通货膨胀实时预测及菲利普斯曲线的适用性

Real-time Inflation Forecasting and Its Applicability of Phillips Curve to China

本文从实时分析的视角,基于多种退势方法的产出缺口最终估计、准最终估计和实时估计序列,分别构建了四类预测模型对我国通货膨胀率进行预测,分析了产出缺口修正效应和滞后阶数变化效应对通胀预测的影响,并进一步考察了产出缺口在通胀预测中的作用及菲利普斯曲线在通胀预测中的适用性。研究结论表明,通胀率的实时预测效果要明显比基于最终数据的差,其中滞后阶数变化效应对实时预测精度的影响大于产出缺口修正效应;尤为重要的是,尽管在最终数据的预测分析中,产出缺口的引入能够提高通胀率的预测精度,但是在实时预测中,产出缺口没有提供有价值的信息,因此"产出—通胀"型菲利普斯曲线在我国通胀实时预测中并不适用。

From the perspective of real-time analysis, this paper, based on the final estimates, quasi-final estimates, and real-time estimates of China＇s quarterly output gap derived from six detrending methods, constructs four kinds of forecasting models to predict China＇s inflation rate, analyzes the effects of output gap revision and variations in lag length on inflation forecasting, and further evaluates the usefulness of output gap for predicting inflation and the applicability of Phillips curve in inflation forecasting. The results show that the predicted inflation values in real time are less accurate than those based on final data, and the introducing of time-varying lag lengths has more important effects than output gap revision on forecast accuracy. More importantly, although output gap estimates appear to be quite useful for predicting inflation in final data analysis, it does not provide any useful information in real time forecast, which means the ＂ output-inflation＂ type of Phillips curve is not suitable for China＇s inflation forecast.