# The Nexus of Energy Consumption, Foreign Direct Investments, and Inclusive Growth in Sub-Saharan Africa

## Description

The paper focused on how energy consumption could promote inclusive growth and which type of energy consumption either renewable or non-renewable can lead to inclusive growth. The paper then proceeds to test empirically if the presence of foreign direct investment could help energy consumption to further enhance inclusive growth. The paper used 32 sub-Saharan African (SSA) countries over a 25-year time spanning 1995 to 2019. The is primarily based on the availability of data which was obtained from the World Bank (World Development Indicators), and the United States Energy Information Agency (EIA). The dependent variable was proxied with an inclusive growth index where principal component analysis (PCA) was used in the generation of the index. The paper used 21 variables to create the index and these include access to clean fuels and technologies for cooking (% of the population); access to electricity (% of the population); mobile cellular subscriptions (per 100 people); contributing family workers, total (% of total employment); employment to population ratio, 15+, total (%) (modelled ILO); Immunization, DPT (% of children ages 12-23 months); mortality rate, under-5 (per 1,000 live births); nurses and midwives (per 1,000 people); physicians (per 1,000 people); the prevalence of underweight, weight for age (% of children under 5); primary education, duration (years), the proportion of seats held by women in national parliaments (%); pupil-teacher ratio, primary; school enrolment, primary (gross); gender parity index (GPI); school enrolment, secondary (gross), gender parity index; school enrolment, tertiary (gross) gender parity index; people using at least basic drinking water services (% of the population); people using at least basic sanitation services (% of the population); domestic general government health expenditure (% of general government expenditure); government expenditure on education, total (% of government expenditure); Gross Domestic Product per Capital (Constant, 2017, US$ PPP). All these variables were sourced from WDI. To ensure that the index created is robust, the paper conducts various diagnostic tests such as determinant of the correlation matrix; Kaiser-Meyer-Olkin Measure of Sampling (KMO) and Bartlett test of sphericity. The main variable of interest is energy consumption which was disaggregated into renewable energy consumption (% of total final energy consumption) and fossil fuel comprises coal, oil, petroleum, and natural gas products (non-renewable) sourced from EIA. The moderating variable is net Foreign Direct Investment Inflow (% GDP). The control variables we sourced from WDI include gross fixed capital formation (% of GDP), labour force participation rate, total (% of total population ages 15-64) (modelled ILO estimate), trade (% of GDP) and GDP per person employed (constant 2017 PPP $).

## Files

## Steps to reproduce

Some of the data sets have missing values, hence we used STATA 17 ipolation method which finds the average of the values to replace the missing values. The command for the ipolation can be stated as "ipolate variable name Year, gen(new name for variable) epolate by( id)". After, we take the logarithm of data that have high values, for example, both renewable and non-renewable are in log form. Similarly, GDP per persons employed is also in log form. We realised that FDI has outliers so we worked on it to cater for outliers using the Winsor approach in STATA. STATA command is "winsor2 FDI, replace cut(10 95)'" The paper computed the index using PCA; PCA command " pca list all the variables". "factortest list all the variables" which produces the diagnostic results for index. "estate loadings" and predict components with egenvalues greater than "predict pc1 pc2 pc3 pc4, score". Find the weighted average of the scores and normalise it to fall between the range of 0-1. The study computed the descriptive and correlation matrix to know the raw nature of the values. Pre-estimation for Pooled Mean Group was also conducted. For instance, cross-sectional dependence test was conducted using this Pesaran (2004) test of cross-sectional dependence. SATA command: ""xtreg dependent variable independent variable control variable, fe r" "xtcsd, pesaran abs". Pesaran (2007) Cross-sectional Augmented Dickey-Fuller (CADF) panel unit root test to check the stationarity of the data. Command: "xtcips variable, maxlags(5) bglags(1). Alternative unit root test; Harris-Tzavalis (HT) test and Levin-Lin-Chu’s (LLC) test. Kao (1999) cointegration test {command " xtcointtest kao dependent variable independent variable control variable, lags(aic 5) demean ar(same )} and Pedroni (2004) cointegration test {command " xtcointtest pedroni dependent variable independent variable control variable, lags(aic 5) demean ar(same )} was conducted to see if variables have long run relationship. Hausman test to select pooled mean group or mean group command: ""xtpmg first difference of dependent variable first difference of independent variable first difference of control variable, lr(lag of dependent variable independent variable control variable) ec(ect) replace mg" "est sto mg" ""xtpmg first difference of dependent variable first difference of independent variable first difference of control variable, lr(lag of dependent variable independent variable control variable) ec(ect) replace pmg"" "est sto pmg"" ""hausman mg pmg, sigmamore""