Until 2010's, Japan and other East Asian countries have chronically accumulated current account surpluses. In 2006, the Japanese current account surplus reached a 2.6% of its GDP, while the Chinese surplus reached 7.2%. As the counterpart, the U.S. was running a current account deficit of a 6.9% of its GDP. These imbalances had accumulated for almost 30 years. As a result, Japan held a net foreign asset of $1500 billion (a 33% of its GDP), while the U.S. net foreign debts ran up to $2500 billion. However, this pattern has changed recently. As of July 2013, Japan has been running trade deficit for 13 months in a row, although the current account is still in surplus because of the positive net factor income flow from abroad. China's trade surplus is also shrinking from 10 percent of its GDP in 2007 to only 2.8 percent in 2011. The U.S. current deficit has been almost halved compared to its peak in 2006. Such a change in the trend, without a major catastrophe, was not expected by earlier studies mentioned above. Why such huge current account imbalances had been continuing for as long as 30 years? Does the recent shrinkage of the imbalance suggest that the direction of the capital flow will be reversed in the near future, or is it just a temporary phenomenon? To explain these phenomena, most studies focus on the changes in the pattern of international trade. Instead, I interpret the imbalances as reflecting the difference in the optimal consumption-saving pattern of households over the process of economic growth. From the national income identity, the current account surplus is identical to the difference between the national saving and the national investment. Given other conditions, countries that save more and invest less will run current account surpluses. However, earlier studies faced difficulties in applying this theory to the reality, because standard growth models (which do not consider the finiteness of the lifetime) predict that the most patient country will collect all the wealth in the world. I correct this unrealistic situation by explicitly considering the transition of generations. Intuitively, a country's wealth would be bounded within a realistic range because people will eventually be replaced by new generations. The project consists of three steps: I develop the model, collect the relevant data, and simulate the growth process of actual economy by numerical analysis. I first develop an international macroeconomic model with transitions of generations, which enables us to examine how impatience, as well as technology, demographics, and intergenerational income distribution affects the dynamics of current account imbalances. I then collect data on demographics and income distribution from official statistics and household surveys and apply to the model. However, the most important parameter in our project, impatience of American and Japanese consumers, is difficult to measure directly. I estimate it using a numerical procedure called "calibration." Through this procedure, I obtain the dynamics of current account surpluses implied by the optimal saving behavior of consumers.