Most empirical evidence suggests that there is a strong and robust positive correlation between the rate of national saving and GDP growth.47 Although the direction of causality is a debated issue, saving and growth are likely to reinforce each other, resulting in a virtuous circle for countries that are successful in achieving a sustained increase in saving.48 Appropriate policies can help achieve such a result. An improvement in public saving appears to be the most effective policy instrument for increasing national saving, despite some partial offset in private saving.49 But conventional measures of public saving can be misleading, since they tend to exclude a substantial share of expenditures devoted to human capital accumulation. For example, fiscal consolidation that raised measured public saving at the expense of reduced current spending on primary education would be likely to harm growth prospects.50 Other potential policy-related influences on private saving include changes in taxation;51 interest rates; changes in the severity of liquidity constraints; and the possible redistributive effects of devaluation (from labor income to profits) if the propensity to save out of profits is higher. However, it has generally been difficult to identify many of these effects empirically for developing countries.
Evidence of the Links Between Policies and Saving
Panel data estimates from an IMF staff study provide some indications of the influences at work in the countries of the present study.52 In addition to reaffirming that an increase in public saving tends to be associated with higher national saving, the estimates also suggest that a decrease in the dependency ratio raised private saving substantially and that increases in per capita income raised private saving rates in developing countries. Changes in real interest rates had no significant effect on private or national saving. Increases in foreign saving affected national saving negatively; the offset was only partial, estimated to be between 40 percent and 50 percent, suggesting that an increased availability of external financing typically supports both higher consumption and higher investment. Finally, terms of trade windfalls were found to have a positive, but transitory, effect on national saving.
The estimated equations were able to track the path of saving reasonably well for most of the countries of the present study, suggesting that the above explanatory variables account for much of the variation in saving in these cases. However, the actual decline in national saving in Mexico beginning in the late 1980s and the substantial increase in Chile and Thailand during the same period were both considerably larger than predicted.53
Evidence for Individual Countries
Developments in saving in the eight countries can be summarized as follows (Chart 16).54 or of consumption smoothing, or both in the face of adverse terms of trade shocks.
Saving: Program Targets and Outcomes
Most IMF-supported programs in the eight countries aimed for an increase in saving, but this failed to materialize in about half of the cases.1 Part of the shortfall seems to be accounted for by interactions between public and private saving that were different than assumed under the programs. Thus, in over half of the years covered, increases in both public and private saving were programmed, whereas actual outcomes were more in line with the empirically observed tendency for increases in public saving to be partially offset by a fall in private saving (see chart). The largest shortfalls occurred in Mexico (see discussion in main text).
One should avoid drawing direct conclusions on the effectiveness or consistency of IMF-supported programs on the basis of this data. Exogenous shocks influenced the outcomes; moreover, observed empirical associations hold over a longer time period than that of IMF-supported programs, and variable lags can affect the contemporaneous associations, especially following periods of serious macroeconomic imbalances. Nevertheless, this exercise highlights the need to pay more attention to the implicit assumptions on private saving and its interaction with public saving, when formulating program targets for the external current account and fiscal and monetary policies. Evidence presented in Savastano (1995) suggests there was little shortfall, on average, in total national saving in the more recent programs that were covered by the last review of IMF conditionality (Schadler and others (1995)).
• In most countries, the national saving rate typically declined prior to adjustment as a result of expansionary policies that led to a fall in public saving, or of consumption smoothing, or both in the face of adverse terms of trade shocks.
• Several countries (Bangladesh, Ghana, Senegal, and Morocco) achieved a moderate increase in saving rates during adjustment, although the gain was short-lived for Ghana.55
• Only two countries (Chile and Thailand) achieved large and lasting increases in national saving rates, suggesting a shift to a path where higher saving and higher growth are mutually reinforcing.
• In marked contrast, saving rates in Mexico declined substantially during the second adjustment period.
How did policies contribute to these developments? Evidence from the country studies and Masson and others (1995) suggests that, in most cases, higher public saving was a major factor behind the increases in saving during the adjustment periods. However, non-policy-related factors also played a substantial role in a number of countries—notably the large fluctuations in the terms of trade as well as supply shocks in Ghana, Morocco, and Senegal.
Did specific aspects of the design of policies in Thailand and Chile contribute to their especially favorable saving performance? In Thailand, the only unusual factor was the size of the increase in public saving—to over 10 percent of GDP by the early 1990s—as part of a deliberate policy choice to offset demand pressures arising from capital inflows. Private saving was also buoyant, but country-specific econometric studies suggest that, rather than any policies designed specifically to promote higher private saving, the main factors were the beneficial effects of rapid real income growth, the favorable macroeconomic environment, as well as demographic changes associated with a large decline in the dependency ratio.56
Chile did implement a comprehensive set of policy measures in the 1980s designed to boost the national saving rate. The improvement in public saving brought about by the major fiscal adjustment and the reduction of central bank losses was reinforced by the funneling of increased export earnings from the copper price boom of the late 1980s into the Copper Stabilization Fund. Public enterprise reforms, including large-scale privatization and a hardening of budget constraints, raised enterprise profitability and also reduced the drain on the budget.57 A major pension reform began in 1981, involving a shift to a fully privatized saving plan and an increase in the retirement age. The initial impact of the reform was to reduce public saving and increase private saving; in addition, the reform was a major factor in deepening the capital market.58 Finally, important income tax reforms, designed specifically to increase saving through reduced rate progressivity and various incentives were implemented in 1984, but their impact on saving is difficult to assess.
The case of Mexico illustrates the potential for sizable unanticipated effects on private saving behavior as a result of changes in macroeconomic and structural policies. These effects can be especially large if there is some doubt about the consistency and sustainability of policies at the same time that financial sector liberalization and capital inflows ease liquidity constraints. In fact, program targets for private, and hence total, saving under the 1989-92 extended arrangement with the IMF were missed by wide margins.59 More generally, whereas most IMF-supported programs in the eight countries targeted an increase in saving, this did not occur in about half of the programs (Box 4). Although the causes of the saving decline in Mexico are not yet well established, several factors may have been responsible. The initial decline in 1988 resulted from a consumption boom fueled by purchases of imported durables that may have been brought forward because of doubts about the sustainability of the announced exchange rate anchor. But it is unlikely that this can explain the prolonged nature of the decline. Rather, there are indications that substantial capital inflows at a time of financial market liberalization eased consumers’ liquidity constraints.60