Earlier sections have documented a trend decline in the saving rate of the industrial countries in the 1980s and have identified a number of factors that might have contributed to this decline. A natural follow up would be to consider the question of whether the level of saving is too low in individual countries. This section draws on the economic literature to develop a framework for answering this question. As will become apparent, however, it is not possible to devise a simple criterion for assessing the adequacy of saving in general.
The following discussion considers the intertemporal relationship between saving and economic welfare. This relationship is initially explored in the context of traditional models of economic growth in which technological advancement is determined by an exogenous process. Empirical work based on these models suggests that all major industrial countries are dynamically efficient—that is, the saving rate in none of the countries can be said to be too high. It would not be possible, however, to deduce that the saving rate in these countries is too low, unless the argument can be made that distortions have driven a wedge either between the social and private rates of time preference or between the market and social rates of return to capital.
The discussion subsequently considers more recent models of growth in which technological advancement is determined endogenously. While these models are in an early stage of development, they have important implications for assessing the adequacy of saving. In particular, they suggest that the social rate of return to the current generation from saving and investment may be far above the private return. If this result turns out to be true, a strong case can be made that current saving is too low.
Growth Models with Exogenous Technical Progress
The standard benchmark for analyzing longterm effects of saving on capital accumulation and growth is provided by Solow’s (1956) classic model. A striking implication of this model and its descendants is that the longrun rate of growth is unaffected by the saving rate. An increase in saving increases the ratio of the capital stock to labor supply (which is given exogenously) and results in a faster growth of output per capita initially. With diminishing returns to capital, however, the growth rate of per capita output eventually converges to the exogenously given rate of technological advancement.
In this framework, an increase in the saving rate will always raise the level (although not the growth rate) of long-run per capita output, but not necessarily per capita consumption. An increase in the capital-labor ratio at long-run equilibrium has two offsetting effects on consumption per capita: the desirable effect of a high capital-labor ratio is high per capita output; the undesirable effect is the high rate of saving required to maintain this capital-labor ratio. The higher the capital-labor ratio, the larger the fraction of output that must be saved and invested to keep the ratio constant. At some point, the benefit of a higher capital-labor ratio in terms of higher per capita output just offsets the cost of the higher required saving. The saving rate associated with this capital-labor ratio, which maximizes consumption per capita at long-run equilibrium, is called the golden rule rate. Saving rates above the golden rule rate are said to be dynamically inefficient in the sense that, in such a state, the current generation can consume more without depriving future generations. Thus, when society is dynamically inefficient, a consumption binge by the current generation is appropriate and the only issue is the distributional one of how best to allocate the efficiency gains.
Identifying cases of dynamic inefficiency is clearly important. A recent study by Abel and others (1989) has proposed a simple criterion for judging dynamic efficiency. They show that saving is above the golden rule rate if, in each period, total profits are less than investment.14 In such cases, it would follow that saving is too high and it would be efficient to increase consumption. The results of the study show that all of the seven major industrial countries easily pass the test for dynamic efficiency. For example, in Japan, the country with the highest saving rate, profits exceed investment, on average, by over 10 percent of GNP; the smallest difference between profits and investment during 1960–84 is more than 5 percent of GNP. Similar results hold for the other industrial countries.
The next obvious question is whether any of the major industrial countries can be said to save too little. To the extent that these economies are dynamically efficient, the assessment of the adequacy of saving must ultimately be based on the rate of time preference for current and future consumption. A case for raising the saving rate can be made if one believes individuals’ saving decisions may conflict with maximizing social welfare. Three broad categories of factors can be identified that give rise to such a conflict. First, myopic behavior on the part of the present generation could put excessive weight on current consumption and make the next generations worse off. In this context, it should be noted that, in the presence of technological progress, future generations always “live better” than the present generation. The relevant question is then whether the saving rate should be raised to widen even further the gap between the living standards of the present and future generations. The issue of whether the present generation can be trusted to adequately take into account the welfare of future generations is basically a philosophical one and beyond the scope of the present discussion.
A second reason for questioning the appropriateness of individuals’ saving decisions is distortions arising from government policy and market rigidities. For example, many aspects of the corporate and personal tax codes cause the private return to saving to differ from the social return (see Section V). In addition, all measures that affect government saving also directly affect national saving since private saving only partially offsets shifts in government saving.
Finally, the saving rate may be too low if individuals do not take into account external benefits to the capital accumulation that result from saving. It can be argued that economic growth leads to important social benefits (see Tobin, 1971). For example, it may be easier to achieve low inflation in a growing economy with rising real wages than in an economy with stable real wages. Similarly, it may be easier to redistribute income in a growing economy than in a stagnant economy. These social benefits of capital accumulation do not accrue to the firms making the investment decision, nor to the savers who finance investment. To the extent that these externalities are important, individuals’ choices may lead to inefficiently low levels of saving and investment.
