RENEWABLE ENERGY CONSUMPTION, EDUCATION AND ECONOMIC GROWTH IN BRAZIL, RUSSIA, INDIA, CHINA, SOUTH AFRICA

The study investigated two aspects, namely, (1) the impact of renewable energy consumption on economic growth in BRICS (Brazil, Russia, India, China, South Africa) and (2) whether education is a channel through which renewable energy consumption affects economic growth in BRICS. Panel data analysis such as fully modified ordinary least squares (FMOLS), pooled ordinary least squares (POLS) and fixed effects methods were used with data ranging from 1994 to 2015. Both models across all the three estimation techniques show that renewable energy consumption had a significant negative effect on economic growth in support of the findings by Silva et al (2012) and Lee and Jung (2018). What is also clear across all the three panel data analysis methods used is that education reduced the size of the negative effect of renewable energy consumption on economic growth in BRICS. In other words, education is a channel through which renewable energy consumption’s influence on economic growth is enhanced, in support of views by Dunn and Mutti (2004), Ozcicek and Agpak (2017) and Lawrence et al. (1991). The implication of the study is that BRICS countries are therefore urged to invest more in education as that is more likely to enhance the impact of renewable energy consumption on economic growth.


INTRODUCTION
, energy is a key component in the economic growth process of any country as it drives not only day to day household activities but also industrial activities that forms the basis upon which the economy is built. However, the rate at which the energy sources are depleting has led to most governments resorting to renewable energy sources, which also are clean sources, cheaper and have a sustainable impact on economic growth (Apergis and Danuletiu, 2014). It is for this reason that an increasing number of empirical researchers in the last decade has investigated what role renewable energy consumption plays in the economy.
The non-linearity hypothesis is an emerging view which has not yet explicitly been empirically tested to the best of the authors' knowledge. Despite literature availability which says that education increases the use of renewable energy and consequently boost economic growth (Lawrence et al., 1991;Ozcicek and Agpak, 2017;Dunn and Mutti, 2004), there is no study that the authors are aware of that explored if education is a channel through which renewable energy consumption stimulates economic growth. The current study seeks to fill this gap.
The story of the influence of renewable energy consumption on economic growth has been told for several economic groupings such as Central and Eastern European countries, G7 countries, developed countries in Europe, Organization for Economic Cooperation and Development countries, Organization of Islamic Cooperation countries, developing countries, European Union, Western Europe, Asia, Latin America and African countries, Europe, European Union countries, Gulf Cooperation Council countries, Middle East and North African countries, United States of America, Denmark, Portugal and Spain. The study on the impact of renewable energy consumption on the economy also focused on individual countries such as South Korea, South Africa, Vietnam, Tunisia, Lithuania, India, Brazil and Indonesia, among others. Clearly, the story of Africa with regards to the relationship between renewable energy consumption and economic growth is still not yet told. Another notable economic grouping which the topic has not yet been explored is the Brazil, Russia, India, China, South Africa (BRICS). The current study fills in that void.
Six more sections constitute the rest of the paper. Section 2 is the literature review on the impact of renewable energy consumption on economic growth, section 3 discusses the education-led growth hypothesis whilst section 4 focuses on the relationship between education and renewable energy consumption. Methodological framework, data analysis and interpretation is discussed in section 5 whereas section 6 concludes the study and highlight policy implications of the study. Section 7 is the bibliography.

