RELATIONSHIP BETWEEN GDP AND EXPORT IN TURKEY

RELATIONSHIP BETWEEN GDP AND EXPORT IN TURKEY MÜCAHIT AYDIN RESEARCH ASSİSTANT, DEPARTMENT OF FİNANCİAL ECONOMETRİCS, FACULTY OF ECONOMICS AND BUSINESS ADMINISTRATION, SAKARYA UNIVERSITY, SAKARYA 54187, TURKEY MURAT SARI ASSOC. PROF. DR., PAMUKKALE UNIVERSITY, DEPARTMENT OF MATHEMATICS, DENIZLI 20070, TURKEY e-mail: msari@pau.edu.tr Abstract This article investigates the causal relationship between GDP and export of Turkey using Augmented Dickey Fuller (ADF) test and Granger causality test based on Toda-Yamamoto Analysis through EViews 7.1. To carry out the present study, the quarterly time series data reported by the OECD were used in the time period of 1980 2012. The results revealed that the series were not stationary.the obtined results concluded that there is a unidirectional causal relationship from the GDP to the export.the current study is expected to provide a better understanding of possible effects of the GDP in a country related to sustainable development in terms of the export. This work could also help in targeting beginner groups in economical society for economical modelling. Key words: Export, GDP, Turkish Economy, Toda-Yamamoto Test, Econometric Analysis JEL Classification: E60, F40, C10, C80 1. Introduction Causality between economic growth and export of empirical literature continues to be debated. The relationship between economic growth and export in developing countries has drawn the attention of researchers both to theoretically and empirically. There is a vast amount of empirical literature on this issue. Export provides a positive contribution to the economic growth in different ways such as; the removal of foreign exchange constraint (McKinnon, 1964), increasing in the technical informing (Grossman and Helpman, 1991), with the increasing competition to increase productivity (Bhagwati and Srinivasan, 1979; Krueger, 1980), simplifying the exploitation of economies of scale (Helpman and Krugman, 1985). Export directly affects economic growth through the increasing of production goods. In other words, increasing in both production and services will give rise to increasing export. Some analysts claim that causality flows from exports to economic growth and this causes to the export-led growth (ELG) hypothesis (Balassa, 1978; Bhagwati, 1978; Edwards, 1998). Expansion of export sales and opening up foreign markets with positive externalities, the basis for economic growth and export expansion is seen as a determinative parameter. The econometric relationship between export and economic growth in Turkey has been rarely analyzed as given by Takim (2010).The present differs from the corresponding study with the use of various data and the accepting various statistical approach.this study empirically investigates the relationship between economic growth and export through an econometric analysis over the period 1980q1-2012q4. 2. Export in Turkey Significant changes have been made for the export of Turkey in 1960. With these changes, Turkey followed 'import substitution policies' to export. This year is called the planned period. However, these plans have not become successful because, for many years, Turkey was influenced by both political and economic crisis. One of the most important issues on the corresponding topics was the embargo imposed by the United States to Turkey in 1974. In the time period 1980 83, Turkey was governed by a military system. At the beginning of this period, Turkey removed the import substitution policies in 1980 with 24 January decisions. In this period, the tight fiscal policy and low wages were observed to lead to relief of the deteriorated public balances. Controls on price controls and interest rates were gradually liberalized. The import regime was gradually liberalized. As underlined by Kandil et al. (2007), thus, the 1980s were the years of integration with the world economy for Turkey. Import was liberalized by 282

