Journal Issue

Dynamics of Asian Savings: The Role of Growth and Age Structure

International Monetary Fund. Research Dept.
Published Date:
January 1989
  • ShareShare
Show Summary Details

The capacity of Asian countries to mobilize savings since 1960 has been impressive (see World Bank (1984 and 1985, Table 5)). The general buoyancy of the savings rate in Asia, however, masks a diversity of experiences. For example, from an insignificant savings rate in the early 1960s, Singapore achieved a rate as high as 47 percent of gross domestic product (GDP) in 1984. By contrast, the Philippines recorded a modest rise in the savings rate during the latter part of the 1970s, but this was followed by a precipitous fall that culminated in an average rate in the 1980s lower than that in the 1960s. This paper examines savings behavior during 1960-85 in eight Asian countries—India, Indonesia, the Republic of Korea, Malaysia, the Philippines, Singapore, Sri Lanka, and Thailand—with a view to explaining their diverse experiences.

Given the close nexus of savings, capital accumulation, and growth, the need to understand the determinants of savings rates can hardly be overemphasized. Just as the savings rate has a salutory impact on capital accumulation and thereby on growth, prosperity influences savings through a “growth dividend” on the savings rate. The interdependence of growth and savings is at the root of theories of self-generating growth and development—the “virtuous circle” of development. Modigliani (1970), in his outstanding international cross-sectional study involving 36 countries, found growth and demographic factors to be of profound importance in explaining the intercountry differences in private savings rates during 1952-60.1 Growth and the changing demographic structure affect aggregate savings by making young savers both more affluent and more numerous than older dissavers. One of the major aims of this paper is to quantify the effect of growth and the changing age structure on savings in Asian countries. The Asian data have sufficient variation to enable one to test various hypotheses and theories of savings.

Although there have been quite a few studies on savings in Asia—for example, Fry (1978, 1986), Giovannini (1983, 1985), Gupta (1987), Ravallion and Sen (1986), and Rossi (1988)—most of these have been carried out with pooled cross-sectional and time-series data and have used the questionable assumption of a uniform marginal response of the savings rate to various factors across countries.2 Further, they have paid very little attention to the dynamic adjustment pattern of savings to various stimuli and have analyzed total domestic savings without distinguishing between private and public savings. The latter is a valid procedure only under the strong assumption of Ricardian equivalence, which yields automatic and complete neutralization of any variation in government savings by opposite movements in private savings.3 This study differs, therefore, from its precursors in several important respects. First, a distinction is made between private and public savings, and empirical findings are obtained from individual time-series analyses of countries, carried out in an integrated framework incorporating the effects of growth, demography, inflation, the terms of trade, and export orientation.4 Second, explicit attention is given to the dynamic lag structure of the adjustment of savings, and an attempt is made to benefit from the advances in the literature made by Hall (1978). Davidson and others (1978). Flavin (1981). and Blinder and Deaton (1985).

The plan of the paper is as follows. Section I presents the developments in savings rates and related characteristics of the eight countries, the issues, and the basic framework of analysis. Section II reports the empirical findings. The conclusions and some policy issues are presented in Section III. An Appendix contains a brief discussion of data issues and the empirical estimates.

I. Developments, Issues, and Basic Framework

This section reviews developments in savings in the Asian region over the past twenty-five years, raises issues germane to a study of savings behavior in these countries, and presents the framework for analysis.

Review of Developments and Issues

Table 1 presents some summary statistics relevant to an analysis of savings. During 1960-85, the savings rate showed considerable buoyancy in India, Indonesia, the Republic of Korea, Malaysia, and Singapore, but it fell in the Philippines and registered a very modest rise in Sri Lanka and Thailand, The secular inertia of the rate in the last three countries, however, conceals considerable variation over time. In Thailand, for example, the savings rate registered a healthy rise in the early 1970s before declining to 19.5 percent in the early 1980s.

According to the pattern of direct surplus generation by the central, state, and local governments, the countries can be divided into three distinct groups: (1) Indonesia and Singapore, where the governments generate 11-12 percent of GDP in savings; (2) the Republic of Korea, Malaysia, and the Philippines, where the governments account for 2.0-3.5 percent of GDP as savings; and (3) India, Sri Lanka, and Thailand, where the governments saved an insignificant proportion of GDP. Although differences in fiscal policies can explain part of the variation in national savings rates across countries, there remains a large diversity in private behavior that needs explanation. In this study, the government savings rate is taken as a policy variable, and the focus is on the determinants of private savings.

Growth and changing age structure are the two main determinants of the savings rate in the life-cycle approach. The growth rates achieved in Asia have ranged from the modest to the spectacular. For example, during 1960-85 the Republic of Korea and Singapore grew at an average annual rate of over 6.0 percent in per capita terms, but India managed an annual rate of only 1.7 percent. Similarly, the demographic experience has been quite diverse. During the same period, for instance, the proportion of population of working age—that is, aged 15-64—increased by as much as 11 and 16 percentage points in Korea and Singapore, respectively, while it remained practically stationary in India and Indonesia. It is interesting to note that in the context of developing countries, although Leff (1969, 1980) found a significant inverse relationship between dependency and savings rates, his findings have been challenged by Bilsborrow (1980), Ram (1982), and Rossi (1989), among others, on grounds of specification bias and sample selection bias.

There is considerable disagreement about the effect of inflation on savings in developing countries. Juster and Wachtel (1972) have found that inflation dampens consumer confidence and increases savings in a developed country. Deaton (1977) has presented a price-confusion effect—individual consumers buying one thing at a time cannot distinguish unanticipated inflation from relative price increases and, in response, tend to postpone purchases; unanticipated inflation therefore leads to involuntary savings.5 Inflation may also depress the savings rate in developing countries, however, because of a flight from currency and increased purchases of consumer durables, which bias consumption figures upwards.6 There are two further reasons why inflation may discourage savings in such economies. First, according to McKinnon (1973), there is the “conduit” effect of money, and complementarity between money and physical capital because of lumpiness of investment and a preponderance of self-investment in a developing economy with highly imperfect capital markets. The return on money goes down with inflation, and so do savings and investment. Second, the administered deposit rates of interest are frequently not adjusted appropriately in response to inflation, so that the real rate of interest falls. Although there remains a debate on the interest sensitivity of savings, there is reason to believe that lower real interest rates reduce savings.

With the exception of Indonesia in the early 1960s, there have been no instances of high and sustained inflation in the sample countries by international standards. There are, however, perceptible differences in the inflationary profiles of the eight countries: Malaysia and Singapore had an average annual inflation rate of 3.5 percent between 1960 and 1985; in India, Sri Lanka, and Thailand, the average varied between 5.6 percent and 7.4 percent; whereas in Korea, the Philippines, and Indonesia, inflation averaged 12.0-14.5 percent a year.7 Data on interest rates are extremely scanty. From whatever data are readily available, there seems to be some evidence of inertia in revising the largely administered deposit rates of interest appropriately in response to inflation.

