Tests for unit roots in macroeconomic time series data

Much attention has been paid in recent years to the study of the unit root tests in time series. The relevance of the subject arises because most of the time series analysis in economic, business and finance are based on the unit roots hypothesis. This paper examines time series behavior of two important macroeconomic variables; GDP and GDP Per Capita of the People's Republic of China over the past five decades. The unit root tests done by ADF and PP tests have revealed that both series are found to be integrated of order one suggesting shocks to it will have permanent effects. Information criteria used in model specification have suggested an ARIMA(2,1,2) and ARIMA(1,1,0) models for the series GDP and GDPPC, respectively. Since no tests are statistically significant at the 5% or higher confidence levels the models are adequate, data are completely random, and the models often describe the behavior of the growth rates of macroeconomic time series.