carring away quotas and decreasing the tariffs, thereby increasing the share of import to GDP. Until 1989, the economy grew well and high export growth showed to be the trigger element of growth. In the early 1990s, Turkey has experienced economic crises due to many internal and external factors. One of the crucial economic crises broke out in 1994. During the crisis, inflation, unemployment, and many macroeconomic indicators were deteriorated. Thus, the stabilization measures were applied to a series known as 5 April decisions. One of the most important of them hold high rate of devaluation. With this devaluation, the exportation of Turkey increased in 1995. Fluctuations in the world economy affected negatively the carried out export and the rate of increase in the export started to decline in 1996. To prevent this negative trend in 1996, the customs union agreement was signed with the EU. As a result of this agreement, Turkish export reachecd to $ 31 billion in 2001.The export rose to $ 85.5 billion at the end of 2006.To increase and sustain export, the export strategic plan' started to be applied in 2004. As a result of this plan, the export reached to $ 152 billion in 2012. 3. Literature Review The relationship between export and economic growth has been studied with the use of correlation techniques for developing countries. In this analysis, time series and cross sectional data were used (Michaely, 1977; Balassa, 1978; Tyler, 1981; Kavoussi, 1984). After those studies, the total production of export function is an additional input of labor and capital were examined in various studies (Moschos, 1989; Tyler, 1981; Feder, 1983; Ram, 1985; Burney, 1996; Balassa, 1978, 1985; Amirkhalkhali and Dar, 1995; Dodaro, 1991; Salvatore and Hatcher, 1991; De Gregorio, 1992; Greenaway and Sapsford, 1994; Rana, 1988). All the aforementioned studies showed that economic growth and export have meaningful relation. However, it is recognized that they do not address the issue of causality, while the cross-country regressions provide little insight into the way the various explanatory variables affect growth and the dynamic behaviours within countries (Giles and Williams, 2000). These deficiencies led to the recognition of all the studies which using time series analysis and causality tests (Jung and Marshall, 1985; Chow, 1987; Ahmad and Kwan, 1991; Bahmani-Oskooee et al., 1991; Dodaro, 1993). Contrary to common findings in the literature, some researchers (Akbar and Naqvi, 2000; Ahmad et al., 2000; Panas and Vamvoukas, 2002) could not find a causal link between export and economic growth. Various investigations of export in Turkey were carried out in the literature. Cointegration and error model were constructed using data from the study and ralized that two-way between export and economic growth has a positive and long-term relationship (Bahmani-Oskooee and Domac, 1995). Real exports of 26 countries with a GNP of variables in the study period 1953-1991, causal relationship is true for export of Turkey (Dutt and Ghosh, 1996). In the period of 1951-1990, a two-way causal relationship was found between real GDP and real exports in the study variables of Turkey (Xu, 1996). However, Granger causality test based on the analysis covers the period 1980-1996, Yigidim and Kose (1997) obtained results that do not support the hypothesis that export-led growth. After 1970, based on the seasonal data, Johansen cointegration technique does not support the hypothesis that the study results have export-led growth (Özmen and Furtun, 1998). In the periods of 1980-2003, Granger causality test was accepted to see the effects of Turkish total exporton economic growth. As a result of the analysis, a significance level of 5% of GDP and export growth was found to be a reciprocal relationship (Ay et al., 2004). In other words, export was found to be one of the main determinants of economic growth and it was also found to be an effective factor led to increase of the GDP. 4. Materials and Methods The data used in this study are for the variables the GDP and export, which are collected from the OECD, IMF and the Central Bank of the Republic of Turkey. The data set is quarterly and covers the period 1980q1-2012q4. In this study, the relationship between export and economic growth is examined using Augmented Dickey Fuller (ADF) test and the Granger causality test based on Toda-Yamamoto, but firstly we must check the time series properties of data in order to avoid spurious results. For this purpose, the ADF test is used to find unit roots if the variables are proven to contain unit roots.the ADF test is considered to realize that if the series are stationary or not. Note that the corresponding test determines the rankings of integration. Using the Granger causality test method based on Toda-Yamamoto is trying to determine the direction of the relationship between export and the GDP. 5. Technical Analysis The stability will be used in the analysis of the equation needs to be done in determining the stability of the series. Most of the macroeconomical data contain a unit root. In other words, the mean and variance of the series 283