Table 1.Savings Rates and Related Characteristics, Selected Asian Countries, 1960-85
TimeRep. of
VariablePeriodaIndiaIndonesiaKoreaMalaysiaPhilippinesSingaporeSri LankaThailand
Savings rateI14.923.533.8517.0815.5415.8113.9315.65
(percent of GDP)II23.1224.1028.9725.1813.9643.6016.8219.46
Government savingsbIII0.8912.533.603.423.059.01-1.520.41
(percent of GDP)IV0.4811.333.463.732.1212.510.11-0.63
Per capita incomecI180.54239.83476.39951.31423.911,415.03140.95297.33
(1980 U.S. dollars)II270.02551.692,253.771,963.27618.216,141.71302.60d815.30
Percentage of populationI54.0553.1953.2849.8551.8954.8754.3252.11
aged 15-64II56.3156.0664.9258.7556.4070.3760.8359.72
Consumer price indexI25.736.726.0752.7514.7650.3932.5231.53
(1980 = 100)II155.90158.91140.96125.51253.81117.35176.32127.09
Terms of tradeI139.93110.95e103.37207.86285.60132.60
(1980= 100)II140.00105.8878.2044.99101.1978.26
Exports (percent ofI5.248.883.3548.9610.7730.2917.52
Source: See the Appendix.

I: 1960 for India, Republic of Korea, Philippines, Sri Lanka, and Thailand; 1965 for Malaysia and Singapore; and 1967 for Indonesia. II: 1985. III: average for 1976-80 for all countries except Philippines, for which it is 1978-80. IV: 1981-84 for India and Singapore; 1981-83 for Indonesia; 1981-85 for Republic of Korea, Philippines, Sri Lanka, and Thailand; and 1981 for Malaysia.

Has been derived as current revenues less current expenditure of central, state, and local governments.

Has been derived as per capita GDP at 1980 prices divided by the average official exchange rate per U.S. dollar during 1979-81.



Source: See the Appendix.

I: 1960 for India, Republic of Korea, Philippines, Sri Lanka, and Thailand; 1965 for Malaysia and Singapore; and 1967 for Indonesia. II: 1985. III: average for 1976-80 for all countries except Philippines, for which it is 1978-80. IV: 1981-84 for India and Singapore; 1981-83 for Indonesia; 1981-85 for Republic of Korea, Philippines, Sri Lanka, and Thailand; and 1981 for Malaysia.

Has been derived as current revenues less current expenditure of central, state, and local governments.

Has been derived as per capita GDP at 1980 prices divided by the average official exchange rate per U.S. dollar during 1979-81.



A change in the terms of trade leads to three different effects on consumption: first, a direct effect through revaluation of exports and imports; second, a wealth effect through revaluation of wealth; and third, a pure substitution effect from relative price changes within and between periods. The possibility of a reduction in the savings rate in response to an adverse movement in the terms of trade has come to be known as the Laursen-Metzler-Harberger effect. Svensson and Razin (1983) and Persson and Svensson (1985) have shown the ambiguous nature of the effect of changes in the terms of trade on the savings rate. There are no readily available data on the terms of trade for Indonesia and Singapore for the period 1960-85. Among the other six countries, the terms of trade deteriorated considerably in Malaysia, the Philippines, Sri Lanka, and Thailand during 1960-85.

There are substantial differences in the export orientation and performance among the eight countries in the sample. According to Maizels (1968) and Lee (1971). export orientation may affect the savings rate in a developing economy. Apart from augmenting government revenue from export taxes and, hence, government savings, an increase in the ratio of exports to GNP may also increase private savings because exporters are supposed to have a relatively high propensity to save.

In sum, the Asian countries in the sample have undergone vast changes during 1960-85. Growth, demographic changes, various degrees of inflation, movements in the terms of trade, and the opening up of export markets are some of the dimensions of the process of their transformation. The speed of these changes has differed between subperiods as well as across countries. The experience of the countries in mobilization of savings has also varied over time as well as in comparison with one another. In the light of these other changes, the variety of savings experience merits a careful scrutiny.8

Empirical Specification

Even if successive cohorts of the population have identical tastes and preferences, and the young save only to dissave when they are old, it is well known that growth and demographic composition can affect the aggregate savings rate. With reference to Diamond (1965), it is fairly straightforward to construct models in the life-cycle-overlapping-generations framework to illustrate this dependence of the savings rate on growth and demography.

For example, consider a perfectly certain and competitive world with only one commodity and overlapping generations of agents who live for two periods—working during the first and in retirement during the second—with constant returns to scale. Cobb-Douglas production and utility functions, no bequest motive, population growing at the rate of n per period, and labor-augmenting technology progressing at θ per period. It is easy to show that, in such an economy,

where S is savings, Y is output, α the marginal product of capital, and β the coefficient of time preference. Equation (1) shows that the savings rate in this economy depends solely on productivity (1 - α), time preference (β), and growth (ΔYt/Yt).

In the steady state, growth in income (ΔYt/Yt) will equal 1 - 1/ (1 + θ)(1 + n), and the savings rate will be (1 - α)(1-β) times the steady-state growth rate. The equilibrium savings rate thus depends on productivity, time preference, technical progress (θ), and population growth (n). Despite the constancy of the average propensity to save (APS) of the individual, productivity growth and technical progress increase the aggregate savings rate by making the young savers more affluent than the young savers of last period, who constitute the older dissavers of today. The age-dependency ratio, defined as the ratio of old to young, is 1/(1 + n). An increase in population growth decreases the ratio, increases the number of savers relative to dissavers, and increases the APS in the economy.

The relationship between growth and the savings rate, as given in equation (3). does not explicitly bring out the dependence of the savings rate on the age structure of the population. For this, the evolution of the savings rate has to be derived from an arbitrary initial configuration. Note that in the simple illustrative economy described above, any arbitrary initial level of capital automatically determines the level of output by the condition of full employment. Given the level of output, the preference relation determines the level of savings and, hence, the capital stock for the next period. Thus, the entire future path of output in this economy can be traced for any given initial configuration. A bit of algebra applied to this simple economy, and approximation of log (1 - x) by -x, leads to

After substitution of (S/Y) by -log[1 - (S/Y)] = -log(C/Y), one has

as a complete description of consumption behavior of the economy starting from any arbitrary initial configuration. Because C and Y always come together in the form of a ratio, it is possible to transform them to per capita terms and rewrite equation (3) as

where c and y are the natural logarithms of per capita real consumption and income,9 and

There is a direct one-to-one relationship between the rate of growth of population and the age-dependency ratio in this simple illustrative economy, but in reality the age-dependency ratio also depends on a host of other demographic factors such as age-specific mortality rates and life expectancies (Clark and Spengler (1980)). Allowing for an adjustment period of two years, one may replace the last term in equation (4)—that is, β4n—by β4PR + β4ΔPR, a linear combination of the percentage of the population aged 15-64 (PR) and the change in that percentage (ΔPR) Furthermore, to test for the effects of changes in the terms of trade, of inflation, and of the degree of export orientation, variables are added to equation (4) for the change in the terms of trade (ΔT), the rate of inflation (π), and the ratio of exports to GNP (x). For estimation purposes, a random disturbance term, є, is also included to obtain

which is an empirical specification of private consumption behavior. Note that the expression is very similar to the form proposed by Blinder and Deaton (1985).

