MODELING SEASONAL G7 STOCK MARKET RETURN SERIES

Time series properties of economic and financial series are routinely inspected in applied works for proper designing of models and to avoid misleading inferences. Depending on which property series under examination possesses, its occurrence has far-reaching importance on modeling, testing economic theories, forecasting and others. Therefore, testing whether series is stationary, long-range dependence or non-stationary is routinely carried out prior to modeling and conducting reliable statistical inference. Over the years, many testing procedures have been developed to test for properties of time series data. Testing for the presence of a stationarity against a seasonal long memory process has been given substantial concentration in current literature. This paper modeled quarterly G7 stock return time series data using seasonal fractional unit root tests. It is found that seasonal fractional unit roots are present in most of the stock return time series data at zero and seasonal frequencies. We believe that findings of this paper will be helpful to applied researchers when they will work on quarterly G7 stock return time series and/or on test long-memory property of the seasonal series.