Analiysing income diversty in Turkey by quantile regression model: Comarsion of 2002 and 2010 years

Selecting the correct functional structure of model and providing the basic assumptions of the selected model are very important subjects in econometrics. Generally, relationship among variables is examined using parametric methods. Ordinary Least Squares is most known and frequently used method among these methods. But this method not gives good results because of reasons such as used data set structure and not provided the assumption of normality. In such cases, we will need to use alternative regression methods.The purposes of this study compare the Quantile regression method which is popularized recent years with Ordinary Least Squares and reveals the advantages of Quantile regression method. In the first chapter, alternative methods such as least absolute deviations and m regression have been introduced. In the second chapter, Quantile regression which is not require the assumption of normality and for user unlike averages dividing the data set to the many different quantiles has been introduced. In the last chapter as an empirical part, we compare the income differences in Turkey between 2002 and 2010 with using Quantile regression model. Keywords: Ordinary Least Squares (OLS), Quantile Regression, Least Absolute Deviations(LAD), M Regression