Unobserved components models: The investigation of relationship GDP, unemployment and inflation in Turkey

Unobservable component models are models that are directly formulated in terms of unobservable components such as trend, cyclical, seasonal, and irregular components that represent important features of the series. The main advantages of these models are to obtain unobservable components such as potential output, output gap and natural unemployment rate using existing observed variables and to estimate by preserving the originality of the series. The main purpose of the study is to examine key macroeconomic variables in terms of their components using unobservable component models and to estimate variables that are not observable but important for economic decision makers. In line with the aim of study, economic macroeconomic variables such as GDP, unemployment and inflation has been examined in terms of its components using unobserved component model methods and statistical theories on which they are based. The Okun Law and the Phillips Curve estimates were made using the output gap, potential output and natural unemployment rate variables obtained from the so-called methods. In the GDP calculations, the base year is set as 2009 after 2016: Q2 period and the chained volume index is taken at the place of calculation with fixed prices. The selected base year, the global financial crisis that began in 2008 in the United States carry the impact on Turkey's economy. Therefore, it is thought that it is not the correct base year and it is effective for values in the next year (Eğilmez, 2016: 1). In view of these considerations, GDP, unemployment and inflation data for Turkey analysis was performed using the related methods for 1998:Q1-2016:Q2 period. With the results obtained, GDP is evaluated in terms of the components of the change, and findings related to output gap and potential output are interpreted. It is found that Okun coefficient satisfied theoretical expectations and the Phillips Curve for Turkey is not valid.