containing a unit root varies over time. Therefore, analyzes of this series carried out using the established hypothesis test will be invalid. Stability in the time series can be classified as weak and strong stability. A stationary series has the property that the mean, variance and autocorrelation structure do not change over time. In practice, the weak stability is sufficient. The properties of a series are: where and stand for mean, variance and covariance, respectively. (1) (2) (3) As can be seen from figures 1-3 of the series, it was found that the figures fluctuate around atrend. Likewise ln(export) series also have an upward trend, and it has been shown that the two series are not stationary at the level. Fig.1. Export of Turkey (in billions of US $) Source: OECD stat Fig.2. ln(export) of Turkey (in billions of US $) Source: OECD stat 284

Fig.3. ln(gdp) of Turkey (in billions of US $) Source: OECD stat 5.1 Augmented Dickey Fuller (ADF) Test The most familiar stability tests for the detection arethe Dickey-Fuller and ADF retests. Standard Dickey-Fuller test is based on the assumption that the error terms are random and homogeneous. In the absence of these assumptions of the error term, this test is not used. When thenecessary assumptions related to error terms are not satisfied, the ADF test is used. Now then p-th order autoregressive process is given by, The term expression is obtained: is added to and removed from both sides of the previous equation the foolowing More details of the derivation of the regressive model can be found in Johnston and DiNardo (1997). To find out an acceptable way to determine the stability, one needs to apply unit root test. The most valid method to analyzethe stationary of a variable is to use the unit root test. To be able to determine either the static graphs Gina or the autocorrelation function is best way, the correlogram is considered. If the series does not fluctuate in a balanced way around a certain average of the series, then it is considered to be stationary (Gujarati, 2004: 802).The ADF test was applied to the series and the following findings were obtained. The 5% significance level was taken into account in the analysis. H 0 = have unit roots, H A = not have unit roots Prob-value is less than 0.05 at the 5% significance level.so H 0 hypothesis is rejected.namely the series does nothave a unit root, but prob-value is more than 0.05 at the 5% significance level, so H 0 hypothesis is not rejected. The series has a unit root. As understood from Tables 1-3, it could be noted that the series contain a unit root at the level so series are not stable at the level, but series does not contain a unit root at the first difference therefore series are stable at the first difference.on all models except the second of a series of export, unit roots have been identified. Looking at the series level of the figure has been found to contain the unit root. Table 1. ADF test: statistic and prob-value for ln(gdp) (4) (5) Table 2. ADF test: statistic and prob-value for ln(export) 285

Table 3. Granger Test based on Toda-Yamamoto CausalityTest: Statistic and prob-value 5.2 Toda Yamamota Causality Analysis Toda and Yamamoto (1995), to investigate the Granger causality, developed a method based on the estimated the adjusted VAR model. The degree of integration of the series or potential cointegration relationship does not affect the validity of this test (Yılancı and Özcan; 2010). The length of the delay of VAR model (k) and the maximum degree of integration of the series (d max ) are important points in this test. The VAR model is created with these two criteria (k+ d max ). The Granger causality test is done by adding constraints on the model parameters.the VAR model is estimated with the use of the following unrelated regression analysis: (6) (7) H : β = 0(The basic hypothesis is shown that