.987
LREX -1.57 0.117 -0.101 0.336
CHE_GDP -2.64 0.008 -0.171 0.327
PGR 2.86 0.004 0.184 0.445
Source: Computed by the Author based on the World Bank WDI (2025)

4.2.6b Unit Root Test for Income and Infant Mortality

The null hypothesis of cross-sectional dependence was not denied thus, the cross-
sectionally augmented IPS (Im, Pesaran, and Shin) unit root test of Pesaran (2007) was
performed, and the data ranged from 1995 to 2023. Table 4.6b presents the results of unit root
tests for all variables using panel data. The null hypothesis of the presence of unit roots is not
rejected for all the variables when utilising the level form of the variables. Using the variables
in their first-differenced form, however, the null hypothesis is rejected. This means that the
variables have been integrated to the first order, indicating that each variable is a
nonstationary variable I(1). This result led to the use of panel data cointegration techniques.

Table 4.6b,
CIPS root tests (Income and Infant Mortality)
Cross-sectionally Augmented Im, Pesaran, and Shin (CIPS) test
Level Stat (critical First Difference Stat Remark
Variables level@5% (critical level@5%
LIMR -1.508(-2.33) -1.43(-2.33) I(1)
LCPI -1.270(-2.33) -2.655 (-2.33) I(1)
LGDPC -2.320(-2.33) -3.816 (-2.33) I(1)

15
 PGR -1.386(-2.33) -2.552 (-2.33) I(1)
LREX -1.997(-2.33) -4.426 (-2.33) I(1)
CHE_GDP -1.477(-2.33) -3.876 (-2.33) I(1)
Source: Computed by the Author based on the World Bank WDI (2025)

4.2.7b Panel data Cointegration Test for Income and Infant Mortality

The Westerlund Panel Cointegration test, which accounts for cross-sectional

dependence, was employed to examine the existence of a long-run relationship among the
variables. In line with Levin, Lin, and Chu (2002), the test was adapted to reflect cross-
sectional dependence. The test evaluates the null hypothesis of no cointegration against two
alternatives: that at least some panels are cointegrated, or that all panels are cointegrated. As
shown in Table 4.7b, the null hypothesis is rejected when testing for cointegration across all
panels, indicating a common long-term relationship among the variables. However, the null is
not rejected under the assumption that only some panels are cointegrated. Overall, the
rejection of the null for all panels confirms the presence of cointegration, suggesting that a
stable long-run relationship exists among the variables.
Table 4.7b,
Modified Westerlund Panel Cointegration Test for Income and Infant Mortality
Some panels are cointegrated All panels are cointegrated
statistics p-value statistics p-value
Variance ratio -2.3497 0.0005 --2.4133 0.0002
Source: Computed based on the data from the World Bank WDI (2025)

4.2.8b Heteroskedasticity Test for Income and Infant Mortality

In this study, heteroscedasticity was examined using a modified Wald test for group-
wise heteroscedasticity. The null hypothesis that the model does not contain
heteroskedasticity is to be rejected when the P-value is less than 5% or 0.05. Table 4.8b
demonstrates that the null hypothesis cannot be rejected because the probability value is less
than 5%. This suggests that the model contains heteroscedasticity.
Table 4.8a,
Modified Wald test for groupwise heteroskedasticity result for Income and Infant Mortality
Modified Wald test Statistics and P-value

16
 chi2 (5) 5566.81

Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.2.9b Testing for Serial Correlation for Income and Infant Mortality

Table 4.9b shows the result of the Breusch-Pagan test, which shows that the
probability value of the Chi-statistics is less than 1%. The null hypothesis that there is no
serial correlation should be rejected if the p-value is less than 5% or 0.05. From Table 4.9b,
the null hypothesis that there is no serial correlation is rejected because the p-value (0.000) is
less than 0.05. This suggests that the residuals of the panel data model suffer from serial
correlation. Hence, the model is best estimated using the feasible generalised least squares
(FGLS).
Table 4.9b,
Breusch-Pagan LM test of independence for Income and Infant Mortality
Breusch-Pagan test of independence Statistics and P-value
chi2(10) 115.066
Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.2.10b Testing for the Endogeneity for Income and Infant Mortality

When a regression model’s error term and the predictor variables are correlated,
endogeneity problems occur. If the endogeneity problem is disregarded, the results could be
skewed, with the true effect of a variable on the intended outcome being over- or under-
estimated depending on how accurate the estimations were. The endogeneity test in this study
was done using the Durbin-Wu-Hausman Test of Endogeneity. The null hypothesis is based
on the fact that endogeneity does not exist. Table 4.10b shows that both the Durbin and Wu-
Hausman tests have p-values that are more than 0.05, which suggests that the null hypothesis
is not rejected. This implies that the model is free of endogeneity issues
Table 4.10b,
Durbin-Wu-Hausman Test of Endogeneity for Income and Infant Mortality
Durbin (score) chi2(1) Wu-Hausman F(1,115)
Statistics 1.61868 1.5212

17
 P-value 0.1819 0.1720
Source: Computed based on the data from the World Bank WDI (2025)

4.2.11b The Effects of Income on Infant Mortality

The regression results based on the feasible generalized least squares (FGLS)
estimation offer important insights into the socioeconomic determinants of Infant Mortality
(LIMR) across five selected African countries Egypt, Nigeria, Cameroon, Kenya, and South
Africa. Table 4.11b indicates that the overall model is statistically significant, as evidenced
by the Wald chi-square statistic (Wald chi2(5) = 461.54, p < 0.01), suggesting that the
explanatory variables collectively explain substantial variation in infant mortality across the
countries in the study. Among the key predictors, GDP per capita (LGDPC) exhibits a
negative and statistically significant relationship with infant mortality rate, with a coefficient
of -0.1606 (p = 0.000). This implies that a 1% increase in per capita income is associated with
a 0.161% reduction in infant mortality rate. This finding suggests that higher income levels
improve access to healthcare, education, nutrition, and other determinants of well-being that
ultimately lead to reduction in infant death rate.
Health expenditure as a percentage of GDP (CHE_GDP) is also negatively and
significantly associated with infant mortality, with a coefficient of -0.0149 (p < 0.000). This
result indicates that increased public investment in healthcare contributes positively to health
outcomes. Specifically, a one-unit increase in CHE_GDP results in a -0.015% decrease in
infant mortality rate. This underscores the critical role of sustained healthcare financing,
especially in countries like Kenya and Cameroon where healthcare infrastructure may still be
underdeveloped. Governments aiming to reduce infant mortality should prioritize health
sector budgetary allocations.

Similarly, the real exchange rate (LREX) has a negative and significant coefficient (-0.0777,
p = 0.007), indicating that exchange rate movements may affect health outcomes.
Specifically, at 1% level of significance, a percentage increase in real exchange rate reduces
infant mortality rate by 0.078%, meaning that improvement in exchange rate reduces death
among the infant

Contrarily, inflation (LCPI) shows a strong and statistically significant positive effect
on infant mortality rate, with a coefficient of 0.2391 (p < 0.000). This suggests that rising
price levels erode purchasing power and make essential goods and healthcare services less

18
affordable, ultimately increasing infant mortality. A 1% increase in inflation is associated
with a 0.239% increase in infant death. In volatile economic environments such as Nigeria
and Egypt, inflation control remains essential not only for economic stability but also for
public health resilience. In the same vein, population growth rate (PGR) also shows positive
and significant relationship with Infant Mortality (coefficient = 0.0263, p = 0.059). This
suggests that high population growth may strain public services and healthcare delivery while
increasing pressure on available health care resource, thus increasing infant death rate. The
effect at 10% level of significance is not robust enough in this model to establish a definitive
link. Nonetheless, the positive sign is consistent with expectations.

Overall, income, real exchange rate and health expenditure have negative and
significant effect on infant mortality, while inflation rate and population growth have positive
and significant effect on infant mortality in the long run. These findings underscore the
importance of economic growth and increased public health expenditure in reducing infant
death in African countries. Governments in the region should invest more in the health sector
while ensuring macroeconomic stability, particularly through inflation and population growth
control, to foster positive health outcomes.

Table 4.11b,
Results on the Effect of Income on Infant Mortality
Lower Upper
LIMR Coef. Std. Err. z P>|z| limit Limit
LGDPC -0.1606 0.0161 -10.01 0.000*** -0.1921 -0.1292
LCPI 0.2391 0.0208 11.5 0.000*** -0.1984 0.2799
LREX -0.0777 0.0288 -2.7 0.007*** 0.1342 -0.0212
CHE_GDP -0.0149 0.0042 -3.53 0.000*** -0.0231 -0.0066
PGR 0.0263 0.0140 1.88 0.059* -0.0011 0.0537
Constant 5.8852 0.1467 40.12 0.000*** 5.5977 6.1728
Wald chi2(5) 461.54
Prob > chi2 0.0000
No of Obs. 120
Source: Computed by the Author based on the World Bank WDI (2025)
Note: P-value < 1%, <5%, <10% indicates significance at 1%, 5%, and 10% respectively

19
4.3 Effect of Education on Health Outcomes in selected African countries (objective 2)

This section presents the results of the effects of employment on health outcomes (life
expectancy and infant mortality) in selected African countries. The analysis starts with
descriptive statistics, which describe the characteristics of the variables of interest used in this
study. The correlation analysis, unit root test, slope heterogeneity, cross-sectional
dependence, and cointegration tests results were all presented as pre-estimation procedures to
the final results of the linear regression with feasible generalised least squares (FGLS).

Part A: The model of Education and Life Expectancy

4.3.1 Descriptive Statistics for Education and Health Life Expectancy

The descriptive statistics provide preliminary insights into the distribution and central
tendencies of key macroeconomic and health-related indicators across selected African
countries, Egypt, Nigeria, Cameroon, Kenya, and South Africa. The variables examined
include life expectancy at birth (LEB), infant mortality rate (IMR), employment to population
ratio (EDU), consumer price index (CPI), real exchange rate (REX), current health
expenditure as a percentage of GDP (CHE_GDP), household consumption expenditure
(CEX), and population growth rate (PGR). These indicators offer foundational understanding
of the development and public health dynamics within these countries during the study
period.

Life expectancy at birth (LEB) has a mean value of 60 years, indicating generally low
longevity outcomes across the sample, with a minimum of 47 years and a maximum of 71
years as shown in Table 4.12a. The standard deviation of 6.69 suggests moderate variability
among the countries. The positive skewness (0.18) implies a slight tendency towards higher
life expectancy, although the kurtosis of 2.00 indicates a relatively flat distribution compared
to a normal curve. The Jarque-Bera statistic (5.63, p = 0.06) suggests marginal non-
normality, which may reflect disparities in healthcare infrastructure and socioeconomic
factors among the countries.

Infant mortality rate (IMR) averages 50.53 deaths per 1,000 live births, a reflection of
ongoing public health challenges. The wide range from a minimum of 15.50 to a maximum
of 109.60 and high standard deviation (24.27) show substantial disparities in infant health
outcomes across the sample. The skewness value (0.56) suggests a right-skewed distribution

20
with some countries experiencing persistently high infant mortality. A kurtosis of 2.20
implies a flatter distribution, and the Jarque-Bera test (9.33, p = 0.01) confirms non-
normality, underscoring the need for targeted health interventions in countries like Nigeria
and Cameroon. Education (EDU) averages 44.50, indicating moderate educational outcomes
across the countries observed. The standard deviation of 21.95 and a range from 8.70 to 86.90
show a notable spread in educational access or attainment levels. The distribution is nearly
symmetric, as indicated by a skewness of 0.09, and the kurtosis of 2.04 suggests a shape close
to the normal distribution. The Jarque-Bera statistic (4.81, p = 0.09) does not reject normality
at the 5% level. This suggests that education is evenly distributed, providing a relatively
stable foundation for human capital development. This could be attributed to sustained efforts
in education reforms or global development agendas promoting universal access to schooling.

