论文部分内容阅读
对当前和未来经济状况的评估是及时制定宏观经济政策的基础,但由于数据不完整和噪音数据的存在,使得评估变得困难。这个问题在新兴经济体较为突出,因为其大部分经济数据不能周期性地公布,而且还存在时间滞后的问题。本文用5种模型对10个拉美国家的实际GDP增长进行实时预测和提前预测,并对预测效果进行评估。结果表明,对于本文涉及的大部分拉美国家而言,月度数据有助于提高预测的准确性;相对于其他几种模型,动态因子模型可以持续地产生较为准确的实时预测和提前预测结果。
Assessment of current and future economic conditions is the basis for the timely formulation of macroeconomic policies, but assessment is made difficult by the lack of data and the existence of noise data. This issue is more prominent in emerging economies, because most of its economic data can not be published periodically, but there are still time lags. In this paper, five models are used to predict the real GDP growth of 10 Latin American countries in real time and predict them in advance, and evaluate the forecasting results. The results show that for most of the Latin American countries covered in this paper, the monthly data can help improve the accuracy of the prediction. Compared with other models, the dynamic factor model can consistently produce more accurate real-time prediction and early prediction results.