Dynamic implication of biomass energy consumption on economic growth in Sub-Saharan Africa: evidence from panel data analysis

GeoJournal DOI 10.1007/s10708-016-9698-y

Dynamic implication of biomass energy consumption on economic growth in Sub-Saharan Africa: evidence from panel data analysis Hamisu Sadi Ali . Siong Hook Law . Zulkornain Yusop . Lee Chin

 Springer Science+Business Media Dordrecht 2016

Abstract The present article examines the dynamic linkages between biomass energy consumption, capital stock, human capital and economic growth across selected Sub-Saharan African countries based on dynamic heterogeneous panels of a mean group (MG) and pooled mean group (PMG) techniques. The finding based on PMG as the preferred method reveals that biomass energy consumption, capital stock and human capital are statistically significant, which means aforementioned variables have positive significant impact on economic growth in the countries studied. When an alternative panel estimation techniques of panel cointegration, dynamic OLS (DOLS) and fully modified OLS (FMOLS) are applied, the result based on panel cointegration technique reveals that biomass energy consumption, capital stock, human capital and economic growth are cointegrated as null hypothesis of most statistics are rejected at 1 % level of significance. The finding based on FMOLS shows that biomass energy consumption, capital stock and human capital positively influences economic growth at 1 % level and same result is obtained from panel OLS. The result based on DOLS however reveals that biomass energy consumption and capital stock are significant at 1 % on economic

H. S. Ali (&)  S. H. Law  Z. Yusop  L. Chin Department of Economics, Faculty of Economics and Management, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia e-mail: [email protected]

growth while human capital is insignificant. Considering its positive effect on economic growth with little or no environmental degradation when compared with fossil fuel uses, consumption of biomass energy is more preferable in these countries therefore is the best option to adopt by the policy makers of Sub-Saharan African countries. Keywords Biomass energy  Economic growth  Panel cointegration  Capital stock  Human capital  Sub-Saharan Africa

Introduction Energy consumption become necessary in the global economy, this is related to its diverse functions in accelerating various economic functions that made it one of the backbone of every economy in the world. Currently, fossil fuel is the highest energy consumed globally therefore its limited quantity, energy insecurity, energy price instabilities as well as environmental degradation caused by non-renewable energy necessitated the global policy makers to find an alternative source of energy. Because of the aforementioned problems the world now focused on renewable energy considering its unique advantages over the other energy sources, one of the main targets is the biomass energy. Biomass refers to the source of energy that is stored via plants absorption from the sun through the

123

GeoJournal

process of photosynthesis; it is the type of energy that is sourced from plants and animals. These types of energy sources are known as biofuels and usually contain wood chips, rooted trees, manure, mulch and components of tree. Globally the number of people that depends on traditional biomass as the major source of heating and cooking fuel is expected to rises from estimated 2.4 billion in 2002 to 2.6 billion in 2030 which shows an increment of 8 %. This increase is as a result of increase uses of biomass energy in Africa as the figure is expected to increase from 646 to 996 million in the same time period which is the 54 % increase.1 While biomass energy consumption in OECD countries is only 3 %, the rate of its usage in Africa differs as it used by the majority of the populace in order to meet their energy needs. For example, its consumption in Kenya reach 68 %, also 94 % of energy consumed in Burundi is from biomass. People in Sub-Saharan Africa heavily relied on biomass source of energy particularly combustibles for fundamental energy generation for household cooking and heating purposes. The percentage of energy sources out of total primary energy supply in Sub-Sahara Africa according to IEA and Stecher et al. (2013) in 2009 comprised of 61 % of solid biofuels, 0.2 % of other renewables, 19.7 % of coal and peat, 14.1 % of oil, 2.7 % of natural gas, 0.6 % of nuclear and 1.4 % of hydro. Biomass energy is therefore regarded as the primary source of total energy supply in the midst of other energy sources in the region (Stecher et al. 2013). Biomass energy use in SubSahara Africa differs from one country to another depending on the accessibility of electricity, availability of resources as well as renewable energy policy. Its uses in the region cannot be compared with any other source because it is regarded as the simple and easy to get and hence is available for lower level household (Mohammed et al. 2013) which constitutes the majority in the African economies. It is therefore essential to study the dynamic impact of biomass energy consumption on economic growth in this region because biomass is considered as the main source of energy in the continent. This study empirically examines the dynamic impact of biomass energy consumption on economic

1

http://ec.europa.eu/europeaid/where/acp/regional-cooperation/ energy/documents/biomass_position_paper_en.pdf.