There has been extensive discussion in the literature about the precise functional form and dynamic structure of consumption functions such as that specified in equation (5). Two general types of models stand out: the rational expectations-permanent income hypothesis by Hall (1978) and Flavin (1981), among others, and the error-correction model by Davidson and others (1978). According to the rational expectations-permanent income framework of Hall, foreseen changes in income should not affect private savings or consumption because these changes are already embodied in past savings decisions made in an intertemporal framework. In this formulation, growth should not affect savings so long as it is anticipated. Alternatively, however, it is often argued that, particularly in developing countries, because of liquidity constraints even anticipated growth may induce changes in savings behavior. For example, even if a consumer knows with certainty that his income is going to double next year, he may not be able to increase his current consumption adequately if he cannot obtain the credit to do so.

Let superscripts a and u distinguish between anticipated and unanticipated categories, and let us rewrite (5) as

From equation (6), it is easy to see that, according to Hall’s (1978) hypothesis. β1 = β2 = β3 = 0. Furthermore, in the rational expectations-permanent income framework, there is the assumption of an “immortal” household (through the operation of a bequest motive). As Rossi (1989) has shown, in this framework variations in per capita real consumption may be related to variations in demographic composition through “cost of children” or overhead costs, but it is independent of the level of the age-composition variable; that is, β4 = 0,β5 ≷0. Thus, a test of the rational expectatons-permanent income hypothesis in the context of the present model is obtained by testing the hypothesis

H : β1 = β2 = β3 = β4 = 0.

From Blinder and Deaton (1985), this will be called the “surprise only” hypothesis.

Thus, there is a clear need to verify whether anticipated changes in income did matter in the determination of Asian savings rates in the light of Hall’s (1978) hypothesis and liquidity constraints. If they did, then one must find out whether anticipated changes in income affected the savings rate in the same way as unanticipated income; that is, one must test the hypothesis

when H1 stands rejected. An explanation of savings rates in terms of measured income is valid if and only if anticipated and unanticipated income affect savings in the same way.

Let us suppose that H2 is not rejected and that the analysis of savings can be legitimately carried out in terms of measured income. The adjustment of savings to changes in income is presumably far from instantaneous, and there is a need to specify the dynamic adjustment pattern—that is, the lag structure. In many developed countries the savings rate has been observed to vary procyclically in the short run but to remain constant over the long run. Is a similar pattern of adjustment of savings rates evident in Asian countries? Davidson and others (1978) have proposed an “error-correction” model that has an attractive dynamic property: the consumption or savings ratio is constant in the long run and any short-term “error,” or divergence of this ratio from its long-run value, is corrected over time to ensure the constancy of the ratio in the steady state. A test of the error-correction model in the context of equation (6) takes the following form:

H3 : β1 + β2 + β3 = 0,

when H2 is not rejected. Note that if the error-correction model is valid and one restricts the parameters such that β1 + β2 + β3 = 0, then equation (6) can be rewritten as

Furthermore, if the random error term in equation (7) is ignored, in a noninflationary steady state with Δct = Δyt = g, PRt=PR, and xt=x, one has

The long-run APS increases with growth and a rise in the proportion of people in the working ages (that is, PR) as long as β1 < 0, 0 < β2 < 1, β4 < 0.

Note that equation (6) involves anticipated and unanticipated variables that are unobservable. The procedure of Blinder and Deaton (1985) is followed to generate these unobserved variables. Accordingly, let m and yus denote money supply and U.S. GNP (at constant 1980 prices) in logarithmic form, and let m stand for a time trend factor. The unobserved anticipations za (for z=y, π, and ΔT) are generated, as forecasts one-period ahead, from

and the unanticipated values, zu are the residuals (v) of equation (9). Note that the formulation of equations (6) and (9) is identical to that proposed by Blinder and Deaton (1985). Equation (9) may be interpreted as an unrestricted reduced-form equation, with yus serving as a proxy for world demand.

As Blinder and Deaton (1985) have pointed out, the functional form (6) “nests” both the Hall hypothesis and the error-correction model. Moreover, a sequential testing procedure can be adopted—proceeding from H1 to H2 to H3—for discriminating among the alternative models. In the following section these sequential tests are carried out to choose the appropriate model for each country in the sample; checks are then made for overparameterization, and a simplification search is conducted. Checks are also conducted for the absence of serial correlation, parametric stability, and the “reliability” of steady-state predictions. Through this process, one equation that best explains private savings is chosen for each country.

II. Empirical Findings

Before the empirical results are presented, a methodological note is in order.

Concepts, Data, and Missing Observations

Private consumption expenditure and disposable income at current prices have been deflated by the implicit GDP deflator to arrive at private consumption and disposable income in real terms. These figures have been divided by population to obtain private consumption and disposable income in per capita real terms. Inflation is measured by the percentage rise in the consumer price index.10 The other variables are self-explanatory, and the Appendix contains details about the sources of data, problems of measurement, and limitations.

For India, data on personal disposable income were readily available and have been used to explain private consumption. The concept of income used for other countries is private, disposable, per capita real income. Aggregate private disposable income at current prices has been deflated by the implicit GDP deflator and divided by population to obtain the figures in per capita real terms. Aggregate private disposable income at current prices has been derived from GNP at current prices by adding (1) private unrequited transfers from abroad and (2) public sector subsidies, and by subtracting (3) all government taxes.11 Because data on taxes and subsidies are not available for much of the 1960s, there are serious nonuniformities and discontinuities in the income series. Two easy options were not to use the concept of private disposable income or to use it with a much shorter sample period. The former would be incorrect, whereas the latter would be a waste of useful information on all the other variables. Thus it has been assumed that, for the earlier period with missing observations, taxes (net of subsidies and as a proportion of total income)—say, k—did not vary from year to year. Let z* be the logarithm of “true” private real disposable income, and let z be its incorrectly measured counterpart, and ignore the net taxes levied by the government, Note that

which under the assumption of constancy of kt can be approximated by

Data for the entire sample period can thus be used by taking zt* as the income variable for the later years and zt as the income variable for the earlier years with missing observations. A dummy variable—taking the value unity for the years with missing observations and zero otherwise—is employed to remedy the problem of data discontinuity for all the countries except India.12

Tests of Alternative Models

Three versions of equation (6) were estimated by ordinary least squares: (1) an unrestricted version; (2) constraining β2=β2* consistent with H2; and (3) constraining β1 + β2 + β3 = 0 (and β2*=β2) consistent with H3. (Tables 3-5 in the Appendix present the estimates.)

Hypothesis H1, that “only surprises matter,” is convincingly rejected by a traditional F-test at the 5 percent level of significance. The test seems to be valid because no signs of serial correlation could be detected in the error structure for any country except Korea. The correlogram of residuals for Korea revealed a high negative value (-0.58) at the second lag. The problem of serial correlation, however, appears to be related to the absence of income terms with lags greater than 2. This seems to weaken the case for the validity of the rational expectations-permanent income hypothesis in Korea. The general result—that even anticipated movements in income or other variables have an impact in Asian countries—is intuitively appealing in view of the likely liquidity constraints in these developing countries.

