This section explores the determinants of saving in developing countries by analyzing ways in which the behavior of saving in such countries is likely to differ from that in industrial countries (see Section II of Part Two). While the discussion is of a theoretical nature, reference is made to supporting empirical evidence in the literature, as well as in Appendix III, which reports the results of a cross-section study of the determinants of the aggregate saving ratio for a large group of developing countries. This section is divided into three parts. Factors affecting the behavior of private and public saving are examined in the first two subsections, while interactions between the public and private sectors are considered in the third.
Private Saving
Fundamental motives for saving in developing countries are similar to the motives of savers in industrial countries. However, the saving decision by households in developing countries takes place in an environment that is very different from that facing households in industrial countries. These differences tend to affect the determinants of household saving behavior, such as resource endowments, the time profile of income streams, the nature of preferences, and opportunities for intertemporal reallocations of consumption available to households.
Household Resources
Household resource endowments in developing countries tend to differ from those in industrial countries in at least two important ways. First, household income is much lower in the developing countries. A large number of households are likely to have incomes that barely exceed subsistence, especially in countries with unequal income distribution. In countries in which a large portion of aggregate consumption is accounted for by households of this type, the response of private saving to increases in inter-temporal rates of return would be weak because many households would find it difficult to reduce consumption. Indeed, as shown in Table 2, among developing countries themselves, high-income countries appear to exhibit much higher saving ratios than low-income countries.26
Second, household incomes in developing countries may be more uncertain than in industrial countries, because a relatively large portion of households in developing countries derive their incomes from agriculture, where incomes are subject to large fluctuations due to variations in world prices of agricultural commodities and to local climatic conditions. This uncertainty is compounded by higher risks of disability, by macroeconomic instability, and by the less important role of income taxation in developing countries.27 While increased uncertainty regarding future labor income would enhance the precautionary motive for saving—see Deaton (1989) and Zeldes (1989)—in unstable macroeconomic environments it is likely that increased uncertainty would affect household saving adversely through its effects on intertemporal rates of return. The higher saving rates registered by Asian and low-inflation countries, for example, suggest the importance of a stable macroeconomic environment (Table 2).28
Time Profile of Income Streams
Retirement practices represent a way in which differing time profiles of income streams for individual households in industrial and developing countries may result in different household saving behavior, but the direction of the effect is unclear. Because of shorter life expectancy, differences in the organization of production (that is, greater importance of household production) and cultural factors, individuals in developing countries may spend a relatively small portion of their lifetimes in retirement. Thus, they would have less of an incentive to save for retirement than do their counterparts in industrial countries. However, this effect may be partially offset by the reduced role of formal social security arrangements in developing countries. Where no such system is in place, or where the reliability of future benefits is in question, provision for retirement may remain an important saving motive, even if the length of retirement is somewhat shorter than in countries with well-established social security systems.
Household Preferences
Several researchers (see, for example, Gersovitz, 1988) have argued that the greater importance of extended family arrangements may imply that intergenerational links are particularly strong in developing countries. If so, then as Barro (1974) has suggested, the effective planning horizons over which households make their consumption decisions would be lengthened. For an empirical support of this proposition, see Haque and Montiel (1989).29
The role of intergenerational links represents one way in which cultural factors can have important effects on differences in household saving behavior across countries. In general, such factors can be expected to operate primarily through the nature of household preferences, for example by affecting the degree with which the marginal utility of consumption declines with the level of consumption or with the rate of time preference. Cultural factors have sometimes been alluded to in explaining the high saving rates of Asian countries (Table 2), although the econometric estimates reported in Appendix III show little evidence of such regional effects, after controlling for other determinants of national saving rates.30
Intertemporal Rates of Return
From a policy standpoint, the most important difference between the environment in which household saving decisions are made in developing and industrial countries probably concerns household opportunities for borrowing and lending. In developing countries, the menu of assets available to households is often limited to cash, deposits in the domestic banking system, and durable goods. For households in rural areas, even this reduced menu of assets may not be available, since there may be no access to the banking system in the countryside. Furthermore, not only is the range of assets limited but the widespread use of ceilings on bank borrowing and lending rates in many developing countries has further constrained the options available to savers in such countries. Low or negative real interest rates on bank deposits have discouraged the accumulation of financial savings and may have represented a disincentive to total saving. On the other hand, nonmarket-clearing interest rates on bank credit have resulted in the rationing of such credit, leaving many house-holds out of the bank credit market altogether, thereby limiting the intertemporal choices available to them. While the limited availability of bank credit to households may encourage saving, the combination of a limited menu of assets and financial repression in developing countries may have on the whole acted to reduce saving in all forms—not just financial saving—in such countries. The evidence for this in the experience of countries that have pursued financial liberalization, however, does not appear to be strong.
