Does Singapore Invest Too Much?


Kenneth Kasa

FRBSF Economic Letter 1997-15 | May 15, 1997

During the past 30 years, the economies of several East Asian nations grew on average by about 8% per year. Such rapid growth over such a long period of time is historically unprecedented.

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.

During the past 30 years, the economies of several East Asian nations grew on average by about 8% per year. Such rapid growth over such a long period of time is historically unprecedented. Recently, Paul Krugman (Foreign Affairs 1994) sparked a controversy by arguing that this growth is unsustainable. He based his argument on the pioneering work of Alwyn Young (1995). Young conducted a detailed empirical analysis of four economies – Hong Kong, Korea, Singapore, and Taiwan. For each of these economies he decomposed output growth into two parts. One part measures growth in the inputs to production, like capital and labor. The other part measures how efficiently these inputs are used. Young’s startling finding was that most of the growth in these economies is accounted for by growth in inputs. Productivity growth played a minor role. This contrasts sharply with the western industrialized countries, where long-term output growth is mostly due to productivity growth.

The relevance of this finding to the sustainability of Asia’s rapid growth is that factor accumulation tends to be self-limiting. Eventually you run out of labor, and supplying a given labor force with more and more capital equipment eventually runs into diminishing returns. In contrast, ideas and human ingenuity, which are the wellsprings of productivity growth, seem to escape the forces of diminishing returns. Breakthroughs in science and engineering are happening as fast today as they were in the 1800s.

This Letter provides a follow-up to Young’s analysis of Asian growth. Like Young, I analyze the efficiency of this growth. However, I shift the focus of analysis from technological efficiency to allocational efficiency. In addition to producing goods, an economy also allocates the goods it does produce to various groups within society. In general, economists cannot judge the efficiency of a particular allocation, since it typically requires weighing one person’s gain against another person’s loss. Balancing these kinds of trade-offs is a task for the political system. However, when it comes to allocating resources over time, i.e., to individuals of different generations, it turns out that it is possible to assess the efficiency of an allocation. In particular, we can say whether current generations are saving and investing too much and consuming too little. Roughly speaking, an economy is investing too much when it consistently gets less back from its investment projects than it puts in.

This Letter applies this intertemporal efficiency criterion to Singapore. Of all the world’s nations, Singapore seems to be a prime candidate for excess saving and investment. Singapore’s saving and investment rates are the highest in the world, and much of this saving and investment is mandated by the government. I find that we cannot rule out the possibility that Singapore invests too much. This is in sharp contrast to western industrialized countries, where previous work has shown (not surprisingly) that there is no evidence of excess saving and investment. If anything, the opposite is more likely to be true. However, to understand what this means we must first understand how it can happen that an economy invests too much, and how economists can diagnose this situation.

Assessing Dynamic Efficiency

In a private enterprise economy, saving and investment are determined by the desire of individuals to provide for their retirement. Diamond (1965) was the first to show that there is no guarantee that these individual saving decisions translate into a socially efficient outcome. In particular, he showed that it is possible for each generation to save too much. This can happen if individuals care only about their own lifetime income and consumption profiles and are not concerned with what happens after they die. If this is the case, then a strong desire to save may depress the rate of return on saving below the economy’s underlying growth rate, which is determined by labor force growth and technological progress. This would be inefficient. From a technological standpoint, the economy has the ability to transfer resources into the future at a rate that exceeds the market rate of return on savings and investment. Diamond argued that if this happens it is desirable for the government to step in and provide an alternative savings vehicle that diverts individuals’ savings out of physical assets, like capital (which are subject to diminishing returns), and into paper assets, like government bonds. Doing this would raise the market rate of return and make everyone wealthier, and therefore better off. (Note, because of diminishing returns, less capital investment produces a higher rate of return.) Alternatively, the government could set up a tax and transfer scheme (like a pay-as-you-go social security system) that would tax young people and give the money to retired people. Since the tax base is growing faster than the interest rate, the government can offer young people a higher return than the market and still run a balanced budget each year!

The theoretical model Diamond used to illustrate this point is highly abstract. He purposely made it this way in order to illustrate the basic point in the clearest manner possible. Unfortunately, the abstract nature of Diamond’s model makes it difficult to apply. For example, according to Diamond dynamic inefficiency is indicated by an investment return that is less than the economy’s growth rate. But what rate of return should we use? Annual (real) returns on short-term U.S. government securities have averaged only a few tenths of a percent during the post-war period, while the economy has grown about 2 to 3% per year. This suggests the U.S. economy saves too much! However, the stock market (real) rate of return has been closer to 7%, so if we use it instead we would conclude that the U.S. economy is indeed dynamically efficient. The problem is that Diamond ignored uncertainty, and in the real world, investment returns reflect uncertainty. Savers require a higher rate of return on riskier investments. Thus, to apply Diamond’s efficiency criterion it is essential to incorporate uncertainty into Diamond’s analysis, which was done by Abel, Mankiw, Summers, and Zeckhauser (1989).