EFFECT OF RENEWABLE ENERGY CONSUMPTION ON ECONOMIC GROWTH: LITERATURE REVIEW
Consistent with Odhiambo (2009), four dominant views on the relationship between energy consumption and economic growth include (1) the growth hypothesis, (2) feedback hypothesis, (3) conservation hypothesis and (4) neutrality hypothesis. The growth hypothesis is of the view that economic growth is enhanced by energy consumption whilst feedback hypothesis argues that both energy consumption and economic growth have got a positive influence on each other. According to the conservation hypothesis, economic growth is the one that drives energy consumption whilst the neutrality hypothesis argues that the relationship between energy consumption and economic growth is non-existent.
On the empirical front, Table 1 shows a summary of the empirical literature which focused on the influence of renewable energy consumption on economic growth.
From the literature, it is clear that the relationship between renewable energy consumption and economic growth has got mixed results. Firstly, renewable energy consumption has a positive impact on economic growth. Secondly, renewable energy consumption negatively affects economic growth. Thirdly, both renewable energy consumption and economic growth influence each other. Fourthly, economic growth has a deleterious effect on renewable energy consumption. Fifth, there is no relationship between renewable energy consumption and economic growth.
Sixth, a non-linear function describes the relationship between renewable nervy consumption and economic growth. These contradictions are evidence that the relationship between these two variables is still inconclusive and needs further research.
In line with Dunning (1988), one of the locational advantages of Foreign direct investment (FDI) and consequently drivers of economic growth in the host country is human capital development. According to Keynes (1936), human capital development as measured by education is one of the key drivers of economic growth, a view which was empirically supported by a study done by Li and Huang (2009).
According to Ozcicek and Agpak (2017), socioeconomic factors that influence renewable energy consumption include education, demographic qualifications, energy prices and carbon intensity. Moreover, Lawrence et al. (1991) argued that education enhances people's ability to adapt to new technology, coordination, learning and self-dependency on economic matters and it also increases the people's chances of getting employed, skills, higher incomes and a differentiated consumption pattern. Dunn and Mutti (2004) also argued that higher income make the people want to live in a clean environment thereby preferring the use of clean energy (renewable energy). In the study of Erdogan et al. (2020), it was observed that there was a negative relationship between oil revenues and educational level for the relevant group of countries in the long term. According to this result, it was concluded that there was no strong awareness of the importance of education in the countries those are included in the analysis, and that the resources were not sufficiently transferred to the increase of human capital power.

Data, its Description and Sources
This paper used annual panel data  to investigate the relationship between renewable energy consumption, education and economic growth in BRICS group of countries. The sources of the secondary data used include African Development Indicators, World Development Indicators, International Financial Statistics and International Monetary Fund database. The study used these sources of data because they are reliable, reputable and consistent.  ARDL -1990ARDL - -2014 Renewable energy consumption and economic growth were found to have had no relationship in the short run in Bulgaria and Romania. In Lithuania, Slovenia and Hungary, renewable energy consumption improved economic growth in the short run. A feedback effect in the long run was detected in the CEE group of countries Tugcu and Topcu (2018) G7 countries NARDL -1980NARDL - -2014 In the long run, renewable energy consumption had a positive effect on the economy in G7 group of nations Lee and Jung (2018) South Korea VECM and ARDL -1990-2012.
ARDL results show that renewable energy consumption had a deleterious impact on economic growth whilst VECM approach shows a positive relationship running from economic growth towards renewable energy consumption in South Korea Ucan et al. (2014) Developed countries in Europe -1990-2011 A uni-directional causality relationship from non-renewable energy consumption to economic growth was detected Hung-Pin (2014) OECD countries ARDL andVECM approaches (1982-2011) In the long run, renewable energy consumption had a positive impact on economic growth in the case of Germany, United Kingdom and Italy. In United Kingdom and Italy, short run results show that economic growth granger caused renewable energy consumption Anwar et al. (2017) OIC countries FMOLS and DOLS -1990DOLS - -2014 A significant positive impact of renewable energy consumption on economic growth was detected both in the short and long run Shakouri and Yazdi (2017) South Africa ARDL (1971ARDL ( -2015 The relationship between renewable energy consumption and economic growth supported the feedback hypothesis both in the long and short run Fotourehchi (2017) Developing countries

