Impacts of Trade Intensity in China and India on the Aggregated Sub-Saharan Africa Economic Growth

Transcription

1 Impacs of Trade Inensiy in China and India on he Aggregaed Sub-Saharan Africa Economic Growh Bédia F. Aka 1 rapca Third Annual Conference 2008 Srenghening and deepening economic inegraion in LDCs: curren siuaion, challenges and way forward 1 Deparmen of Economics and Lampe (Laboraoire d analyse, de modélisaion e de prospecive économique), Cenre de Recherche pour le Développemen, Universiy of Bouaké, Côe d Ivoire. akbdia@yahoo.fr

2 Impacs of Trade Inensiy in China and India on he Aggregaed Sub-Saharan Africa Economic Growh Absrac Recen years have shown he emergence of wo major counries, China and India in he inernaional rade area. This paper aims o examine how he increasing par of hese wo counries in rade impacs he African coninen. The inensiy of rade is measured by he share of expors plus impors in he gross domesic (GDP) while economic growh in he Sub-Saharan Africa is proxies by he GDP per capia. We focus on Sub-Saharan Africa as a whole during he period o see how rade inensiy in China and India and in Sub-Saharan Africa as well has impaced he GDP per capia growh rae in ha African zone. For ha purpose firs we use an Auoregressive disribued lag (ARDL) Time series model o analyze boh he shor and long-run impacs of hese variables on he GDP per capia growh rae of he aggregaed SSA. Secondly using a VAR we perform impulse response analysis o see he impacs of rade inensiy in China, India and Sub-Saharan Africa on he per capia growh in he Sub-Saharan Africa. JEL Classificaion: C22, C32, C51, F13 Keywords: Economic Growh, Trade, Time Series, ARDL Model, VAR Model. 1. Inroducion Two major counries, namely China and India, have emerged in he inernaional rade during he recen years wih a large share of world rade, in expors as well as in impors. More imporanly here is a debae of he increasing par of hese wo gian counries on he African coninen as hey need o impor raw maerial from Africa and expor manufacured goods o Africa coninen and o oher pars of he world.

3 Abou China, Zafar (2007) indicaes ha: China's economic ascendance over he pas wo decades has generaed ripple effecs in he world economy. Is search for naural resources o saisfy he demands of indusrializaion has led i o Sub-Saharan Africa. Trade beween China and Africa in 2006 oaled more han $50 billion, wih Chinese companies imporing oil from Angola and Sudan, imber from Cenral Africa, and copper from Zambia. Demand from China has conribued o an upward swing in prices, paricularly for oil and meals from Africa, and has given a boos o real GDP in Sub-Saharan Africa. In he same ime China and India inves wih someimes rivalry in several secors of Africa s economy (mining, hospial, banking, and elecommunicaion), creaing jobs, and providing aid all over he coninen. A number of auhors have examined specific counries cases or specific quesions or goods. Jenkins and Edwards (2004) invesigaed how China s growh affecs povery reducion in Asia, Africa and Lain America, while Kaplinsky (2006) deals wih erm of rade, indicaing ha he enry of China ino he global marke has played an imporan role in he augmening he demand for many commodiies (as meals). Sreifel (2006) looked a he impac of China and India on global commodiy markes focusing on meals, minerals and peroleum. Jenkins and Edwards (2006) examined he impac of he emergence of China and India as imporan players in he global economy on 21 sub-saharan African counries. They asser ha he ways in which Africa has been affeced differs from counry o counry, wih some such as Angola, Nigeria and Sudan being imporan exporers, ohers such as Ghana, Ehiopia, Kenya, Tanzania and Uganda mainly imporers from Asia, and Lesoho facing compeiion from China in expor markes. Few works have specifically examined he impac of hese growing counries on he aggregaed Sub-Saharan African zone. Here we inend o examine empirically how rade inensiy ino hese o Gians impacs he Sub-Saharan Africa growh by looking a he effecs of he growh in he GDP

4 per capia of hese wo counries on rade and growh in he aggregaed Sub- Saharan Africa. This paper uses alernaively firs an Auoregressive disribued lag (ARDL) Time series model o analyze boh he shor and long-run impacs of hese variables on he GDP per capia growh rae of he aggregaed and secondly a VAR model o perform impulse response analysis o see he impacs of rade inensiy in China, India and SSA on he per capia growh in he SSA Our empirical findings indicae ha for he aggregaed Sub-Saharan Africa area he rade inensiy of he wo emerging counries, namely China and India have differenial shor and long-run impacs on he GDP per capia growh rae in he aggregaed Sub-Saharan Africa. We find ha rade inensiy in China is no significan in he shor-run while an increase in he Indian rade inensiy posiively impacs he shor-run GDP per capia growh in he Sub-Saharan Africa, and rade inensiy in China negaively impacs he GDP per capia growh rae of he Sub-Saharan Africa. Concerning he impac of Trade inensiy on he GDP per capia he wo models (he ARDL and he VAR) deliver conflicing resuls for China and Sub-Saharan Africa while he resuls of he wo models are concordan for India. Regarding he effec of he variables on he rade inensiy in he Sub- Saharan Africa, he VAR model indicaes negaive impacs in he shor and long-run, excep for he impac of he GDP per capia of Sub-Saharan Africa on he rade inensiy in he Sub-Saharan Africa. Overall he findings sugges ha he rade inensiy of he wo emerging counries, namely China and India have differenial shor and long-run impacs on he GDP per capia growh rae in he aggregaed Sub- Saharan Africa.