123

growth in Sub-Saharan Africa. Some studies has earlier been conducted on this issue in Sub-Saharan Africa, but there are still few studies that applied heterogeneous panel data estimation techniques of DFE, MG and PMG as well as Panel cointegration, dynamic ordinary least square (DOLS) and fully modified ordinary least square (FMOLS) to address the impact of biomass energy consumption on economic growth in sub-Saharan Africa. Therefore the present article added to the existing literature in examining such relationship based on the combination of the aforementioned econometric methods.

Literature review The existing empirical literature recorded some studies that investigated this relationship across various economies with different findings. This may not be unconnected with the use of different datasets and econometrics methodologies, as well using different economies in the studies. The studies that captured the implications of overall energy consumption on economic growth will be review first and subsequently we narrowed on those that particularly dealt with renewable sources of energy. For example a study by Ali et al. (2015) on energy consumption and financial development in Nigeria that consider economic growth as one of the control variables reveals that economic growth negatively influences energy consumption in Nigeria. Shahbaz et al. (2013a) analysed the relationship between energy consumption and economic growth by including financial development and trade variables in China for the period of 1971–2011 based on ARDL estimation. The finding shows that energy consumption, financial development and trade positively influences economic growth. Shahbaz et al. (2013b) also examined the relationship between economic growth, energy consumption, financial development, trade openness and CO2 emission based on ARDL framework for 1975Q1-2011Q2 in Indonesia. The finding reported that increase is the consumption of energy and GDP growth leads to more emissions of CO2 , while development of the financial market and trade liberalization reduces it. The VECM causality test reported that there exist bidirectional causal effect between economic growth and CO2 emissions while unidirectional causality exists running from financial development to CO2 emissions.

GeoJournal

Al-Mulali and Sab (2012a) investigated the effect of energy consumption, CO2 emissions on economic growth and financial development across 30 subSaharan African countries for the period of 1980–2008. The outcome shows that energy uses increase GDP and financial sector performance though it increases environmental degradation by polluting the environment. Also the impact of energy consumption on economic growth and financial development is investigated by Al-Mulali and Sab (2012b) across 19 countries for the same period of 1980–2008. The result reported that consumption of energy enhances financial development as well as economic growth, though with a negative implications to the quality of environment. Chaudhry et al. (2012) applied ARDL approach and examined the impact of energy consumption on economic growth in Pakistan for the period of 1972–2012. The outcome reveals that when oil is used as an indicator of energy consumption, the effect on economic growth is negative in Pakistan. Ozturk and Uddin (2012) investigate the long-run and Granger causality relationship among energy consumption, carbon dioxide emission and economic growth for the period of 1971–2007 in India. They used augmented Dickey Fuller test (ADF), Phillips-Perron test (PP) and KPSS in order to test for long-run relationship. The main finding shows the existence of feedback causality between energy consumption and economic performance which suggest the existence of bidirectional causality between economic growth and energy consumption. This means that increase in economic performance stimulates energy consumption in India. In their study on the impact of energy uses on economic growth based on VECM approach in Tunisia for the period of 1980–2007 Abid and Sebri (2011) found that in general perception consumption of energy increases GDP growth, whereas when sectors of the economy are used it reduces it. Islam et al. (2013) investigated the impact of financial development, population and economic growth on energy uses in Malaysia. The result shows that energy uses influence financial development and economic growth in the both short and long-run, whereas its effect on population growth is present in the short-run only. Tugcu et al. (2012) examine long-run and causal nexus between renewable and nonrenewable energy and economic performance for the period of 1980–2009 across G7 economies. They applied autoregressive Distributed Lag approach to cointegration (ARDL) approach as well as causality approaches. The

long-run results reveal that both energy sources have influences on economic growth. The result also shows the existence of bi-directional causal relationship for the sample countries based on classical production function, mixed results is obtained for each country when the production is augmented. However, some studies were conducted on the renewable energy consumptions which includes among others; the study by Uc¸an et al. (2014) that examined the impact of renewable and non-renewable energy on economic growth across 15 EU member countries for the time frame of 1990–2011. The finding based on panel cointegration reveals that there exists long-run relationship among the variables, and Granger-causality result shows unidirectional causality running from non-renewable energy consumption to economic growth. Apergis and Payne (2010) examined the effect of renewable energy uses on economic performance across 20 OECD member countries for the period of 1985–2005. The Granger-causality test result reveals bidirectional causality exist between renewable energy consumption and economic performance in both short and long-run periods. Also Apergis and Payne (2011) examined the effect of renewable energy consumption uses on economic growth across 6 Central American economies for the period of 1980–2006. The causality result shows bidirectional in nature that is between renewable energy consumption and GDP in both short and longrun. Bildirici (2013) analysed the nexus between biomass energy uses and GDP among emerging and developing countries (Argentina, Bolivia, Brazil, Chile, Colombia, Guatemala, and Jamaica) based on ARDL framework. The result shows that there exist unidirectional causality running from economic growth to biomass energy uses for Colombia, while unidirectional causality exist from biomass to economic growth in Bolivia, Brazil and Chile and bidirectional causality exist for Guatemala. However, in the long-run bidirectional causality exists for all the countries. ¨ zaksoy (2013) investigated the Bildirici and O causality between biomass energy consumption and economic growth across 10 European countries based on ARDL and vector error correction models for the period of 1960–2010. The result of causality reveal that there exist unidirectional causal relationship running from economic growth to biomass energy uses for Australia and Turkey, unidirectional causality also exist running from biomass energy uses to