The consumer price index (CPI), which measures inflation, has a mean value of
131.92, ranging from 29.60 to 524.91. A standard deviation of 89.11 highlights volatility in
price levels, particularly in inflation-prone economies like Nigeria. The high skewness (2.01)
and kurtosis (7.81) suggest extreme outliers and leptokurtic behavior, reflecting periods of
macroeconomic instability. The Jarque-Bera statistic (196.21, p < 0.01) confirms significant
non-normality in inflation patterns, warranting caution in monetary policy formulation across
the countries. Real exchange rate (REX) averages 99.37, indicating relative currency
stability, although variation exists as shown by a standard deviation of 15.97. The skewness
(0.43) and kurtosis (3.29) are close to normal thresholds, but the Jarque-Bera statistic (4.10, p
= 0.13) does not indicate a significant departure from normality. This suggests that exchange
rate fluctuations were moderately stable in countries like Egypt and Kenya during the study
period. Current health expenditure as a percentage of GDP (CHE_GDP) averages 4.79%,
suggesting limited government prioritization of healthcare in public spending. The standard
deviation (1.63) indicates variation in commitment to health financing. The variable is right-
skewed (1.00) and moderately peaked (kurtosis = 2.87), with the Jarque-Bera test (20.11, p <
0.01) confirming non-normal distribution. Countries like South Africa and Kenya spend
relatively more, reflecting stronger institutional commitment to public health. Finally, the
population growth rate (PGR) averages 2.23%, reflecting youthful and expanding populations
in the region. The skewness of -0.58 indicates a slight leftward skew, suggesting that most
countries have growth rates above the mean. With a standard deviation of 0.64 and kurtosis
of 2.35, the distribution is relatively close to normal, although the Jarque-Bera statistic (8.81,

21
p = 0.01) implies some non-normality. High growth rates, especially in Nigeria and Kenya,
may exert pressure on healthcare systems and economic infrastructure if not matched with
adequate investment.

The descriptive results reveal deep-rooted structural disparities in health outcomes

and economic performance across the five selected African countries. Egypt and South Africa
generally perform better in terms of life expectancy and GDP per capita, likely due to more
advanced infrastructure and policy environments. Nigeria and Cameroon, however, exhibit
higher infant mortality rates and volatile macroeconomic indicators, suggesting the need for
comprehensive public health reforms and macroeconomic stabilization. Kenya, while
showing moderate performance across most variables, faces challenges related to population
growth and healthcare financing. Overall, the findings underscore the importance of
integrated economic and health policies tailored to the unique conditions of each country.
Overall, most of the variables demonstrate non-normal distributions with positive skewness
and high kurtosis, suggesting asymmetry. These distributional characteristics underscore the
importance of applying logarithmic transformations and robust estimation methods to ensure
more accurate and reliable econometric analysis.

Table 4.12a,
Summary Statistics for Education and Health Life Expectancy
LEB IMR EDU CPI REX CHE_GDP PGR
Mean 59.60 50.53 44.50 131.92 99.37 4.79 2.23
Median 59.95 47.50 45.40 110.00 100.00 4.34 2.26
Maximum 71.37 109.60 86.90 524.91 141.61 8.58 3.15
Minimum 47.19 15.50 8.70 29.60 69.83 2.49 0.65
Std. Dev. 6.69 24.27 21.95 89.11 15.97 1.63 0.64
Skewness 0.18 0.56 0.09 2.01 0.43 1.00 -0.58
Kurtosis 2.00 2.20 2.04 7.81 3.29 2.87 2.35
Jarque-Bera 5.63 9.33 4.81 196.21 4.10 20.11 8.81
Probability 0.06 0.01 0.09 0.00 0.13 0.00 0.01
Observations 120 119 120 120 120 120 120
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.2a Correlation Analysis for Education and Life Expectancy

The correlation analysis presented in Table 4.13a shows that education attainment
(LEDU) has a positive but weak association with life expectancy (LLEB), with a correlation

22
coefficient of 0.319. This suggests that higher level of education is modestly linked to longer
life expectancy across the selected African countries. Similarly, real exchange rate (REX) and
health expenditure as a share of GDP (CHE_GDP) also show positive relationships with life
expectancy, with coefficients of 0.070 and 0.349 respectively, implying that favourable
exchange conditions and increased public health spending may support better health
outcomes. Conversely, inflation rate (LCPI) and population growth rate (PGR) exhibit
negative and weak correlations with life expectancy, at -0.306 and -0.369 respectively. This
indicates that rising prices and rapid population growth may place pressure on health systems,
thereby limiting improvements in longevity. Importantly, all correlation coefficients fall well
below the conventional multicollinearity threshold of 0.8 (Gujarati & Porter, 2008; Hair et
al., 2009). This implies that the independent variables are not highly collinear, supporting the
validity and robustness of further regression analysis.

Table 4.13a,
Correlation Matrix for Education and Life Expectancy
LGDP CHE_G
LLEB C LCPI LREX DP PGR
LLEB 1.000
LEDU 0.319 1.000
LCPI -0.306 0.268 1.000
LREX 0.070 -0.010 0.397 1.000
CHE_GDP 0.349 0.276 0.067 -0.394 1.000
PGR -0.369 -0.491 -0.272 0.092 -0.784 1.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.3a Test for multicollinearity (Variance Inflation and Tolerance Factor Test)

Based on the findings from the previous section, the Variance Inflation Factor (VIF)
and tolerance values offer a robust assessment of potential multicollinearity among the
independent variables in the regression model. The VIF identifies how much the variance of a
regression coefficient is inflated due to multicollinearity, while tolerance, the inverse of VIF
measures the proportion of variance in a predictor not explained by other predictors.
Tolerance values close to 1 indicate low multicollinearity, whereas values near 0 suggest high
multicollinearity. As a rule of thumb, a tolerance value above 0.1 and a VIF below 10 are
acceptable thresholds (Miles, 2014). As shown in Table 4.14a, all VIF values are below 10

23
and tolerance values exceed 0.1, indicating that multicollinearity is not a concern in this
study.

Table 4.14a,
Variance Inflation Factor (VIF) and Tolerance level for Education and Life Expectancy
Variable VIF 1/VIF
PGR 3.90 0.256497
CHE_GDP 3.74 0.26723
LREX 1.65 0.607274
LEDU 1.42 0.705372
LCPI 1.37 0.727307
Mean VIF 2.42

4.3.4 a Slope Heterogeneity Test for Education and Life Expectancy

Table 4.15a presents the results of the Pesaran and Yamagata (2008) slope
heterogeneity test examining the relationship between income and life expectancy. At the 1%
significance level, both the Delta and adjusted Delta statistics yield p-values of 0.000,
indicating strong statistical significance. Consequently, the null hypothesis of slope
homogeneity is rejected. This confirms the presence of slope heterogeneity across the panel
data, implying that the relationship between income and life expectancy varies across the
selected African countries.

Table 4.15a,
Pesaran and Yamagata Slope Heterogeneity Test for Education and Life Expectancy

Pesaran and Yamagata Test statistics p-value

Delta 5.322 0.000
Delta adj 6.324 0.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.5a Cross-Sectional Dependency Test for Education and Life Expectancy

In addition to slope heterogeneity, testing for cross-sectional dependence (CSD) is a

critical step in panel data analysis. Cross-sectional dependence arises when individual units
within a panel, such as countries are not independent of each other due to shared shocks,
spillover effects, or regional integration. The presence of CSD can lead to biased and
inconsistent parameter estimates if not properly addressed. The cross-sectional dependence

24
(CD) test remains valid under both homogeneous and heterogeneous dynamic models, as well
as for non-stationary data structures. A p-value below 5% leads to the rejection of the null
hypothesis of no cross-sectional dependence. Table 4.16a indicates that the null hypothesis of
cross-sectional independence is rejected, confirming the presence of cross-sectional
dependence among panel units. This finding necessitates the use of second-generation panel
unit root and cointegration tests that account for cross-sectional dependence. Accordingly, the
study employs the Pesaran (2007) Cross-Sectionally Augmented IPS (CIPS) unit root test,
which adjusts for common factors across countries. It is noteworthy that all variables
demonstrate cross-sectional independence, except for the real exchange rate (LREX), which
exhibits significant interdependence across the selected African countries.
Table 4.16a,
Pesaran CD Cross-Sectional Dependency Test for Education and Life Expectancy
Variable CD-test p-value corr Abs(corr)
LLEB 14.06 0.000 0.908 0.908
LEDU 14.94 0.000 0.964 0.964
LCPI 15.29 0.000 0.987 0.987
LREX -1.57 0.117 -0.101 0.336
CHE_GDP -2.64 0.008 -0.171 0.327
PGR 2.86 0.004 0.184 0.445
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.6a Unit Root Test for Education and Life Expectancy

The null hypothesis of cross-sectional dependence was not denied, thus, the cross-
sectionally augmented IPS (Im, Pesaran, and Shin) unit root test of Pesaran (2007) was
performed, and the data ranged from 1995 to 2023. Table 4.17a presents the results of unit
root tests for all variables using panel data. The null hypothesis of the presence of unit roots
is not rejected for all the variables when utilising the level form of the variables. Using the
variables in their first-differenced form, however, the null hypothesis is rejected. This means
that the variables have been integrated to the first order, indicating that each variable is a
nonstationary variable I(1). This result led to the use of panel data cointegration techniques.

25
Table 4.17a,
CIPS root tests (Education and Life Expectancy)
Cross-sectionally Augmented Im, Pesaran, and Shin (CIPS) test
Level Stat (critical First Difference Stat Remark
Variables level@5% (critical level@5%
LLEB -2.022(-2.33) -4.097(-2.33) I(1)
LCPI -1.270(-2.33) -2.655 (-2.33) I(1)
LEDU -0.752(-2.33) -2.712(-2.33) I(1)
PGR -1.386(-2.33) -2.552 (-2.33) I(1)
LREX -1.997(-2.33) -4.426 (-2.33) I(1)
CHE_GDP -1.477(-2.33) -3.876 (-2.33) I(1)
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.7a Panel data Cointegration Test for Education and Life Expectancy

The Westerlund Panel Cointegration test, which accounts for cross-sectional

dependence, was employed to examine the existence of a long-run relationship among the
variables. In line with Levin, Lin, and Chu (2002), the test was adapted to reflect cross-
sectional dependence. The test evaluates the null hypothesis of no cointegration against two
alternatives: that at least some panels are cointegrated, or that all panels are cointegrated. As
shown in Table 4.18a, the null hypothesis is rejected when testing for cointegration across all
panels, indicating a common long-term relationship among the variables. However, the null is
not rejected under the assumption that only some panels are cointegrated. Overall, the
rejection of the null for all panels confirms the presence of cointegration, suggesting that a
stable long-run relationship exists among the variables.
Table 4.18a,
Modified Westerlund Panel Cointegration Test for Education and Life Expectancy
Some panels are cointegrated All panels are cointegrated
statistics p-value statistics p-value
Variance ratio -1.8191 0.0327 -1.6777 0. 0467
Source: Computed based on the data from the World Bank WDI (2025)

26
4.3.8a Heteroskedasticity Test for Education and Life Expectancy
In this study, heteroscedasticity was examined using a modified Wald test for group-
wise heteroscedasticity. The null hypothesis that the model does not contain
heteroskedasticity is to be rejected when the P-value is less than 5% or 0.05. Table 4.19a
demonstrates that the null hypothesis cannot be rejected because the probability value is less
than 5%. This suggests that the model contains heteroscedasticity.
Table 4.19a,
Modified Wald test for groupwise heteroskedasticity result for Education and Life
Expectancy
Modified Wald test Statistics and P-value
chi2 (5) 24451.19

Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.3.9a Testing for Serial Correlation for Education and Life Expectancy
Table 4.20a shows the result of the Breusch-Pagan test, which shows that the
probability value of the Chi-statistics is less than 1%. The null hypothesis that there is no
serial correlation should be rejected if the p-value is less than 5% or 0.05. From Table 4.20a,
the null hypothesis that there is no serial correlation is rejected because the p-value (0.000) is
less than 0.05. This suggests that the residuals of the panel data model suffer from serial
correlation. Hence, the model is best estimated using the feasible generalised least squares
(FGLS).