Hypothesis H2, that decomposition of current income between the anticipated and unanticipated parts is not important (that is,β2=β2*) is rejected at the 5 percent level of significance only in the case of Sri Lanka.13 The estimated unrestricted version of the equation for Sri Lanka, however, suffers from some serious deficiencies; for example, the long-run income elasticity is abnormally large (about 3).14 In general, the evidence suggests that the decomposition of current income into anticipated and unanticipated categories is not of great importance in the countries under investigation when lagged income and consumption are also included in the model as explanatory variables. At first sight this may appear to be a refutation not only of the rational expectation version of the permanent income hypothesis but also of the permanent income hypothesis per se. This is not true, and due cognizance must be taken of the presence of lagged income and consumption terms in the relationship when interpreting the result regarding H2:β2=β2* . As Davidson and others (1978) have shown, a distributed lag formulation of the consumption function of the form of equation (6) even under H3 (which assumes the validity of H2:β2=β2*) is consistent with distributed lag variants of the permanent income hypothesis.15 A once-and-for-all increase in income elicits quite different responses in consumption behavior in the model even when β2=β2*, depending on whether the increase is permanent or temporary.

The overall goodness of fit of equation (6) under H2: β2=β2*is fairly impressive, especially because the dependent variable is the rate of growth of per capita private consumption in real terms and does not have a time trend. A straightforward way of testing the error-correction model is to look at the value of long-run income elasticity of consumption obtained by estimating equation (6) under H2 : β2=β2* and to judge whether it is significantly different from unity. The results of this exercise show that, in only three of the countries—Indonesia, the Philippines, and Thailand—is the long-run elasticity significantly different from unity. Thus, at least for five countries there are reasons to believe that the error-correction model is valid. Notice that the error-correction model does not require “that the data satisfy a unit elasticity restriction,” but only “that the model satisfy this restriction and the data are consonant with the model” (Davidson and others (1978. p. 681)). When income increases, the average propensity to consume can decrease in the short run because of an initial disequilibrium effect, but a long-run income elasticity that is less than unity implies that the APS increases ad infinitum in a steady state with growth in income—a result that is contrary to prior assumptions.

The anomalous result in the case of Indonesia, the Philippines, and Thailand could well be because of overfitting—there are variables in the equations that do not really matter in the context of each of these countries. The investigation turns, therefore, to whether the unsatisfactory results for these three countries are due to the presence of such redundant variables.

Factors That Determine the Private Savings Rate in Each Country

Although all the factors included in the model can, in theory, affect a country’s savings rate, for historical, political, or structural reasons they need not have played equally important roles in every country. For example, the rate of inflation in Malaysia and Singapore has not been high or volatile enough for the sample data to reveal the impact of inflation on the private savings rate in these two countries. Similarly, exports as a share of GNP in India have hardly grown over the sample period, and it is therefore difficult to detect the impact of this ratio on savings in India.

A simplification search was conducted to narrow down the list of important factors that influenced private savings in each of our sample countries.16 The results are reported in Table 2. Note that the standard errors have not been corrected for “generated regressors.”17

It is interesting to note that the hypothesis H2 : β2=β2* for Sri Lanka is not rejected at the 5 percent level of significance once the redundant variables are omitted from the equation. Similarly, it may be recalled that the hypothesis that the long-run income elasticity of savings is unity—that is, holds—was rejected for Indonesia, the Philippines, and Thailand when all the factors were included in the model. Once the superfluous variables are dropped from the equation, H3 is not rejected for Indonesia, but it remains rejected for the Philippines and Thailand. Nevertheless, in the specification of the final chosen equation, the assumption of unitary long-run income elasticity is retained for the Philippines and Thailand.

Growth unambiguously leads to increased private savings in all the eight countries, and the coefficient of the income-growth variable is less than unity. According to these calculations, a 1 percentage point increase in the rate of growth of per capita real income leads, on average, to a 1 percentage point rise in the private savings rate in the steady state. The change in the APS, however, differs substantially between the short run and the long run. The short-run behavior of the APS depends on the extent of the initial disequilibrium. The coefficient of the feedback variable (the lagged ratio of consumption to income) is highly significant, indicating that an acceleration in the rate of growth results in a substantially greater increase in the APS in the short run than in the long run.18 In only one country, Korea, do changes in the rate of growth play an important role in determining private savings. Houthakker and Taylor. (1970) and Davidson and others (1978), however, found similar results for other countries.

On average, a 1.0 point increase in the percentage of the population aged 15-64 leads to a 1.6 percentage point rise in the APS in the long run in India, Korea, Malaysia, Singapore, and Sri Lanka. No lasting effect of age composition on savings behavior in Indonesia and the Philippines could be detected.19 In Indonesia, however, where the age structure of the population has shown very little long-term movement, evidence suggests that changes in the age composition of the population (ΔPR) affect private savings in the short run. These short-run movements are also found to be of importance in India. Korea, Malaysia, Singapore, and Thailand, indicating the existence of lags in the adjustment of savings to age composition. Thailand is the only country in which the private savings rate appears to decline in the long run with lower age dependency; however, no satisfactory reasons could be found for the large perverse coefficient of the age-composition variable in that country.20

Although the age composition of the population exerts a significant influence on the private savings rate in all the countries except the Philippines,21 an important question arises regarding the significance of the demographic variables. Primary information on age composition is available only at intervals of ten years or so, when a census is conducted. For the intermediate years, the data are prepared by demographers through interpolation. The correlation between age composition and time trend can, therefore, be expected to be high in any of the countries in the sample. There is then the question of whether the coefficient of the age composition variable is identified or whether it simply captures the effect of a time trend or some other time-trending variable. The time-series property of the age-composition variable has a fairly complicated variation structure in every country—a simple time trend yields a high correlation coefficient but fails to explain the movements in the demographic pattern. Thus, there is no prima facie reason for undue skepticism.22

Table 2.Consumption Functions
Rep. of
IndiaIndonesiaKoreaMalaysiaPhilippinesSingaporeSri LankaThailand
Private disposable income
Rate of growth0.97740.41820.67430.58770.15460.80090.50860.6647
Change in the rate-0.1800
of growth(2.44)
Lagged consumption-1.0931-0.7414-0.4440-0.6543-0.2515-0.1851-0.3075-0.5795
income ratio(4.50)(2.82)(5.44)(4.87)(3.88)(2.40)(2.81)(3.05)
Percentage of population
aged 15-64
Changes in terms
of trade
Exports (percent-0.00380.0036-0.0045
of GNP)(1-77)(2.68)(1.35)
Standard error0.01420.02810.01010.01780.01800.01960.02750.0169
Sum of squared residuals0.00320.00870.00140.00480.00580.00540.01290.0046
Degrees of freedom1611141518141716
Steady-state average
propensity to consume
(APC) at no growth0.69280.80670.85590.59910.84220.56330.77570.8318
No. of observations x sum
of squared first four
Lagrange multiplier (LM)
test for first-order
Chow-test for parametric
Source: See the Appendix.Note: The dependent variable is first difference of logarithm of private consumption expenditure. Figures within parentheses under country names are periods of analysis: those under coefficients are t-values.

The period considered is 1980-85 except for India. Republic of Korea, and Singapore, for which it is 1980-84, 1981-85, and 1980-84, respectively.

Source: See the Appendix.Note: The dependent variable is first difference of logarithm of private consumption expenditure. Figures within parentheses under country names are periods of analysis: those under coefficients are t-values.

The period considered is 1980-85 except for India. Republic of Korea, and Singapore, for which it is 1980-84, 1981-85, and 1980-84, respectively.