variable does not Granger cause ) H1: β1 i 0 (The alternative 0 1i hypothesis is shown that variable Granger cause ) Hypotheses are tested with the Wald test; Chi-square distribution with k degrees of freedom used for this test. Table 3 shows Granger causality between export and GDP. From the result of the first equation, it could be noted that the null hypothesis that export does not Granger causes GDP and is accepted at a significance level of 5%. From the result of the second equation, it could be noted that the null hypothesis that GDP does not Granger causes exportand is not accepted at a significance level of 5%. The alternative hypothesis that GDP Granger causes export is accepted, implying that causality is running from GDP to export (GDP export). 6. Discussion and Conclusion In the literature, the quarterly data was used in very few studies that investigate the relationship between export and GDP. To best of authors knowledge, for the first time, this work has been carried out by using Toda- Yamamoto test based on the quarterly datareported by the OECD covering the period of 1980-2012. The paper tested the series for stationarity using the ADF test, and the variable proved to be integrated of the order one I(1) (1 degree cointegrated) at first difference and found that the series are non-stationary at the level but stationary at first difference. Before testing causality, the optimal lag was found to be 1 with Schwarz Information Criterion. Suitable VAR model was established and Toda-Yamamoto test was used to analyze the causal relationship between the series. The results of the test show that there is a unidirectional causal relationship from GDP to export. Note that export of Turkey has been found to be a variable that elucidates GDP as expected. Our findings are mainly in agreement with the literature on the relation between the export and the GDP. The causal relationship between GDP and export has been discovered using the ADF- Granger test based on Toda-Yamamoto Analysis through EViews 7.1. The results revealed that the series were not stationary. The presented 286

results figured out that there is a unidirectional causal relationship from the GDP to the export. It is believed that the present study could help in targeting beginner groups in economical society for analysing economical problems. Acknowledgements The corresponding author would like to express his sincere gratitude to Assoc. Prof. Dr. Cristina Gafu (Petroleum Gas University of Ploieşti), Prof. Dr. Catalin Popescu (Petroleum Gas University of Ploieşti), Prof. Marian Zaharia (Petroleum Gas University of Ploieşti) and Dr. Aniela Balacescu (Constantin Brancuşi University of Targu-Jiu) for their great hospitality during the international scientific symposium on Information Society and Sustainable Development in Romania. 7. References [1] Ahmad, J. and Kwan, A.C.C., Causality between Exports and Economic Growth: Empirical Evidence from Africa, Economics Letters, Vol. 37, 1991, pp. 243 8; [2] Akbar, M. and Naqvi, Z.F., Export Diversification and the Structural Dynamics in the Growth Process: The Case of Pakistan, The Pakistan Development Review, Vol.39, No.4, 2000, pp.573-589; [3] Amirkhalkhali, S. and Dar, A., A Varying coefficients Model of Export Expansion, Factor Accumulation and Economic Growth: Evidence from Cross-country, Time Series Data, Economic Modelling, Vol.12, No. 4, 1995, pp. 435 41; [4] Ay, A., Erdoğan, S. and Mucuk, M., On export-led growth in Turkey: A Causality Test (1980-2003), S.U.Karaman Journal of Economics and Administrative Sciences, Vol.4, No.1, 2004, pp.107-118; [5] Bahmani-Oskooee, M. and Domac, I., Export Growth and Economic Growth in Turkey: Evidence from Cointegration Analysis, METU Studies in Development, Vol.22, 1995, pp. 67-77; [6] Bahmani-Oskooee, M., Mohtad, H. and Shabsigh, G., Exports, Growth and Causality in LDCs: A Reexamination, Journal of Development Economics, Vol.36, issue 2, 1991, pp. 405 15; [7] Balassa, B., Exports and Economic Growth: Further Evidence, Journal of Development Economics, Vol. 5 issue2, 1978, pp. 181 189; [8] Balassa, B., Exports, Policy Choices, and Economic Growth in Developing Countries After the 1973 Oil Shock, Journal of Development Economics, Vol. 18, issue 1, 1985, pp. 23 35; [9] Bhagwati, J. and Srinivasan, T., Trade Policy and