Table 4.20a

Breusch-Pagan LM test of independence for Education and Life Expectancy

Breusch-Pagan test of independence Statistics and P-value

chi2(10) 86.440
Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.3.10a Testing for the Endogeneity for Education and Life Expectancy
When a regression model’s error term and the predictor variables are correlated,
endogeneity problems occur. If the endogeneity problem is disregarded, the results could be
skewed, with the true effect of a variable on the intended outcome being over- or under-

27
estimated depending on how accurate the estimations were. The endogeneity test in this study
was done using the Durbin-Wu-Hausman Test of Endogeneity. The null hypothesis is based
on the fact that endogeneity does not exist. Table 4.21a shows that both the Durbin and Wu-
Hausman tests have p-values that are more than 0.05, which suggests that the null hypothesis
is not rejected. This implies that the model is free of endogeneity issues
Table 4.21a,
Durbin-Wu-Hausman Test of Endogeneity
Durbin (score) chi2(1) Wu-Hausman F(1,115)
Statistics 0.053325 0.051126
P-value 0.8174 0.8215
Source: Computed based on the data from the World Bank WDI (2025)

4.3.11a The Effects of Education on Life Expectancy

The regression results based on the feasible generalized least squares (FGLS)
estimation offer important insights into the socioeconomic determinants of life expectancy
(LLEB) across five selected African countries Egypt, Nigeria, Cameroon, Kenya, and South
Africa. Table 4.22a indicates that the overall model is statistically significant, as evidenced
by the Wald chi-square statistic (Wald chi2(5) = 156.90, p = 0.000), suggesting that the
explanatory variables collectively explain substantial variation in life expectancy across the
countries in the study. Among the key predictors, educational attainment (LEDU) exhibits a
positive and statistically significant relationship with life expectancy, with a coefficient of
0.1324 (p = 0.000). This implies that a 1% increase in educational level is associated with a
0.132% improvement in life expectancy. This finding aligns with economic theory and
previous empirical evidence, which suggest that higher educational levels improve access to
healthcare, income, nutrition, and other determinants of well-being that ultimately lead to
longer lives.

Health expenditure as a percentage of GDP (CHE_GDP) is also positively and

significantly associated with life expectancy, with a coefficient of 0.0018 (p = 0.000). This
result indicates that increased public investment in healthcare contributes positively to health
outcomes. Specifically, a one-unit increase in CHE_GDP results in a 0.18% increase in life
expectancy. This underscores the critical role of sustained healthcare financing, especially in
countries like Kenya and Cameroon where healthcare infrastructure may still be
underdeveloped. Governments aiming to enhance life expectancy should prioritize health

28
sector budgetary allocations. Likewise, the real exchange rate (LREX) has a positive but
statistically insignificant coefficient (0.0011, p = 0.036), indicating that improvement in
exchange rate lead to reduction in the cost of imported medicines and medical equipment
thereby increasing life span in the selected African countries.

Also, inflation (LCPI) shows a strong and statistically significant negative effect on
life expectancy, with a coefficient of -0.0169 (p = 0.001). This suggests that rising price
levels erode purchasing power and make essential goods and healthcare services less
affordable, ultimately reducing life expectancy. A 1% increase in inflation is associated with
a 0.017% decline in life expectancy. In volatile economic environments such as Nigeria and
Egypt, inflation control remains essential not only for economic stability but also for public
health resilience. Population growth rate (PGR) also shows a positive but statistically
insignificant relationship with life expectancy (coefficient = 0.0065, p = 0.207). The positive
sign is inconsistent with expectations and may warrant further investigation, especially in
countries like Nigeria and Kenya where population growth is rapid.

Overall, education, real exchange rate and health expenditure have positive and
significant effect on life expectancy, while inflation rate has negative and significant effect on
life expectancy in the long run. In addition, population growth has positive and insignificant
effect on life expectancy. These findings underscore the importance of education and
increased public health expenditure in improving life expectancy in African countries.

Table 4.22a,
Results on the Effect of Education on Life Expectancy
Lower Upper
LLEB Coef. Std. Err. z P>|z| limit Limit
LEDU 0.1324 0.0140 9.45 0.000*** 0.1049 0.1598
LCPI -0.0169 0.0051 -3.34 0.001*** -0.0269 0.0070
LREX 0.0114 0.0054 2.09 0.036** 0.0007 0.0220
CHE_GDP 0.0018 0.0010 1.88 0.06* -0.0001 0.0037
PGR 0.0065 0.0051 1.26 0.207 -0.0036 0.0166
Constant 3.4689 0.0579 59.91 0.000*** 3.3554 3.5824
Wald chi2(5) 156.90
Prob > chi2 0.0000

29
 No of Obs. 120
Source: Computed by the Author based on the World Bank WDI (2025)
Note: P-value < 1%, <5%, <10% indicates significance at 1%, 5%, and 10% respectively

Part B: The model of Education and Infant Mortality

4.3.2b Correlation Analysis for Education and Infant Mortality

The correlation analysis presented in Table 4.13b shows that educational attainment
(LEDU) has a negative but weak association with Infant Mortality (LIMR), with a correlation
coefficient of -0.451. This suggests that higher education is modestly linked to decrease in
infant mortality rate across the selected African countries. Similarly, real exchange rate
(REX) and health expenditure as a share of GDP (CHE_GDP) also show negative
relationships with infant mortality, with coefficients of -0.016 and -0.468 respectively,
implying that favourable exchange conditions and increased public health spending may
support reduction in infant death. Conversely, inflation rate (LCPI) and population growth
rate (PGR) exhibit negative and weak correlations with infant mortality, at 0.379 and 0.514
respectively. This indicates that rising prices and rapid population growth may place burden
on health systems, thereby reducing available resources and increase financial burden on
health care consumers . Importantly, all correlation coefficients fall well below the
conventional multicollinearity threshold of 0.8 (Gujarati & Porter, 2008; Hair et al., 2009).
This implies that the independent variables are not highly collinear, supporting the validity
and robustness of further regression analysis.

Table 4.13b,
Correlation Matrix for Education and Infant Mortality
LIMR LGDPC LCPI LREX CHE_GDP PGR
LIMR 1.000
LEDU -0.451 1.000
LCPI 0.379 0.486 1.000
LREX -0.016 -0.057 0.397 1.000
CHE_GDP -0.468 0.762 0.067 -0.394 1.000
PGR 0.514 -0.786 -0.272 0.092 -0.784
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.3b Test for multicollinearity (Variance Inflation and Tolerance Factor Test)

30
 Based on the findings from the previous section, the Variance Inflation Factor (VIF)
and tolerance values offer a robust assessment of potential multicollinearity among the
independent variables in the regression model. The VIF identifies how much the variance of a
regression coefficient is inflated due to multicollinearity, while tolerance, the inverse of VIF
measures the proportion of variance in a predictor not explained by other predictors.
Tolerance values close to 1 indicate low multicollinearity, whereas values near 0 suggest high
multicollinearity. As a rule of thumb, a tolerance value above 0.1 and a VIF below 10 are
acceptable thresholds (Miles, 2014). As shown in Table 4.14b, all VIF values are below 10
and tolerance values exceed 0.1, indicating that multicollinearity is not a concern in this
study.

Table 4.14b,
Variance Inflation Factor (VIF) and Tolerance level for primary school attainment and
longevity Model
Variable VIF 1/VIF
PGR 3.90 0.256497
CHE_GDP 3.74 0.26723
LREX 1.65 0.607274
LEDU 1.42 0.705372
LCPI 1.37 0.727307
Mean VIF 2.42

4.3.4b Slope Heterogeneity Test for Education and Life Expectancy

Table 4.15b presents the results of the Pesaran and Yamagata (2008) slope
heterogeneity test examining the relationship between income and life expectancy. At the 1%
significance level, both the Delta and adjusted Delta statistics yield p-values of 0.000,
indicating strong statistical significance. Consequently, the null hypothesis of slope
homogeneity is rejected. This confirms the presence of slope heterogeneity across the panel
data, implying that the relationship between income and Infant Mortality varies across the
selected African countries.

Table 4.1b,
Pesaran and Yamagata Slope Heterogeneity Test for Education and Infant Mortality

Pesaran and Yamagata Test statistics p-value

Delta 9.319 0.000

31
 Delta adj 11.072 0.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.5b Cross-Sectional Dependency Test for Education and Infant Mortality

In addition to slope heterogeneity, testing for cross-sectional dependence (CSD) is a

critical step in panel data analysis. Cross-sectional dependence arises when individual units
within a panel, such as countries are not independent of each other due to shared shocks,
spillover effects, or regional integration. The presence of CSD can lead to biased and
inconsistent parameter estimates if not properly addressed. The cross-sectional dependence
(CD) test remains valid under both homogeneous and heterogeneous dynamic models, as well
as for non-stationary data structures. A p-value below 5% leads to the rejection of the null
hypothesis of no cross-sectional dependence. Table 4.16b indicates that the null hypothesis of
cross-sectional independence is rejected, confirming the presence of cross-sectional
dependence among panel units. This finding necessitates the use of second-generation panel
unit root and cointegration tests that account for cross-sectional dependence. Accordingly, the
study employs the Pesaran (2007) Cross-Sectionally Augmented IPS (CIPS) unit root test,
which adjusts for common factors across countries. It is noteworthy that all variables
demonstrate cross-sectional independence, except for the real exchange rate (LREX), which
exhibits significant interdependence across the selected African countries.
Table 4.16b,
Pesaran CD Cross-Sectional Dependency for Education and Infant Mortality
Variable CD-test p-value corr Abs(corr)
LIMR 14.86 0.000 0.959 0.959
LEDU 14.94 0.000 0.964 0.964
LCPI 15.29 0.000 0.987 0.987
LREX -1.57 0.117 -0.101 0.336
CHE_GDP -2.64 0.008 -0.171 0.327
PGR 2.86 0.004 0.184 0.445
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.6b Unit Root Test for Education and Infant Mortality

32
 The null hypothesis of cross-sectional dependence was not denied, thus the cross-
sectionally augmented IPS (Im, Pesaran, and Shin) unit root test of Pesaran (2007) was
performed, and the data ranged from 1995 to 2023. Table 4.17b presents the results of unit
root tests for all variables using panel data. The null hypothesis of the presence of unit roots
is not rejected for all the variables when utilising the level form of the variables. Using the
variables in their first-differenced form, however, the null hypothesis is rejected. This means
that the variables have been integrated to the first order, indicating that each variable is a
nonstationary variable I(1). This result led to the use of panel data cointegration techniques.
Table 4.17b,
CIPS root tests (Education and Infant Mortality)
Cross-sectionally Augmented Im, Pesaran, and Shin (CIPS) test
Level Stat (critical First Difference Stat Remark
Variables level@5% (critical level@5%
LIMR -1.508(-2.33) -1.43(-2.33) I(1)
LCPI -1.270(-2.33) -2.655 (-2.33) I(1)
LEDU -0.752(-2.33) -2.712(-2.33) I(1)
PGR -1.386(-2.33) -2.552 (-2.33) I(1)
LREX -1.997(-2.33) -4.426 (-2.33) I(1)
CHE_GDP -1.477(-2.33) -3.876 (-2.33) I(1)
Source: Computed by the Author based on the World Bank WDI (2025)

4.3.7b Panel data Cointegration Test for Education and Infant Mortality

The Westerlund Panel Cointegration test, which accounts for cross-sectional

dependence, was employed to examine the existence of a long-run relationship among the
variables. In line with Levin, Lin, and Chu (2002), the test was adapted to reflect cross-
sectional dependence. The test evaluates the null hypothesis of no cointegration against two
alternatives: that at least some panels are cointegrated, or that all panels are cointegrated. As
shown in Table 4.18b, the null hypothesis is rejected when testing for cointegration across all
panels, indicating a common long-term relationship among the variables. However, the null is
not rejected under the assumption that only some panels are cointegrated. Overall, the
rejection of the null for all panels confirms the presence of cointegration, suggesting that a
stable long-run relationship exists among the variables.

