币种结构调整是一件很困难的事情，一是币种结构调整本身的成本，一种货币换成另一种货币肯定是有成本的；二是我们做出调整之后，国际市场会有相应的调整，这个过程注定是一个缓慢的过程。这也正是本文的研究意义之所在，通过本文的分析研究，可以为这种调整过程提供一个明确的方向，从而降低调整成本，加快调整的速度和效率。
本文基于H. M. Markowitz的资产组合模型，通过对主要储备币种的风险与收益分析，获得了一组优化的币种结构（一系列收益率一定情况下的风险最小点），并对得到的结论进行检验和分析，最终得出我国外汇储备币种结构的调整方向。不仅如此，本文还根据预测期限将模型分为短期模型（一年以内）和中期模型（一年以上）分别进行研究以使其更切合实际。
本文得出的结论为：现阶段中国合理的外汇储备货币结构非最优化，并且提出了币种结构的调整方向——中长期内，我国应尽量增加欧元在储备资产中的比例，减少美元资产在储备中所占的比例；近期，可适当增加日元储备。
关键词：外汇储备 ；币种结构 ；资产组合模型 ；优化
Abstract
The adjustment of currencies structure is a task so difficult, firstly, it’s about the cost of currencies restructuring itself, no doubt there is a cost in shifting form one currency to another; secondly, once we make adjustment, the international market will respond to it and make adjustment too, that’s must be a process which takes a long term. That’s why this paper makes sense, according to the analysis of this paper, the direction of currencies restructuring is decided, thus, the cost of the adjustment can be reduced, and the whole process will speed up and can be more effective.
This paper is based on the portfolio model proposed by H. M. Markowitz , which is mean to find out a series of points which have the minimize risk under certain benefits by analyzing the risks and benefits of the reserve currencies. Then, these results are tested and analyzed for their correctness, and, finally, advices on the direction of adjustment of currencies are made out. What’s more, in order to make the paper more practical, the model is divided into short-time model (less than one year) and middle-time model (more than one year) by the forecast period.
This paper comes to the conclusion as follow: the currency structure of foreign exchange reserves China adopted now is non-optimal, and, what’s more, the direction of currencies restructuring is proposed, that is, on the middle and long run, increasing the proportion of Euro in reserve assets and meanwhile reducing the proportion of US dollar; while, currently, it’s better to increase the proportion of Japanese Yen appropriately.