FMOLS
Panel granger causality tests  GDP was granger caused by renewable energy consumption in developing countries in the long run Saad and Taleb (2018) Europe Panel VECM (1990VECM ( -2014 In the short run, economic growth Granger caused renewable energy consumption whilst the long run is characterized by a bi-directional relationship between renewable energy consumption and economic growth Soava et al. (2018) European union countries Panel data analysis  Economic growth was found to have been enhanced by renewable energy consumption in European Union countries studied Apergis and Danuletiu, (2014) European union, Western Europe, Asia, Latin America and African countries Group mean panel test statistics  Across all the regions studied, renewable energy consumption was found to be a source of economic growth Halkos and Tzeremes (2013) European countries Data envelopment analysis and non-parametric regressions For lower renewable energy consumption levels, economic efficiency was found to have been enhanced by renewable energy consumption. The study also revealed that regional characteristics determined the impact of renewable energy consumption on economic efficiency in European countries Clottey et al. (2018) Vietnam ARDL andVAR models (1985-2017) In Vietnam, the study revealed that renewable energy consumption and economic growth positively affected each other in the long run Hassine and Harrathi (2017) GCC countries FMOLS andDOLS (1980-2012) No causality between renewable energy consumption and output was detected in the short run. However, a strong and statistically significant influence of renewable energy consumption on output was found in the long run Habib (2015) Tunisia ARDL  Renewable energy consumption and economic growth in Tunisia positively influenced each other in the short run. On the contrary, long run results show that economic growth increased renewable energy consumption in Tunisia Farhani, (2013) MENA countries FMOLS andDOLS (1975-2008) In the short run, no causality between renewable energy consumption and economic growth was found. Economic growth was found to have a positive influence on renewable energy consumption in the long run in the MENA region Thombs (2017) Developing countries Panel data analysis Renewable energy consumption had a negative influence on total carbon emissions per unit of GDP Bobinaite et al. (2011) Lithuania Granger causality test (1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009) Renewable energy consumption positively influenced economic growth in the short run in Lithuania. In the long run, the study could not find any relationship between renewable energy consumption and economic growth in Lithuania  Table 2 is a summary of the variables, proxies, expected sign(s) and the author(s).

Econometric Model Specification
The following empirical models were tested.
GROWTH i,t =β 0 +β 1 RENEW i,t +β 2 EDUC i,t +X i,t +µ i +ε it (1) GROWTH, RENEW, EDUC and X represents economic growth, renewable energy consumption, education and explanatory factors (FDI, infrastructural development, trade openness, financial development). Following Goff and Singh (2014), the combination between renewable energy consumption and education is denoted by (RENEW i,t , EDUC i,t ). Ɛ it is error term. Subscripts t and i stands for time and country respectively. β 0 is the unobserved country specific time invariant effect.

Main Data Analysis
Pre-estimation diagnostics (correlation analysis, descriptive statistics, trend analysis), diagnostic tests (panel unit root tests, panel co-integration tests) and main data analysis (fixed effects, fully modified ordinary least squares [FMOLS], pooled ordinary least squares [POLS]) are the three categories dealt with under this sub-section.
The correlation analysis results in Table 3 shows that there is significant negative relationship between (1) renewable energy consumption and economic growth and (2) savings and economic growth. The results also show that the relationship between (1) education and economic growth and (2) infrastructure development and economic growth is significant positive. Moreover, a nonsignificant positive relationship between (1) FDI and economic growth, (2) trade openness and economic growth and (3) financial development and economic growth was detected. The fact that the maximum correlation in Table 3 is 79% (between trade openness and renewable energy consumption) is an indication that there is no multi-collinearity between and among the variables studied, in line with Abel and Le Roux (2016).
According to Table 4  .
For all the countries with the exception of USA, renewable energy consumption had a deleterious impact on economic growth Pao and Fu, (2013) Brazil ECM -1980ECM - -2010 The results on the relationship between renewable energy consumption and economic growth are threefold: (1) Non-hydroelectric renewable energy consumption had a positive influence on economic growth, (2) total renewable energy consumption and economic growth positively affected each other and (3)   China and South Africa are also outliers because their mean financial development values far much exceeded the overall mean financial development value of 94.17% of GDP. Table 5 shows a standard deviation of economic growth at 3655 (above 1000), an indication that there are extreme values in the economic growth data. There are also probabilities of the Jarque-Bera criteria which are equivalent to zero in economic growth and savings data, a sign that data for the variables does not follow a normal distribution, a problem which needs to be addressed before using the data for main analysis in order to avoid spurious findings. This was done using Hair et al.'s. (2014) recommendations. Table 6 produced results which show that the data for all the variables is integrated of order 1 (stationary at first difference).

Panel Unit Root Tests
Employing Johansen fisher panel Co-integration test, at most six co-integrating vectors were established (results in Table 7). It means that the no co-integration null hypothesis is rejected, a finding which allows main data analysis to take place, in line with Tsaurai (2018).