33
Table 4.18b,
Modified Westerlund Panel Cointegration Test for Education and Infant Mortality
Some panels are cointegrated All panels are cointegrated
statistics p-value statistics p-value
Variance ratio --1.7419 0.0467 --1.6911 0.0480
Source: Computed based on the data from the World Bank WDI (2025)

4.3.8b Heteroskedasticity Test for Education and Infant Mortality

In this study, heteroscedasticity was examined using a modified Wald test for group-
wise heteroscedasticity. The null hypothesis that the model does not contain
heteroskedasticity is to be rejected when the P-value is less than 5% or 0.05. Table 4.19b
demonstrates that the null hypothesis cannot be rejected because the probability value is less
than 5%. This suggests that the model contains heteroscedasticity.
Table 4.19b
Modified Wald test for groupwise heteroskedasticity result for Education and Infant
Mortality
Modified Wald test Statistics and P-value
chi2 (5) 30000.00

Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.3.9b Testing for Serial Correlation for Education and Infant Mortality

Table 4.20b shows the result of the Breusch-Pagan test, which shows that the
probability value of the Chi-statistics is less than 1%. The null hypothesis that there is no
serial correlation should be rejected if the p-value is less than 5% or 0.05. From Table 4.19b,
the null hypothesis that there is no serial correlation is rejected because the p-value (0.000) is
less than 0.05. This suggests that the residuals of the panel data model suffer from serial
correlation. Hence, the model is best estimated using the feasible generalised least squares
(FGLS).

Table 4.20b,
Breusch-Pagan LM test of independence for Education and Infant Mortality

34
 Breusch-Pagan test of independence Statistics and P-value
chi2(10) 121.240
Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.3.10b Testing for the Endogeneity for Education and Infant Mortality

When a regression model’s error term and the predictor variables are correlated,
endogeneity problems occur. If the endogeneity problem is disregarded, the results could be
skewed, with the true effect of a variable on the intended outcome being over- or under-
estimated depending on how accurate the estimations were. The endogeneity test in this study
was done using the Durbin-Wu-Hausman Test of Endogeneity. The null hypothesis is based
on the fact that endogeneity does not exist. Table 4.21b shows that both the Durbin and Wu-
Hausman tests have p-values that are more than 0.05, which suggests that the null hypothesis
is not rejected. This implies that the model is free of endogeneity issues
Table 4.21b,
Durbin-Wu-Hausman Test of Endogeneity for Education and Infant Mortality
Durbin (score) chi2(1) Wu-Hausman F(1,115)
Statistics 1.7704 1.3393
P-value 0.1310 0.1710
Source: Computed based on the data from the World Bank WDI (2025)

4.3.11b The Effects of Education on Infant Mortality

The regression results based on the feasible generalized least squares (FGLS)
estimation offer important insights into the socioeconomic determinants of Infant Mortality
(LIMR) across five selected African countries Egypt, Nigeria, Cameroon, Kenya, and South
Africa. Table 4.22b indicates that the overall model is statistically significant, as evidenced
by the Wald chi-square statistic (Wald chi2(5) = 695.71, p = 0.000), suggesting that the
explanatory variables collectively explain substantial variation in infant mortality across the
countries in the [Link] the key predictors, education (LEDU) exhibits a negative and
statistically significant relationship with infant mortality rate, with a coefficient of - -0.5048
(p = 0.000). This implies that a 1% increase in educational level is associated with a 0.505%
reduction in infant mortality rate. This finding suggests that higher education improve access

35
to healthcare, income, nutrition, and other determinants of well-being that ultimately lead to
reduction in infant death rate. Similarly, the real exchange rate (LREX) has a negative and
significant coefficient (-0.0512, p = 0.002), indicating that exchange rate movements may
affect health outcomes. Specifically, at 1% level of significance, a percentage increase in real
exchange rate reduces infant mortality rate by 0.078%, meaning that improvement in
exchange rate reduces death among the infant
Contrarily, inflation (LCPI) shows a strong and statistically significant positive effect
on infant mortality rate, with a coefficient of 0.1993 (p = 0.000). This suggests that rising
price levels erode purchasing power and make essential goods and healthcare services less
affordable, ultimately increasing infant mortality. A 1% increase in inflation is associated
with a 0.199% increase in infant death. In volatile economic environments such as Nigeria
and Egypt, inflation control remains essential not only for economic stability but also for
public health resilience. In addition, health expenditure as a percentage of GDP (CHE_GDP)
and population growth are also negatively and insignificantly associated with infant
mortality. This analysis suggests that high population growth may strain public services and
healthcare delivery while increasing pressure on available health care resource, thus
increasing infant death rate. Also, increased public investment in healthcare contributes
positively to health outcomes.

Overall, education and real exchange rate have negative and significant effect on
infant mortality, while inflation rate has positive and significant effect on infant mortality in
the long run. In addition, health expenditure and population growth have negative and
insignificant effect on infant mortality in the selected African countries. Governments in the
region should invest more in the health sector while ensuring high population control to foster
positive health outcomes.

Table 4.22b,
Results on the Effect of Education on Infant Mortality
LIMR Coef. Std. Err. z P>|z| Lower Upper

36
 limit Limit
LEDU -0.5048 0.0277 -18.21 0.000*** -0.5591 -0.4505

LCPI 0.1993 0.0164 12.14 0.000*** -0.1672 0.2315

LREX -0.0512 0.0168 -3.04 0.002*** -0.0841 -0.0182
CHE_GDP -0.0049 0.0031 -1.59 0.113 -0.0109 0.0012
PGR -0.0075 0.0113 -0.67 0.504 -0.0297 0.0146
_cons 6.6839 0.1348 49.57 0.000*** 6.4197 6.9482
Wald chi2(5) 695.71
Prob > chi2 0.0000
No of Obs. 120
Source: Computed by the Author based on the World Bank WDI (2025)
Note: P-value < 1%, <5%, <10% indicates significance at 1%, 5%, and 10% respectively

4.4 Effect of Employment on Health Outcomes in selected African countries (objective

3)

This section presents the results of the effects of employment on health outcomes (life
expectancy and infant mortality) in selected African countries. The analysis starts with
descriptive statistics, which describe the characteristics of the variables of interest used in this
study. The correlation analysis, unit root test, slope heterogeneity, cross-sectional
dependence, and cointegration tests results were all presented as pre-estimation procedures to
the final results of the linear regression with feasible generalised least squares (FGLS).

Part A: The model of Employment and Life Expectancy

4.4.1a Descriptive Statistics for Employment and Health Outcomes

The descriptive statistics provide preliminary insights into the distribution and central
tendencies of key macroeconomic and health-related indicators across selected African
countries, Egypt, Nigeria, Cameroon, Kenya, and South Africa. The variables examined
include life expectancy at birth (LEB), infant mortality rate (IMR), employment to population
ratio (EDU), consumer price index (CPI), real exchange rate (REX), current health
expenditure as a percentage of GDP (CHE_GDP), household consumption expenditure
(CEX), and population growth rate (PGR). These indicators offer foundational understanding

37
of the development and public health dynamics within these countries during the study
period.

Life expectancy at birth (LEB) has a mean value of 60 years, indicating generally low
longevity outcomes across the sample, with a minimum of 47 years and a maximum of 71
years as shown in Table 4.23a. The standard deviation of 6.69 suggests moderate variability
among the countries. The positive skewness (0.18) implies a slight tendency towards higher
life expectancy, although the kurtosis of 2.00 indicates a relatively flat distribution compared
to a normal curve. The Jarque-Bera statistic (5.63, p = 0.06) suggests marginal non-
normality, which may reflect disparities in healthcare infrastructure and socioeconomic
factors among the countries. Infant mortality rate (IMR) averages 50.53 deaths per 1,000 live
births, a reflection of ongoing public health challenges. The wide range from a minimum of
15.50 to a maximum of 109.60 and high standard deviation (24.27) show substantial
disparities in infant health outcomes across the sample. The skewness value (0.56) suggests a
right-skewed distribution with some countries experiencing persistently high infant mortality.
A kurtosis of 2.20 implies a flatter distribution, and the Jarque-Bera test (9.33, p = 0.01)
confirms non-normality, underscoring the need for targeted health interventions in countries
like Nigeria and Cameroon.

Employment-to-population ratio (EMP) averages 60.81%, indicating that roughly 6

out of 10 working-age individuals are employed across the selected African countries. The
relatively high standard deviation (15.88), along with a range stretching from 36.80 to 80.29,
points to significant disparities in labor market absorption across countries and periods. The
negative skewness (-0.25) suggests a slightly left-skewed distribution, implying that some
countries experience lower-than-average employment levels. A kurtosis of 1.32 reflects a
flatter-than-normal distribution, and the Jarque-Bera statistic (15.36, p = 0.00) confirms a
significant departure from normality, indicating heterogeneity in employment dynamics that
may require tailored labor market reforms. Household consumption expenditure (CEX)
records a high average of 138.98, though with a wide dispersion as reflected in the high
standard deviation (117.34), and values ranging from 7.40 to 410.00. The positive skewness
(0.55) suggests that while most observations are clustered at lower consumption levels, a few
countries or periods exhibit extremely high household spending. A kurtosis value of 1.96
implies a relatively flat distribution with light tails, and the Jarque-Bera test (11.46, p = 0.00)
confirms non-normality. These results reflect pronounced inequalities in household welfare

38
and living standards across the sample. Employment and consumption are more unevenly
distributed across the sample, reflecting underlying disparities in labor market opportunities
and household welfare.

The consumer price index (CPI), which measures inflation, has a mean value of
131.92, ranging from 29.60 to 524.91. A standard deviation of 89.11 highlights volatility in
price levels, particularly in inflation-prone economies like Nigeria. The high skewness (2.01)
and kurtosis (7.81) suggest extreme outliers and leptokurtic behavior, reflecting periods of
macroeconomic instability. The Jarque-Bera statistic (196.21, p < 0.01) confirms significant
non-normality in inflation patterns, warranting caution in monetary policy formulation across
the countries. Real exchange rate (REX) averages 99.37, indicating relative currency
stability, although variation exists as shown by a standard deviation of 15.97. The skewness
(0.43) and kurtosis (3.29) are close to normal thresholds, but the Jarque-Bera statistic (4.10, p
= 0.13) does not indicate a significant departure from normality. This suggests that exchange
rate fluctuations were moderately stable in countries like Egypt and Kenya during the study
period. Current health expenditure as a percentage of GDP (CHE_GDP) averages 4.79%,
suggesting limited government prioritization of healthcare in public spending. The standard
deviation (1.63) indicates variation in commitment to health financing. The variable is right-
skewed (1.00) and moderately peaked (kurtosis = 2.87), with the Jarque-Bera test (20.11, p <
0.01) confirming non-normal distribution. Countries like South Africa and Kenya spend
relatively more, reflecting stronger institutional commitment to public health.

Finally, the population growth rate (PGR) averages 2.23%, reflecting youthful and
expanding populations in the region. The skewness of -0.58 indicates a slight leftward skew,
suggesting that most countries have growth rates above the mean. With a standard deviation
of 0.64 and kurtosis of 2.35, the distribution is relatively close to normal, although the
Jarque-Bera statistic (8.81, p = 0.01) implies some non-normality. High growth rates,
especially in Nigeria and Kenya, may exert pressure on healthcare systems and economic
infrastructure if not matched with adequate investment.

The descriptive results reveal deep-rooted structural disparities in health outcomes

and economic performance across the five selected African countries. Egypt and South Africa
generally perform better in terms of life expectancy and GDP per capita, likely due to more
advanced infrastructure and policy environments. Nigeria and Cameroon, however, exhibit
higher infant mortality rates and volatile macroeconomic indicators, suggesting the need for

39
comprehensive public health reforms and macroeconomic stabilization. Kenya, while
showing moderate performance across most variables, faces challenges related to population
growth and healthcare financing. Overall, the findings underscore the importance of
integrated economic and health policies tailored to the unique conditions of each country.
Overall, most of the variables demonstrate non-normal distributions with positive skewness
and high kurtosis, suggesting asymmetry. These distributional characteristics underscore the
importance of applying logarithmic transformations and robust estimation methods to ensure
more accurate and reliable econometric analysis.

