Get Ready for Japanese Trade Deficits


Kenneth Kasa

FRBSF Economic Letter 1997-28 | October 3, 1997

From 1896 to 1970 the United States had a continuous string of surpluses in its balance of trade (in goods and services). Since the late 1970s it has had a continuous string of deficits.

Pacific Basin Notes. This series appears on an occasional basis. It is prepared under the auspices of the Center for Pacific Basin Monetary and Economic Studies within the FRBSF’s Economic Research Department.

From 1896 to 1970 the United States had a continuous string of surpluses in its balance of trade (in goods and services). Since the late 1970s it has had a continuous string of deficits. In Japan, the trade balance was in deficit for the first couple of decades following the war, and since the late 1970s it has been continuously in surplus. This kind of persistence suggests that there is more to trade balances than exchange rates, trade barriers, and business cycles. After all, Japan and the United States have both experienced several ups and downs in their business cycle, several massive swings in the value of their currency, and several major changes in their trade policy, without a reversal in their underlying trade imbalance.

This Letter argues that demographic trends are a fundamental force behind trade imbalances, and on the basis of these trends (which are easier to forecast than most economic variables) we can expect Japan to start running trade deficits sometime within the next decade. There are two steps to the argument. The first is to understand how demographics affect national saving. The second is to understand how national saving affects the trade balance.

Demographics and National Saving

National saving consists of three parts — household saving, corporate saving, and government saving. Of the three parts, the household sector is more “fundamental.” Corporations and the government are just intermediaries. That is, they are institutional mechanisms that facilitate transactions and transfers between households (e.g., between those who want to borrow and those who want to lend). Thus, economists often argue that the only important savings rate is the household savings rate. For example, if households have a desired savings rate, any changes in corporate or government saving will simply be offset by opposite changes in household saving. Again, this is based on the assumption that households recognize that at a fundamental level they own the corporations and the government. If this assumption is valid, economists say that households “see through the corporate (or government) veil.”

Of course, the extent to which households see through the corporate and government veils is an empirical question and, as usual, economists do not agree on an answer. From the standpoint of demographics, however, the more important veil is the government veil (because demographic changes influence government budgets more than they influence corporate profits), and the preponderance of evidence suggests that households do not see through the government veil. This means we need to analyze the effects of demographics on the household and government sectors separately.

The conventional vehicle for understanding the link between household saving and demographics is Modigliani’s “Life-Cycle Model.” The essence of this model is portrayed in Figure 1. The key assumption is that individuals desire a relatively constant level of consumption throughout their lifetime. To achieve this constancy, fluctuations in income are smoothed over by borrowing and lending (or selling-off previously accumulated assets). This assumption, when combined with the empirical fact that most individuals’ age-earnings profiles are hump-shaped (as shown in Figure 1), produces a straightforward and intuitive prediction about the relationship between demographics and the household savings rate, i.e., the household savings rate should be inversely related to the relative number of “young” and “old” people in the population. Although there are some puzzling disparities (e.g., old people don’t seem to decumulate their assets fast enough), for the most part the life-cycle model appears to be reasonably consistent with the data.

To use the life-cycle model to make predictions about household savings rates, we therefore need data on the age-structure of the population. Figure 2 contains plots for the U.S. and Japan of one commonly used measure of the age-structure of the population — the Elderly Dependency Ratio. This is defined as the ratio of the population aged 65 and over to the population aged 15 to 64. (Note, plots of the Total Dependency Ratio, which adds the number of children to both the numerator and denominator, look qualitatively similar.) To interpret the magnitude of this ratio, note that its inverse gives the number of young workers and savers that are available to support each retired dissaver.

Two features of Figure 2 are immediately apparent. First, dependency ratios in both the United States and Japan have been trending up over time. (In fact, they are trending up in all industrialized countries). For example, in the United States there are currently about five workers available to support each retired person. However, this number begins to decline rapidly around 2010, when the first baby-boomers begin to retire. By 2030, it is projected that there will be fewer than three workers per retired person.

The second feature to notice in Figure 2 is that the U.S. and Japanese dependency ratios recently crossed. This has occurred for two reasons. First, Japan has experienced a more rapid growth in life expectancy (i.e., the numerator has grown faster). For example, between 1960 and 1990 Japanese life expectancy increased by about 40 percent, whereas in the United States the gains were only about 19 percent for women and 16 percent for men. Second, fertility rates in Japan have been lower than in the United States (so the denominator has grown more slowly). For example, the current fertility rate (i.e., the average number of children per woman) is only about 1.5 in Japan, whereas in the United States it is about 2.1.