On average, inflation has an adverse impact on private savings in the countries under study, but the effect seems to vary considerably from country to country. Inflation, both anticipated and unanticipated, results in lower private savings in Indonesia. Unanticipated inflation reduces private savings in India, but anticipated inflation does not have an impact. In Korea and the Philippines, however, only anticipated inflation matters: in Korea it has an adverse impact on private savings; in the Philippines the effect seems to be the reverse. Inflation, whether anticipated or not, does not affect private consumption-savings behavior in the two low-inflation economies of Malaysia and Singapore or in Sri Lanka and Thailand.

In general, adverse movements in the terms of trade tend to depress private savings in the sample countries. Any anticipated deterioration in the terms of trade affects private savings adversely in India, Malaysia, Sri Lanka, and Thailand. Similar movements, when they are unanticipated, also affect private behavior in Malaysia, the Philippines, and Sri Lanka. As noted earlier, the effect of movements in the terms of trade on private consumption could not be estimated for Indonesia and Singapore, although Indonesia, as a major oil exporter, may have witnessed some sharp changes in its terms of trade during the reference period. For Korea, where there has been a modest and gradual deterioration in the terms of trade, we could not find any evidence of the Laursen-Metzler-Harberger effect.

The Maizels-Lee hypothesis, that increased export orientation augments savings, is for the most part rejected by the evidence. Even for Korea, exports as a proportion of GDP do not seem to have played a direct role in the determination of private consumption or savings. As can be verified from the sign of the relevant coefficient in Table 2, whereas private savings seems to have risen in line with increasing share of exports in GDP in Indonesia and Thailand, for Malaysia it seems to have declined. This may point to the possibility that exporters do not have a higher propensity to save relative to others in every country, or that the Maizels-Lee hypothesis is sensitive to the pattern of ownership of export industries.

To verify the reliability of the results, the final equations for each country were tested for (1) their error structure, (2) parametric stability, and (3) implied steady-state values. None of the equations reported in Table 2 seems to suffer from any first-order serial correlation. The Lagrange-multiplier (LM) test statistic for first-order serial correlation, which is distributed as x21, is not significant at the 5 percent level of significance. No prima facie evidence of higher-order serial correlation could be detected for any country except Korea. For Korea, the statistic obtained by multiplying the sum of squares of the first four autocorrelations by the number of observations (see Godfrey (1978)) is a significantly high 14.47. The statistic is high for all the specifications tried out in this study. The source of the misspecification could not be discovered.

The final equations are parametrically stable. The Chow-test carried out with out-of-sample forecasts for the subperiod of the early 1980s did not reveal any nonconstancies.

The steady-state average propensity to consume (APC) was computed under the assumption that there is no growth in per capita income, no inflation, and no movements in the terms of trade. Exports were assumed to constitute 40 percent of GNP, as is the present case for Korea. Furthermore, 65 percent of the population was assumed to be aged 15-64 in the steady state. On the basis of the steady-state APS (that is, 1 - steady-state APC), the countries fall into three groups: Indonesia, Korea, the Philippines, Sri Lanka, and Thailand with a steady-state APS of roughly 0.2; Malaysia and Singapore with a steady-state APS of 0.4 approximately; and India with a steady-state APS of 0.3.23

III. Conclusions

Savings behavior in Asian countries is diverse in two ways: first, there have been large changes in savings rates over time in many of the countries; second, the rates have varied markedly across countries.

Both diversities are partially explained by government behavior. Except for the Philippines and Thailand, national savings rates rose in all the countries in the sample between the early 1970s and early 1980s—a development that is intimately connected to government behavior. In Indonesia, Malaysia, and Singapore, government savings as a proportion of GDP rose by 3.5-4.0 percentage points between the early 1970s and early 1980s. During the same period, the Korean Government generated an additional 1.5 percent of GDP, whereas in Sri Lanka it was only 0.6 percent. In India and Thailand, in contrast, the additional contribution was negative: -1 percent and -2 percent, respectively. Similarly, as is evident from Table 1, the differences in national savings rates across countries are related to the differences in direct surplus generation by governments.

The primary explanation for the two diversities, however, is found in the behavior of private savings and its determinants. Growth in per capita income and lower dependency rates stimulate private savings. The countries that rank high in terms of these two criteria—for example, Korea and Singapore—have higher rates of private savings as well. Similarly, the variations in growth rates and demographic changes over time explain a large part of the changes in the savings rate in any particular country in the sample. For example, the deceleration in the rate of income growth in the Philippines accounted for a large part of the fall in that country’s savings rate. The dependency ratio in Korea fell at a much faster rate during the 1970s than it did a decade earlier, and this partially explains the country’s higher savings rate in the 1970s compared with the 1960s. According to these findings, on average a 1 percentage point increase in the rate of growth of per capita income leads to an equivalent increase in the private savings rate in the long run. A 1.0 percentage point fall in the dependency ratio results, on average, in a 1.6 percentage point rise in the long-run private savings rate in India, Korea, Malaysia, Singapore, and Sri Lanka.

Inflation and movements in the terms of trade are found to be two additional determinants of private savings. Inflation, both anticipated and unanticipated, is found to have an adverse effect on savings in some of the countries. It is interesting to note that the effect is quite different from that found in a developed economy. Apart from the measurement errors attributable to “the flight from currency” during an inflationary situation and McKinnon’s (1973) hypothesis of complementarity between the demand for money and savings resulting from the lumpiness of investment and preponderance of self-finance, the difference could be due to the repressed nature of financial markets in a developing economy. In a developed economy the effect of anticipated inflation is weak when the nominal rate of interest adjusts to maintain the equilibrium real rate of interest. In developing economies, however, because of financial repression through administered rates of interest, anticipated inflation may depress savings by reducing the real rate of interest. In view of this, although cross-sectional results (for a much shorter period for which the data on interest rates are readily available) fail to reveal any interest sensitivity of savings in line with findings by Giovannini (1983, 1985) and Ravallion and Sen (1986), the inflation sensitivity of savings may be interpreted as weak partial evidence of the responsiveness of savings to changes in the interest rate.

Countries with price stability, such as Singapore, have tended to be high savers as well. Inflation, which reached crisis proportions by Asian standards in the early 1970s, declined markedly in the early 1980s in all the countries except the Philippines and Sri Lanka. The deceleration in inflation has contributed to the increase in Asian savings rates.

In conformity with Fry (1986), who found evidence of the savings rate declining in the face of adverse movements in the terms of trade, the present analysis finds consistent evidence in favor of the Laursen-Metzler-Harberger effect in a majority of the sample countries. Terms of trade movements during 1980-85 were much more favorable than a decade earlier for most countries except Malaysia, the Philippines, and Thailand. As a result, the savings rate rose in the other countries but actually declined in these three countries.

The evidence from the eight Asian countries is somewhat negative with regard to the Maizels (1968) and Lee (1971) hypothesis: although export orientation did not affect private savings in at least five of the countries, there is some support for the hypothesis in the cases of Indonesia and Thailand. Finally, for Malaysia the direction of the effect is perverse: an increased export orientation actually serves to reduce the private savings rate.

Some of the short-run variations in the savings rate in a particular country can be attributed to the adjustment lags present in savings behavior. The adjustment of savings to an economic stimulus is far from instantaneous and differs between the short run and the long run. Although there is clear evidence of procyclical behavior of the savings rate over the short run, the evidence is consistent with secular constancy of the savings rate in the long run in the error-correction framework of Davidson and others (1978).