Table 4.23a,
Summary Statistics for Employment and Health Life Expectancy
CHE_GD
LEB IMR EMP CEX CPI REX PGR
P
Mean 59.60 50.53 60.81 138.98 131.92 99.37 4.79 2.23
Median 59.95 47.50 66.97 86.00 110.00 100.00 4.34 2.26
Maximum 71.37 109.60 80.29 410.00 524.91 141.61 8.58 3.15
Minimum 47.19 15.50 36.80 7.40 29.60 69.83 2.49 0.65
Std. Dev. 6.69 24.27 15.88 117.34 89.11 15.97 1.63 0.64
Skewness 0.18 0.56 -0.25 0.55 2.01 0.43 1.00 -0.58
Kurtosis 2.00 2.20 1.32 1.96 7.81 3.29 2.87 2.35
Jarque-Bera 5.63 9.33 15.36 11.46 196.21 4.10 20.11 8.81
Probability 0.06 0.01 0.00 0.00 0.00 0.13 0.00 0.01
Observation
120 119 120 120 120 120 120 120
s
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.2a Correlation Analysis for Employment and Life Expectancy

The correlation analysis in Table 4.24a shows that employment (EMP) has a strong
positive association with life expectancy (LLEB), with a correlation coefficient of 0.780. This
suggests that higher employment rates are closely linked to longer life expectancy, likely
reflecting the role of income and job-related benefits in improving access to healthcare and
overall well-being. Household consumption expenditure (LCEX) shows a weak positive
correlation with life expectancy (0.202), indicating that increased household spending may be
modestly associated with improved living conditions and health outcomes. Health
expenditure as a percentage of GDP (CHE_GDP) displays a moderate positive relationship

40
with life expectancy (0.349), reinforcing the importance of public health investment in
enhancing longevity.

In contrast, inflation (LCPI) and population growth rate (PGR) exhibit negative correlations
with life expectancy, at -0.306 and -0.369 respectively, implying that economic instability
and demographic pressures may strain healthcare systems and reduce life expectancy. The
real exchange rate (LREX) has a weak positive correlation with life expectancy (0.070),
suggesting a limited direct influence on health outcomes. Importantly, all correlation
coefficients are well below the multicollinearity threshold of 0.8 (Gujarati & Porter, 2008;
Hair et al., 2009), indicating the absence of serious multicollinearity among the variables.
This supports the reliability of these independent variables for subsequent regression analysis.

Table 4.24a,
Correlation Matrix for Employment and Life Expectancy
CHE_G
LLEB EMP LCEX LCPI LREX DP PGR
LLEB 1.000
EMP 0.780 1.000
LCEX 0.202 -0.458 1.000
LCPI -0.306 -0.123 0.536 1.000
LREX 0.070 0.254 0.055 0.397 1.000
CHE_G
DP 0.349 -0.719 0.441 0.067 -0.394 1.000
PGR -0.369 0.752 -0.621 -0.272 0.092 -0.784 1.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.3a Test for multicollinearity (Variance Inflation and Tolerance Factor Test)

Based on the findings from the previous section, the Variance Inflation Factor (VIF)
and tolerance values offer a robust assessment of potential multicollinearity among the
independent variables in the regression model. The VIF indicates how much the variance of a
regression coefficient is inflated due to collinearity, while tolerance (the inverse of VIF)
measures the proportion of variance in a predictor not explained by other predictors.
Tolerance values closer to 1 suggest low multicollinearity, whereas values near 0 indicate
potential problems. According to the standard threshold, VIF values below 10 and tolerance
values above 0.1 are considered acceptable (Miles, 2014). As shown in Table 4.25a, all

41
variables have VIF values well below 10, ranging from 1.66 (LREX) to 4.50 (PGR), and
corresponding tolerance values all exceed 0.1, ranging from 0.222 to 0.603. The mean VIF is
2.73, further supporting the absence of severe multicollinearity. These results confirm that the
explanatory variables used in the model are statistically independent of each other to a
satisfactory degree, enhancing the reliability of subsequent regression estimates.

Table 4.25a,
Variance Inflation Factor (VIF) and Tolerance level for Employment and Life Expectancy

Variable VIF 1/VIF

PGR 4.5 0.222152
CHE_GDP 3.73 0.267887
EMP 2.62 0.382274
LCEX 2.15 0.465319
LCPI 1.73 0.577029
LREX 1.66 0.603315
Mean VIF 2.73

4.4.4a Slope Heterogeneity Test for Employment and Life Expectancy

Table 4.26a presents the results of the Pesaran and Yamagata (2008) slope
heterogeneity test examining the relationship between income and life expectancy. At the 1%
significance level, both the Delta and adjusted Delta statistics yield p-values of 0.000,
indicating strong statistical significance. Consequently, the null hypothesis of slope
homogeneity is rejected. This confirms the presence of slope heterogeneity across the panel
data, implying that the relationship between income and life expectancy varies across the
selected African countries.

Table 4.26a,
Pesaran and Yamagata Slope Heterogeneity Test for Employment and Life Expectancy

Pesaran and Yamagata Test statistics p-value

Delta 8.503 0.000
Delta adj 10.103 0.000
Source: Computed by the Author based on the World Bank WDI (2025)

42
4.4.5a Cross-Sectional Dependency Test for Employment and Life Expectancy

In addition to slope heterogeneity, testing for cross-sectional dependence (CSD) is a

critical step in panel data analysis. Cross-sectional dependence arises when individual units
within a panel, such as countries are not independent of each other due to shared shocks,
spillover effects, or regional integration. The presence of CSD can lead to biased and
inconsistent parameter estimates if not properly addressed. The cross-sectional dependence
(CD) test remains valid under both homogeneous and heterogeneous dynamic models, as well
as for non-stationary data structures. A p-value below 5% leads to the rejection of the null
hypothesis of no cross-sectional dependence. Table 4.27a indicates that the null hypothesis of
cross-sectional independence is rejected, confirming the presence of cross-sectional
dependence among panel units. This finding necessitates the use of second-generation panel
unit root and cointegration tests that account for cross-sectional dependence. Accordingly, the
study employs the Pesaran (2007) Cross-Sectionally Augmented IPS (CIPS) unit root test,
which adjusts for common factors across countries. It is noteworthy that all variables
demonstrate cross-sectional independence, except for the real exchange rate (LREX), which
exhibits significant interdependence across the selected African countries.
Table 4.27a,
Pesaran CD Cross-Sectional Dependency Test for Employment and Life Expectancy
Variable CD-test p-value corr Abs(corr)
LLEB 14.06 0.000 0.908 0.908
EMP 8.58 0.000 0.554 0.554
LCEX 13.85 0.000 0.894 0.894
LCPI 15.29 0.000 0.987 0.987
LREX -1.57 0.117 -0.101 0.336
CHE_GDP -2.64 0.008 -0.171 0.327
PGR 2.86 0.004 0.184 0.445
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.6a Unit Root Test for Employment and Life Expectancy

The null hypothesis of cross-sectional dependence was not denied thus, the cross-
sectionally augmented IPS (Im, Pesaran, and Shin) unit root test of Pesaran (2007) was
performed, and the data ranged from 1995 to 2023. Table 4.28a presents the results of unit
root tests for all variables using panel data. The null hypothesis of the presence of unit roots

43
is not rejected for all the variables when utilising the level form of the variables. Using the
variables in their first-differenced form, however, the null hypothesis is rejected. This means
that the variables have been integrated to the first order, indicating that each variable is a
nonstationary variable I(1). This result led to the use of panel data cointegration techniques.
Table 4.28a,
CIPS root tests (Employment and Life Expectancy)
Cross-sectionally Augmented Im, Pesaran, and Shin (CIPS) test
Level Stat (critical First Difference Stat Remark
Variables level@5% (critical level@5%
LLEB -2.022(-2.33) -4.097(-2.33) I(1)
LCPI -1.270(-2.33) -2.655 (-2.33) I(1)
EMP -1.521 (-2.33) -2.834(-2.33) I(1)
LCEX -1.745 (2.33) -3.754(-2.33) I(1)
PGR -1.386(-2.33) -2.552 (-2.33) I(1)
LREX -1.997(-2.33) -4.426 (-2.33) I(1)
CHE_GDP -1.477(-2.33) -3.876 (-2.33) I(1)
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.7a Panel data Cointegration Test for Employment and Life Expectancy

The Westerlund Panel Cointegration test, which accounts for cross-sectional

dependence, was employed to examine the existence of a long-run relationship among the
variables. In line with Levin, Lin, and Chu (2002), the test was adapted to reflect cross-
sectional dependence. The test evaluates the null hypothesis of no cointegration against two
alternatives: that at least some panels are cointegrated, or that all panels are cointegrated. As
shown in Table 4.29a, the null hypothesis is rejected when testing for cointegration across all
panels, indicating a common long-term relationship among the variables. However, the null is
not rejected under the assumption that only some panels are cointegrated. Overall, the
rejection of the null for all panels confirms the presence of cointegration, suggesting that a
stable long-run relationship exists among the variables.

Table 4.19a,
Modified Westerlund Panel Cointegration Test for Employment and Life Expectancy

44
 Some panels are cointegrated All panels are cointegrated
statistics p-value statistics p-value
Variance ratio 1.7087 0.0457 1.6206 0. 0477
Source: Computed based on the data from the World Bank WDI (2025)

4.4.8a Heteroskedasticity Test for Employment and Life Expectancy

In this study, heteroscedasticity was examined using a modified Wald test for group-
wise heteroscedasticity. The null hypothesis that the model does not contain
heteroskedasticity is to be rejected when the P-value is less than 5% or 0.05. Table 4.30a
demonstrates that the null hypothesis cannot be rejected because the probability value is less
than 5%. This suggests that the model contains heteroscedasticity.

Table 4.30a,
Modified Wald test for groupwise heteroskedasticity result for Employment and Life
Expectancy
Modified Wald test Statistics and P-value
chi2 (5) 1127.57

Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.4.9a Testing for Serial Correlation for Employment and Life Expectancy
Table 4.31a shows the result of the Breusch-Pagan test, which shows that the
probability value of the Chi-statistics is less than 1%. The null hypothesis that there is no
serial correlation should be rejected if the p-value is less than 5% or 0.05. From Table 4.31a,
the null hypothesis that there is no serial correlation is rejected because the p-value (0.000) is
less than 0.05. This suggests that the residuals of the panel data model suffer from serial
correlation. Hence, the model is best estimated using the feasible generalised least squares
(FGLS).

Table 4.31a,
Breusch-Pagan LM test of independence for Employment and Life Expectancy
Breusch-Pagan test of independence Statistics and P-value
chi2(10) 62.753

45
 Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.4.10a Testing for the Endogeneity for Employment and Life Expectancy
When a regression model’s error term and the predictor variables are correlated,
endogeneity problems occur. If the endogeneity problem is disregarded, the results could be
skewed, with the true effect of a variable on the intended outcome being over- or under-
estimated depending on how accurate the estimations were. The endogeneity test in this study
was done using the Durbin-Wu-Hausman Test of Endogeneity. The null hypothesis is based
on the fact that endogeneity does not exist. Table 4.32a shows that both the Durbin and Wu-
Hausman tests have p-values that are more than 0.05, which suggests that the null hypothesis
is not rejected. This implies that the model is free of endogeneity issues

Table 4.32a,
Durbin-Wu-Hausman Test of Endogeneity
Durbin (score) chi2(1) Wu-Hausman F(1,115)
Statistics 2.51357 2.46037
P-value 0.1129 0.1195
Source: Computed based on the data from the World Bank WDI (2025)

4.3.11a The Effects of Employment on Life Expectancy

The regression results based on the feasible generalized least squares (FGLS)
estimation offer important insights into the socioeconomic determinants of life expectancy
(LLEB) across five selected African countries Egypt, Nigeria, Cameroon, Kenya, and South
Africa. Table 4.33a indicates that the overall model is statistically significant, as evidenced
by the Wald chi-square statistic (Wald chi2(5) = 594.66, p = 0.000), suggesting that the
explanatory variables collectively explain substantial variation in life expectancy across the
countries in the study. Among the key predictors, employment (EMP) exhibits a positive and
statistically significant relationship with life expectancy, with a coefficient of 0.0207 (p =
0.000). This implies that a 1% unit in employment is associated with a 2.07% improvement in
life expectancy. This finding aligns with economic theory and previous empirical evidence,
which suggest that job opportunities improve access to healthcare, income, nutrition, and
other determinants of well-being that ultimately lead to longer lives. Similarly, household

46
consumption expenditure (LCEX) is also positively and significantly associated with life
expectancy, with a coefficient of 0.0207 (p = 0.000). This result indicates that increased
Household consumption contributes positively to health outcomes. Specifically, a percentage
increase in LCEX results in a 2.07% increase in life expectancy. This suggest that increased
household spending on nutritious food is beneficial to their health, thus prlong their life.
Likewise, the real exchange rate (LREX) has a positive and statistically significant
coefficient (0.0300, p = 0.008), indicating that improvement in exchange rate lead to
reduction in the cost of imported medicines and medical equipment thereby increasing life
span in the selected African countries.

