Emerging Trends in Trade in Turkey

TABLE OF CONTENTS

TABLE OF CONTENTS

INTRODUCTION

THE LINK BETWEEN TRADE OPENNESS AND HUMAN DEVELOPMENT IN TURKEY

Abstract

Introduction

Literature review

Methodology, data and variables

Results of the research

Conclusion

INTRODUCTION

The empirical research to date in general explores the link between trade liberalization and economic growth in the case of both, developed and developing countries. Most of the authors agree on the positive link between the variables of interest. This is since, outward-oriented economies consistently have higher growth rates compared to inward-oriented economies and thus tend to have a greater ability to catch up to leading technologies of the rest of the world. The most commonly used proxy measure of trade openness to date is the sum of exports and imports of goods and services divided by GDP. However, some of the authors have criticized these measures. In addition, most of the papers to date, analyze the macroeconomic role of the terms of trade by employing the regression analysis. This analysis is commonly used to explore whether or not one variable depends on other variables. However, it does not necessarily imply causation i.e. even though the regression analysis gives the evidence on the existence of the relationship between two variables, this is not enough evidence to prove the causality or the direction of the influence. In addition, it is important to emphasize that the macroeconomic role of trade openness is not explored quite intensively in the case of Turkey what was the motivation to conduct this research. This research aimed to fill in the gap in literature, by first, using the most recent data on the variables of interest. Besides that, two different measures of the trade openness are employed in order to test whether or not the results are sensitive to the selection of the proxy variable. In addition, we have employed the ARDL model to estimate whether or not there exits the link between the variables of interest. In addition to regression analysis, we have employed the VECM to provide the evidence on the causality. Thus, this book consists of the four essays treating the macroeconomic role of the trade openness. In the first essay, we have explored the trade openness-human development nexus by employing VECM in the period 1990-2017. The second chapter employs ARDL to display the evidence on the trade openness-education nexus. However, the second chapter was looking for the answer whether or not the better education can contribute to the trade openness. In the third chapter we have explored the link between trade openness and economic growth while employing the VECM model, and lastly we have explored the link between one of the fastest growing industries in the World (tourism industry) and trade openness. The research aims to give an important insight for the key policy makers in Turkey. It is suggested that in order to increase the trade openness, the policy makers need to make a necessary effort in order to promote the education and tourism that will lead economic growth and better statistics in terms of HDI. These essays have been written in the time-span 2017-2019.

Turkey, 2019

THE LINK BETWEEN TRADE OPENNESS AND HUMAN DEVELOPMENT IN TURKEY

Abstract

This research aims to explore the link, if any, between trade openness (TO) and human development (HDI) while analyzing the role of investment (GFC). Thus, the time-series data are collected in the time-span 1990-2017 in the case of Turkey. To increase the interpretability, the log values are calculated and used in research to follow. The variables are found to be stationary in the first difference. The findings of this paper suggest that a percentage change in HDI will results in a 1.16% increase in trade openness in the long-run. With regard to GFC, the coefficient is not found to be significant. The evidence on the long-run causality can be also inferred from to error terms that are negative and statistically significant ?n the case of TO. The short-run coefficients are not reported to be significant implying that the reaction of the trade openness on the increase in HDI is not instant and it takes a long for the trade openness to use the advantage of the increase in human development. The diagnostic tests advocate the stability of the model. This paper summarizes in detail the policy implications as well as the recommendations for future research.

Key words: human development, trade openness, Turkey, VECM

Introduction

One of the main characteristics of the last 20 years of 20th century is the effort of countries at the global level to open their borders for international trade. This is due to the fact that trade openness tends to significantly contribute to the economic growth. Trade liberalization tends to increase the efficiency of production process which is of crucial importance especially in developing economies. Thus many of the authors provide the various definitions of the trade openness. For instance, Squalli and Wilson (2006) define the trade openness as the share of exports in total income. Besides this, Alcala and Ciccone (2004) define the trade openness as the sum of export and import as the share of GNP. Following these definitions, this paper approximates the trade openness using the sum of exports and imports as the share of GDP in the case of Turkey.

With regard to Turkey, it is important to outline the fact that Turkey has reported an exponential increase in macroeconomic performance at the beginning of 21st century. The fiscal stability has increased the employment rate as well as the income what has resulted in the improvement in macroeconomic performance. Besides that, the extreme poverty has been almost eliminated. Consequently, the urbanization rate in Turkey has been increased and it has aimed to increase the inclusion in international trade and finance. As indicated above, the trade liberalization process has been in constant progress since the end of 20th century at the global level. This holds true for Turkey as well. It has implemented the growth strategy that is based on trade. Thus, Turkey has reported a tremendous economic growth and the increase in trade openness. Graph 1 gives a summary of the trade openness in Turkey in the period of interest.

