IS SECTORAL SHIFTS HYPOTHESIS VALID IN THE TURKISH ECONOMY?

Abstract

The aim of this study is to analyze reasons of unemployment and the validity of sectoral shift hypothesis developed by Lilien for the Turkish economy between years 2005 and 2014. The causal linkage between variables is analyzed by Toda–Yamamoto Granger causality, frequency domain causality and asymmetric causality test methods.

The results obtained from all tests show that the main reasons of unemployment in the Turkish economy are the cyclical factors. It is not possible to speak about validity of sectoral shifts hypothesis in the economy. The intensity and direction of interactions between macroeconomic variables may differ due to time frequencies. The existence of causality may disappear in the long run while it exists in the short run. Also conventional causality analysis methods do not decompose the causality into positive and/or negative shocks and so they could not explain the direction of causality. In order to test causation linkage between variables and to find direction of causality in different shock types we employ recently developed causality analysis methods which would increase the quality of the study.