Also, inflation (LCPI) shows a strong and statistically significant negative effect on
life expectancy, with a coefficient of -0.0512 (p = 0.001). This suggests that rising price
levels erode purchasing power and make essential goods and healthcare services less
affordable, ultimately reducing life expectancy. A 1% increase in inflation is associated with
a 0.051% decline in life expectancy. In volatile economic environments such as Nigeria and
Egypt, inflation control remains essential not only for economic stability but also for public
health resilience. Similarly, population growth rate (PGR) also shows a negative and
statistically significant relationship with life expectancy (coefficient = -0.0238, p = 0.001).
This suggest that overpopulation is detrimental to people’s health and reduce lifespan,
especially in countries like Nigeria and Kenya where population growth is rapid.

Overall, employment, consumption expenditure and real exchange rate have positive
and significant effect on life expectancy, while inflation rate and population growth have
negative and significant effect on life expectancy in the long run. These findings underscore
the importance of employment opportunities and increased household consumption
expenditure in improving life expectancy in African countries.

47
Table 4.33a,
Results on the Effect of Employment on Life Expectancy

Lower Upper
LLEB Coef. Std. Err. z P>|z| limit Limit
EMP 0.0207 0.0018 11.3 0.000*** 0.0171 0.0243
LCEX 0.0207 0.0037 5.67 0.000*** -0.0135 0.0279
LCPI -0.0512 0.0043 -11.91 0.000*** -0.0597 0.0428
LREX 0.0301 0.0113 2.66 0.008*** 0.0079 0.0522
CHE_GDP -0.0018 0.0018 -1.02 0.307 -0.0053 0.0017
PGR -0.0238 0.0075 -3.19 0.001*** -0.0385 0.0092
Constant 4.6043 0.0944 48.78 0.000*** 4.4193 4.7893
Wald chi2(6) 594.66
Prob > chi2 0.0000
No of Obs. 120
Source: Computed by the Author based on the World Bank WDI (2025)
Note: P-value < 1%, <5%, <10% indicates significance at 1%, 5%, and 10% respectively

Part B: The model of Employment and Infant Mortality

4.4.2b Correlation Analysis for Employment and Infant Mortality

The correlation analysis presented in Table 4.24b shows that employment (EMP) has
a strong negative association with infant mortality rate (LIMR), with a correlation coefficient
of -0.724. This suggests that higher employment levels are strongly linked to reductions in
infant mortality, possibly due to improved household income and access to health-related
goods and services. Similarly, household consumption expenditure (LCEX) and health
expenditure as a share of GDP (CHE_GDP) also show negative relationships with infant
mortality, with coefficients of -0.354 and -0.468 respectively, indicating that increased
consumption and greater public investment in health are associated with better infant health
outcomes. In contrast, inflation rate (LCPI) and population growth rate (PGR) exhibit
positive correlations with infant mortality, at 0.379 and 0.514 respectively. This implies that
rising price levels and rapid population growth may strain healthcare systems, reduce the
quality of maternal and child health services, and increase financial burdens on households,
thereby contributing to higher infant mortality rates. The real exchange rate (LREX) shows a

48
weak negative correlation with infant mortality (-0.016), indicating minimal direct impact.
Importantly, all correlation coefficients are well below the conventional multicollinearity
threshold of 0.8 (Gujarati & Porter, 2008; Hair et al., 2009). This implies that the independent
variables are not highly collinear, supporting the validity and robustness of subsequent
regression analysis.

Table 4.24b,
Correlation Matrix for Employment and Infant Mortality
CHE_G
LIMR EMP LCEX LCPI LREX DP PGR
LIMR 1.000
EMP -0.824 1.000
LCEX -0.354 -0.458 1.000
LCPI 0.379 -0.123 0.536 1.000
LREX -0.016 0.254 0.055 0.397 1.000
CHE_G
DP -0.468 -0.719 0.441 0.067 -0.394 1.000
PGR 0.514 0.752 -0.621 -0.272 0.092 -0.784 1.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.3b Test for multicollinearity (Variance Inflation and Tolerance Factor Test)

Based on the findings from the previous section, the Variance Inflation Factor (VIF)
and tolerance values offer a robust assessment of potential multicollinearity among the
independent variables in the regression model. The VIF identifies how much the variance of a
regression coefficient is inflated due to multicollinearity, while tolerance, the inverse of VIF
measures the proportion of variance in a predictor not explained by other predictors.
Tolerance values close to 1 indicate low multicollinearity, whereas values near 0 suggest high
multicollinearity. As a rule of thumb, a tolerance value above 0.1 and a VIF below 10 are
acceptable thresholds (Miles, 2014). As shown in Table 4.25b, all VIF values are below 10
and tolerance values exceed 0.1, indicating that multicollinearity is not a concern in this
study.

49
Table 4.25b,
Variance Inflation Factor (VIF) and Tolerance level for primary school attainment and
longevity Model

Variable VIF 1/VIF

PGR 4.5 0.222152
CHE_GDP 3.73 0.267887
EMP 2.62 0.382274
LCEX 2.15 0.465319
LCPI 1.73 0.577029
LREX 1.66 0.603315
Mean VIF 2.73
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.4b Slope Heterogeneity Test for Employment and Life Expectancy

Table 4.26b presents the results of the Pesaran and Yamagata (2008) slope
heterogeneity test examining the relationship between income and life expectancy. At the 1%
significance level, both the Delta and adjusted Delta statistics yield p-values of 0.000,
indicating strong statistical significance. Consequently, the null hypothesis of slope
homogeneity is rejected. This confirms the presence of slope heterogeneity across the panel
data, implying that the relationship between income and Infant Mortality varies across the
selected African countries.

Table 4.26b,
Pesaran and Yamagata Slope Heterogeneity Test for Employment and Infant Mortality

Pesaran and Yamagata Test statistics p-value

Delta 8.979 0.000
Delta adj 10.669 0.000
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.5b Cross-Sectional Dependency Test for Employment and Infant Mortality

In addition to slope heterogeneity, testing for cross-sectional dependence (CSD) is a

critical step in panel data analysis. Cross-sectional dependence arises when individual units
within a panel, such as countries are not independent of each other due to shared shocks,

50
spillover effects, or regional integration. The presence of CSD can lead to biased and
inconsistent parameter estimates if not properly addressed. The cross-sectional dependence
(CD) test remains valid under both homogeneous and heterogeneous dynamic models, as well
as for non-stationary data structures. A p-value below 5% leads to the rejection of the null
hypothesis of no cross-sectional dependence. Table 4.27b indicates that the null hypothesis of
cross-sectional independence is rejected, confirming the presence of cross-sectional
dependence among panel units. This finding necessitates the use of second-generation panel
unit root and cointegration tests that account for cross-sectional dependence. Accordingly, the
study employs the Pesaran (2007) Cross-Sectionally Augmented IPS (CIPS) unit root test,
which adjusts for common factors across countries. It is noteworthy that all variables
demonstrate cross-sectional independence, except for the real exchange rate (LREX), which
exhibits significant interdependence across the selected African countries.

Table 4.27b,
Pesaran CD Cross-Sectional Dependency for Employment and Infant Mortality
Variable CD-test p-value corr Abs(corr)
LIMR 14.86 0.000 0.959 0.959
EMP 8.58 0.000 0.554 0.554
LCEX 13.85 0.000 0.894 0.894
LCPI 15.29 0.000 0.987 0.987
LREX -1.57 0.117 -0.101 0.336
CHE_GDP -2.64 0.008 -0.171 0.327
PGR 2.86 0.004 0.184 0.445
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.6b Unit Root Test for Employment and Infant Mortality

The null hypothesis of cross-sectional dependence was not denied thus, the cross-
sectionally augmented IPS (Im, Pesaran, and Shin) unit root test of Pesaran (2007) was
performed, and the data ranged from 1995 to 2023. Table 4.28b presents the results of unit

51
root tests for all variables using panel data. The null hypothesis of the presence of unit roots
is not rejected for all the variables when utilising the level form of the variables. Using the
variables in their first-differenced form, however, the null hypothesis is rejected. This means
that the variables have been integrated to the first order, indicating that each variable is a
nonstationary variable I(1). This result led to the use of panel data cointegration techniques.
Table 4.28b,
CIPS root tests (Employment and Infant Mortality)
Cross-sectionally Augmented Im, Pesaran, and Shin (CIPS) test
Level Stat (critical First Difference Stat Remark
Variables level@5% (critical level@5%
LIMR -1.508(-2.33) -1.43(-2.33) I(1)
LCPI -1.270(-2.33) -2.655 (-2.33) I(1)
LEDU -0.752(-2.33) -2.712(-2.33) I(1)
PGR -1.386(-2.33) -2.552 (-2.33) I(1)
LREX -1.997(-2.33) -4.426 (-2.33) I(1)
CHE_GDP -1.477(-2.33) -3.876 (-2.33) I(1)
Source: Computed by the Author based on the World Bank WDI (2025)

4.4.7b Panel data Cointegration Test for Employment and Infant Mortality

The Westerlund Panel Cointegration test, which accounts for cross-sectional

dependence, was employed to examine the existence of a long-run relationship among the
variables. In line with Levin, Lin, and Chu (2002), the test was adapted to reflect cross-
sectional dependence. The test evaluates the null hypothesis of no cointegration against two
alternatives: that at least some panels are cointegrated, or that all panels are cointegrated. As
shown in Table 4.29b, the null hypothesis is rejected when testing for cointegration across all
panels, indicating a common long-term relationship among the variables. However, the null is
not rejected under the assumption that only some panels are cointegrated. Overall, the
rejection of the null for all panels confirms the presence of cointegration, suggesting that a
stable long-run relationship exists among the variables.
Table 4.29b,
Modified Westerlund Panel Cointegration Test for Employment and Infant Mortality
Some panels are cointegrated All panels are cointegrated

52
 statistics p-value statistics p-value
Variance ratio --1.7419 0.0467 --1.6911 0.0480
Source: Computed based on the data from the World Bank WDI (2025)

4.4.8b Heteroskedasticity Test for Employment and Infant Mortality

In this study, heteroscedasticity was examined using a modified Wald test for group-
wise heteroscedasticity. The null hypothesis that the model does not contain
heteroskedasticity is to be rejected when the P-value is less than 5% or 0.05. Table 4.30b
demonstrates that the null hypothesis cannot be rejected because the probability value is less
than 5%. This suggests that the model contains heteroscedasticity.
Table 4.30b,
Modified Wald test for groupwise heteroskedasticity result for Employment and Infant
Mortality
Modified Wald test Statistics and P-value
chi2 (5) 30000.00

Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

4.4.9b Testing for Serial Correlation for Employment and Infant Mortality

Table 4.31b shows the result of the Breusch-Pagan test, which shows that the
probability value of the Chi-statistics is less than 1%. The null hypothesis that there is no
serial correlation should be rejected if the p-value is less than 5% or 0.05. From Table 4.31b,
the null hypothesis that there is no serial correlation is rejected because the p-value (0.000) is
less than 0.05. This suggests that the residuals of the panel data model suffer from serial
correlation. Hence, the model is best estimated using the feasible generalised least squares
(FGLS).
Table 4.31b,
Breusch-Pagan LM test of independence for Employment and Infant Mortality
Breusch-Pagan test of independence Statistics and P-value
chi2(10) 121.240
Prob>chi2 0.0000
Source: Computed based on the data from the World Bank WDI (2025)

53
4.4.10b Testing for the Endogeneity for Employment and Infant Mortality

When a regression model’s error term and the predictor variables are correlated,
endogeneity problems occur. If the endogeneity problem is disregarded, the results could be
skewed, with the true effect of a variable on the intended outcome being over- or under-
estimated depending on how accurate the estimations were. The endogeneity test in this study
was done using the Durbin-Wu-Hausman Test of Endogeneity. The null hypothesis is based
on the fact that endogeneity does not exist. Table 4.32b shows that both the Durbin and Wu-
Hausman tests have p-values that are more than 0.05, which suggests that the null hypothesis
is not rejected. This implies that the model is free of endogeneity issues
Table 4.32b,
Durbin-Wu-Hausman Test of Endogeneity for Employment and Infant Mortality
Durbin (score) chi2(1) Wu-Hausman F(1,115)
Statistics 1.7704 1.3393
P-value 0.1310 0.1710
Source: Computed based on the data from the World Bank WDI (2025)

4.4.11b The Effects of Employment on Infant Mortality

The regression results based on the feasible generalized least squares (FGLS)
estimation offer important insights into the socioeconomic determinants of Infant Mortality
(LIMR) across five selected African countries Egypt, Nigeria, Cameroon, Kenya, and South
Africa. Table 4.33b indicates that the overall model is statistically significant, as evidenced
by the Wald chi-square statistic (Wald chi2(5) = 648.88, p = 0.000), suggesting that the
explanatory variables collectively explain substantial variation in infant mortality across the
countries in the study. Among the key predictors, employment (EMP) exhibits a negative
and statistically significant relationship with infant mortality rate, with a coefficient of -
0.0194 (p = 0.000). This implies that a 1% increase employment is associated with a 0.019%
reduction in infant mortality rate. This finding suggests that employment improve access to
healthcare, income, nutrition, and other determinants of well-being that ultimately lead to
reduction in infant death rate. Similarly, household consumption expenditure (LCEX) has a
negative and significant coefficient (-0.0603, p = 0.001), indicating that exchange rate
movements may affect health outcomes. Specifically, at 1% level of significance, a

54
percentage increase in household consumption expenditure reduces infant mortality rate by
0.06%, meaning that improvement in exchange rate reduces death among the infant
Contrarily, inflation (LCPI) shows a strong and statistically significant positive effect
on infant mortality rate, with a coefficient of 0.2345 (p = 0.000). This suggests that rising
price levels erode purchasing power and make essential goods and healthcare services less
affordable, ultimately increasing infant mortality. A 1% increase in inflation is associated
with a 0.235% increase in infant death. In volatile economic environments such as Nigeria
and Egypt, inflation control remains essential not only for economic stability but also for
public health resilience. In addition, health expenditure as a percentage of GDP (CHE_GDP),
exchange rate and population growth have insignificantly impact on infant mortality.

Overall, employment and household consumption spending have negative and

significant effect on infant mortality, while inflation rate has positive and significant effect on
infant mortality in the long run. In addition, health expenditure, real exchange rate and
population growth have insignificant effect on infant mortality in the selected African
countries.

Table 4.33b,
Results on the Effect of Employment on Infant Mortality
Lower Upper
LIMR Coef. Std. Err. z P>|z| limit Limit
EMP -0.0194 0.0017 -11.43 0.000*** -0.0227 0.0160
-
LCEX -0.0603 0.0188 -3.21 0.001*** -0.0970 0.0235
LCPI 0.2345 0.0200 11.74 0.000*** -0.1953 0.2736
LREX 0.0213 0.0364 0.59 0.558 -0.0500 0.0927
CHE_GDP 0.0008 0.0058 0.14 0.893 -0.0107 0.0122
PGR -0.0208 0.0186 -1.12 0.264 -0.0572 0.0157
Constant 5.2994 0.4277 12.39 0.000*** 4.4611 6.1376
Wald chi2(5) 648.88

55
 Prob > chi2 0.0000
No of Obs. 120
Source: Computed by the Author based on the World Bank WDI (2025)
Note: P-value < 1%, <5%, <10% indicates significance at 1%, 5%, and 10% respectively

4.5 Joint Effect of Income, Education and Employment on Health Outcomes (objective 4)

This section presents the results of the joint effect of income, education and
employment on health outcomes (life expectancy and infant mortality) in selected African
countries. The analysis starts with the multicollinearity as pre-estimation procedure to the
final results of the linear regression with feasible generalised least squares (FGLS).

4.5.1 Test for multicollinearity (Variance Inflation and Tolerance Factor Test)

Based on the findings from the previous section, the Variance Inflation Factor (VIF)
and tolerance values offer a robust assessment of potential multicollinearity among the
independent variables in the regression model. The VIF indicates how much the variance of a
regression coefficient is inflated due to collinearity, while tolerance (the inverse of VIF)
measures the proportion of variance in a predictor not explained by other predictors.
Tolerance values closer to 1 suggest low multicollinearity, whereas values near 0 indicate
potential problems. According to the standard threshold, VIF values below 10 and tolerance
values above 0.1 are considered acceptable (Miles, 2014). As shown in Table 4.34, all
variables have VIF values well below 10, ranging from 1.66 (LREX) to 4.50 (PGR), and
corresponding tolerance values all exceed 0.1, ranging from 0.222 to 0.603. The mean VIF is
2.73, further supporting the absence of severe multicollinearity. These results confirm that the
explanatory variables used in the model are statistically independent of each other to a
satisfactory degree, enhancing the reliability of subsequent regression estimates.

Table 4.34,
Variance Inflation Factor (VIF) and Tolerance level for Income, Education, Employment and
Life Expectancy
Variabl
e VIF 1/VIF
LEDU 7.93 0.126103
LGDPC 2.21 0.452489
ED_GD 8.07 0.123916

56
 ED_EM 6.02 0.166113
EM_GD 4.26 0.234742
EMP 5.88 0.170068
CHE_G
DP 7.42 0.134771
PGR 4.73 0.211416
LREX 2.9 0.344828
LCPI 2.2 0.454545
Mean
VIF 5.815 0.171969
Source: Computed by the Author based on the World Bank WDI (2025)

4.5.2 The Joint Effects of Income, Education, Employment and Life Expectancy

The regression results based on the feasible generalized least squares (FGLS)
estimation provide valuable insights into the determinants of life expectancy (LLEB) across
the selected African countries. In Table 4.35, the overall model is statistically significant, as
indicated by the Wald chi-square statistic (Wald chi2(10) = 650.55, p = 0.000), suggesting
that the explanatory variables jointly explain a significant portion of the variation in life
expectancy. GDP per capita (LGDPC) is positively associated with life expectancy (coef. =
0.2070, p = 0.002), indicating that higher income levels contribute to longer life spans,
possibly through better living standards and access to healthcare. Moreover, educational
attainment (LEDU) has a strong and significant positive impact on life expectancy (coef. =
0.5536, p < 0.001). This underscores the crucial role of education in improving health
awareness, healthcare access, and general well-being. Employment (EMP) has a positive but
statistically insignificant effect on life expectancy, with a coefficient of 0.2018 (p = 0.169).
While the sign implies that increased employment may enhance longevity, the lack of
statistical significance suggests that the relationship is not robust across the sample.

The joint effect reveal that education-GDP interaction (ED_GD) is positively and
significantly associated with life expectancy (coef. = 0.2140, p = 0.016), suggesting that the
impact of education on longevity is stronger in higher-income settings. Likewise, education-
employment interaction (ED_EM) is also positive and significant (coef. = 0.2860, p = 0.005),
indicating that the health benefits of education are enhanced when employment levels are
higher. In the same vein, employment-GDP interaction (EM_GD) shows a significant
positive effect (coef. = 0.2603, p = 0.002), highlighting that the effect of employment on
health outcomes improves with income growth. The results indicate that while employment

57
alone does not have a statistically significant direct effect on life expectancy, its impact
becomes more pronounced in combination with education and income. Thus, multi-
dimensional policy approaches that simultaneously promote education, income growth, and
employment quality are critical to improving health outcomes.

Furthermore, real exchange rate (LREX) has a positive and significant impact on life
expectancy (coef. = 0.0324, p = 0.001), possibly reflecting improved affordability of
imported medical goods or stability in economic fundamentals. Public health expenditure
(CHE_GDP) is positively but not significantly related to life expectancy (coef. = 0.0015, p =
0.106), suggesting a weak or context-specific effect of health spending on longevity.
However, inflation (LCPI) is negatively and significantly related to life expectancy (coef. = -
0.0405, p < 0.001), suggesting that macroeconomic instability may erode household
purchasing power and reduce access to essential health services. Similarly, population growth
rate (PGR) is negatively and significantly associated with life expectancy (coef. = -0.0241, p
< 0.001), implying that high population pressures may dilute healthcare resources and
constrain public service delivery. Thus, addressing inflation and managing population growth
will be essential to sustain gains in longevity.

Table 4.35,
Results on the Joint Effect of Income, Education, Employment on Life Expectancy
Lower Upper
LLEB Coef. Std. Err. z P>|z| limit Limit
LGDPC 0.2070 0.0654 3.16 0.002*** -0.0788 0.3351
LEDU 0.5536 0.1375 4.03 0.000*** 0.2841 0.8232
EMP 0.2018 0.1466 1.38 0.169 -0.0855 0.4891
ED_GD 0.2140 0.0958 2.23 0.016** 0.1655 0.2526
ED_EM 0.2860 0.1008 2.84 0.005*** 0.2297 0.3424
EM_GD 0.2603 0.0904 2.88 0.004*** 0.2340 0.3865
LCPI -0.0405 0.0053 -7.64 0.000*** 0.0508 -0.0301
LREX 0.0324 0.0097 3.34 0.001*** 0.0134 0.0514
CHE_GDP 0.0015 0.0009 1.62 0.106 -0.0003 0.0034
PGR -0.0241 0.0066 -3.67 0.000*** 0.0369 -0.0112
Constant 3.9104 0.6913 5.66 0.000*** 2.5555 5.2653
Wald chi2(10) 650.55

58
 Prob > chi2 0.0000
No of Obs. 120
Source: Computed by the Author based on the World Bank WDI (2025)
Note: P-value < 1%, <5%, <10% indicates significance at 1%, 5%, and 10% respectively

4.5.3 The Joint Effects of Income, Education, Employment on Infant Mortality

The regression results using feasible generalized least squares (FGLS) estimation
provide insight into the drivers of infant mortality (LIMR) across selected African countries.
Table 4.36 reveal that the model is statistically significant overall, as indicated by the Wald
chi-square statistic (χ² = 818.46, p = 0.000), implying that the explanatory variables
collectively explain a significant portion of the variation in infant mortality. GDP per capita
(LGDPC) has a statistically significant negative effect on infant mortality (coef. = -0.042, p =
0.000). This suggests that as income per capita increases, infant mortality declines, likely due
to better nutrition, healthcare access, and living standards. Similarly, educational attainment
(LEDU) also shows a negative but statistically insignificant relationship with infant mortality
(coef. = -0.626, p = 0.254). Although the direction is as expected, the result indicates that
education alone may not have a strong direct effect on infant mortality in this model.
Employment (EMP) has a statistically significant negative effect on infant mortality (coef. = -
0.026, p = 0.000), indicating that higher employment levels are associated with lower infant
deaths, possibly due to improved household income and access to essential services.