Development, Johns Hopkins University Press, 1978, pp. 1-38; [10] Bhagwati, J., Anatomy and Consequences of Exchange Control Regimes: Liberalization Attempts and Consequences, 1978; [11] Burney, N.A., Exports and Economic Growth: Evidence from Cross Country Analysis, Applied Economics Letters, Vol. 3, issue 6, 1996, pp. 369 73; [12] Chow, P.C.Y., Causality between Export Growth and İndustrial Development: Empirical Evidence from the NICs, Journal of Development Economics, Vol.26, issue 1, 1987, pp. 55 63; [13] De Gregorio, J., Economic Growth in Latin America, Journal of Development Economics, Vol. 39, issue 1, 1992, pp. 59 84; [14] Dodaro, S., Comparative Advantage, Trade and Growth: Export-led Growth Revisited, World Development, Vol.19, issue 9, 1991, pp. 1153 65; [15] Dutt, S.D. and Ghosh, D., The Export Growth-Economic Growth nexus: A Causality Analysis, The journal of Devoloping Areas, Vol.30, 1996, pp.167-182; [16] Edwards, S., Openness, Productivity and Growth: What Do We Really Know? Economic Journal Vol.108, issue, 447, 1998, pp. 383 398; [17] Feder, G., On Exports and Economic Growth, Journal of Development Economics, Vol. 12, issue 1-2, 1983, pp. 59 73; [18] Greenaway, D. and Sapsford, D., Exports, Growth, and Liberalization: An Evaluation, Journal of Policy Modelling, Vol. 16, issue 2, 1994, pp. 165 86; [19] Grossman, G.M. and Helpman, E., Innovation and Growth in the Global Economy, MIT Press, Cambridge, 1991; [20] Gujarati, D.N., Basic Econometrics. New York: McGraw-Hill Comp., 2004; [21] Helpman, E. and Krugman, P., Market Structure and Foreign Trade, MIT Press, Cambridge, 1985; [22] Isıgıcok, E., Causality in Time Series Analysis, Bursa: Uludag University Press, 1994; [23] Johnston, J. and DiNardo, J., Econometric Methods, Fourth Edition, McGraw-Hill, 1997; [24] Jung, S.W. and Marshall, P.J., Exports, Growth and Causality in Developing Countries, Journal of Development Economics, Vol. 18, issue 1, 1985, pp. 1 12; [25] Kavoussi, R.M., Export Expansion and Economic Growth: Further Empirical Evidence, Journal of Development Economics, Vol.14, issue 1, 1984, pp. 241 50; 287

[26] Krueger, A., Trade Policy as an İnput to Development. American Economic Review Vol. 70, No. 2, 1980, pp. 188 292; [27] McKinnon, R., Foreign Exchange Constraint in Economic Development and Efficient Aid Allocation. Economic Journal, vol.74, no.294, 1964, pp.388 409; [28] Michaely, M., Exports and Economic Growth: An Empirical İnvestigation, Journal of Development Economics, vol.4, 1977, pp. 49 53; [29] Moschos, D., Export Expansion, Growth and the Level of Economic Development: An Empirical Analysis, Journal of Development Economics, Vol.30, issue 1, 1989, pp. 93 102; [30] Ozmen, E. and Furtun, G., Export-Led Growth Hypothesis and the Turkish Data:An Emprical Investication, METU Studies in Development,vol.25, no.3, 1998, pp.491-503; [31] Panas, E. and Vamvoukas, G., Further Evidence on the Export-Led Growth Hypothesis,Applied Economics Letter, vol. 9, issue 11, 2002, pp.531-555; [32] Ram, R., Exports and Economic Growth: Some Additional Evidence, Economic Development and Cultural Change, vol. 33, no. 2, 1985, pp. 415 25; [33] Rana, P.B., Exports, Policy Changes and Economic Growth in Developing Countries After the 1973 Oil Shock, Journal of Development Economics, vol. 28, issue 2, 1988, pp. 261 4; [34] Salvatore, D. and Hatcher, T., Inward and Outward Oriented Trade Strategies, Journal of Development Studies, vol.27, issue 3, 1991, pp. 7 25; [35] Takim, A., Türkiye de GYSIH ile Ihracat Arasındaki İlişki: Granger Nedensellik Testi, Atatürk Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 2010, 14(2):1-16; [36] Tyler, W., Growth and Export Expansion in Developing Countries: Some Empirical Evidence, Journal of Development Economics, vol. 9, issue 1, 1981, pp. 121 30; [37] Xu, Z., On the Casuality between Export Growth and GDP Growth: An Emprical Reinvestigation, Review Internatıonal Economics, vol.4, issue 2, 1996, pp.172-184; [38] Yilanci, V. and Ozcan, B., Analyzing the Relationship Between Defense Expenditures and GNP for Turkey Under Structural Breaks, C.U. Journal of Economics and Administrative Sciences, vol. 11, No. 1, 2010; [39] Yigidim, A. and Kose, N., Relationship between GDP and Export in Turkey, Role of Imports: (1980-1996), Economic Approach, C.8, S.26, 1997, pp.71-85. 288