Test on the Roles of Sino-us Inflation Levels to the Other in Their Predictions

The paper researches on the inflation level forecasting for China and the United States in one kind of univariate and three kinds of bivariate cases, using Vector Autoregressive and Bayesian Vector Autoregressive Models, based on the rolling-sample forecasts and the mean absolute percentage error standard. Empirical tests show that, both China and the United States’ inflation level series are stationary; China and the United States’ inflation level are both able to provide additional information besides itself to the other in the inflation level forecasting, but the effects of the United States’ inflation level to China’s inflation level is greater than that of China’s inflation level to the United States’ inflation level; In addition, the influence of the United States’ inflation level to China's inflation level maybe is mainly through the expectation gotten by its past values. Finally, some suggestions are given.