基于组合漂移模型的订单流变化条件下汇率波动研究

通常,宏观名义汇率模型的预测效果很不理想,其R2统计量最高10%左右.针对这一现象,本文对汇率决定基础模型中的一个考虑微观因素的新模型做了实证分析,结果表明,组合漂移模型成功地解释了汇率变动,R2统计量超过了60%.在样本外预测中,本模型预测效果也明显优于随机游走模型.同样,本文估计即期汇率与订单流的敏感程度是相似的.这个结果将引导作者在研究过程中同时引入微观和宏观方法,而这种方式将可以更广泛地应用于资产定价方面的研究.

Exchange Rate Fluctuation Study Based on a Cluster Drifting Model of Orders

Based on the fact that the conventional exchange rate fluctuation model has a poor predicting capacity, the authors propose a new model,on a more microscopic scale,that may predict the exchange fluctuation with high accuracy,namely,above 60%in terms of R~2,the experiments of which reveal further superiority of the model.