In the context of cross-country differences in savings behavior, one of the interesting findings is that, even under identical conditions, in the long run the countries can be divided into three distinct categories: the first—Malaysia and Singapore—with a long-run private savings rate of 0.4 of private income: the second—India—with a rate of 0.3: and the third—Indonesia. Korea, the Philippines. Sri Lanka, and Thailand—with a rate of 0.2. Although this categorization could be due to misspecification—for example, bias from omitted variables—it could well be due to particular differences in taste, which is another way of saying that some countries save more because their residents want to. For any individual country, could it be argued, on similar lines, that its residents have developed a preference (taste) for increased savings over time? The parametric stability of the estimated equations rules out such a possibility.

Movements in the savings rate are of great interest to economists and policymakers because of the close connection between savings and the current account of the balance of payments and the close links between savings, capital accumulation, and growth. Although many crucial policy instruments—such as provident funds, tax incentives, and innovative savings schemes—could not be incorporated in the analysis because of data problems, there are three important policy conclusions that can be drawn from the analysis in this paper. First, governmental initiative in generating savings on its own account can lead to a substantial increase in the national savings rate in many developing countries. There are three countries in the sample—Indonesia, Malaysia, and Singapore—that demonstrate this adequately. Second, just as savings leads to capital accumulation and hence growth, growth leads to savings by making the young savers more affluent than the older dissavers. Technical progress and a more efficient allocation of resources can foster growth even without any augmentation in the supply of physical factors. Such increases in the growth rate bring forth an increase in savings and can set off the process of self-generating growth. Thus, the importance of appropriate policies for increasing the efficiency of resource allocation and providing incentives for technical innovation can hardly be overemphasized. Third, population planning has been widely advocated for economic development, and it has been pursued in many Asian countries. The consequent reduction in dependency rates has resulted in a bonus in the form of increases in the private savings rate. A fall in the birthrate can increase the savings rate in a developing country by increasing the rate of growth of per capita income for an unchanged rate of growth of aggregate national income and by reducing the dependency rate in the short run and the medium run.

Can the Asian countries in the sample sustain their high savings rates? The tentative answer varies by country. Singapore and Korea have sustained remarkably high growth rates and low age-dependency ratios for quite some time. The age structure of the population in these two countries leans more heavily toward those aged 15-64 than in a demographically stable country such as Sweden. Any deceleration in growth and rise in dependency rates will adversely affect savings in these two countries. For the rest of the countries in the sample, the levels of per capita income and the proportions of the population in the working ages are far below those in Korea and Singapore. In the process of catching up with their more prosperous neighbors, these countries should be increasing their savings rate over time, provided that there are no adverse movements in the terms of trade and no inflationary buildup.

Table 3.Empirical Estimates of the Unrestricted Consumption Function: Equation (6)
Rep. of
IndiaIndonesiaKoreaMalaysiaPhilippinesSingaporeSri LankaThailand
Lagged consumption-1.1242-1.0826-0.5369-0.8831-0.6307-0.6041-0.2450-1.0537
Private disposable income
Percentage of population
aged 15-64
Changes in terms of trade
Exports (percent0.0079-0.00100.00060.0025-0.00010.0053-0.0048
of GNP)(1.01)(0.42)(0.45)(1.53)(0.04)(1.91)(1.48)
Standard error0.01430.02480.01260.01700.01830.01850.02740.0129
Sum of squared residuals0.00220.00490.00140.00320.00370.00340.00830.0018
Degrees of freedom11891111101111
Long-run income elasticity0.93460.35521.12190.75380.61070.52272.99490.8210
No. of observations × sum
of first four auto-
correlations squared0.71203.081015.28203.93602.03908.41102.44404.2090
Source: See the Appendix.Note: The dependent variable is first difference of logarithm of private real consumption expenditure. Figures within parentheses under country names are periods of analysis; those under coefficients are t-values.
Source: See the Appendix.Note: The dependent variable is first difference of logarithm of private real consumption expenditure. Figures within parentheses under country names are periods of analysis; those under coefficients are t-values.
Table 4.Empirical Estimates, of the Consumption Function: Equation (6), under H2 : β2=β2*
Rep. of
IndiaIndonesiaKoreaMalaysiaPhilippinesSingaporeSri LankaThailand
Lagged consumption-1.0275-1.1073-0.5134-0.8361-0.6335-0.6097-0.3461-1.0561
Private disposable income
Percentage of population
aged 15-64
Changes in terms of trade
Exports (percent of-0.0050-0.00080.00030.0025-0.00010.0001-0.0057
Summary statistics
Standard error0.01430.02380.01240.01680.01760.01760.03240.0126
Sum of squared residuals0.00250.00510.00150.00340.00370.00340.01260.0019
Degrees of freedom129101212111212
Long-run income elasticity0.89130.35851.15040.78260.61220.52080.94610.8329
No. of observations × sum
of first four auto-
correlations squared1.68602.208014.48103.43102.14608.67801.63704.3350
Source: See the Appendix.Note: The dependent variable is first difference of logarithm of private real consumption expenditure. Figures within parentheses under country names are periods of analysis: those under coefficients are t-values.
Source: See the Appendix.Note: The dependent variable is first difference of logarithm of private real consumption expenditure. Figures within parentheses under country names are periods of analysis: those under coefficients are t-values.
Table 5.Empirical Estimates of the “ECM’-Type Consumption Function: Equation (6) under H312 + β3 = 0
Rep. of
IndiaIndonesiaKoreaMalaysiaPhilippinesSingaporeSri LankaThailand
Rate of growth of private0.96550.41440.45950.61O00.24970.78780.52450.7064
disposable income(12.39)(1.64)(6.21)(4.27)(0.81)(7.55)(2.40)(2.75)
Lagged consumption-1.0163-0.7296-0.6435-0.7665-0.2122-0.1813-0.3423-0.6225
income ratio(3.20)(2.57)(7.44)(4.33)(1.37)(2.35)(1.58)(1.90)
Percentage of population
aged 15-64
(0.08)(0.87)(5.10)(0.90)(0.80)(0 96)(0.08)(0.09)
Changes in terms of trade
Exports (percent-0.0024-0.00380.0009000410.00070.0005-0.0052
of GNP)(0.34)(1.68)(0.75)(2.79)(0.24)(0.18)(1.27)
Summary statistics
Standard error0.01450.02940.01240.01820.02090.01950.03110.0185
Sum of squared residuals0.00270.00860.00170.00430.00570.00460.01260.0045
Degrees of freedom1310111313121313
Steady-state APC at no growth0.89131.04940.79710.48510.74750.53030.74841.1504
No. of observations × sum
of squared first four
Source: See the Appendix.Note: The dependent variable is first difference of logarithm of private real consumption expenditure. Figures within parentheses under country names are periods of analysis: those under coefficients are t-values.
Source: See the Appendix.Note: The dependent variable is first difference of logarithm of private real consumption expenditure. Figures within parentheses under country names are periods of analysis: those under coefficients are t-values.
APPENDIX: Data Sources and Empirical Estimates

The concept of savings used in this study is that of gross national savings (GNS).24 Figures for estimated GNS were obtained by adding the current account balance to gross domestic investment (GDI). The current account data were taken from the balance of payments account, whereas GDI figures were taken from national accounts. Note that the savings figures include net current transfers, exclude statistical discrepancies, and use two different data accounts (national accounts and balance of payments). Net current transfers have been included to make the savings figures representative of available resources for investment purposes. Statistical discrepancies have been excluded on the presumption that consumption figures are perhaps less reliable than other national accounts categories and may be underestimated. Exclusion of the statistical discrepancies produces underestimates of savings, which are not as damaging as overestimates in a discussion of high savings rates in Asia. Data from balance of payments accounts are not always consistent with data from national accounts; this was the case with two countries in our study—India and Indonesia (see Sigit (1985)). The uniform practice of taking the figures of foreign savings for all the eight countries as the current account balance (with sign reversed) was followed. The average official exchange rate for every year was used to convert foreign currency figures into local currency.