123

GeoJournal

economic growth for Hungary and Poland and bidirectional causality exist among Spain, Sweden and France. Payne (2011) investigated the causal link between biomass energy consumption and real GDP using multivariate framework for US for the period of 1949–2007. The finding shows that unidirectional causal relationship exist that run from biomass energy consumption to real GDP. Using Canning and Pedroni long-run causality test Apergis and Danuletiu (2014) examined the relationship between renewable energy consumption and economic growth across 80 countries. The result shows the existence of positive longrun causality running from renewable energy to real GDP for the entire sample and on regional basis as well. Ozturk and Bilgili (2015) investigated the longrun relationship between biomass energy consumption and economic growth using dynamic panel technique across 51 Sub-Saharan African countries for the period of 1980–2009. The finding indicated that biomass energy consumption, openness and population positively affected economic growth significantly in the sample countries. Ozturk (2015) investigates the dynamic relationship between biofuels production and its sustainable indicators i.e. carbon dioxide emissions, energy intensity, renewable energy generation and total population based on dynamic heterogeneous panel econometric technique-Generalized Method of Moments across 17 developed and developing economies. The finding reveals that a renewable electricity generation and carbon emission increases biofuels productions. The finding based on robust least square regression ratified that the four sustainable indicators have all significant link with the biofuels productions. Ben Jebli et al. (2014) applied panel cointegration approach based on Environmental Kuznets Curve (EKC) hypothesis and examine both short and long-run relationship among carbon emissions, economic growth, renewable energy consumption and trade openness across 24 Sub-Saharan Africa countries over the time frame of 1980–2010. The EKC hypothesis is not supported in the countries investigated. The main result reveals the presence of negative and significant impact of real GDP per-capita and real imports per-capita on carbon emissions. However, the squares of real GDP per-capita and real exports per-capita have positive and significant impact on carbon emissions in the sample countries. Bhattacharya et al. (2016) investigates the impact of renewable energy consumption on the economic

123

growth for the main renewable energy consuming economies in the global economy for the period of 1991–2012. The panel estimation result shows that variables have long-run relationship. The long-run result also shows that 1 % increase in renewable energy consumption could positively increase economic growth by 0.57 %. Jebli et al. (2016) used 25 OECD countries and examines the causal relationship between per capita CO2 emissions, GDP, renewable and non-renewable energy consumption and international trade for the period of 1980–2010. The long-run causality result reveals the presence of bi-directional causal relationship among all variables. Environmental Kuznets curve (EKC) hypothesis is tested based on FMOLS and DOLS and the result shows that higher consumption of non-renewable energy increase carbon emissions, while trade and renewable energy consumption reduces it.

The Data The data for this study was derived from various sources, economic growth as measured by real GDP is obtained from world development indicators dataset, World Bank (WDI CD ROM, 2015), biomass energy consumption as measured by used extraction of biomass is obtained from global material flow data base 2012 version, while capital stock as measured by capital stock at current PPPs (in mil. 2005US$) and human capital as measured by human capital index per individual based on years of schooling and returns to education were both obtained from Penn World Table 2013 version 8.1 respectively. Econometric model and methodology Following Pesaran and Smith (1995) and Pesaran et al. (1999) we assume that given data on time periods, t = 1, 2…,T, and groups, I = 1, 2…, N, the aim is to estimate an ARDL (p, q, q,…q) model, yit ¼

p X j¼1

kij xi;tj yi;tj þ

q X

dij xi;tj þ li þ eit

ð1Þ

j¼0

where xit (k  1) refers to the vector of independent variables for group i; li signifies the fixed effects; the coefficients of lagged dependent variable xij are scalars; and dij are k  1 coefficient vectors. T must