5 The remaining of he paper organized as follows. Secion 2 presens he model and he economeric mehodology used in his sudy. Secion 3 presens he esimaions resuls and final remarks are provided in Secion The Model Specificaion The ARDL model We adop a simple model aiming o esimae he following relaionship where he GDP per capia of he Sub-Saharan Africa (SSA) is explained by rade inensiy in China, India and SSA: GDP = f ( TI Chin,, TI Ind,, TI ) + ζ (1) In his equaion he variable GDPSSA denoes he per capia GDP of he while TIChin, TIInd and TISSA indicae rade inensiy as measured by he share of expors and impors in GDP, for China, India and SSA respecively, ζ is a sochasic error erm, and is year index. Expressing Eq. 1 in a naural logarihm form gives he following long run equaion: ln GDP = α + γ lnti + γ lnti + γ lnti, ) + ζ 1 Chin, 2 Ind, 3 SSA (2) The assumpions concerning he preceding Eq. 2 are ha rade inensiy variables are supposed o posiively impac he GDP per capia of he SSA. The esimaion will be performed using an unresriced general o specific Hendry (1995) ype error correcion model (ECM) where he long-run

6 relaionship is embedded wihin he dynamic specificaion, including he lagged dependen and independen variables as follows: Δ lngdp = α + γ Δ lnti + γ (ln GDP 4 1 Chin, SSA + β lnti 2 β lnti Ind + γ Δ lnti 1 2 Chin + β lnti 3 Ind, ) SSA 1 + γ Δ lnti 3 + ν (3) This model is re-parameerized in he following esimable form: Δ lngdp = α + γ Δ lnti + γ lngdp 4 + γ lnti 6 1 Chin, 1 Ind, 1 + γ Δ lnti + γ lnti + γ lnti Chin, 1 1 Ind, + ν + γ Δ lnti 3 (4) Alhough his model helps o explain he link beween he variables by expressing he impacs of he exogenous righ-hand side variables on he endogenous lef-hand side variable, i does no give any cross relaionships beween all hese variables. To allow for iner relaionships analyses a VAR model appears more suiable. The VAR Model Mehodologically wihou a formal explicaive model he VAR models appear an appropriae alernaive o esimae he relaionships beween variables in presence (see Bagliano and Favero, 1998 ; and Chrisiano e al., 2005). The VAR model of 6 variables is expressed as followed: [ GDPpc GDPpc, GDPpc, TI TI TI ] ' X = (5), Chin, Ind,, Chin, Ind, wih GDPpci, represening he GDP per capia of China and India a ime, and TIi, is he rade inensiy in China and India.

7 Nex rade inensiies and he GDP per capia in China and India are modelled in Eq. 6 bellow as a funcion of he informaion I, and a rade shock ε orhogonal o he informaion conained in I. Y = f ( I) + ε i, (6) Hypohesis: GDPpcSSA does no respond conemporaneously o shocks ε. The VAR conains only one lag due o daa consrains, and has he following general form: X = AX 1 + Cη (7) The variable C in Eq. 7 represens a riangular 6x6 marix wih diagonal elemens equal one, and η is a vecor of dimension 6 wih mean zero. The shocks he vecor η. ε are he 2 nd, he 3 rd, 4 h, he 5 h and he 6 h elemens of This represenaion is useful o es he dynamic response of he GDP per capia in SSA o rade inensiy and GDP shocks. These responses will be performed using he impulse response funcions of he VAR. 3. Daa and Resuls Daa We esimae he model using daa from World Developmen Indicaors (WDI 2007) and covering he period Figure 1 shows he series in level and Figure 2 in firs difference (growh rae).