FINDINGS AND ANALYSIS
Tables 8-10 show results of the fixed effects, FMOLS and the POLS respectively.
In model 1 and 2 under the fixed effects, FMOLS and POLS, renewable energy consumption was found to have a significant negative effect on economic growth, consistent with Lee and Jung (2018) in the case of South Korea and Silva et al. (2012) in the case of Portugal, Denmark and Spain. Education was however observed to have a non-significant negative impact on economic growth in model 1 and 2 under the fixed effects and FMOLS approaches, in contradiction to majority theoretical predictions on the subject matter.
On the other hand, model 2 under the fixed effects, FMOLS and POLS shows that education positively but insignificantly influenced  King and Levine (1993), Goldsmith (1969), Townsend (1983), Kiprop et al. (2015) Source: Author's compilation. GDP: Gross domestic product  economic growth and the same finding was observed in model 1 under POLS approach. These results resonate with literature (Keynes, 1936;Li and Huang. 2009) whose view is that economic growth is enhanced by human capital development (education, skills, health).
The interaction between renewable energy consumption and education had a negative but insignificant impact on economic growth in model 2 under the fixed effects, FMOLS and POLS, a finding which contradicts available literature (Ozcicek and Agpak, 2017;Lawrence et al., 1991;Dunn and Mutti, 2004) which says that educated people are more likely to use renewable energy because they prefer clean energy. Consistent with Inglesi-Lotz (2016) and Khobai (2018), increased use of renewable energy leads to higher economic growth. In other words, the interaction between renewable energy consumption and education is expected by literature to have a significant positive impact on the growth of the economy.
In both models 1 and 2 under the fixed effects, FMOLS and POLS, economic growth was found to have been positively and significant affected by FDI, results which follows an argument by Romer (1986). Financial development was also observed to have had a significant positive influence on economic growth in both models 1 and 2 across all the three estimation approaches used in the study (King and Levine, 1993).
Across all the three estimation approaches, trade openness had a significant negative impact on economic growth in both models 1 and 2, results which support a view by Baltagi et al. (2009). A non-significant positive relationship running from infrastructural development towards economic growth was detected in both models under the fixed effects and FMOLS whilst infrastructural development had a significant positive effect on the economy in BRICS. The findings resonate with Tsaurai (2018).
Savings were found to have had a significant positive impact on the economy of BRICS in both models under the fixed effects and FMOLS, in support of the view by McKinnon (1973). Moreover, a significant negative relationship running to economic growth from savings was observed in both models under the POLS approach, in contradiction with the available literature.    Levin et al. (2002); Im et al. (2003); ADF Fisher Chi-square and PP Fisher Chi-square tests respectively. * , ** and *** denote 1%, 5% and 10% levels of significance, respectively. Source: Author's compilation from e-views

CONCLUSION AND POLICY RECOMMENDATIONS
The study investigated two aspects, namely, (1) the impact of renewable energy consumption on economic growth in BRICS and (2) whether education is a channel through which renewable energy consumption affects economic growth in BRICS. Panel data analysis (FMOLS, POLS, fixed effects) methods were used with data ranging from 1994 to 2015. Both models across all the three estimation techniques show that renewable energy consumption had a significant negative effect on economic growth in support of the findings by Silva et al. (2012) and Lee and Jung (2018). What is also clear across all the three panel data analysis methods used is that education reduced the size of the negative effect of renewable energy consumption on economic growth in BRICS.
In other words, education is a channel through which renewable energy consumption's influence on economic growth is enhanced, in support of views by Dunn and Mutti (2004), Ozcicek and Agpak (2017) and Lawrence et al. (1991). The implication of the study is that BRICS countries are therefore urged to invest more in education as that is more likely to enhance the impact of renewable energy consumption on economic growth. Adjusted R-squared -0.89 GDP per capita is the dependent variable. *** , ** and * denote 1%, 5% and 10% levels of significance, respectively. Source: Author's compilation from e-views Adjusted R-squared -0.88 GDP per capita is the dependent variable. *** , ** and * denote 1%, 5% and 10% levels of significance, respectively. Source: Author's compilation from e-views