GeoJournal

be adequately large in order to estimate each group independently. It is suitable to work with the following reparameterization of Eq. (1); 0

Dyit ¼ /i yi;t1 þ bi þ

p1 X

ki  j Dyi;tj þ

j¼1

q1 X

di

0

/ MG ¼ ð2Þ

bi ¼

j¼0

dij xij ¼

p X

kim ;

m¼jþ1

^

d jMG ¼ ð3Þ

j ¼ 1; 2; . . .; p  1 and dij q X ¼ dim j ¼ 1; 2; . . .; q  1 When the time series observation for each group was stacked Eq. (2) can be written as; p1 X

ki  j Dyi;j

j¼1

þ

q1 X

ð4Þ

0

Dxi;j di  j þ li l þ ei

j¼0

I = 1, 2,…N, where yi ¼ ðyi1; . . .yiT Þ0 is T  1 vector of observations on the dependent variable of ith group, xi ¼ ðxi1; . . .xiT Þ0 is T  k matrix of the observations on the independent variables that differ both through the group and time periods, i = (1,…,1)0 refers to T  1 vector of 1s, yij and xij are j lagged values of yi and xi , Dyi ¼ ðyi  yi1 Þ, Dxi ¼ ðxi  xi1 Þ. Dyij and Dxij are j period lagged values of Dyi and Dxi , and ei ðei1; . . .eiT Þ. Both group specific short run and long run coefficients are calculated by the pooled maximum likelihood estimation. The estimators are represented by; PN

PN ~ _ ^ ^ b ;~i / PMG ¼ i¼1 ; b PMG ¼ i¼1 i ; N N PN ~_ ^ kij k jPMG ¼ i¼1 ; j ¼ 1; . . .; N PN ^ dij P  1 k jPMG ¼ i¼1 ; N ^



j ¼ 0; . . .; q  1; h PMG ¼ h

i¼1

;^i

N PN

i¼1

^

b MG

; kij

N PN

j ¼ 1; . . .; P  1

^

i¼1 d ij

N

;

PN ^_ b ¼ i¼1 i ; N

; ^

j ¼ 0; . . .; q  1; h PMG ¼

  1 XN ^ b =/ i i¼1 N

^

m¼jþ1

Dyi ¼ /i yi;1 þ xi bi þ

^

k jMG ¼

I ¼ 1; 2. . .N; and t ¼ 1; 2. . .T; where /i   Xp ¼ 1 k ; ij j¼1 Xp

PN

^

j¼0

 tj Dxi;tj þ li þ eit

However, the MG estimator suggested by Pesaran and Smith (1995) permits the heterogeneity of all parameters and the below estimates of short run and long run parameters:

where ;i , bi , kij and c ij refers to OLS estimators obtained independently from Eq. (2). Additionally, MG technique entails estimating separate regression for each sample unit and calculating averages of each sample specific coefficients. The MG estimators may probably be inefficient when the sample is small, because each sample outlier could strictly affect the averages of the sample coefficients. The long-run mean coefficients of MG estimator considers efficient and consistent, however it will be inefficient if the slope is homogeneous. The pooled estimators are consistent and efficient under the longrun slope homogeneity. Homogeneity hypothesis of the long-run policy parameters cannot presumed expected result and is empirically tested in overall specifications. The Hausman-type test (Hausman 1978) is used to detect the existence of heterogeneity in the means of the coefficients; the test is applied to differentiate between MG and PMG. Based on the null hypothesis the difference in the MG and PMG estimated coefficient is insignificant and PMG is considered more efficient. The empirical model of the study is therefore specified as follows; yit ¼ a0i þ a1 bimecit þ a2 cstit þ a3 hmcit þ eit

ð5Þ

When the variables were logged the following Eq. (6) was obtained below; lnyit ¼ a0i þ a1 lnbimecit þ a2 lncstit þ a3 lnhmcit þ eit ð6Þ The restricted version to estimate PMG based on 25 Sub-Saharan African countries for the period of 1980– 2011 is specified as follows;

123

GeoJournal

DlnYit ¼ ;i ðlnYit1  h1 lnBIMECit  h2 lnCSTit þ h3 lnHMCit  a4it  h0i Þ þ b1i DlnBIMECit þ b2i DlnCSTit þ b3i DlnHMCit þ eit ð7Þ where; lny refers to the log of economic growth, lnbimec is the log of biomass energy consumption, lncst is the log of capital stock and lnhmc is the log of human capital and eit is the unobservable error term, all the variables are in log form for normalization purposes.

