本项目探讨在开放资本账户的条件下，如何衡量跨国资本流动的金融风险以及相应的最优政策管理工具。现有理论往往从代表性消费者模型出发，得出的最优税制是线性的，即对各类资本流动均征收同样的税率。事实上，不同种类、期限、规模的资本流动所隐含的金融风险有显著差别。单一的税率会造成经济效率损失，并且往往无法避免金融危机的发生。本项目基于微观数据，采用异质性经济主体理论框架构建资本流动管理模型，测度不同类别资本流动的金融风险，并针对性的制定资本管理政策。特别的，最优资本流动税制是非线性的（累进的），平均税率随着资本流动的变化自动调整，起到自发稳定器的作用。本项目进一步探讨了该税制在我国的具体实现方式。这对于我国达到资本市场开放与避免金融危机的双重目标具有现实意义。

We study with capital account openness, how to measure the systemic risk of capital flows and the corresponding optimal capital flow management tools. The current literature on capital flow management commonly use the representative-agent framework, and thus their optimal capital tax scheme is linear, i.e. tax rate is constant for all kinds of capital flows. In fact, different types, sizes and lengths of capital flows vary significantly in their externalities, and thus the linear tax scheme is inefficient and usually cannot prevent financial crisis from happening. On the contrary, we uses a heterogeneous-agent framework to study the optimal capital management highlighting capital flow compositions. Especially, we find the optimal tax scheme should be nonlinear, i.e., it is progressive in the sizes of capital flows, and this tax scheme can be an automatic stabilizer as capital flows fluctuate. We also discuss how to put this tax scheme into practice in China to help achieve both capital market openness and financial stability.