Aggregate Household Saving
In addition to factors that affect individual savers in developing countries, other factors at the aggregate level may also contribute to differences in the saving behavior of the private sector in developing countries. These include, among others, the distribution of income, the rate of population growth, and the rate of growth of per capita incomes. Their effects on aggregate saving in developing countries are essentially the same as those found in industrial countries (see Section II of Part Two) and therefore will not be examined in detail here. It is worth noting, however, that while the role of income distribution in determining the aggregate saving rate has long been a controversial topic in developing countries, the case for a systematic, unambiguous link between the distribution of income and the aggregate saving ratio in a developing-country context lacks both theoretical basis and empirical evidence (see Gersovitz, 1988). Empirical research has been hampered by inadequate time-series and cross-country data on income distribution.
Public Saving
As in the case of private saving, public sector saving in developing countries tends to involve factors that differ markedly from those in industrial countries. Among those to be discussed here are the role of nonfinancial public enterprises, tax administration and the composition of tax revenue, the importance of external transfers and external debt, and the role of financial repression.
In a large number of developing countries, the public sector is heavily involved in productive activities through nonfinancial public enterprises. The net income of such enterprises represents a source of public sector revenue. In many cases, however, these enterprises have been managed with public policy objectives in mind other than profit maximization, thus necessitating transfers from the central government and reducing public sector saving. Although there may be valid microeconomic reasons for such enterprises to run deficits under some circumstances, in many cases losses have occurred because of a failure to adjust public sector prices upward with inflation. Further-more, in cases where these enterprises produce traded goods, exchange rate overvaluation has also contributed to public enterprise losses. Finally, poor management practices allowed by monopoly positions, low productivity, featherbedding, and inefficient investment have sometimes magnified these losses.
In many developing countries the collection of tax revenue is beset by severe administrative problems, although experience varies considerably across countries. In countries where these problems are prevalent, tax evasion is common and sources of tax revenues that place strong claims on administrative resources—such as income taxes—are relied on less intensively than others that are easier to administer, such as trade taxes and excise taxes. More important, perhaps, is that administrative problems may reduce public sector saving in many circumstances. When exogenous negative shocks to public sector saving cannot be offset by increases in tax revenue, the burden of adjustment falls on public consumption. If reductions in public consumption are subject to increasing marginal (economic or political) costs, then undue reliance may be placed on deficit financing in response to such shocks.31
External factors are an important source of shocks to public sector saving in developing countries. This is not only because export revenues are often an important source of public sector income through direct production, trade taxes, and the activities of marketing boards but also because external transfers and external debt service figure prominently in the budgets of developing country governments. Fluctuations in such transfers and in international interest rates will thus have large effects on public sector saving in developing countries, particularly where rigidities in current expenditures or tax revenues make such shocks difficult to offset.
Finally, since the public sector in the majority of developing countries is a net debtor to the domestic financial system, the maintenance of interest rates below equilibrium levels lowers public borrowing costs, at the expense of distorting the composition and possibly the level of private saving.
Interactions Between Public and Private Saving
The nature of the interaction between private and public saving in developing countries depends crucially on the extent to which anticipations of future taxes affect private saving behavior. For example, where the public sector is a substantial net debtor, the private sector may anticipate significant increases in taxation. The substitution effect of this expectation tends to favor current consumption at the expense of saving and investment, while the income effect would tend to reduce consumption in all periods. A second example is that of perfect “Ricardian equivalence,” as described in Part Two, section on “Government Saving.” Anticipations of future taxes have indeed been inferred to play an important role in the macroeconomics of developing countries. For example, such expectations have been referred to in explanations of capital flight and effects of the debt overhang. However, there is very little evidence in support of the Ricardian equivalence proposition itself for such countries. While individuals in developing countries may form expectations about their future tax liabilities in a systematic way, liquidity constraints may prevent them from acting on these expectations by adjusting their consumption-saving behavior as would be predicted by the Ricardian equivalence proposition—see Rossi (1988) and Haque and Montiel (1989). One implication of this result is that increases in taxation are indeed likely to have a significant effect on national saving in developing countries, as long as public consumption can be restrained.