Results Estimation results The below table 1 presented three alternative panel estimation techniques; dynamic fixed effect; PMG which enforces common long-run coefficient, and MG which enforces restrictions. The table therefore presented the long-run coefficient estimates, adjustment coefficients and joint Hausman test statistics. To choose the best model between MG and PMG Hausman test statistic is used, therefore based on the result PMG estimator is more suitable as we cannot reject Hausman statistic p value which is greater than 5 % (0.69). The main focus is now on PMG and dynamic fixed effect estimation techniques. Three alternative panel estimators are reported based on

Eq. (7) and the sign of long-run coefficients for the three models is all negative and significant which is in conformity with the theory. The coefficients of biomass energy consumption of PMG and dynamic fixed effect models is statistically significant at 1 %, and capital stock coefficient of PMG is also statistically significant at 1 % while it is statistically significant at 5 % based on dynamic fixed effect model. Human capital is also statistically significant at 5 % based on PMG model while it is statistically insignificant in case of dynamic fixed effect model. Since we rely on PMG as suggested by the Hausman pvalue, our result is consistent with that of Ozturk and Bilgili (2015) and Bhattacharya et al. (2016). In order to confirm the previous findings of dynamic fixed effect and PMG estimation techniques, an alternative heterogeneous panel techniques was applied which includes; panel cointegration, panel OLS, dynamic OLS (DOLS) and fully modified OLS (FMOLS). The first stage of panel cointegration is to test the unit roots so as to identify the order of integration of the variables which is reported in Table 2. Various panel unit root tests of Breitung (2000), Im et al. (2003), Levin et al. (2002) and Maddala and Wu (1999) are carried out and the results are reported in Table 2 above. The unit root test result which contained constant and trend terms indicated that almost unanimously shows that variables are nonstationary at level but become stationary after taking first difference. However, Breitung (2000) unit root

Table 1 DFE, MG and PMG estimation results, dependent variable: ln of real GDP (25 countries, 1980–2011) DFE

MG

PMG

lnBIMEC

0.93 (0.33)***

0.25 (0.29)

0.42 (0.08)***

lnCST

0.31 (0.14)**

0.62 (0.28)**

0.31 (0.04)***

lnHMC Error correction adjustment

0.14 (0.94) -0.07 (0.13)***

0.64 (1.39) -0.44 (0.06)***

0.44 (0.17)** -0.18 (0.04)***

DlnBIMEC

2.62 (1.23)**

0.00 (0.00)

0.00 (0.00)

DlnCST

-2.16 (4.69)

0.00 (0.00)

0.00 (0.00)

DlnHMC

0.42 (0.29)

-0.50 (0.69)

0.30 (0.82)

Maximum log likelihood

–

–

1451.6

Number of parameters

3

3

3

Hausman test

–

1.47 (0.69)

1.47 (0.69)

Country specific term is included in the equations, figures in parentheses shows t statistics while p values is used for Hausman test which indicated the level of significance at ***1 %, ** 5 % and * 10 % levels respectively DFE dynamic fixed effect, MG mean group, PMG pooled mean group

123

GeoJournal Table 2 Panel unit root estimation results dependent variable: ln of real GDP (31 countries, 1980–2011)

BIMECit

Y it

CST it

HMCit

Level No trend LLC

4.41 (1.00)

-1.51 (0.06)*

5.09 (1.00)

-15.2 (0.00)***

IPS

10.3 (1.00)

3.00 (0.99)

10.3 (1.00)

-6.45 (0.00)***

ADF

29.7 (0.99)

67.2 (0.30)

15.1 (1.00)

181 (0.00)

Breitung

–

–

–

–

With trend

Y economic growth, BIMEC, biomass energy consumption, CST capital stock, HMC human capital, LLC Levin, IPS Im, Pesaran and Shin (2003) panel unit root test, ADF augmented Dickey Fuller, Maddala and Wu (1999), Breitung (2000) The values in parentheses are respective p-values, ***, ** and * are used as a benchmark for null hypothesis rejection of nonstationary at 1, 5 and 10 % respectively

LLC IPS

-1.98 (0.02)** 0.29 (0.62)

ADF

62.9 (0.44)

Breitung

2.46 (0.99

-3.15 (0.00)*** -3.22 (0.00)***

2.56 (0.99) 0.94 (0.82)

113 (0.00)***

68.2 (0.27)

78.9 (0.07)*

0.05 (0.52)

13.1 (1.00)

-0.49 (0.31)

1.37 (0.99) -1.16 (0.12)

First difference No trend LLC

-17.5 (0.00)***

-22.9 (0.00)***

-27.7 (0.00)***

3.79 (0.99)

