Türkiye’de Ekonomik Büyüme ve Dış Ticaret Arasındaki Nedensellik İlişkisi Zamanla Değişiyor mu?
Gülçin Kendirkıran, Furkan EmirmahmutoğluBu çalışma, Türkiye’de ekonomik büyüme, ihracat büyümesi ve ithalat büyümesi arasındaki ilişkiyi zamanla değişen nedensellik yöntemleriyle araştırmaktadır. Nedensellik ilişkisinin incelenmesinde çıkış noktası Granger’ın 1969 yılındaki çalışmasıdır. Granger (1969) nedensellik çalışması, örneklem dönemini bir bütün olarak ele aldığından dolayı yapısal değişimleri veya kırılmaları dikkate almamaktadır. Bu sebeple, örneklem aralığında yapısal kırılma veya değişimler meydana geldiğinde geleneksel Granger nedensellik analizi değişen nedensellik ilişkilerini belirleyememektedir. Dolayısıyla bu çalışmada zamanla değişen parametreli ve tekrarlamalı gelişen pencereli nedensellik yöntemlerinden yararlanılmaktadır. Öncelikle örneklemin tamamı için uygulanan geleneksel Granger nedensellik analizi sonucunda ithalat büyümesi ve ekonomik büyüme arasında çift yönlü nedensellik ilişkisine ulaşılmaktadır. Ardından dirençli istatistiklere sahip zamanla değişen parametreli nedensellik ilişkisi incelenmektedir. Son olarak, değişen varyans altında tutarlı tekrarlamalı gelişen pencereli nedensellik yöntemi değerlendirilmektedir. Elde edilen sonuçlardan bahsedilen nedensellik ilişkilerinin parametreleri zamanla değişmektedir. Aynı zamanda, ekonomik büyümeden ihracat büyümesine ve ithalat büyümesine doğru, ihracat büyümesinden ve ithalat büyümesinden ekonomik büyümeye doğru belirli kriz dönemlerine karşılık gelen nedensellik ilişkileri tespit edilmektedir. Hem zamanla değişen parametreli hem de tekrarlamalı gelişen pencereli nedensellik yöntemleri, Türkiye’de yapısal değişimler meydana geldiğinde bu etkinin değişkenler arasındaki nedensellik ilişkisine yansıdığını doğrulamaktadır. Sonuç olarak bu çalışmada her iki nedensellik yöntemi, geleneksel Granger nedensellik yönteminin eksik yönlerini belirlemek ve Türkiye’de nedensellik ilişkilerinin değiştiği dönemleri tespit etmek için kullanılmaktadır.
Does Change over Time the Causal Relationship between Economic Growth and Foreign Trade in Turkey?
Gülçin Kendirkıran, Furkan EmirmahmutoğluThis paper investigates the relation between economic growth, export growth, and import growth in Turkey using time-varying causality methods. The starting point in examining the causality relationship is Granger’s study in 1969. Granger’s(1969) causality study does not considerstructural changes or breaks because it examinesthe sample period as a whole. For this reason, traditional Granger causality analysis cannot determine changing causality relations when structural changes or breaks occur in the sample period. Therefore, this study uses the time-varying parameters and the recursive evolving window causality methods. First, a traditional Granger causality analysis applied for the entire sample period shows a bidirectional causality relationship between import growth and economic growth. Subsequently, the paper examines the time-varying parameters causality method that has robust statistics. Finally, it evaluates the consistency under the heteroscedasticity recursive evolving window causality method. The results of the parameters of the discussed causality relationships change over time. Further, numerous causality relationships corresponding to certain crisis periods are reached from economic growth to export growth and import growth and from export growth and import growth to economic growth. Both time-varying parameters and recursive evolving window causality methods confirm the reflection of the causality relationship between the variables when structural changes occur in Turkey. Consequently, the authors used two different causality methods to identify the inadequacies of the traditional Granger causality method and determine the periods of changing causality relations in Turkey.
The relations between economic growth and foreign trade have been widely investigated in the econometrics literature. These relationships can be evaluated using causality analysis, the starting point of which is often considered Granger’s (1969) study. The study provides the traditional definition of causality. The Granger causality relationship is obtained from the whole sample. This means that the effects of social, economic, and political shocks cannot be expressed clearly when the causal relationship between variables is evaluated. Therefore, even if structural changes have occurred in the interval of the sample, it may not be possible to identify these changes in the causality relationship.
When structural changes impact the value of the variables, methods that help identify these changes should be used. Therefore, this study evaluates two different time-varying causality methods. The first is the time-varying parameters (TVP) causality. According to the TVP method, the effect of structural changes on parameter stability should be evaluated. For this, a joint null hypothesis is used. The joint null hypothesis is tested such that both the parameters are stable and there is no causal relationship. The other method is the recursive evolving window (REW) causality. In the latter method, structural shocks can change the causality relationship, and therefore, causality should be examined in sub-samples.
This study evaluated the causality relationship between economic growth, export growth, and import growth in Turkey. First, it examines the relationships with the traditional Granger causality analysis. The study found a bidirectional causality relationship between economic growth and import growth in the sample. This result indicates that, in fact, increasing income levels cause more attention to be given to information and technology in imports. Further, there is an increase in purchasing power and input requirements.
The causality analysis of the whole sample indicates a uniform relationship. The opposite conclusion was based on investigating the effects of structural changes over time. The authors conclude that Turkey should closely follow the events that affect the commercial activities—both within the country and in neighboring countries— to maintain balance between foreign trade and economic growth. Therefore, it is inevitable for Turkey to be affected by the results of structural changes. Therefore, the results of structural changes are the prime subject of this study, and the changes in the causality relationship were observed with time-varying causality methods.
Further, this study shows the stability of the parameters while examining the timevarying causality relationship. For this reason, the results of time-varying causality relationships are presented by examining parameter stability, robust standard errors, and consistent estimators under heteroscedasticity.
Implementing the TVP method, the results show that the parameters are not stable and that there is a bidirectional causality relationship between import growth and economic growth, and between export growth and economic growth. According to the results of the REW method, there are causality relationships arising from economic growth with both export growth and import growth between 2011–2016. There are also causality relationships from export growth and import growth to economic growth between 2006–2008. Additionally, there are bidirectional causality relationships between import growth and economic growth between 2007–2008 and in 2016. Therefore, even if there is a causal relationship between the variables in different periods, it sharply prevents the rejection of the null hypothesis that there is no causality. These results provide evidence that the causal relationship may change over time.
Consequently, according to the results of the time-varying causality relationship, the relationship between foreign trade and economic growth in Turkey varies intensively between 2006–2008 and 2011–2016. Thus, both the global economic crisis and the European debt crisis have affected Turkey. The estimated parameters while examining the causality relationship did not show stability.