Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 © Scholarlink Research Institute Journals, 2015 (ISSN: 2141-7024) jetems.scholarlinkresearch.com Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016)

Impact of Exports and Imports on Economic Growth: Evidence from Tunisia Afaf Abdull J. Saaed and Majeed Ali Hussain Associate Professor of Applied Economics, College of Business Administration American University in the Emirates, United Arab Emirates Associate Professor of Econometrics, College of Business Administration American University in the Emirates, United Arab Emirates Corresponding Author: Afaf Abdull J. Saaed ________________________________________________________________________________________ Abstract In this paper we investigated the impact of exports and imports on the economic growth of Tunis over the period 1977-2012. The study used Granger Causality and Johansen Cointegration approach for long run relationship Using Augmented Dickey-Fuller (ADF) and Phillip-Perron (PP) stationarity test, the variable proved to be integrated of the order one 1(1) at first difference. Johansen and Juselius Cointegration test was used to determine the presence or otherwise of a cointegrating vector in the variables. To determine the direction of causality among the variables, at least in the short run, the Pairwise Granger Causality was carried out. Economic growth was found to Granger Cause import and Export was found to Granger Cause import. The results show that there is unidirectional causality between exports and imports and between exports and economic growth. These results provide evidence that growth in Tunisia was propelled by a growth -led import strategy as well as export led import. Imports are thus seen as the source of economic growth in Tunisia. __________________________________________________________________________________________ Keywords: Co integration, Granger causality; Exports; Imports; Economic growth; additional relevant variable, imports. Second, this study will use cointegration test to investigate for the presence of a long-run relationship. To investigate for cointegrating relationship between exports and economic growth, previous papers use mainly Engle and Granger (1987), Johansen (1988) and Johansen and Juselius (1990) approaches. However, this study will test of the impact of exports and imports on the economic growth of Tunis over the period 19772012.

INTRODUCTION It has been theoretically argued that both export and import may play a crucial role in economic development. The theoretical and empirical studies mainly concentrate on either the relationship between export and growth or between import and growth or the association between export, import and economic growth. The export-led growth hypothesis (ELGH) assumes that export advancement is one of the key indicators of growth. It encourages that the overall progress of countries can be achieved not only by mounting the quantity of manpower and investment within the economy, but also by increasing exports.

The aim of this paper, therefore, is to econometrically investigate direct linkages among trade and economic growth for Tunis. by employing yearly data for the period 1977-2012. In particular, this work tries to empirically find an answer for the question of whether export leads economic growth or of whether import lead economic growth or economic growth leads export and import.

Another relationship of causality from growth to export is called growth-led exports and it tells that there is unidirectional causality from economic growth to exports but not vice versa. There is also a possibility of two way causality link from exports to growth and from growth to exports.

To achieve this objective the paper is structured as follows. We discuss the Methodology, Model Specification and Data used in this study in Section 3. Section 4 presents the empirical results as well as the analysis of the findings. Section 5 provides our conclusion.

This paper contributes to the literature in the following ways. First, previous studies focus mainly on the interactions between exports and economic growth. Recognizing the role of imports on economic growth and possibly on export activities of a country, this paper empirically examine the relationship between exports and economic growth in a multivariate framework with the introduction of an

LITERATURE REVIEW THEORY AND APPLICATION The relationship between import, export and economic, has been a subject matter for a substantial 13

Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016) Carbajal, Canfield and De la Cruz (2008) examined both the existence of causality, in the Granger Sense, and its direction between Gross Domestic Product (GDP), Exports, Imports and Foreign Direct Investment in Mexico (FDI).

body of empirical work. Their nexus is usually investigated in the empirical literature in two different lines: The first line of the existing empirical research attempt to separately examine the importance of export or import on economic growth, the second line of the empirical works examines the relationship between export and import collectively. With regard to methods haven used to determine the importance of export and/or import to economic growth, there are two main methods. The first one employs simple or multiple regressions, while the second method employs the causality technique. Recently, most of studies have attended to focus on VAR and VEC models and cointegration approach. Our review of literature is limited to studies that focus on the joint impact of both export and import on economic growth.