The joint effects reveal that education-GDP interaction (ED_GD) is significantly

negative (coef. = -0.040, p = 0.000), suggesting that the combined effect of education and
income further reduces infant mortality. Also, education-employment interaction (ED_EM) is
negative and significant (coef. = -0.022, p = 0.000), implying that the benefit of education in
reducing infant mortality is greater when employment is high. Employment-GDP interaction
(EM_GD) is significantly negative (coef. = -0.012, p = 0.000), showing that employment has
a stronger effect on reducing infant mortality in wealthier contexts. The findings highlight
that income, employment, and their interactions with education are significant drivers of
reductions in infant mortality. Although education alone is not significant, its synergistic
effects with income and employment are crucial. These results emphasize the need for
integrated economic and social policies that improve labor market conditions and education
in tandem with poverty reduction strategies to lower infant mortality rates across the region

59
 However, inflation (LCPI) has a strong positive and significant effect on infant
mortality (coef. = 0.176, p = 0.000), meaning rising prices may undermine household welfare
and access to healthcare, leading to more infant [Link] significant role of inflation points
to the importance of macroeconomic stability in ensuring health access and affordability.
Real exchange rate (LREX) and health expenditure as a share of GDP (CHE_GDP) show
statistically insignificant effects on infant mortality (p = 0.819 and p = 0.949, respectively),
suggesting that in this sample, these macroeconomic variables do not directly influence infant
survival rates. Also, population growth rate (PGR) has a positive but insignificant
relationship with infant mortality (coef. = 0.026, p = 0.213), implying that while higher
population growth may place pressure on healthcare systems, the effect is not statistically
conclusive.

Table 4.36,
Results of Joint Effects of Income, Education, Employment and Infant Mortality
Lower Upper
LIMR Coef. Std. Err. z P>|z| limit Limit
LGDPC -0.042 0.011 -3.75 0.000 -0.065 -0.020
LEDU -0.626 0.549 -1.14 0.254 -1.702 0.450
EMP -0.026 0.006 -4.07 0.000 -0.039 -0.014
ED_GD -0.040 0.011 -3.48 0.000 -0.062 -0.017
ED_EM -0.022 0.004 -4.90 0.000 -0.030 -0.013
EM_GD -0.012 0.002 -5.00 0.000 -0.017 -0.007
LCPI 0.176 0.021 8.38 0.000 0.135 0.217
LREX -0.011 0.049 -0.23 0.819 -0.106 0.084
CHE_GDP -0.001 0.009 -0.06 0.949 -0.018 0.017
PGR 0.026 0.021 1.25 0.213 -0.015 0.066
_cons 1.837 1.157 1.59 0.112 -0.431 4.105
Wald chi2(10) 818.46
Prob > chi2 0.0000
No of Obs. 120
Source: Computed by the Author based on the World Bank WDI (2025)
Note: P-value < 1%, <5%, <10% indicates significance at 1%, 5%, and 10% respectively

4.6 Discussion of Findings

60
 The findings of this study provide compelling evidence of the significant impact of
socioeconomic variables on health outcomes, especially on life expectancy and infant
mortality across selected African countries. This study investigated the determinants of life
expectancy and infant mortality in selected African countries, with specific emphasis on the
effects of income, education, and employment. The results provide valuable insights into how
socioeconomic and macroeconomic variables contribute to health outcomes, particularly life
expectancy at birth and the infant mortality rate. The findings are discussed in line with the
four key objectives of the study. The findings reveal that income, measured as GDP per
capita, has a statistically significant and positive effect on life expectancy and a negative
effect on infant mortality. This aligns with theoretical expectations and empirical studies
indicating that higher income levels enable better access to nutrition, clean water, improved
housing, and healthcare services, all of which contribute to longer life spans and reduced
child mortality (Momoh et al., 2024; Awoyemi, Makanju, and Duru 2024; Awoyemi et al.,
2023; Aristovnik et al., 2020). The negative coefficient for GDP per capita in the infant
mortality model (-0.042) suggests that as income increases, the incidence of infant deaths
declines. This is consistent with the findings of Esseau-Thomas et al. (2022), who
documented the inverse relationship between income and early childhood deaths in low-
income sub-Saharan African economies. Furthermore, recent global health literature
corroborates that economic growth contributes to health improvements when it leads to
increased public and private health investments (WHO, 2023; Akpan & Madu, 2023).
However, the marginal effect observed suggests that income alone is not sufficient, especially
if not equitably distributed or reinvested in health-enhancing sectors.

Education remains a critical determinant of health, as reflected in the significant

coefficients of education on life expectancy and infant mortality. While education positively
influences life expectancy, its direct effect on infant mortality appears statistically
insignificant. This contrast could be attributed to the lagged effects of educational
investments or disparities in the quality and accessibility of education systems across the
studied countries. Nonetheless, disaggregated variables such as education-GDP and
education-employment interaction are significant in both models. These results support the
argument by Momoh and Okwu (2022) and Ogu (2023), who emphasize that beyond general
literacy, health outcomes respond more directly to education targeted toward health literacy,
women’s education, and maternal health awareness. Education enhances individuals’ ability

61
to make informed health decisions, utilize medical services, and adopt healthier behaviors,
which in turn improve longevity and reduce child mortality (Awoyemi, Makanju, and Duru
2024; UNESCO, 2024).

Employment has a positive but statistically insignificant effect on life expectancy and
a negative and highly significant effect on infant mortality. These findings indicate that while
employment may not immediately influence longevity in the studied regions, it plays a
crucial role in improving child survival. Employment boosts household income and stability,
which supports better prenatal care, nutrition, and early childhood development (Esseau-
Thomas et al., 2022; Bakare et al., 2023). The significant interaction terms further reveal that
employment gains translate into health improvements when accompanied by investment in
education and equitable economic development. As Aristovnik et al. (2020) assert,
employment without corresponding increases in wage levels, job security, or health coverage
may fail to influence adult health outcomes. Nonetheless, improved employment prospects
create enabling environments for families to prioritize healthcare and hygiene, particularly for
infants and young children. The combined model shows that the joint effects of income,
education, and employment are more significant in explaining variations in both life
expectancy and infant mortality than any of the variables individually. All interaction terms
are statistically significant in both health outcome models. This highlights the
multidimensional nature of health determinants, suggesting that holistic policies targeting
economic growth, educational access, and employment generation are necessary for
sustainable health improvements in Africa.

The significance of interaction terms aligns with the findings of recent studies by
Momoh et al. (2024) and Agyeman et al. (2023), which suggest that the synergy between
social and economic capital determines the extent to which development leads to better
health. For example, educational gains are more impactful when the population is gainfully
employed and the returns to education are high, thereby reinforcing the link between
education and improved health behaviors. Similarly, income contributes more to health when
used to finance public goods and services, such as hospitals and clean water infrastructure
(Awoyemi and Nwibe, 2023). Additionally, the inflation is negatively associated with life
expectancy and positively with infant mortality, implying that macroeconomic instability can
reverse health gains. Population growth rate (PGR) also has a detrimental effect, especially

62
on life expectancy, reflecting the strain that rapid population expansion places on health
infrastructure and social services.

Overall, this study finds robust evidence that income, education, and employment
both individually and collectively significantly influence health outcomes in selected African
countries. The policy implication is clear: governments and development partners must
pursue integrated development strategies that promote inclusive economic growth, improve
educational systems, and create sustainable employment opportunities. This holistic approach
will not only extend life expectancy but also reduce the burden of preventable infant deaths
across the continent.

References (Other ones are in your work)

1. Gujarati, D. N., & Porter, D. C. (2008). Basic Econometrics (5th ed.). McGraw-Hill.

2. Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2009). Multivariate Data
Analysis (7th ed.). Pearson Prentice Hall.

3. Agyeman, P. O., Dankyi, E., & Osei, S. (2023). Socioeconomic determinants of

health in West Africa: Evidence from structural equation modeling. African
Development Review, 35(1), 23–45.
4. Akpan, U., & Madu, C. (2023). Health expenditure and infant mortality in Africa: A
disaggregated analysis. Health Economics Review, 13(7), 1–18.
5. UNESCO (2024). Education and Health: Building Sustainable Futures in Africa.
Paris: UNESCO Press.
6. World Health Organization (WHO). (2023). Tracking Universal Health Coverage:
Global Monitoring Report 2023. Geneva: WHO.
7. Awoyemi, B. O., Makanju, A. A., & Duru, C. (2024). The Analysis of Human
Capital Development, Economic Growth and Longevity in West African Countries.
Scientific Annals of Economics and Business.

63
 8. Awoyemi, B.O., Makanju, A.A., Mpapalika, J., Ekpeyo, R.S (2023).A time series
analysis of government expenditure and health outcomes in [Link] of
Public Health in Africa. doi:10.4081/jphia.2023.1409.
9. Awoyemi, B.O., Nwibe, D.A (2022) A Causal Assessment of Nigeria’s Crude Oil
Revenue, Health Expenditure, and Economic Growth. International Journal of
Energy Economics and Policy, 2022, 12(5), 420-424. DOI:
[Link]

Literature Review (Add them to you empirical literature in chapter two)

Recent literature highlights the critical interplay between socioeconomic factors and health
outcomes in Africa, with growing attention on education, income, and healthcare financing.
Agyeman, Dankyi, and Osei (2023) employed structural equation modeling to investigate the
socioeconomic determinants of health in West Africa. Their study found that education,
income, and employment significantly influence health status, both directly and indirectly
through mediating factors like access to healthcare and environmental conditions. The
authors emphasized that policies promoting human capital development yield substantial
long-term health benefits.
Akpan and Madu (2023) provided a disaggregated analysis of health expenditure and infant
mortality across African countries. They concluded that public and private health
expenditures reduce infant mortality, but the effectiveness varies by region. Their study
underscored the importance of efficient allocation and governance of health budgets,
particularly in low-income settings, to ensure measurable outcomes in child health.
UNESCO (2024), in its comprehensive report "Education and Health: Building Sustainable
Futures in Africa", emphasized the bidirectional relationship between education and health.
Educated populations are more likely to make informed health decisions, utilize healthcare
services, and demand accountability from governments. The report advocates for integrated
development planning, arguing that investments in education systems should be viewed as
co-investments in public health.
The World Health Organization (2023), in its Global Monitoring Report on Universal Health
Coverage (UHC), assessed progress toward health equity globally. In the African context, the
report found that financial risk protection and service coverage remain limited, especially for
marginalized groups. The report emphasized that socioeconomic inequalities, including gaps
in education, employment, and income continue to drive health disparities, reinforcing the
need for multisectoral approaches to health system strengthening.

Awoyemi, Makanju, and Duru (2024) explored the long-term impact of human capital
development on economic growth and longevity in four West African countries, Nigeria,
Ghana, Burkina Faso, and the Benin Republic. Using panel data techniques such as POLS,
FM-OLS, and DOLS, their findings revealed that human capital development significantly
enhances both economic growth and life expectancy. Specifically, they found a 347.5%
positive effect on economic growth and a 26.8% increase in life expectancy at birth

64
associated with improvements in health and education. These results affirm the centrality of
human capital in sustainable development, reinforcing calls for increased public investment
in education and healthcare across West Africa to boost productivity and improve population
health outcomes.
Awoyemi et al. (2023) conducted a time-series analysis to assess the relationship between
government health expenditure and health outcomes in Nigeria, using the ARDL approach to
capture both short- and long-run dynamics. Health outcomes were measured using life
expectancy at birth and mortality rate. The findings indicate that increased government health
spending is significantly associated with reduced mortality rates and higher life expectancy,
suggesting that public investment in health has tangible benefits. The study highlights the
importance of not just increasing health budgets but also improving the efficiency and
partnership structures, especially through Public-Private Partnerships, to enhance healthcare
delivery and outcomes in Nigeria.

Awoyemi and Nwibe (2023) examined the causal relationship between crude oil revenue,
health expenditure, and economic growth in Nigeria, employing robust econometric
techniques to determine both short- and long-term dynamics. The study revealed that crude
oil revenue significantly influences health expenditure and economic growth, suggesting that
fluctuations in oil income impact government capacity to fund the health sector and drive
development. The results underscore the need for Nigeria to diversify its revenue base to
ensure more stable and sustainable investment in health and human capital, which are
essential for long-term growth.

65