As a component of human capital, health has been acknowledged to play significant role in the growth and development process of any economy. Improved health status enhances capacity and strength of the labour force thereby raising their work participation and productivity. As rightly posited by [1], investment in health, raises efficiency in human capital which in turn boosts individuals’ productivity and consequently economic growth. Similarly, [2] argued that productivity growth should be positively correlated with the level of health, in particular with the average level of life expectancy in a country. These have triggered the development of new theoretical models [3–6] incorporating health as a determinant of economic growth such as Grossman’s Health Capital Theory and Demand for Health Model.

Also, high rate of mortality due to epidemiological and other external causes halt economic progress as evidenced by the recent case of Covid-19 pandemic. Huge resources that could have been used for investment in other productive ventures have been channeled towards curbing the spread of the disease in order to reduce fatalities. Going by the record of World Health Organization [7], increasing mortality from infectious diseases such as HIV/AIDS, Ebola, Malaria and tuberculosis have not only swept away improvements in life expectancy over the years but also contributed significantly to the slow growth of African countries. This has been buttressed by the report of the Economic Commission for Africa [8] which posited that countries affected by the Ebola epidemic like Guinea and Liberia have found it difficult to adjust the serious health crisis, which affected the social and production systems of the two countries. For instance, in 2014, the economic growth rates in the two countries were 1.1% and 1.7% respectively but dropped to 0.9% for each, in 2015. This is more intense in Sierra Leone where GDP growth rate was -21.5% in 2015, compared to 4.6% in 2014. This decline according to the report, stemmed from the closing down of two iron ore manufacturing plants, following the Ebola epidemic just to avoid fatality.

Again, high fertility rates have been reported to have significant influence on economic growth particularly in Africa. According to [9], increase in fertility rate leads to increase in population in the long-run; and where there are no job opportunities to absorb the increased population, it may result to increase in unemployment, dependency ratio, social vices and insecurity which cumulatively hamper the economic growth and development process of African economies.

Following the above line of arguments, various studies have been put forward on both single and multi-country basis to examine the contribution of health indices to economic growth particularly in Africa [10–14]. However, despite these efforts, there are at best, few studies that investigated the regional differences in the impact of health indices in Africa. For instance, [15] has analysed the dynamic relationships between healthcare spending and health indices on a panel of 41 Sub-Saharan African countries from 2000 to 2011 using fixed effect, random effect and GMM techniques. But, even this study cantered on the regional differences in the impact of healthcare spending on child mortality thereby saying nothing on the consequential effect on regional economic growth.

Going by the above submissions, it is obvious that poor health and growing burden of diseases are detrimental to economic growth and development as well as a major source of disparities between poor and rich countries. However, with the existing positive correlation between human development and economic growth, it is likely that a gradual improvement in health status in form of increase in life expectancy and decrease in mortality rates can have significant impact on the economic growth of Africa and its sub-regions, just as income brings about improved health status.

Therefore, this study intends to investigate and analyze the impact of health indices proxied by life expectancy, mortality rate and fertility rate on economic growth in Africa taking regional differences into account. This is because different countries and regions may have varied health policies, programs, and interventions as well as different socioeconomic conditions and institutional quality. All these may likely influence the extent to which health indices influence economic growth across African regions.

Empirical evidence

Various studies have been put forward on both single and multi-country basis to examine the contribution of health indices to economic growth. For instance, [10] examined the link between per capita health expenditure and health outcomes in 47 African countries from 1999 to 2004. The results of the study based on Robust OLS; Robust 2SLS; and Fixed Effect Estimator showed that total health expenditures (as well as the public component) are statistically significant contributors to improved health indices and economic growth in Africa. In another study on African countries, [11] examined how increased prevalence of HIV/AIDS and other diseases caused high mortality, retarded human capital development and consequently led to slow economic growth. Using a panel data for 38 African countries in the years 1980, 1985, 1990, 1995, 2000 and 2004, findings based on three-period Overlapping Generations Model revealed that prevalence of HIV/AIDS is responsible for a substantial decline in life expectancy in African countries which results in lower educational attainment and slow pace of economic growth.

Furthermore, [12] examined the contribution of health inputs and outcomes to the growth process in the Sub-Saharan Africa using a panel data of 30 countries spanning from 1995 to 2011. Using Dynamic Generalized Method of Moment (GMM) modeling framework, the results of the study showed that education has statistically significant positive effect on economic growth while health expenditure and mortality rate have statistically significant negative effects. Moreover, [13] analyzed the relation between economic growth and human capital between 1990 and 2011 for 15 MENA (Middle East and North Africa) countries. Using fixed effect, random effect, and Generalized Least Square (GLS), the results showed that life expectancy and fertility rate have statistically significant effect on GDP per capita while public health expenditure has no significant effect on economic growth over the study period.

