| Abstract | Many studies show that sudden-stop of capital inflows significantly reduces the welfare of consumers and sufficient foreign exchange reserves can slow or even offset the impact of capital inflows. Jeanne and Rancière had established a dynamic optimization model of optimal international reserves to capture sudden-stop of capital inflow. we improved Jeanne and Rancière's and added more items such as long-term debt capital inflows, foreign direct investment to it, aim to more closed to reality. Through analysis from the new model, we can see that there is an optimal scale of a country's foreign exchange reserves, in an open economy. Optimal scale of foreign exchange reserves to GDP ratio is depend on the following factors: the probability of sudden-stop; output losses after crisis; short and long-term external debt to GDP ratio, foreign direct investment to GDP ratio; and consumer risk aversion. With the increase of all the factors above, the optimal size of a country's foreign exchange reserves has correspondingly expanded. And the long-term growth rate, the cost of foreign capital almost has no effect on the optimal size of foreign reserves. We adapt the model to estimate the optimal size of foreign exchange reserves in China, from 1988 to 2009. From the research, we found that the optimal size of China's foreign exchange reserves to gross domestic product is about 20.4%, and have a conclusion about the actual foreign exchange reserves of China over the past 22 years: reserve is insufficient from 1988 to 2001; it is optimal in the year of 2002; it is over-sufficient after 2003. The problem of over-sufficientness of reserves particularly serious after 2007, the actual size of reserves is more the twine of optimal. The scale of foreign exchange reserves is too large to that its adverse affection cannot be measured. |