8 The variable LGDPpc indicaes he naural logarihm of he Gross domesic produc per capia and LTI indicaes he naural logarihm of rade inensiy as measured by he share of expors plus impors in GDP. From Figure 1 we could see ha populaion in he aggregaed SSA is lower han in boh China and India. While he GDP per capia of he Sub- Saharan Africa was higher han hose of China and India from 1969 o 1991, he GDP per capia of China overcomes hose of he Sub-Saharan Africa and India since 1991, and he GDP per capia in India overcomes he one of SSA from The rade inensiy of he Sub-Saharan Africa is higher han boh of China and India all over he period, bu since 2001 rade inensiy in China is he highes, overcoming he Sub-Saharan Africa one. The growh rae of he GDP and GDP per capia in China is higher han hose of India and Sub- Saharan Africa. Figure 1: Trade as percenage of GDP; GDP (Consan 2000 US$); Populaion; and GDPpc (Consan 2000 US$) 60 TI_CHN TI_IND TI_SSA 1.5e12 GDP_CHN GDP_IND GDP_SSA 40 1e e e9 1e9 POP_CHN POP_SSA POP_IND 1500 GDPpc_CHN GDPpc_IND GDPpc_SSA e8 5e e8

9 Figure 2: Trade as percenage of GDP; GDP (Consan 2000 US$); Populaion; and GDPpc (Consan 2000 US$); Growh Rae DLTI_CHN DLTI_SSA DLTI_IND DLPOP_CHN DLPOP_IND DLPOP_SSA DLGDP_CHN DLGDP_SSA DLGDPpc_CHN DLGDPpc_SSA DLGDP_IND DLGDPpc_IND Economeric Resuls - The ARDL Model The final esimaed dynamic ECM of Eq. 4 is repored in Table 1 along wih he diagnosic ess and he long-run elasiciies of he variables and heir - raio regarding which all variables are significan. Figure 3 depics he esimaes and residuals. The dynamic ECM equaion is repored as follows: ΔlnGDP = Δ lnti 0.095lnGDP Ind, Δ lnti lnTI Chin, lnti Ind, 1 (8) This equaion indicaes ha in he shor run, a 1% increase in he rade inensiy of India enhances he Sub-Saharan Africa s GDP per capia growh by 0.055%, while a 1% increase in he rade inensiy of he Sub-

10 Saharan Africa leads o 0.14% increase in he Sub-Saharan Africa growh. The coefficien of rade inensiy in China appears insignifican in he esimaions meaning ha rade inensiy in China has no shor run significan impac on he Sub-Saharan Africa s GDP per capia growh. The coefficien of he GDP per capia (-0.095) in Eq. 5 indicaes he speed of he adjusmen of he sysem o he long-run. The long-run equaion derived from he preceding dynamic ECM in Eq. 5 is he following: ln GDPpc 0.31lnTI Ind, 0.10 lnti Chin, = (9) In his long-run equaion he rade inensiy in India has a posiive sign expressing hus he fac ha he rade inensiy in India does have a posiive impac on he Sub-Saharan Africa s GDP per capia growh; whereas he rade inensiy in China has a negaive sign indicaing ha rade inensiy in China is no beneficial o he long run GDP per capia growh in he aggregaed Sub-Saharan Africa area. The rade inensiy in he Sub- Saharan Africa is no significan in he long run. Table 1: Model esimaion, OLS 1971 o 2005 (Dependen variable: DLGDPpc_SSA) 2 Coefficien Sd. Error -value Consan DLTI_IND DLTI_SSA LGDPpc_SSA_ LTI_CHN_ LTI_IND_ DUM sigma RSS R^ F(6,28) = [0.001]** log-likelihood DW 1.51 no. of observaions 35 no. of parameers 7 Noe: LGDPpc_SSA_1 indicaes he log of GDPpc_SSA one period lag. DUM92 is a dummy variable. 2 Esimaions are performed using PcGive 10 (Hendry and Doornik 2001).

11 Figure 3: Esimaes and Residuals DLGDPpc_SSA Fied r:dlgdppc_ssa (scaled) The VAR Model The impulse responses from he esimaion of he VAR model are depiced in he following Figure 4. The responses of he GDP per capia in he SSA o a shock on Trade inensiy in he Sub-Saharan Africa, in China and in India are repored in hese Figures. We could see ha following a shock on he rade inensiy in he Sub- Saharan Africa, he GDP per capia of he Sub-Saharan Africa increases immediaely (one year period) and decreases in he shor run (2 years), and hen increases in he long run (afer 3 years), while a shock on Trade inensiy in China leads o an increase of he GDP per capia of he Sub- Saharan Africa over 9 years before decreasing during 10 years and increasing in he long run. On he oher hand, a shock on he rade inensiy in India has a posiive growing impac on he GDP per capia of he Sub- Saharan Africa during 3 years before declining during 6 years. In he long run (afer 9 years) he shock is posiive.