Abel, et al., show that with uncertainty we can be sure an economy is dynamically efficient if gross capital income consistently exceeds gross investment (where capital income is defined as the sum of profit, rental, and interest income). If this is the case, then the financial sector is making more resources available for future consumption than it is using. (Of course, this doesn’t rule out the possibility that distortions cause investment to be inefficiently too low.) Conversely, if investment consistently exceeds capital income then the financial sector is draining resources from the economy. This is inefficient, since the whole point of investing is to augment future consumption possibilities. As a caveat, however, one should realize that historical observations on investment and capital income can never definitively establish dynamic inefficiency, since it is always possible that future capital income becomes much higher.

Abel, et al., apply their efficiency criterion to seven industrialized nations (i.e., the G-7). They find that in all seven, capital income consistently exceeds investment. Hence, these economies appear to be dynamically efficient. This is true even for Japan, which has a much higher investment rate than the others. As an example, Figure 1 reproduces their results for the U.S. It plots investment and capital income (both expressed as a percentage of GNP) for the period 1975-1985. Note that capital income in the U.S. has consistently been about 27% of GNP while the investment rate has only been about 16% (although it fluctuates somewhat more). Thus, the investment sector consistently yields a dividend of about 11% to U.S. consumers, and thus, there is no evidence of dynamic inefficiency.

Is Singapore Dynamically Efficient?

Applying the Abel, et al., dynamic efficiency criterion to Singapore is not as straightforward as it is for the G-7 countries. Unlike the G-7, Singapore does not publish detailed historical data on labor and capital income. As a result, I take an indirect approach. I infer labor and capital income (as a share of GDP) from estimates of a production function, which relates output to inputs. These data are readily available. The main drawback to this indirect approach is that it requires assumptions about the nature of the production function, and it presumes competitive market conditions.

The particular estimates I use are taken from van Elkan (1995). Although many researchers have estimated production functions for Singapore, van Elkan’s estimates are the only ones I found that include land as a separate factor of production. This is important because if land is ignored then the productivity of labor and capital are overstated, and hence, so are their incomes. This would bias the results toward dynamic efficiency. It turns out that land rents consistently account for about 9% of Singapore’s national income.

Figure 2 contains a plot of Singapore’s investment rate along with the (implied) share of capital income in GDP. The first thing to notice is the extremely high investment rate in Singapore. During the past 20 years it has averaged about 37%, and for several years during the 1980s it approached 50%! Thus, Singapore typically invests between two and three times as much of its income as the United States. Much of this investment comes from other countries. For example, during the 1980s foreign investment accounted for about 25% of total fixed capital formation, and for over 60% of investment in manufacturing.

At the same time Singapore’s government and businesses have been undertaking massive investment, its citizens have been saving at unrivaled rates. As reflected by its persistent current account surplus, Singapore’s saving rate is even higher than its investment rate! However, much of this saving is “forced.” Since 1955, the government has operated a compulsory savings program called the Central Provident Fund. Like the U.S. social security system, this program requires a “contribution” from both employees and their employers. The current contribution rates are 21.5% for employees and 18.5% for employers. This sort of forced savings policy, along with tax incentives that lure in foreign investment, raise the suspicion that Singapore could be a dynamically inefficient economy.

However, the second thing to notice in Figure 2 is that capital income in Singapore has also been quite high. During the past 20 years it has averaged about 35.5% of GDP, although capital’s share is steadily falling as labor’s share rises (presumably due to improvements in education). Overall, the evidence in Figure 2 does not allow us to make a

definitive judgement about dynamic efficiency. However, unlike the G-7 countries, we cannot rule out the possibility that Singapore invests too much and is dynamically inefficient. During the early 1980s, investment consistently exceeded capital income, and over the full sample the average investment rate has slightly exceeded capital’s average income share.

Kenneth Kasa


Abel, Andrew B., N. Gregory Mankiw, Lawrence H. Summers, and Richard J. Zeckhauser. 1989. “Assessing Dynamic Efficiency: Theory and Evidence.” Review of Economic Studies 56, pp. 1-20.

Diamond, Peter A. 1965. “National Debt in a Neoclassical Growth Model.” American Economic Review 55, pp.1126-1150.

van Elkan, Rachel. 1995. “Accounting for Growth in Singapore.” In Singapore: A Case Study in Rapid Development. IMF Occasional Paper No. 119.

Young, Alwyn. 1995. “The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth Experience.” Quarterly Journal of Economics 110, pp.641-680.

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