Graph 1: Trade index

Abbildung in dieser Leseprobe nicht enthalten

Source: Author

The graph indicates an increase in the sum of exports and imports of goods and services divided by GDP. The increase is exponential in the period between 1990 and 2000. Due to the economic crisis in 2000, this increase has been stagnated. However, after the macroeconomic issues have been resolved, the trade index increases exponentially. Meschi et al. (2008) suggests that in 2014, Turkey was one of the leaders in terms of the trade index at the global level. The great macroeconomic performance did not only have a positive impact on trade openness, but also on the standard of living by increasing urbanization as well as the school enrollment. Thus, we explore the link between the proxy of standard of living (Human Development Index-HDI) and trade openness in the case of Turkey. Graph 2 outlines the growth of HDI in Turkey in the period 1990-2017.

Graph 2: HDI

Abbildung in dieser Leseprobe nicht enthalten

Source: Author

Kabadayi (2013) suggests that the link between trade openness and human development can be observed from the two points of view. At first, trade liberalization tends to increase the welfare in economies. This is due to the fact that open economies tend to specialize in the sectors they have a comparative advantage. Moreover, these countries are also expected to grow faster compared to closed economies. Thus, the impact of trade openness on the human development is actually indirect due to the fact that trade openness affects economic growth and consequently human development. On the other hand, the author suggests the direct link between the trade openness and human development. This is since international trade does not only exchange goods and service but also for human capital, ideas and technology enabling the citizens to choose the products they consider the best. Bumann et al. (2013) and Carrieri et al. (2013) have also displayed the positive link between trade openness and human development indicating that the trade openness is a prerequisite to the progress of society as a whole. Thus, Hill (2010) suggests that the liberalization acts like an invisible force stimulating not only the economic growth but also the human development of the economies worldwide. The supportive evidence to the link between trade openness and human development is also given by Sachs and Warner (1995). Taking the previous paragraphs into account it is important to emphasize that trade openness has attracted the great interest of researchers to date. Thus, various empirical researches have been conducted to explore whether or not the trade openness matters for economic growth. However, not many studies treat the link between human development and economic growth what was the motivation to conduct this study in the case of Turkey.

In the following parts of this paper we will present the recent literature on the trade openness-investment-HDI nexus. Moreover, we will present the variables used as well as the employed methodology. In the results section, we will report and interpret the findings of VECM. Lastly, we will provide concluding remarks together with the recommendations for future research in the last part.

Literature review

Mustafa et al. (2017) have explored the human development-trade openness-growth nexus for Asian countries. They have collected the annual panel data for twelve countries in the time span between 1970 and 2011. The findings of this paper suggest a great role of human development in economic growth. Besides that, the authors outline the great role of trade openness in both, economic growth and human development. The positive role of trade openness in growth process has also been recognized by Frankel and Romer (1999) and Wang et al. (2004).

Zahonogo (2016) has explored the link between trade liberalization and growth in the case of developing economies. Special attention is given to sub-Africa countries. He has collected the annual panel data in the period between 1980 and 2012 and has employed the PMG technique. The findings of this paper suggest a beneficial impact of trade openness on economic growth. However, author indicates that the link between variables of interest is not linear in the case of the observed economies. Thus, the author suggests the necessity to improve the effectiveness of the trade openness by controlling the import. The evidence on the positive link between the variables of interest is also given by Kim and Lin (2009) and Herzer (2013) in the case of developed economies. However, it is important to emphasize that trade openness can hardly contribute to the economic growth without macroeconomic stability and favorable investment climate (Newfarmer and Sztajerowska, 2012). Thus, the country cannot automatically use the benefit of trade openness. Benefits become certain in the long-run.

Hamid and Amin (2013) have investigated the link between trade openness and the social development in the case of OIC countries. They have employed the GMM technique. They have collected the annual panel data in the periods 1980-2005 and 2000-2009. The findings of this paper suggest a positive link between the variables of interest. This association is found to be highly reliant on the income level. The authors also indicate that the empirical evidence to date on the link between trade openness and human development is very scarce. For instance, Davies and Quinlivan (2006) and Gunduz et al. (2009) suggest a positive link between social development approximated using HDI and trade openness.

Peneva and Ram (2013) have explored the link between trade restrictiveness index and human development approximated using the six different proxy measures of the well-being. The trade restrictiveness index is the relatively new measure of trade openness developed by Kee et al. (2009). The results of this paper suggest no evidence on the negative impact of restrictive trade policy on human development. Hence, the authors emphasize the policy makers need to be careful while selecting the trade openness policy. These results are in accordance with Pritchett (1996) suggesting the low correlation between the variables of interest.

Kabadayi (2013) suggests that countries that are more open to international trade tend to be more effective in the allocation of the resources. Thus, this paper aims to answer the question whether or not the more inclusion in the international trade contributes to the society as whole. For this purpose, the author has collected the annual panel data in the time span 1995-2010. The empirical findings suggest the positive link between standard of living and trade openness in the sample of observed countries. These findings are supported by Nourzad and Powell (2003) in terms of developing countries in the period between 1965 and 1990.

