动态避险模型之建立—以远期外汇与一篮子货币避险策略为例

Abstract

本研究建构一动态避险模型，由传统避险工具与一篮子货币避险策略所构成。由于传统避险工具的好处在于风险的完全规避，而一篮子货币避险之主要目的则在于降低避险成本，因此本研究之动态避险模型，系以一篮子货币避险策略为主，于每期期初比较两种策略之避险成本，择其低者作为该期之避险策略。另外，为建立更为精确的一篮子货币避险部位，本研究亦导入人工智能工具于模型中，试图利用人工智能于汇率预测上的优秀表现，优化篮子内货币之权重，进而达到更佳的避险绩效。经实证，本研究之动态避险模型，确能显着地优于仅执行单一传统避险工具或者一篮子货币避险策略。其中，以两年作为权重估计期间并纳入预测技术之模型，于实验期间内表现最佳，是为本研究所推荐使用之模型。 This paper proposes a “dynamic hedging model”, which adjusts the hedging strategies by the time, to increase the hedging performance. The dynamic hedging model is combined with the traditional hedging strategy (e.g. Delivery Forward) and the basket currency hedging. The traditional hedging strategy covers the whole risk, in the mean time, the basket currency reduce the hedging cost. Therefore, we establish the position of basket currency first, and execute this strategy when its hedging cost is lower than the traditional hedging strategy; otherwise, we execute the traditional hedging strategy. Besides, for establishing a more precise position of basket currency, we also use artificial intelligence to forecast the exchange rate, which is expected to estimate the currency weight in a basket more precisely. Empirically, the dynamic hedging model we propose performs much better than either the traditional hedging strategy or basket currency hedging strategy. In addition, due to the way of using two-year estimation period and adding forecast technology to correct the estimation got the best performance, we recommend this model to be a reference of a company’s hedging behavior.