Wong, (2008) examined the importance of exports and domestic demand to economic growth in ASEAN-5, namely Indonesia, Malaysia, the Philippines, Singapore and Thailand before Asia financial crisis, 1997- 1998. The results of the Granger causality test show some evidence of bidirectional Granger causality between exports and economic growth. A successful sustained economic growth requires growth in both exports and domestic demand. Moreover, economic growth will increase domestic demand and exports. There is no strong evidence to suggest that the export-led growth (ELG) strategy is a main cause to Asia financial crisis

Yuhong,Li and et. al. (2010) did co-integration analyses with the data of import, export and economic, and the results suggests that growth of import greatly promoted economic growth of China, while that of export performed an opposite one.

Ramos (2002) investigated the Granger-causality between exports, imports, and economic growth in Portugal over the period 1865-1998. The role of the import variable in the investigation of exports output causality is emphasized, enabling one to test for the cases direct causality, indirect causality, and spurious causality between export growth and output growth. The empirical results do not confirm a unidirectional causality between the variables considered. There is a feedback effect between exports output growth and imports output growth. More interestingly, there is no kind of significant causality between import export growths. Both results seem to support the conclusion that the growth of output for the Portuguese economy during that period revealed a shape associated with a small dual economy in which the intra-industry transactions were very limited.

Asafu-Adjaye et al (1999) consider three variables: exports, real output and imports (for the period 19601994). They do not find any evidence of the existence of a causal relationship between these variables for the case of India and no support for the ELG hypothesis, which is not too surprising given India’s economic history and trade policies Ullah et al (2009) investigated Export-led-growth by time series econometric techniques (Unit root test, Co-integration and Granger causality through Vector Error Correction Model) over the period of 1970 to 2008 for Pakistan. In this paper, the results reveal that export expansion leads to economic growth. They also checked whether there is uni-directional or bidirectional causality between economic growth, real exports, real imports, real gross fixed capital formation and real per capita income. The traditional Granger causality test suggests that there is unidirectional causality between economic growth, exports and imports. On the other hand Granger causality through vector error correction was checked with the help of F-value of the model and t-value of the error correction term, which partially reconciles the traditional Granger causality test.

Yuhong Li, Zhongwen Chen & Changjian San (2010), Research on the Relationship between Foreign Trade and the GDP Growth of East China. Empirical Analysis Based on Causality, Modern Economy. Hussain M and Saaed A.(2014) examined the nexus of Exports, Imports and Economic growth in Saudi Arabia, using annual data for the period 1990- 2011. Granger Causality and Cointegration test were employed in the empirical analysis. Both Trace and Maximum Eigenvalue indicated cointegration at 5% level of significance pointing to the fact that the variables have a long-run relationship. Also, economic growth was found to Granger Cause import. There was a unidirectional causality existing between export and import. But the result of the causation between Exports and economic growth and imports and economic growth was statistically insignificant.

Vohra (2001) tested the relationship between the export and growth in India, Pakistan, the Philippines, Malaysia, and Thailand for 1973 to 1993. The empirical results indicated that when a country has achieved some level of economic development than the exports have a positive and significant impact on economic growth. The study also showed the importance of liberal market policies by pursuing export expansion strategies and by attracting foreign investments. 14

Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016) The estimation methodology employed in this study is the cointegration and error correction modeling technique. The entire estimation procedure consists of five steps: first, unit root test; second, cointegration test; third, the error correction model estimation, forth Granger Causality and fifth VAR stability model.

Hussain M.(2014) examines the relationship between exports and GDP in Pakistan using, annual data collected from 1976 to 2011. Co integration and Granger causality test were employed in the empirical analysis, using Augmented Dickey Fuller stationarity test, the variable proved to be integrated of the order one (1) at first difference. The paper is based on the following hypotheses for testing the co integration and causality between GDP and export as to whether there is short run causality between GDP and export or whether there exists a long run association hip between GDP and Export. The co integration test indicating an existence of long run equilibrium relationship between the two as confirmed by the Johansen cointegration test results. The ECM estimates gave evidence that there is SR causality coming from GDP to Export. The findings indicate that there is unidirectional causality from GDP to exports in Pakistan but not vice versa