IPS

-17.8 (0.00)***

-24.9 (0.00)***

-14.5 (0.00)***

1.91 (0.97)

406 (0.00)***

583 (0.00)***

142 (0.00)***

38.9 (0.99)

ADF Breitung

–

–

–

–

With trend LLC

-16.3 (0.00)***

19.1 (0.00)***

-26.1 (0.00)***

7.79 (0.03)**

IPS

-16.7 (0.00)***

-23.2 (0.00)***

-14.7 (0.00)***

4.20 (0.00)***

367 (0.00)***

501 (0.00)***

389 (0.00)***

36.4 (0.00)***

-9.79 (0.00)***

-12.6 (0.00)***

-4.86 (0.00)***

13.8 (0.00)***

ADF Breitung

Table 3 Panel cointegration estimation results Pedroni cointegration

Panel v-statistic Panel rho-statistic

Model 1a without trend 2.45 (0.00)*** -1.26 (0.10)

Model 1b with intercept and trend

Model 1c without trend or intercept

-0.28 (0.61)

2.18 (0.01)**

1.39 (0.92)

-1.71 (0.04)**

Panel PP-statistic

-5.32 (0.00)***

-4.59 (0.00)***

-4.10 (0.00)***

Panel ADF-statistic

-7.27 (0.00)***

-8.73 (0.00)***

-5.39 (0.00)***

Group rho-statistic

0.92 (0.82)

3.05 (0.99)

-0.47 (0.31)

Group PP-statistic

-4.33 (0.00)***

-4.41 (0.00)***

-4.83 (0.00)***

Group ADF-statistic

-6.22 (0.00)***

-7.21 (0.00)***

-6.29 (0.00)***

Number of countries (N) = 31 and sample period (T) = 31 ***, ** and * indicated the significance level of null hypothesis rejection at 1, 5 and 10 % respectively

reject null hypothesis when linear trend is incorporated. Table 3 presented Pedroni (1999, 2004) panel cointegration results, for model 1a without trend the result shows that five out of seven statistics reject null hypothesis of no cointegration, for the second model 1b which contains intercept and trend it shows that

four out of seven statistics reject null hypothesis of no cointegration and the third model 1c without trend or intercept six out of seven statistics reject null hypothesis of no cointegration. This means there is strong long-run relationship among biomass energy consumption, capital stock, human capital and economic growth in the countries under investigation and this

123

GeoJournal Table 4 Panel OLS, DOLS and FMOLS estimation results Model 1a: Panel OLS

Model 1b: DOLS (lag = 1, lead = 1)

Biomass energy consumption

1.32 (24.5)***

1.97 (33.1)***

Capital stock

0.85 (13.4)***

0.27 (4.08)***

Human capital

1.52 (6.00)***

0.04 (0.09)

Model 1c: FMOLS

2.07 (41.1)*** 0.27 (5.06)*** -1.49 (-6.18)***

Panel OLS panel ordinary least square, DOLS dynamic ordinary least square; FMOLS fully modified ordinary least square Values in parentheses are t-statistics and ***, ** and * are significance values at 1, 5 and 10 % respectively

result is the same with the finding of dynamic fixed effect model above. In order to determine the long-run equilibrium relationship, DOLS and FMOLS approaches for heterogeneous cointegrated panels are estimated since the long-run relationship among the variables is already established. The result is reported in Table 4. The panel OLS and fully modified OLS results shows that biomass energy consumption, capital stock and human capital are statistically significant at 1 %, this means that the three variables highly influenced economic growth in the countries under investigation, this reconfirmed our earlier finding based on PMG which also corroborate the findings of Ozturk and Bilgili (2015) and Bhattacharya et al. (2016). While dynamic OLS result shows that biomass energy consumption and capital stock are statistically significant at 1 % and human capital is positive but insignificant on economic growth.

Conclusion This article examines the impact biomass energy consumption on the performance of the selected African economies for the period of 1980–2011 based on heterogeneous panel estimation techniques of MG and PMG. The result based on PMG which is the preferred model selected reveals that consumption of biomass energy, capital stock and human capital positively influence economic growth for the countries under investigation. To confirm the consistency of our main finding, an alternative

123

heterogeneous panel techniques of panel cointegration, DOLS, FMOLS was employed and the results shows that; based on panel cointegration variables have long-run relationship as null hypothesis of no cointegration was rejected at 1 % level of significance. FMOLS and panel OLS models results shows that biomass energy consumption, capital stock and human capital influence economic growth positively as null hypothesis was rejected at 1 % level of significance also, while the result of DOLS shows that biomass energy consumption and capital stock positively influence economic growth and human capital is insignificant. These findings mainly corroborated our earlier results based on other estimation technique used. The policy recommendation remains that authorities in these countries should promote the consumption of green energy as it enhance economic growth without problems of environmental degradation, this will definitely reduce the level of carbon emissions that pollute the environment and cause global warming which is among the main environmental problems that presently disturb the global economic environment. Therefore, policy makers should target more consumptions of renewable energy (biomass) in order to combat the problems of global warming that have serious adverse effects in the global economic environment.