With regard to Turkey, Gulaliyev et al. (2016) have explored the link between economic liberalization and human development. These results are compared in a great detail with those obtained for Azerbaijani economy. The findings suggest a positive link between the variables of interest. However, Azerbaijani economy is found to be less regulated and more open compared to Turkey. Taking into account these results, and the fact that both, trade liberalization and human development are increasing substantially in the case of Turkey, we expect the positive link between these two variables of interest.

Methodology, data and variables

Vector Error Correction Model (VECM) is the restriction of Vector Autoregressive Model (VAR). It enables the VAR model to be employed in the case of integrated variables. In theory VECM is considered to be the illustration of VAR model that is supposed to becointegrated. However, praxis requires the number of relationships (cointegrating) to be determined. VAR model assumes the variables to be covariance stationary. In addition, first two moments of variables need to be finite and cannot change with the time. Thus, if the variables do not meet the covariance stationarity criteria, but the first difference meets, such models are estimated using VECM. Thus, these models are well suited in the cases when variables in levels do not meet the stationarity criteria but their first difference does (Baum, 2013). The leading advantage of Vector Error Correction Model is that it enables the estimation and interpretation of short- and long-run equations.

Regression analysis is commonly used to explore whether or not one variable depends on other variables. However, it does not necessarily imply causation i.e. even though the regression analysis gives the evidence on the existence of the relationship between two variables, this is not enough evidence to prove the causality or the direction of the influence. Hence, to explore the causality we employ the VECM. The model requires the series to be integrated of the same order and the error terms to be uncorrelated. It is sensitive to the number of lags and asks for the evidence on cointegration. To derive a VECM, there is a need to difference a VAR. By differencing a VAR, one lag will be lost. Hence, when specifying the VECM, all of the variables have to take k-1 lags, while the dependent variables need to be specified with the difference operators. Thus, the model to be estimated in this paper can be formalized as following (Baum, 2013):

Abbildung in dieser Leseprobe nicht enthalten

The notes to support the specification can be presented as following (Baum, 2013):

- k-1 = the optimal number of lags reduced by one (due to the first differencing).
- ?_i = denotes the speed of adjustment; it always comes with the negative sign to show that in the long-run the model will reward to the equilibrium.
- ECT_(t-1) = is simply the residual that you obtain from the cointegrating regression and it contains the long-run information.
- u_it = represent the stochastic error terms or simply shocks.
- ?_i,?_j,?_m = represent the short-run coefficients.
- HDI = is the abbreviation for Human Development Index.
- TO = sum of exports and imports of goods and services divided by GDP.
- GFC = gross capital formation (% of GDP) is used as a proxy of investment.

Thus the step-by-step procedure can be summarized as following. It is first necessary to specify the model. Besides that, we will test for the stationary properties of the variables and determine the optimal number of lags by using the selection criteria. The next step will apply the Johansen cointegration test to test for the cointegration. If this is true, the restricted VAR (VECM) w?ll be employed. In addition to the VECM, the causality will be tested using the additional tests. Lastly, the diagnostic tests will be performed to check for the normality, autocorrelation and the stability of the model.

Results of the research

The empirical part of this research will be outlined following the steps explained in the methodological overview. Thus, this analysis displays first the most important measures of summary statistics in the Table 1.

Table 1: The description of the data

Abbildung in dieser Leseprobe nicht enthalten

Source: Author

The maximum value of the sum of export and import to GDP of 54.97% is recorded in the year 1997, while the minimum of 30.48% is reported in the year 1991. TO is found to be 45.23% on average. With regard to human development, it is important to advocate the tremendous increase in the period of interest. Thus, the maximum value of 79.1% is reported in the year 2017 while the lowest value of 57.9 is found in the year 1990. On average, human development index in Turkey equals 68.24%. At last, gross capital formation as a percentage of GDP equals 25.71% on average. The maximum value of 31.27% is reported in the year 2011 while the lowest value of 18.14% is reported in the year 2001. This result is quite expectable taking into account the economic conditions in Turkey in the year 2001.

To ease the comparison, the natural logarithmic transformation is employed. Thus, log levels instead of level variables are used in the research to follow. The research proceeds to the determination of the optimal number of lags. Using the commonly used selection criteria (AIC, HQIC and SBIC) the number of lags is suggested to be 1. This is expected taking into account the fact that this paper operates with the annual data. Two of the three criteria agree on the 1 lag and it is thus considered to be optimal (Table 2).

In the third step, we have tested whether or not the variables contain the unit-root. To check for the stationarity properties, we have employed the Augmented Dickey-Fuller (ADF) test in log levels and the first difference. Table 3 summarizes the obtained results.

[...]