Model Specification This study examines the causal relationship among Economic growth, Export and Import, in Tunis. Granger-causality test in Vector Error Correction Model (VECM) framework is employed to examine causal relationship among Economic growth, Export and Import in Tunis. The primary model showing the causal relationship among Economic growth, Export and Import in Tunis can be specified thus: GDPt=f (export, import) (1) The function can also be represented in a log-linear econometric format thus: LGDPt=α + βLexportt +β1 Limportt + εt (2) Where: LGDP=LogGDP is economic growth as a proxy for Gross Domestic Product Lexport=LogEx . Limport=LogIm . α is the constant term, ‘t’ is the time trend, and ‘ε’ is the random error term assumed to be normally, identically and independently distributed. Here, GDPt, EXt and show the Gross Domestic Product export and import at a particular time respectively while εt represents the “noise” or error term; α and β1 and β2 represent the slope and coefficient of regression. The coefficient of regression, β1 and β2 indicates how a unit change in the independent variable (export and import) affects the dependent variable (gross domestic product). The error, εt, is incorporated in the equation to cater for other factors that may influence GDP. The validity or strength of the Ordinary Least Squares method depends on the accuracy of assumptions. In this study, the GaussMarkov assumptions are used and they include; that the dependent and independent variables (GDP ,EXPORT and IMPORT) are linearly co-related, the estimators (α, β1 and β2 ) are unbiased with an expected value of zero i.e., E (εt) = 0, which implies that on average the errors cancel out each other. The procedure involves specifying the dependent and independent variables; in this case, GDP is the dependent variable while EXPORT and IMPORT are the independent variable.

Hatemi (2002) studied causality between export growth and economic growth in Japan by performing augmented Granger-causality tests using the bootstrap simulation technique. The results show that the Granger-causality is bidirectional, which means the expansion of exports is an integral part of the economic growth process in Japan. However, they point to a causal relationship between international trade and exports and economic growth. Finally and crucially, for the purpose of this paper, the strong correlations of export, import and GDP growth rates has nothing to say about a relationship between the export (import) and the GDP trend development, as it may arise from a purely short-run relationship. In order to test for the existence of a long-run relationship among GDP, exports and imports, the theory of cointegration developed by Engle and Granger (1987). Johansen (1988) and Stock and Watson (1988), among others, has to be applied. To this end, we analyze annual data for Pakistan using a vector autoregressive (VAR) framework. DATA, METHODOLOGY SPECIFICATION

AND

MODEL

The Data The analysis used in this study cover annual time series of 1977 to 2012 or 36 observations which should be sufficient to capture the short run and long run correlation between Export, Import and economic growth in the model. The data set consists of observation for GDP, exports of goods and services (current US$), and imports of goods and services (current US$). All data set are taken from World Development Indicators 2014.

The paper is based on the following hypotheses for testing the causality and co-integration between GDP, Ex and IM. (i) whether there is bi-directional causality between GDP growth and export and Import, 15

Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016) (ii) whether there is unidirectional causality between the three variables, (iii) whether there is no causality between GDP and export and import in Tunis, (iv) whether there exists a long run relationship between GDP and EX and IM in Tunis.

autoregressive and distributed lag structures in the estimable VAR model. Therefore, lag of 1is used for estimation purpose. Table: 1 Unit Root Tests (ADF, PP) on LGDP, Lexport and Limport :1977-2012 Variables

EMPIRICAL ANALYSIS Tests for Integration This involves testing the order of integration of the individual series under consideration. Several procedures for the test of order of integration have been developed. The most popular ones are Augmented Dickey-Fuller (ADF) test due to Dickey and Fuller (1979, 1981), and the Phillip-Perron (PP) due to Phillips (1987) and Phillips and Perron (1988). Augmented Dickey-Fuller test relies on rejecting a null hypothesis of unit root (the series are nonstationary) in favor of the alternative hypotheses of stationarity. The tests are conducted without adeterministic trend for each of the series.

ADF

PP

Level 0 0.3216 0.0023*** 0.5231 0.0002*** 0.1668 0.0003***

LGDP Lexport Limport

∆

Level 0.2151 0.0026*** 0.7329 0.0001*** 0.1142 0.0003***

Order of integration

∆ I(1) I(1) I(1)

Note: (1) *** denotes significant at 1% level respectively. and in PP test it is based on NeweyWest using Bartlett kernel Source: Eviews version 8. Table 2: Lag Order Selection Criteria Lag 0