Appendix See Table 5.

GeoJournal Table 5 Sample countries according to the econometric method applied

DFE, MG and PMG techniques

Panel cointegration technique

Benin

Benin

Botswana

Botswana

Burundi Cameroun

Burundi Cameroun

Central Africa Republic

Central Africa Republic

Cote d’ivoire

Cote d’ivoire

DR Congo

DR Congo

Gabon

Gabon

Gambia

Gambia

Ghana

Ghana

Kenya

Kenya

Lesotho

Lesotho

Liberia

Liberia

Malawi

Malawi

Mali

Mali

Mauritania

Mauritania

Mauritius

Mauritius

Mozambique

Mozambique

Namibia Niger

Namibia Niger

Rep Congo

Rep Congo

Rwanda

Rwanda

Senegal

Senegal

Sierra Leone

Sierra Leone

South Africa

South Africa Sudan Swaziland Togo Uganda Zambia Zimbabwe

References Abid, M., & Sebri, M. (2011). Energy consumption-economic growth nexus: Does the level of aggregation matter? International Journal of Energy Economics and Policy, 2(2), 55–62. Ali, H. S., Yusop, Z. B., & Hook, L. S. (2015). Financial development and energy consumption nexus in Nigeria: An application of ARDL bound testing approach. International Journal of Energy Economics and Policy, 5(3), 816–821. Al-Mulali, U., & Sab, C. N. B. (2012a). The impact of energy consumption and CO2 emission on the economic growth and financial development in the Sub Saharan African countries. Energy, 39(1), 180–186. doi:10.1016/j.energy.2012.01.032.

Al-Mulali, U., & Sab, C. N. B. (2012b). The impact of energy consumption and CO2 emission on the economic and financial development in 19 selected countries. Renewable and Sustainable Energy Reviews, 16(7), 4365–4369. doi:10.1016/j.rser.2012.05.017. Apergis, N., & Danuletiu, D. C. (2014). Renewable energy and economic growth: evidence from the sign of panel long-run causality. International Journal of Energy Economics and Policy, 4(4), 578–587. Apergis, N., & Payne, J. E. (2010). Renewable energy consumption and economic growth: evidence from a panel of OECD countries. Energy policy, 38(1), 656–660. doi:10. 1016/j.enpol.2009.09.002.

123

GeoJournal Apergis, N., & Payne, J. E. (2011). The renewable energy consumption—Growth nexus in Central America. Applied Energy, 88(1), 343–347. doi:10.1016/j.apenergy.2010.07.013. Ben Jebli, M., Ben Youssef, S., & Ozturk, I. (2014). The role of renewable energy consumption and trade: Environmental kuznets curve analysis for Sub-Saharan Africa Countries. Bhattacharya, M., Paramati, S. R., Ozturk, I., & Bhattacharya, S. (2016). The effect of renewable energy consumption on economic growth: Evidence from top 38 countries. Applied Energy, 162, 733–741. Bildirici, M. E. (2013). Economic growth and biomass energy. Biomass and Bioenergy, 50, 19–24. doi:10.1016/j. biombioe.2012.09.055. ¨ zaksoy, F. (2013). The relationship between Bildirici, M. E., & O economic growth and biomass energy consumption in some European countries. Journal of Renewable and Sustainable Energy, 5(2), 023141. doi:10.1063/1.4802944. Breitung, J. (2000). The local power of some unit root tests for panel data. In B. H. Baltagi, T. B. Fomby, & R. C. Hill (Eds.), Nonstationary panels, panel cointegration, and dynamic panels, advances in econometrics (pp. 161–178). Amsterdam: JAI Press. Chaudhry, I. S., Safdar, N., & Farooq, F. (2012). Energy consumption and economic growth: Empirical evidence from Pakistan. Pakistan Journal of Social Sciences, 32(2), 371–382. Feenstra, R. C., Inklaar, R., & Timmer, M., Penn, P. World table, version 8.0 2013. The next generation of the Penn world table. Accessed www.ggdc.net/pwt. Hausman, J. A. (1978). Specification tests in econometrics. Econometrica: Journal of the Econometric Society, 1251–1271. IEA. Energy balance for Africa (2009). Paris: OECD; 2009. Im, K. S., Pesaran, M. H., & Shin, T. (2003). Testing for unit roots in heterogeneous panels. Journal of Econometrics, 115(1), 53–74. doi:10.1016/S0304-4076(03)00092-7. Islam, F., Shahbaz, M., Ahmed, A. U., & Alam, M. M. (2013). Financial development and energy consumption nexus in Malaysia: A multivariate time series analysis. Economic Modelling, 30, 435–441. doi:10.1016/j.econmod.2012.09. 033. Jebli, M. B., Youssef, S. B., & Ozturk, I. (2016). Testing environmental Kuznets curve hypothesis: The role of renewable and non-renewable energy consumption and trade in OECD countries. Ecological Indicators, 60, 824–831. Levin, A., Lin, C. F., & Chu, C. S. J. (2002). Unit root tests in panel data: asymptotic and finite-sample properties. Journal of econometrics, 108(1), 1–24. doi:10.1016/S03044076(01)00098-7. Maddala, G. S., & Wu, S. (1999). A comparative study of unit root tests with panel data and a new simple test. Oxford Bulletin of Economics and Statistics, 61(S1), 631–652. doi:10.1111/1468-0084.0610s1631. Mohammed, Y. S., Mustafa, M. W., & Bashir, N. (2013). Status of renewable energy consumption and developmental challenges in Sub-Sahara Africa. Renewable and

