A Non-linear Study On Macroeconomic And Financial Series In China

Since the study of Nelson and Plosser (1982), who found that most of the macroeconomic series can be characterized as a unit root process nearly two decades ago, people have had many controversial points of view towards the characterize of macroeconomic and financial series, whether as a stochastic unit root non-stationary process or a segmented trend stationary process around one or two structural breaks. The conclusion of the point has a great influence on the research of the relationship between the series or the series dynamics.It is the basic of all the analysis of the relevant issues. It can also help us analyze the series changes under policy-driven and the great influence on the long-term or short-term effects under unexpected external shocks.In this paper, the macroeconomic and financial series in China over the period 1952-2008 is investigated with a new testing procedure, i.e., the nonlinear test based on the ESTAR model,the KSS non-linear unit root test and cointegration test. We find that all of 9 series are nonlinear,and 7out of 9 series , which are real GDP , nominal wages, real wages, real final consumption, Total import and export, bank credit and deposit liabilities, can be more accurately characterized as a nonlinear stationary process as opposed to a unit root process. The other 2 of 9 series,which are GDP per capita and total number of employed persons ,are characterized as a nonlinear unit root process and have no cointegration relationship between them. Moreover,on the consideration of the limited sample size, this paper applies new critical values by Block Bootstrap method which can be use without distribution assumptions or adding sample information...