SC HQ -1.164252 -1.251547 1 142.3097 206.5526* 1.17e-07* -7.446269* -6.913007* 7.262187* 2 146.8471 7.259837 1.54e-07 -7.191263 -6.258055 -6.869120 3 152.4515 8.006324 1.93e-07 -6.997231 -5.664075 -6.537026 * indicates lag order selected by the criterion LR: sequential modified LR test statistic (each test at 5% level) FPE: Final prediction error AIC: Akaike information criterion SC: Schwarz information criterion HQ: Hannan-Quinn information criterion

At first this is important to know about the stationary properties of variables. Therefore, unit root tests applied to examine the null hypothesis of having unit root This involves testing the order of integration of the individual series under consideration. Several procedures for the test of order of integration have been developed. The most popular ones are Augmented Dickey-Fuller (ADF) test due to Dickey and Fuller (1979, 1981), and the Phillip-Perron (PP) due to Phillips (1987) and Phillips and Perron (1988). Augmented Dickey-Fuller test relies on rejecting a null hypothesis of unit root (the series are nonstationary) in favor of the alternative hypotheses of stationarity. The tests are conducted with intercept for each of the series. The general form of ADF test is estimated by the following regression

LogL 25.70744

LR NA

FPE 5.48e-05

AIC -1.297568

Source: Eviews version 8 The Error Correction Model If cointegration is proven to exist, then the third step requires the construction of error correction mechanism to model dynamic relationship. The purpose of the error correction model is to indicate the speed of adjustment from the short-run equilibrium to the long-run equilibrium state. The greater the co-efficient of the parameter, the higher the speed of adjustment of the model from the shortrun to the long-run We represent equation (2) with an error correction form that allows for inclusion of long-run information thus, the Error Correction Model (ECM) can be formulated as follows;

n

∆Υt =αo +α1Υt−1 +∑ΡiΥt−1 +εi −−−−−−−−−−−−−(3) i=1

Y is a time series, t is a linear time trend, ∆ is the first difference operator, αo is a constant, n is the optimum number of lags in the dependent variable and e is the random error term; and the Phillip-Perron (PP) is equation is thus: ∆Υt = α +

n

n

n

∆GDP t =∑αo∆GDP t−i +∑α1∆exportt−i + ∑α2∆import t−i +δ1EC1t−1 +ε1t

αΥt −1 + εt − − − − − − − − − − − − − − − − − − − −(4)

i=1

The results of Table 1 show that all variables are nonstationary in levels, but stationary in first difference. Since the variables are 1(1) the next step is to test if they are cointegrated using the Johansen full information maximum likelihood. It is clear from Table 2 that LR, FPE, AIC, SC, HQ and HQ statistics are chosen lag 1for each endogenous variable in their

i−1

(5)

i−1

where ∆ is the difference operator; n, is the numbers of lags, α 1 α 2 are short run coefficients to be estimated, EC1ti-i represents the error correction term derived from the long-run co integration relationship and ε 1t the serially uncorrelated error terms in equation (5). Table 5 shows that the result did not 16

Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016) conform to our prior expectation. The adjustment behavior of Export and Import have positive coefficient or the speed of adjustment of GDP is relationship in adjusting to long-run disequilibrium deviated from its long run equilibrium is EC term given the ECM value and are statistically not 0.018721 and P-value is 0.8695 (see appendix A1) significant. Thus, in the long run, the null hypothesis greater than 0.05 level of significant. Also the error is not rejected for all explanatory variables correction estimate equation shows that the long run Method: Least Squares Sample (adjusted): 1978 2012 Included observations: 35 after adjustments D(LGDP) = C(1)*( LGDP(-1) + 2.5201138114*LEXP(-1) 3.27375312934*LIMP(-1) - 6.1705633084 ) + C(2)*D(LGDP(-1)) + C(3)*D(LGDP(-2)) + C(4)*D(LEXP(-1)) + C(5)*D(LEXP(-2)) + C(6)*D(LIMP(-1)) + C(7) *D(LIMP(-2)) + C(8)

C(1) C(2) C(3) C(4) C(5) C(6) C(7) C(8) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

Coefficient 0.018721 0.202587 -0.103302 0.173745 -0.192307 -0.187897 0.281967 0.046553 0.123807 -0.103354 0.078100 0.164691 44.12020 0.545020 0.792705