12 Figure 4: Response of he SSA per Capia Growh o Trade and GDPpc Shocks LTI_SSA (LGDPpc_SSA eqn) LTI_CHN (LGDPpc_SSA eqn) 0.04 LTI_IND (LGDPpc_SSA eqn) LGDPpc_CHN (LGDPpc_SSA eqn) LGDPpc_IND (LGDPpc_SSA eqn) 0.01 Figure 5: Response of Trade Inensiy o he SSA per Capia GDP Shock LGDPpc_SSA (LTI_SSA eqn) 0.00 LTI_CHN (LTI_SSA eqn) LTI_IND (LTI_SSA eqn) LGDPpc_IND (LTI_SSA eqn) LGDPpc_CHN (LTI_SSA eqn)

13 A shock on he GDP per capia in he Sub-Saharan Africa leads o a 2 years increase in rade inensiy in he Sub-Saharan Africa and a decrease in he long run. A shock on he GDP per capia in China and India has a decreasing effec on he rade inensiy in he Sub-Saharan Africa. A shock on he rade inensiy in China has a decreasing effec on he shor-run Sub- Saharan Africa s rade inensiy (2 years), and increasing effec over 6 years and a decreasing impac in he long run. A shock on he rade inensiy in India has a decreasing effec on he Sub-Saharan Africa s rade inensiy in he shor and long-run. Table 2: Models Resuls Comparison Model 1% Increase in Shor-run Long-run Impac on GDPpcSSA TISSA Increase (+) non significan ARDL TIChin non significan Decrease (-) TIInd Increase (+) Increase (+) TISSA Decrease (-) Increase (+) TIChin Decrease (-) Increase (-) VAR TIInd Increase (+) Increase (+) Impac on Trade-InensiySSA GDPpcChin Increase Decrease GDPpcInd Decrease Decrease TIChin Decrease Decrease TIInd Decrease Decrease 4. Final remarks This paper examined he impac of rade inensiy in China and India proxies by he share of impors plus expors in he GDP per capia growh in he Sub-Saharan Africa over he period These relaionships are analyzed using firsly an auoregressive disribues lag ECM approach, and secondly using impulse responses from a VAR model.

14 We find diverse impacs of he Chinese and Indian rade inensiy on he Sub-Saharan Africa s GDP per capia growh. While rade inensiy in China is no significan in he shor run, an increase in he Indian rade inensiy posiively impacs he shor run GDP per capia growh in he Sub- Saharan Africa, while he rade inensiy in China negaively impacs he shor run Sub-Saharan Africa s GDP per capia growh rae. Concerning he impac of rade inensiy on he GDP per capia he wo models (ARDL and VAR) deliver conflicing resuls for China and Sub- Saharan Africa while he resuls are concordan for India. In he case of he effec of he variables on he rade inensiy in he Sub-Saharan Africa, he VAR model indicaes negaive impacs in he shor and long-run, excep for he impac of he GDP per capia of sub-saharan Africa on he Trade inensiy in he Sub-Saharan Africa. The findings sugges ha he rade inensiy of he wo emerging counries, namely China and India have differenial shor and long-run impacs on he GDP per capia growh rae on he aggregaed SSA. The inensiy of rade, which is an aspec of he increasing globalizaion phenomenon of Sub-Saharan Africa economies and China and India as well, does have a significan and beneficial long-run impac on he GDP per capia in his zone. In he long-run, he impac of he variables on he rade inensiy will decrease. The beneficial long-run aspecs of he rade inensiy in he Sub-Saharan Africa, China and India on growh in he Sub- Saharan Africa should be considered by policy markers concerned wih new growh sraegies in he Sub-Saharan Africa. References Bagliano, F.C. and C.A. Favero (1998) Measuring Moneary Policy wih VAR Models: An Evaluaion, European Economic Review 42, pp

15 Chrisiano, L.J., M. Eichenbaum and C.L. Evans (2005) Nominal Rigidiies and he Dynamic Effecs of a Shock o Moneary Policy, Journal of Poliical Economy 113, pp Hendry, D.F. (1995) Dynamic Economeric, Oxford Universiy Press, Oxford. Hendry, D.F. and J.A. Doornik (2001) Empirical Economeric Modeling Using PcGive, Timberlake consulan Ld. Jenkins, R. and C. Edwards (2006) The economic impacs of China and India on sub-saharan Africa: Trends and prospecs, Journal of Asian Economics, pp Jenkins, R. and C. Edwards (2004) How does China s growh affec povery reducion in Asia, Africa and Lain America? Mimeo, Norwich: Universiy of Eas Anglia, Overseas Developmen Group. Kaplinsky, R., 2006, Revisiing he revisied erms of rade: will China make a difference? World Developmen 34, No. 6, pp Zafar, A. (2007) The Growing Relaionship Beween China and Sub-Saharan Africa: Macroeconomic, Trade, Invesmen, and Aid Links, The World Bank Research Observer 2007, World Bank. Sreifel, S. (2006) Impac of China and India on Global Commodiy Markes Focus on Meals & Minerals and Peroleum, Developmen Prospecs Group, World Bank.