Testing Linearity in Time Series by using Eviews Software (Case Study: Banks ROA) Amiri Javada, Rad Najmeh Nakhaeib aDepartment of statistics, Qaemshahr Branch, Islamic Azad University, Qaemshahr, Iran bDepartment of statistics, Mashhad Branch, Islamic Azad University, Mashhad, Iran Online published on 10 March, 2015. Abstract In near all analysis methods of linear or non linear time series, stationary time series are assumed; while most time series are non-stationary in terms of trend, frequency, or jump due to different reasons. Moreover, distinguishing linearity or non linearity of the related process is highly important for modeling. This paper has studied annual time series of return on bank assets. This paper is aimed at studying existence and intensity of common non linearity in time series of ROA level of banks in an annual time scale (2007–2012). In so doing, ADF, KPSS, and BDS tests have been used. ROA data of this paper has been derived from Eghtesad Novin Bank, Arian Electronic Bank, Ansar, Iran Zamin, Parsian, Pasargad, Tat, Tejarat, Hekmat Iranian, Dey, Saman, Sarmayeh, Sina, Saderat, Gardeshgari, Mellat, and Kar Afarin Banks. The results of tests carried out by Eviews software indicated that time series of banks ROA are stationary. Top Keywords Test of linearity, stationarity, non linear time series, return on bank assets, Eviews software. Top |