Std. Error 0.112896 0.211243 0.227235 0.316506 0.250768 0.282502 0.229706

t-Statistic 0.165826 0.959025 -0.454602 0.548945 -0.766870 -0.665117 1.227510

Prob. 0.8695 0.3461 0.6530 0.5876 0.4498 0.5116 0.2302

0.036245 1.284381 0.2099 Mean dependent var 0.062308 S.D. dependent var 0.074353 Akaike info criterion -2.064011 Schwarz criterion -1.708503 Hannan-Quinn criter. -1.941290 Durbin-Watson stat 2.060426

Source: Eviews version 8

Table 5: Result of Granger Causality Granger Causality Results Since there is cointegration between the variables, the next step is to test for the direction of causality using the vector error correction model (VECM). The presence of a cointegrating vector allows for the use of a vector error correction model to test causality. The results of the Granger causality test are presented in Table 5 shows that the economic growth led to import. It is shown that economic growth Granger causes import. Also export granger causes import in Tunis. The results show that there is bi-directional causality between exports and imports and between economic growth and import but export does not Granger cause GDP. The coefficient of the lagged error correction term for all models is positive and not significant and this implies that there is no longrun causal relationship between exports and economic growth in Tunisia. These results provide evidence that growth in Tunisia was propelled by a growth -led import strategy. Imports are thus seen as the source of economic growth in Tunisia.

Pairwise Granger Causality Tests Date: 01/17/15 Time: 15:01 Sample: 1975 2012 Lags: 2 Null Hypothesis:

Obs

LEXP does not Granger Cause LGDP 36 LGDP does not Granger Cause LEXP LIMP does not Granger Cause LGDP 36 LGDP does not Granger Cause LIMP LIMP does not Granger Cause LEXP 36 LEXP does not Granger Cause LIMP

F-Statistic

Prob.

0.79877

0.4589

1.52459

0.2336

0.73490

0.4877

4.21667

0.0240

1.35081

0.2739

6.93138

0.0032

Source: Eviews version 8 Finally, we have to check the model efficiency, whether the model has ARCH affect, histogramnormal, serial correlation and heterscedasticity. First we check for histogram-normal, if Probability = pvalue >0.05, meaning that the residual is normal, so Jarque-Bera p-value=0.770 which is greater than 0.05,meaning that the residual is normally 17

Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016) distributed. Now we check for ARCH affect. We found that R2 probability = p-value=0.2901 which is greater than 0.05, meaning that we cannot reject HN, rather accept HN, meaning that there is no ARCH affect. Now we check for serial correlation. We run the autoregressive model with the dependent variable as independent variable with lag (-1), we find that the model has no serial correlation, when obs’ R2, pvalue =0.4029 which is greater than 0.05, we cannot reject HN, rather accept HN, meaning that this model does not have serial correlation. Finally, we check for Heteroscedasticity we find that the model free from heteroscedasticity when obs’R2 corresponding to pvalue=0.7533 greater than 0.05, meaning that the residuals are free from Heteroscedasticity. (See Appendix Table A3,A4 and A1 as well as figure A1). Figure 1 Figure 2

VAR Stability To confirm the stability of the estimated model, the tests of CUSUM and CUSUMSQ are employed in this study. Figure 1 and 2 respectively provide the graphs of CUSUM and CUSUMSQ tests. Figure 1and 2 indicates that the plot of CUSUM is completely stable within 5% of critical bands indicating the stability of VAR parameters. The test results show that the Modulus of all roots are less than unity and lie within the unit circle. Accordingly we can conclude that our model the estimated VAR is stable or stationary.

16

1.4 1.2

12

1.0

8

0.8

4

0.6

0

0.4

-4 0.2

-8

0.0

-12

-0.2

-16

-0.4 86

88

90

92

94

96

98

CUSUM of Squares

00

02

04

06

08

10

12

86

88

90

92

94

96 CUSUM

5% Significance

98

00

02

04

06

08

10

12

5% Significance

Note: The straight lines represent critical bounds at 5% significance level. Figure 1 and 2. Plot of Cumulative Sum of Squares of Recursive CUSUMSQ Residuals and CUSUM Source: Eviews version 8 VAR stability show that the Modulus of all roots are less than unity and lie within the unit circle. Accordingly we can conclude that our model the estimated VAR is stable or stationary.

CONCLUSION The aim of this study was to test Granger causality between export, import and GDP growth of Tunis during the period 1977-2012. The cointegration, error correction model and Granger's causality tests are applied to investigate the relationship between the export, import and GDP The unit root properties of the data were examined using the Augmented Dickey Fuller test (ADF) after which the cointegration and causality tests were conducted. The error correction models were also estimated in order to examine the short -run and long run between GDP and Exports.

