论文部分内容阅读
本文在实时数据基础上选取金融变量作为预测因子并通过混频数据抽样(MIDAS)模型对GDP增长率进行短期预测。结果表明:短期预测时MIDAS模型预测效果甚佳而且嵌入自回归项的MIDAS模型明显降低预测误差;数据修正对MIDAS模型的预测精度有负面影响;货币供应量等预测因子在包含自回归项MIDAS模型中预测精度较高,投资和出口依旧是拉动我国经济增长的重要因素;SPA检验及组合MIDAS模型的较好预测精度说明组合MIDAS模型预测能力占优。
This paper selects financial variables as predictors on the basis of real-time data and forecasts short-term GDP growth rate through the mixed data sampling (MIDAS) model. The results show that the MIDAS model is very effective in short-term prediction and the MIDAS model embedded in the regression model significantly reduces the prediction error; the data correction has a negative impact on the prediction accuracy of the MIDAS model; predictors such as the money supply in the MIDAS model The prediction accuracy of SPA test and combined MIDAS model shows that the combined MIDAS model has predominant ability of forecasting.