More recently, [14] have empirically examined the relationship between health indices proxied by life expectancy and mortality rates, and economic growth in West African countries over the period of 1970 to 2016. However, the results obtained using panel unit root and panel co-integration techniques developed by Pedroni revealed that there is evidence of long run relationship among life expectancy, infant mortality rates and economic growth. In addition, the results of Granger causality tests revealed the existence of unidirectional causality running from both life expectancy and infant mortality rates to economic growth in West African countries.

In another recent study, [9] empirically examined the long-run relationship among health indices proxied by life expectancy, mortality rates and fertility rates, and economic growth in African countries over the period of 1990 to 2017. The study employed both first and second generation panel unit root as well as panel ARDL co-integration techniques to examine these relationships. On the whole, the overall results revealed the existence of long-run equilibrium relationship between economic growth and health indices as indicated by the coefficient of the Error Correction Term (ECT) which was negative and statistically significant. In addition, life expectancy and mortality rate were reported to have significant impact on economic growth in Africa both in the long-run and short-run whereas fertility rate has significant impact only in the long-run based on the results of PMG estimator. Thus, the research concluded that a long-run equilibrium relationship exists between health indices and economic growth and that life expectancy, mortality rate and fertility rate are among the factors influencing economic growth in Africa.

Going by the above literatures, it is obvious that health indices play significant role in the development process of Africa. However, this role may significantly differ across countries and regions mainly due to varied socioeconomic conditions and institutional quality among other factors which are critical in ensuring quality of life and individuals’ productivity. Thus, the novelty of this study is to examine and analyze these differences in the impact of health indices on economic growth across African regions.

Methods

The main objective of this study is to examine and analyze the regional variations in the impact of health indices on economic growth in Africa. The data of the study were sourced from the World Bank’s publication of World Development Indicators [16] and the International Country Risk Guide [17] spanning from 1990 to 2017. Based on data availability, the study selected Thirty Three (33) countries to form the sample size of the study. Out of this sample, Five (5) countries were from Central Africa namely: Angola, Cameroon, Congo, Democratic Republic of Congo, and Gabon. Eight (8) countries were from Eastern Africa which comprised: Ethiopia, Kenya, Madagascar, Malawi, Mozambique, Uganda, Zambia, and Zimbabwe. Northern Africa consists of Five (5) countries namely: Algeria, Egypt, Morocco, Sudan, and Tunisia. On the other hand, Three (3) countries were chosen from Southern Africa comprising Botswana, Namibia and South Africa. Finally, Twelve (12) countries were drawn from Western Africa which comprised: Burkina Faso, Ivory Coast, Gambia, Ghana, Guinea-Bissau, Liberia, Mali, Niger, Nigeria, Senegal, Sierra Leone, and Togo.

Therefore, in order to achieve the stated objective, the model below is specified following the work of Odhiambo [15]:

Where stands for economic growth for country i at time t, represents life expectancy for country i at time t, stands for mortality rate for country i at time t, and stands for fertility rate for country i at time t. However, is the vector of control variables representing Quality of Government Index (QOG) and Socioeconomic Conditions (SOCIOECO).

The choice of these control variables is based on their role in enhancing quality of life and per capita output. For instance, both the two variables depict not only different regional characteristics but also determine the conduciveness of the environment in which we live and standard of living which are critical to achieving economic growth [9] and [15].

The above equation is known as additive model which investigates only the “main” effects of the predictors. It assumes that the relationship between each variable of health indices and economic growth is independent of the other predictor variables in the model. In other words, the model assumes that the slope coefficient for each variable of health indices is the same for all countries across all regions. Hence, there is no any difference in the impact of life expectancy, mortality rates, and fertility rates on the economic growth across African regions. But this assumption might not be true simply because of differences in regional characteristics which may positively or negatively influence the effectiveness of health indices in impacting on economic growth.

To account for the regional differences, equation (1) above can be estimated separately for each region and then compare the results. However, running separate models for each group can be quite unwieldy because it results to estimating many coefficients than may be necessary. As a result, to account for the differences, the study introduced regional dummy variables and their interaction terms with independent variables. This is in line with the work of [18] who argued that to avoid the difficulties involved in estimating separate models, an alternative approach for testing whether parameter coefficients differ across groups is to introduce dummy variables and their interaction terms with independent variables.

Therefore, in line with the above argument, equation (1) is extended to include regional dummy variables as well as their interaction terms with health indices (life expectancy, mortality rates and fertility rates) across African sub-regions, to yield equation (2) as follows:

Equation (2) above is the baseline model for examining the regional difference in the impact of health indices on economic growth in Africa. This is because it allows the slope coefficients of health indices to differ across the regions. In this case, the model relates economic growth (LRGDP_it) to health indices, the regional dummy variables: Central Africa (CA), Eastern Africa (EA), Northern Africa (NA), and Western Africa (WA), the interactions between regional dummy variables and health indices, and a vector of control variables () which are: Quality of Government Index (QOG) and Socioeconomic Conditions (SOCIOECO) as earlier defined.