By itself, Figure 2 suggests a decline in both the U.S. and Japanese household savings rates, but a greater decline in Japan. However, the relevant concept of saving for expaining trade imbalances is the national savings rate, and for this we need to factor in the likely course of government saving in the U.S. and Japan. Although there has recently been much discussion in the U.S. about the adverse budgetary consequences of an aging population (because of the need to finance the largely pay-as-you-go social security and medicare programs), it turns out that the problem is even worse in Japan. For example, the IMF estimates that the unfunded portion of the public pension system in the U.S. is about 25 percent of GDP. In Japan, however, it is about 110 percent of GDP! This difference arises from both the greater generosity of the Japanese social security system and the relative increase in the Japanese dependency ratio.

Thus, from the perspective of both the household and government sectors, current demographic trends point to a relative decline in Japan’s savings rate. In fact, Horioka (1991) surveys a variety of studies on the future of Japan’s savings rate and concludes that the household savings rate will approach zero sometime around 2010. Now, the next step is to understand how this decline in saving will affect the trade balance.

National Saving and the Trade Balance

While nearly everyone agrees that demographic trends point to a relative decline in Japan’s savings rate, the effect of this decline on the trade balance is more controversial. This is because a country’s trade balance represents the difference between its saving and investment, and one can argue that demographic trends also point to a relative decline in Japan’s investment rate. (More precisely, the difference between a country’s saving and investment equals its “current account” surplus, which includes the interest on its net foreign asset position. For a country like Japan, with a substantially positive net foreign asset position, this means that we can expect its trade balance to turn negative before its current account, because for a while Japan will be able to finance its savings deficiency from the interest on its foreign assets.)

The basic mechanism behind the negative effect of aging on investment is the fact that a declining labor force will tend to lower the rate of return on capital. Why? Because at a broadly defined level, labor and capital tend to complement each other in the production process, in the sense that having more capital makes labor more productive, and having more labor tends to enhance the productivity of capital. So if aging reduces the labor force, capital’s productivity, and therefore its rate of return, will fall. However, in an open economy, a declining rate of return on capital will manifest itself as a capital outflow, as Japanese investors seek higher returns abroad. This capital outflow will then tend to sustain Japan’s current account surplus. Although declining investment will slow down the erosion of Japan’s current account surplus, it will not stop it. For example, Yashiro and Oishi (1997) simulate a number of models and conclude that Japan’s current account will start registering deficits sometime during the decade 2000-2010, the exact date depending on assumptions about productivity and elderly labor force participation. Masson and Tryon (1990) incorporate demographic variables into the IMF’s multicountry model and find similar but less drastic results. They predict that over the next two decades Japan’s current account as a share of GNP will decline by about 4.0 percentage points, from a surplus of 2 percent to a deficit of 2 percent. In contrast to these econometric approaches, Noguchi (1990) simulates a small-scale neoclassical model, which adheres more closely to Modigliani’s life-cycle model. He predicts that the current account won’t start falling until after 2010. Overall, a broad spectrum of alternative simulation methodologies all point to a declining Japanese trade balance. The only debate concerns the timing. An optimist could argue that deficits are at least two decades away, while a pessimist could argue that they will arrive within a decade.


Forecasting anything other than planetary orbits more than a decade in advance is a brave exercise. In the social sciences, there are just too many intervening asteroids. In this case, perhaps the most important orbit-disturbing asteroid is the ambiguous notion of what it means to be “old.” “Old” is a relative concept, of course, and as life expectancy approaches 80, it may not be reasonable to think of someone who is 65 as “old”– when he has nearly 20 percent of his life ahead of him. For example, applying this definition in 1920 would have meant that 43 was old! So if society begins to adapt to longer lifetimes by changing the economic activities of “elderly” people (e.g., an increase in labor force participation), then these sorts of projections could go seriously astray.

Kenneth Kasa


Horioka, Charles Yuji. 1991. “Future Trends in Japan’s Saving Rate and the Implications Thereof for Japan’s External Imbalance.” Japan and the World Economy 3, pp. 307-330.

Masson, Paul R., and Ralph W. Tryon 1990. “Macroeconomic Effects of Projected Population Aging in Industrial Countries.” IMF Staff Papers 37, pp. 453-485.

Noguchi, Yukio. 1990. “The Age Structure of the Population and Saving/Investment: An Analysis Based on Cross-Country Comparisons.” Financial Review (Ministry of Finance) 17, pp. 39-50.

Yashiro, Naohiro, and Akiko Sato Oishi. 1997. “Population Aging and the Savings-Investment Balance in Japan.” In The Economic Effects of Aging in the United States and Japan, eds. M. Hurd and N. Yashiro. Chicago: Univ. of Chicago Press.

Opinions expressed in FRBSF Economic Letter do not necessarily reflect the views of the management of the Federal Reserve Bank of San Francisco or of the Board of Governors of the Federal Reserve System. This publication is edited by Anita Todd and Karen Barnes. Permission to reprint portions of articles or whole articles must be obtained in writing. Please send editorial comments and requests for reprint permission to