123

Sustainable Energy Reviews, 27, 453–463. doi:10.1063/1. 4802944. Ozturk, I. (2015). Utilizing biofuels for sustainable development in the panel of 17 developed and developing countries. GCB Bioenergy. Ozturk, I., & Bilgili, F. (2015). Economic growth and biomass consumption nexus: Dynamic panel analysis for Sub-Sahara African countries. Applied Energy, 137, 110–116. doi:10.1016/j.apenergy.2014.10.017. Ozturk, I., & Salah Uddin, G. (2012). Causality among carbon emissions, energy consumption and growth in India. Ekonomska istrazˇivanja, 25(3), 752–775. Payne, J. E. (2011). On biomass energy consumption and real output in the US. Energy Sources, Part B: Economics, Planning and Policy, 6(1), 47–52. doi:10.1080/ 15567240903160906. Pedroni, P. (1999). Critical values for cointegration tests in heterogeneous panels with multiple regressors. Oxford Bulletin of Economics and Statistics, 61(S1), 653–670. doi:10.1111/1468-0084.0610s1653. Pedroni, P. (2004). Panel cointegration: Asymptotic and finite sample properties of pooled time series tests, with an application to the PPP hypothesis: New results. Econometric Theory, 20(3), 597–627. doi:10.1017/ S0266466604203073. Pesaran, M. H., Shin, Y., & Smith, R. P. (1999). Pooled mean group estimation of dynamic heterogeneous panels. Journal of the American Statistical Association, 94(446), 621–634. Pesaran, M. H., & Smith, R. P. (1995). Estimating long-run relationships from dynamic heterogeneous panels. Journal of Econometrics, 68(1), 79–113. doi:10.1016/03044076(94)01644-F. SERI, & Dittrich, M. Global material flow database. 2012 version. Can be accessed www.materialflows.net. Shahbaz, M., Hye, Q. M. A., Tiwari, A. K., & Leita˜o, N. C. (2013a). Economic growth, energy consumption, financial development, international trade and CO2 emissions in Indonesia. Renewable and Sustainable Energy Reviews, 25, 109–121. doi:10.1063/1.4802944. Shahbaz, M., Khan, S., & Tahir, M. I. (2013b). The dynamic links between energy consumption, economic growth, financial development and trade in China: Fresh evidence from multivariate framework analysis. Energy economics, 40, 8–21. doi:10.1016/j.eneco.2013.06.006. Stecher, K., Brosowski, A., & Thra¨n, D. (2013) Biomass potential in Africa. International Renewable Energy Agency (IRENA). Tugcu, C. T., Ozturk, I., & Aslan, A. (2012). Renewable and non-renewable energy consumption and economic growth relationship revisited: evidence from G7 countries. Energy economics, 34(6), 1942–1950. Uc¸an, O., Aricioglu, E., & Yucel, F. (2014). Energy consumption and economic growth nexus: Evidence from developed countries in Europe. International Journal of Energy Economics and Policy, 4(3), 411–419.