In terms of measurement, the dependent variable of this study is economic growth which is proxied by Real Gross Domestic Product (RGDP) and measured in US Dollars at constant price taking 2015 as the base year. For independent variables, life expectancy is measured as the average number of years a new-born infant would live if prevailing patterns of mortality of the total population at the time of his/her birth were to stay the same throughout his/her life. Also, total mortality rate is measured as the number of deaths occurring during a given year per 100,000 populations estimated at midyear. Again, fertility rate is measured as the average number of children a woman would bear during her lifetime, assuming her childbearing conforms to her age-specific fertility rate every year (i.e. 15 to 44 years). Finally, the regional dummy is measured as a binary variable between 0 and 1. The dummy variable is 1 if a selected country falls within the region of interest, otherwise it is 0.

For control variables, quality of government is measured as the mean value of three indicator variables: “Corruption”, “Law and Order” and “Bureaucracy Quality”, scaled between 0–1. Higher values indicate higher quality of government, and vice versa [17]. On the other hand, socioeconomic condition is measured as the sum of three sub-components: unemployment, consumer confidence, and poverty. The socioeconomic condition is scaled between 0 and 12. According to [17], each component has a maximum score of 4 points and a minimum score of 0 points. A score of 4 points equates to Very Low Risk (high socioeconomic condition) and a score of 0 points to Very High Risk (low socioeconomic condition).

Note that the coefficients of the interaction terms (, , …, and ) indicate how the effect of life expectancy, mortality rates, and fertility rates differs across African sub-regions. For example, if the coefficient of the interaction term between Central Africa dummy and life expectancy is positive, this means that life expectancy has a larger effect (i.e. more positive or less negative) in Central Africa than it does in the other regions. This can be interpreted as the increase or decrease in the effectiveness of life expectancy on economic growth in Central Africa when its regional dummy is 1. Where the regional dummy is 0, it is said to have no any influence on the effectiveness of life expectancy on economic growth in Central Africa. The same applies to other regions in the model. It should be noted that, sometimes, it may be the interaction term that is significant but not the main effects. However, the hierarchical principle states that, if we include an interaction in a model, we should also include the main effects, even if the p-values associated with their coefficients are not significant [18].

However, for the purpose of this study, is the coefficient of life expectancy which is expected to be positive. is the coefficients of mortality rates which is expected to be negative. Equally, , , , …, and are the coefficients of fertility rates, regional dummy variables, and the interaction terms between regional dummy and health indices respectively and each can take positive or negative sign. is the coefficient of the vector of control variables. is the composite error terms which consists of country specific effects and time-specific effects. In addition, the composite error terms are assumed to be normally distributed and homoskedastic.

It should be noted that the Southern Africa (SA) dummy variable and its interactions were dropped from the model (that is, equation 2) to avoid the dummy variable trap and to serve as reference point. Taking a region with higher number of countries (sample size) will significantly reduce the total observations which may affect the accuracy and reliability of the overall results. Hence, being a region with lowest number of countries in the study, Southern Africa was chosen as a reference point in order to avoid small sample bias.

Finally, giving the baseline model above, the study employed Fixed-Effect (FE) and Random-Effect (RE) estimation techniques to estimate and analyse the regional differences in the impact of health indices on economic growth in Africa. However, to choose the most preferred model between the two estimators, the study used Hausman test. The test examines the null hypothesis (H₀) that random effects are consistent and efficient, versus the alternative (H₁) that random effects are inconsistent (as the fixed effects will be always consistent). This choice strongly depends on the behaviour of the coefficients of country-specific effect and time-specific effect. According [19], where time-specific effects are absent but there is a country-specific effect, the estimation results obtained using FE is more appropriate. However, if the time-specific effects are absent but the country-specific effects are characterized as random term, the results of RE model will be chosen.

Results

Descriptive analysis of the data

This section describes the basic statistical elements of all variables used in the study across the African regions as shown in Table 1.

The result shows that the mean value of log of real GDP was US$ 23.33 across the regions with a standard deviation of US$ 1.52; whereas US$19.30 and US$26.8 represent the minimum and maximum values of log of real GDP respectively and both were in Western Africa (specifically in Liberia and Nigeria). Also, the average life expectancy across the regions was about 56 years with a standard deviation of about 8 years. Over the study period, the highest life expectancy of about 76.1 years was recorded in Northern Africa (in Algeria) while the minimum expected years of life was about 35.7 which was recorded in Western Africa (in Sierra Leone).

Furthermore, the average mortality rate in the regions was about 11.92 per 100,000 populations. Interestingly, Northern Africa experienced the lowest mortality rate of about 4.71 (precisely in Algeria again) which justify its position of being the region with highest life expectancy on the African continent.

Table 1: Descriptive Analysis of Variables for All Regions (Combined Panel)