The test of the model efficiency using Wald residuals statistics found that the model has no ARCH affect, the residual is normally distributed and the model does not have serial correlation and free from hetroscedasticity. REFERENCES Asafu-Adjaye, J and D Chakraborty. (1999), ‘Exportled Growth and Import Compression: Further Time Series Evidence from LDCs’, Australian Economic Papers, 38, pp. 164-75.

The finding is clarified that export, import and GDP are found stationary at the first differences.Therefore, the variables were found to be integrated of order one. The cointegration test confirmed that GDP ,export and import are cointegrated, indicating an existence of long run equilibrium relationship between all the variables under study confirmed by the Johansen cointegration test results.

Carbajal, E., Canfield, C., & De la Cruz, J. L. (2008). Economic Growth, Foreign Direct Investment and International Trade: Evidence on Causality in the Mexican Economy. Paper presented at the annual meeting of the BALAS Annual Conference, Universidad de los Andes School of Management, Bogota, D.C., and Colombia.2009-05-23.

The error correction models test confirmed that there exist short run causality between GDP and imports and between export and import. The Granger causality test finally confirmed the presence of unidirectional causality Unidirectional relationship between GDP to imports and between export and import, but not the other way. The test results of

Dickey, D. A. & W. A. Fuller (1981) “Likelihood ratio Statistics for autoregressive time series with a unit root,” Econometrica, 49(4):1057-72. 18

Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016) Dickey, D. A. & W. A. Fuller (1979), “Distribution of Estimators of Autoregressive Time Series with a Unit Root,” Journal of the American Statistical Association, 74, 427-31.

Ramos, F. F. R. (2002). Exports, imports, and economic growth in Portugal: evidence from causality and cointegration analysis. Economic Modeling, 18, 613-623

Engle, R. F. & Granger C. W. (1987), “Cointegration and Error Correction: Representation, Estimation and Testing,” Econometrica, 55, 251-276.

Stock, J. & Watson, M. (1988), “Testing for Common Trends,” Journal of the American Statistical Association, 83, 1097-107.

Hatemi, J. A. (2002), “Export Performance and Economic Growth Nexus in Japan: A Bootstrap Approach,” Japan and the World Economy, 14, 2533.

Vohra, R.(2001),“Export and Economic Growth: Further Time Series Evidence from Less Developed International Affairs and Global Strategy www.iiste.org ISSN 2224-574X (Paper) ISSN 22248951 (Online)

Hussain M (2014)” Export and GDP in Pakistan: Evidence from Causality and Cointegration Analysis” International Journal of Management Cases (IJMC).Vol.16, issue 1.

Ullah, Zaman, Farooq & Javid (2009), Cointegration and Causality between Exports and Economic Growth in Pakistan. European Journal of Social Sciences – Volume 10, Number 2.

Hussain M and Saaed Afaf (2014)”Relationship between Exports, imports, and economic growth in Saudi Arabia: 1990-2011. Evidence from cointegration and Granger causality analysis” Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 5(3):364-370.

Yuhong Li, Zhongwen Chen & Changjian San (2010), Research on the Relationship between Foreign Trade and the GDP Growth of East China— Empirical Analysis Based on Causality, Modern Economy, Vol. 1, pp. 118-124 Wong, Hock Tsen, (2008) Exports and Domestic Demand: Some Empirical Evidence In Asean.

Johansen, S. (1988), “Statistical Analysis of Cointegration Vectors,” Journal of Economic Dynamics and Control, 12, 231-54.

World Development (www.worldbank.org )

Phillips, P. C. B. & Perron, P. (1988), “Testing for a Unit Root in Time Series Regression,” Biometrika, 75(2), 335-46. APPENDIX A Table A1:Wald Test: Equation: Untitled Test Statistic

Value

df

Probability

F-statistic Chi-square

0.677308 1.354617

(2, 27) 2

0.5164 0.5080

Normalized Restriction (= 0)

Value

Std. Err.