The population figures were mid-year estimates. The price index used to deflate figures at current prices to obtain their constant price counterparts was the implicit GDP deflator. For some data-deficient countries (for example, Indonesia), the consumer price index was used to extrapolate the available data on the implicit GDP deflator.

All the data used in the study were taken from the International Monetary Fund’s International Financial Statistics (IFS), with three major exceptions. Data on age composition were from the World Bank’s demographic data base; the data on personal disposable income for India were from the Government of India’s National Accounts Statistics (Central Statistical Organization, January 1987); and the data on consolidated government revenue and expenditure were from the Fund’s Government Finance Statistics Yearbook. “Government” refers to the consolidated government sector, comprising central, state, and local governments.

The inflation variable is the rate of rise in the consumer price index. The terms of trade is the unit value index of exports as a percentage of the unit value index of imports. Exports as a percentage of GNP were calculated by taking the relevant figures from the national accounts data in IFS. Money supply figures used in the analysis refer to narrow money. The data for GNP in the United States are at constant prices.

Empirical estimates of the consumption function (equation (6)), with and without restrictions, are presented in Tables 3-5.


    BilsborrowRichard E.“Dependency Rates and Aggregate Savings Rates Revisited: Corrections, Further Analysis, and Recommendations for the Future,”in Research in Population Economicsed. by JulianL. Simon and JulieDaVanzoVol. 2 (Greenwich, Connecticut: JAI1980) pp. 183204.

    BlinderAlan S. and AngusDeaton“The Time Series Consumption Function Revisited” Brookings Papers on Economic Activity: 2 (1985) The Brookings Institution (Washington) pp. 465511.

    BransonWilliam H. and AlvinKlevorick“Money Illusion and the Aggregate Consumption Function,”American Economic Review (Nashville, Tennessee) Vol. 59 (December1969) pp. 83250.

    ClarkRobert L. and JosephJ. Spengler“Dependency Ratios: Their Use in Economic Analysis,”in Research in Population EconomicsVol. 2ed. by JulianL. Simon and JulieDa Vanzo (Greenwich, Connecticut: JAI1980).

    DattaGautam and ParthasarathiShome“Social Security and Household Savings: Asian Experience,”Developing Economies (Tokyo) Vol. 19 (June1981) pp. 14360.

    DavidsonJames E. H.DavidF. HendryFrankSrba and StephenYeo“Econometric Modelling of the Aggregate Time-Series Relationship Between Consumers’ Expenditure and Income in the United Kingdom,”Economic Journal (Cambridge, England) Vol. 88 (December1978) pp. 66192.

    DeatonAngus“Involuntary Saving Through Unanticipated Inflation,”American Economic Review (Nashville, Tennessee) Vol. 67 (December1977) pp. 899910.

    DiamondPeter A .“National Debt in a Neoclassical Growth Model,”American Economic Review (Nashville, Tennessee) Vol. 55 (December1965) pp. 112650.

    FlavinMarjorie A .“The Adjustment of Consumption to Changing Expectations about Future Income,”Journal of Political Economy (Chicago) Vol. 89 (October1981) pp. 9741009.

    FryMaxwell J.“Money and Capital or Financial Deepening in Economic Development?”Journal of Money Credit and Banking (Columbus, Ohio) Vol. 10 (November1978) pp. 46475.

    FryMaxwell J.“Terms of Trade Dynamics in Asia: An Analysis of National Saving and Domestic Investment Responses to Terms-of-Trade Changes in 14 Asian LDCs,”Journal of International Money and Finance (Guildford, England) Vol. 5 (March1986) pp. 5773.

    GiovanniniAlberto“The Interest Elasticity of Savings in Developing Countries: The Existing Evidence,”World Development (Oxford, England) Vol. 11 (July1983) pp. 601607.

    GiovanniniAlberto“Saving and the Real Interest Rate in LDCs,”Journal of Development Economics (Amsterdam) Vol. 18 (August1985) pp. 197217.

    GodfreyL.G.“Testing Against General Autoregressive and Moving Average Error Models when the Regressors Include Lagged Dependent Variables,”Econometrica (Evanston, Illinois) Vol. 46 (November1978) pp. 12931301.

    GrahamJohn W.“International Differences in Saving Rates and the Life Cycle Hypothesis,”European Economic Review (Amsterdam) Vol. 31 (December1987) pp. 150929.

    GuptaKanhaya L.“Aggregate Savings, Financial Intermediation, and Interest Rate,”The Review of Economics and Statistics (Cambridge, Massachusetts) Vol. 69 (May1987) pp. 30311.

    HallRobert E.“Stochastic Implications of the Life Cycle-Permanent Income Hypothesis: Theory and Evidence,”Journal of Political Economy (Chicago) Vol. 86 (December1978) pp. 97187.

    HaqueNadeem Ul and PeterMontiel“Ricardian Equivalence Liquidity Constraints and the Yaari-Blanchard Effect: Tests for Developing Countries” (unpublished; Washington: International Monetary Fund1987).

    HouthakkerHendrik S. and LanceD. TaylorConsumer Demand in the United States 1929-70 (Cambridge, Massachusetts: Harvard University Press1970).

    HowardDavid H.“Personal Saving Behavior and the Rate of Inflation,”Review of Economics and Statistics (Cambridge, Massachusetts) Vol. 60 (November1978) pp. 54754.

    JusterF. Thomas and PaulWachtel“A Note on Inflation and the Saving Rate” Brookings Papers on Economic Activity: 3 (1972) The Brookings Institution (Washington) pp. 76579.

    LeeJoong-Koon“Exports and the Propensity to Save in LDCs,”Economic Journal (London) Vol. 81 (June1971) pp. 34151.

    LeffNathaniel H.“Dependency Rates and Savings Rates,”American Economic Review (Nashville, Tennessee) Vol. 59 (September1969) pp. 88696.

    LeffNathaniel H.“Dependency Rates and Savings Rates: A New Look,”in Research in Population Economicsed. by JulianL. Simon and JulieDa VanzoVol. 2 (Greenwich, Connecticut: JAI1980) pp. 20514.

    McKinnonRonald I.Money and Capital in Economic Development (Washington: The Brookings Institution1973).

    MaizelsAlfredExports and Economic Growth of Developing Countries (New York: Cambridge University Press1968).

    ModiglianiFranco“The Life-Cycle Hypothesis of Saving and Intercountry Differences in the Saving Ratio,”in Induction Growth and Trade: Essays in Honour of Sir Roy Harroded. by W. A.EltisM.FG.Scott and J.N.Wolfe (Oxford, England: Clarendon Press1970).