It should be noted that in the context of exogenous growth models, the external benefits are enjoyed only for a limited period because higher saving leads to higher growth only during the transition from one long-run equilibrium to another. This result, however, changes dramatically if the assumption of exogenous growth is abandoned. The implications of endogenous growth models for assessing the adequacy of saving are explored in the next section.
Growth Models with Endogenous Technical Progress15
It has long been recognized that in the traditional models of exogenous growth, most of the observed growth in per capita income is attributed, even in the short run, to technological progress. Thus, the widely different growth rates of industrial countries observed since the end of World War II would be attributed mainly to their differing rates of technological progress, which are assumed to be given exogenously. This explanation clearly is not very satisfactory. Furthermore, the traditional models imply hat, in the long run, there should be no relationship between saving and growth. Yet, growth rates in the industrial countries are highly correlated with saving rates (see Carroll and Summers, 1989). These empirical anomalies have prompted a number of recent studies that build on the earlier work of Arrow (1962) and others to develop alternative models in which the long-run growth rate is endogenous and related to the saving rate.
Endogenous growth models assume that there are externalities associated with capital accumulation, which lessen or eliminate the effects of diminishing returns. One approach assumes that the endogenous accumulation of human capital is the means by which labor productivity grows over time (Lucas, 1988). The most important implication of the human-capital model is that both capital and labor (at least labor quality) can be produced. Thus, an increase in saving leads to a permanent increase in the rate of growth of both capital and effective labor supply. With both factors of production growing more rapidly, the economy avoids diminishing returns.
While the prediction that human capital per worker can grow without limit is questionable, the model yields important insights by highlighting the existence of plausible channels through which saving can affect long-run growth. The main implications are that measures of saving that ignore spending on human capital may give a misleading -prognosis for growth; that even with unchanged saving a more efficient mix of human and physical investment could have a sustained impact on growth; and that the social payoff to education and training may be higher than the private payoff. It then follows that government policy to stimulate human capital accumulation could improve efficiency while raising growth. Even if the prediction of unbounded human capital accumulation is not literally correct, the humancapital model still implies more persistent growth effects from saving than does the traditional model.
Other sources of endogenous growth are externalities associated with research and development (Romer, 1986), or with specialization of production (Romer, 1987a). These models contain clear policy prescriptions for subsidies to investment activities that involve externalities. To the extent that firms do not internalize the social product of their investments, government subsidies can increase both growth and welfare. The practical problem lies in identifying the externalities and estimating the warranted subsidies.
Grossman and Helpman (1989) have developed a model in which international trade interacts with new product development and learning. Because countries are assumed to have different factor productivities in research and development and in intermediate production, trade allows each country to specialize. By increasing world output of product designs and of intermediate goods, trade leads to faster growth.
Financial intermediation provides yet another channel for endogenous growth (Greenwood and Jovanovic, 1988). Financial intermediaries would spur growth by inducing individuals to undertake risky projects with high return that would be forgone in the absence of risk diversification. The higher growth from such investments allows the network of financial intermediation to broaden over time and promote further investment.
Sustained growth can also be derived from public sector capital accumulation, via expenditures on infrastructure, schools, law enforcement, and so forth (Barro, 1989a). In this framework, growth is generated by a fiscal policy rule that effectively makes government capital formation (financed through increases in taxes) proportional to the private capital stock. Because of the negative effects of tax distortions, there is a tradeoff from higher public sector capital formation: public sector capital raises output and productivity, but the high tax rates needed to finance this capital could create distortions that retard growth.
Any policy prescriptions made on the basis of endogenous growth theories will require accurate empirical information about the forces determining growth. The empirical testing of these theories, however, is in its infancy. The available evidence on the role of human capital accumulation for national growth is mainly impressionistic (Maddison, 1987). The evidence on the importance of other sources of endogenous growth basically consists of the large unexplained residuals that result from studies trying to account for observed growth without relying on external effects.
Notwithstanding the early stage of theoretical and empirical work on endogenous growth models, these models have important implications both for assessing the adequacy of saving and for the relationship between saving and growth. As regards the adequacy of saving, to the extent that these models rely on externalities associated with the investment process, they suggest that the returns available to private savers may not fully capture the social returns to investment, implying that the market-generated level of aggregate saving may be suboptimal. Concerning the relationship between saving and growth, it is important that any discussion of trends in national saving rates should take into account not just the accumulation of physical or financial assets but also human capital accumulation, research and development expenditure, and other nonconsumption expenditures that enhance future productivity. In addition, in an open economy, the distinction between saving and investment becomes crucial and policies to promote saving may have little effect on growth if the factors fueling growth are associated primarily with domestic investment.
This last consideration becomes more important as the degree of capital mobility increases. In a world of growing capital mobility, increases in saving lead to capital outflows and thus to smaller increases in domestic investment. While national wealth will still increase, the externality associated with domestic investment is no longer captured. Thus, in a world with mobile capital (both human and financial), a country that increases saving will not necessarily enjoy all of the external benefit from the new investment. This problem suggests the need for international cooperation for stimulating saving to provide efficient levels and distribution of world investment.