C(4) C(5)

0.173745 -0.192307

0.316506 0.250768

Null Hypothesis: C(4)=C(5)=0 Null Hypothesis Summary:

Restrictions are linear in coefficients. Table A2: Wald Test: Equation: Untitled Test Statistic

Value

df

Probability

F-statistic Chi-square

1.219285 2.438570

(2, 27) 2

0.3112 0.2954 19

Indicator

(WDI),

2014

Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016)

Null Hypothesis: C(6)=C(7)=0 Null Hypothesis Summary: Normalized Restriction (= 0)

Value

Std. Err.

C(6) C(7)

-0.187897 0.281967

0.282502 0.229706

Restrictions are linear in coefficients. Figure A2: 7

Series: Residuals Sample 1978 2012 Observations 35

6 5

Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis

4 3 2

Jarque-Bera Probability

1 0 -0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

Table A3:Breusch-Godfrey Serial Correlation LM Test: F-statistic Obs*R-squared

0.684944 1.818214

Prob. F(2,25) Prob. Chi-Square(2)

0.5133 0.4029

Test Equation: Dependent Variable: RESID Method: Least Squares Date: 01/21/15 Time: 23:02 Sample: 1978 2012 Included observations: 35 Presample missing value lagged residuals set to zero. Variable

Coefficient

Std. Error

t-Statistic

Prob.

C(1) C(2) C(3) C(4) C(5) C(6) C(7) C(8) RESID(-1) RESID(-2)

-0.017742 0.736797 -0.093649 0.123791 -0.063780 -0.110722 0.106930 -0.047445 -0.815506 -0.042032

0.118662 0.763495 0.588704 0.339240 0.264374 0.310785 0.251563 0.078262 0.789282 0.668730

-0.149514 0.965032 -0.159077 0.364908 -0.241248 -0.356267 0.425061 -0.606232 -1.033225 -0.062854

0.8823 0.3438 0.8749 0.7182 0.8113 0.7246 0.6744 0.5498 0.3114 0.9504

R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

0.051949 -0.289349 0.079028 0.156136 45.05377 0.152210 0.997045

Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat

20

5.39E-17 0.069598 -2.003073 -1.558687 -1.849671 1.995920

5.39e-17 0.003055 0.159256 -0.134523 0.069598 0.073052 2.420648 0.520617 0.770814

Journal of Emerging Trends in Economics and Management Sciences (JETEMS) 6(1):13-21 (ISSN: 2141-7016) Table A4: Heteroskedasticity Test: Breusch-Pagan-Godfrey F-statistic Obs*R-squared Scaled explained SS

0.559193 5.865124 2.479278

Prob. F(9,25) Prob. Chi-Square(9) Prob. Chi-Square(9)

0.8168 0.7533 0.9814

Test Equation Dependent Variable: RESID^2 Method: Least Squares Date: 01/21/15 Time: 23:02 Sample: 1978 2012 Included observations: 35 Variable

Coefficient

Std. Error

t-Statistic

Prob.

C LGDP(-1) LEXP(-1) LIMP(-1) LGDP(-2) LGDP(-3) LEXP(-2) LEXP(-3) LIMP(-2) LIMP(-3)

0.140201 0.005139 0.002346 0.003188 -0.012919 -0.002412 0.017314 0.007524 -0.007036 -0.018717

0.128419 0.016134 0.020172 0.025019 0.023257 0.018235 0.025909 0.021142 0.025369 0.019071

1.091747 0.318497 0.116275 0.127403 -0.555479 -0.132250 0.668283 0.355880 -0.277361 -0.981443

0.2854 0.7528 0.9084 0.8996 0.5835 0.8958 0.5101 0.7249 0.7838 0.3358

R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

0.167575 -0.132098 0.006055 0.000916 134.9685 0.559193 0.816836

Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat

0.004705 0.005690 -7.141056 -6.696670 -6.987654 2.785236

Table A5:Heteroskedasticity Test: ARCH F-statistic Obs*R-squared

1.089349 1.119329

Prob. F(1,32) Prob. Chi-Square(1)

0.3044 0.2901

Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 01/21/15 Time: 23:03 Sample (adjusted): 1979 2012 Included observations: 34 after adjustments Variable

Coefficient

Std. Error

t-Statistic

Prob.

C RESID^2(-1)

0.005526 -0.181531

0.001279 0.173927

4.320132 -1.043719

0.0001 0.3044

R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

0.032921 0.002700 0.005766 0.001064 128.0830 1.089349 0.304437

Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat

21

0.004679 0.005774 -7.416649 -7.326863 -7.386029 2.030607