    PaganAdrian“Econometric Issues in the Analysis of Regressions with Generated Regressors,”International Economic Review (Philadelphia) Vol. 25 (February1984) pp. 22147.

    PerssonTorsten and LarsE.O. Svensson“Current Account Dynamics and the Terms of Trade: Harberger-Laursen-Metzler Two Generations Later,”Journal of Political Economy (Chicago) Vol. 93 (February1985) pp. 4365.

    RamRati“Dependency Rates and Aggregate Savings: A New International Cross-Section Study,”American Economic Review (Nashville, Tennessee) Vol. 72 (June1982) pp. 53743.

    RavallionMartin and AbhijitSen“On Some Estimates of Asian Savings Function,”Economic Letters (Amsterdam) Vol. 20 (1986) pp. 121—24.

    RossiNicola“Government Spending, the Real Interest Rate, and the Behavior of Liquidity-Constrained Consumers in Developing Countries,”Staff PapersInternational Monetary Fund (Washington) Vol. 35 (March1988) pp. 10440.

    RossiNicola“Dependency Rates and Private Savings Behavior in Developing Countries,”Staff PapersInternational Monetary Fund (Washington) Vol. 36 (March1989) pp. 16681.

    SigitHananto“Estimates of Comparable Savings in Selected DMCs,”Statistical Report SeriesNo. 4 (Manila: Economics Office, Asian Development Bank1985).

    SvenssonLars E.O. and AssafRazin“The Terms of Trade and the Current Account: The Harberger-Laursen-Metzler Effect,”Journal of Political Economy (Chicago) Vol. 91 (February1983) pp. 97125.

    TobinJames“Notes on Optimal Monetary Growth,”Journal of Political Economy (Chicago) Vol.76 (July/August1968) pp. 83359.

    World BankThe World Development Report (New York: Oxford University Press for the World Bank1984 1985).

Mr. Lahiri, an economist in the European Department of the Fund, is a graduate of Calcutta University and the Delhi School of Economics, University of Delhi. This paper was written while he was visiting the Asian Department, on leave From the Delhi School of Economics. He thanks Kajal Lahiri and colleagues in the Fund for helpful comments.

In more recent times, Graham (1987, p. 1509) found that these two factors explain almost two thirds of the variations observed in the savings rates of major industrial economies during the 1970s.

Ravallion and Sen (1986) have shown the questionable nature of the assumption of uniform response of savings across countries.

Serious doubts have been raised about the validity of Ricardian equivalence. For a demonstration of its invalidity in developing countries, see Haque and Montiel (1987).

Most existing studies use gross national product (GNP) as the explanatory variable in the analysis of savings behavior. Here, however, private disposable income is estimated and used in the empirical investigations, since it is the appropriate concept of income in the context of private savings.

Branson and Klevorick (1969). however, found evidence of a negative effect of inflation on savings in the United States. Similarly. Howard (1978) found that, although inflation led to increased savings in Canada, the United Kingdom, and the United States, expected inflation discouraged savings in Japan.

Inflation leads to substitution of nominal assets by real assets, including consumer durables. Because what is measured in developing countries is consumption expenditure and not consumption, measured consumption will rise and measured savings will be adversely affected.

For Indonesia, the period covered is 1967-85—that is, the period after hyperinflation.

Social security institutions, such as the Employee Provident Fund in India. Malaysia, and Sri Lanka or the Central Provident Fund in Singapore, have played a crucial role in mobilizing contractual forced savings in Asian countries. There are reasons to believe that such compulsory savings were not neutralized by voluntary dissavings; see, for example, Datta and Shome (1981). The importance of these schemes could not be examined here because of the lack of readily available data.

Note the unfortunate change in the use of the symbols c and y. Although they have been used in the earlier subsection to denote consumption of a representative agent and his or her youth, in what follows they are used to denote the natural logarithms of per capita private consumption and income. This is mostly in deference to convention.

Use of the implicit GDP deflator for deriving real private consumption, on the one hand, and use of the consumer price index to obtain the rate of inflation, on the other, needs a word of explanation. It is preferable to deflate nominal income and consumption, two variables that appear on two sides of the consumption function, by the same price index. The implicit GDP deflator being the natural choice for the income variable, this same index has been used rather than the consumer price index for deflating consumer expenditure. In the case of the inflation variable, the preference runs in the opposite direction because of the more important role that durable goods play in the determination of the consumer price index relative to that in the GDP deflator.

Note that the neglect of governmental property income, interest subsidies, retained earnings of the corporate sector, and depreciation may have introduced a bias in the income series.

Note that, because of the presence of a dynamic lag structure in the model, consumption reacts to current as well as past incomes. Thus, strictly speaking, there should be a dummy variable to take account of the imprecision in the past income figures even in years when the contemporaneous income is measured correctly but the relevant past incomes are not. This procedure, however, would lead to the introduction of more than one dummy variable, to nonlinear parametric restrictions, and to considerable complications.

The hypothesis that β2=β2*, however, is not rejected for Sri Lanka at the 1 percent level of significance.

As reported later, for Sri Lanka the hypothesis β2=β2* is not rejected at the 5 percent level of significance once the redundant variables are dropped from the equation.

The present model departs from Davidson and others (1978) because of the presence of variables other than only income and consumption. Some bold uniformity assumptions about the dynamic response of consumption to changes in inflation, the terms of trade, and export orientation have been made because of the limited number of observations.

Appropriate F-tests were carried out before dropping one or more variables from the equation for any country.

See Pagan (1984) for the complication arising from generated regressors. Exports as a proportion of GNP (x) are present as an explanatory variable in equation (6) but not in equation (9). This renders the calculation of the correct standard errors particularly complex.

For the Philippines, the short-run increase in the APS appears to be too large relative to other countries. This, along with the rejection of and the insignificance of demographic variables, shows the poor performance of the model for the Philippines.

The insignificance of the age-dependency variable in the Philippines’ consumption function could be due in part to the high spurious correlation (-0.92) between the age-composition variable and the dummy for data discontinuity.

Much as for the Philippines, the performance of the model is relatively poor for Thailand because of the perversity of the coefficient for the demographic variable and rejection of H3

Ram (1982) and Rossi (1989) found that dependency rates are not important for savings rates in developing countries. Our results are sharply different. Their studies differ from the present one in three respects: the country coverage, their use of a static framework without any dynamic lag structure, and the functional form. Ram and Rossi’s results are obtained from a very large cross section of time-series data. Although the results reported in the present study relate to individual time-series analysis, an unambiguously positive impact of low dependency rates on private savings was found, even in a pooled analysis of the eight countries.

Admittedly, this study has not investigated whether the composition of dependents in terms of children and elderly people has any impact on savings over and above that of the total proportion of dependents in the population. This question merits further attention.

The high APS in India in the long run may seem surprising. It must be recalled, however, that India has had a relatively low growth rate and little change in the age composition of the population, vet its savings rate has been high.

The difference between GNS and gross domestic savings (GDS) is net factor income from abroad and net current transfers. Because the main interest was to understand the amount of resources available for investment, the analysis concentrated on GNS rather than GDS.

Other Resources Citing This Publication