-
Revised
SAVING IN THE WORLD: STYLIZED FACTS
Norman LoayzaHumberto Lpez
Klaus Schmidt-HebbelLuis Servn
November 1998
Abstract
The worlds average saving rate has declined for the last two
decades but country savingrates exhibit a large dispersion,
especially in developing regions. While in a smallnumber of
developing countries saving has risen substantially (together with
growth),saving has stagnated or even declined in most other
developing countries. Countries thatsave more exhibit, on average,
more stable saving rates over time. Country saving ratestend to
rise with per capita real income levels and growth rates, domestic
investment,public saving, money stocks, and private financial
wealth. However substantialheterogeneity in correlations between
saving and related variables is observed betweenindustrial and
developing countries. These and other findings on the performance
ofnational, private, and household saving in industrial and
developing economies aredocumented in this paper.
JEL class.: E21, O16.
Lpez, Loayza, and Servn: DECRG, The World Bank; Schmidt-Hebbel:
Central Bank of Chile.We are grateful to Mita Chakraborty, George
Monokroussos, and Douglas Smith for excellentresearch
assistance.
-
1
1. Introduction
The worlds average saving rate has been declining since the
first oil shock andthrough the early 1990s. However this trend
conceals a large and increasing dispersion ofsaving rates,
particularly among developing countries. The large heterogeneity in
savingbehavior is associated to country and time differences in
levels of development, growthperformance, and fiscal and financial
policies.
This paper provides a detailed statistical account of
international trends in savingand its association with
saving-related variables. The data used here come from theSaving
Project World Database described in a companion paper (Loayza,
Lpez,Schmidt-Hebbel, and Servn 1998). This dataset represents the
largest and mostsystematic collection to date of annual time-series
on country saving and saving-relatedvariables. It is an unbalanced
data pool spanning a maximum of 35 years (from 1960 to1994) and 112
developing and 22 industrialized countries. Thus the facts
documentedbelow are based on very large data samples; for instance,
the number of available annualcountry observations on the gross
national saving rate is 3464.
Several previous saving studies have described world-wide saving
trends.Examples include Masson, Bayoumi, and Samiei (1995) and
Schmidt-Hebbel and Servn(1997, 1999), which have typically relied
on smaller datasets and a smaller number ofvariants with regard to
saving measures and saving-related variables, as well as time
andregional aggregation. In this paper we exploit the Saving
Project World Database bydescribing trends and correlations based
on: (i) different saving aggregates (national,public, private,
household, and corporate), (ii) different public saving measures
(centralgovernment and public sector, with and without adjustment
for capital gains due toinflation), (iii) different world regions
and country groups, (iv) different summarystatistics (unweighted
averages, medians, GNP-weighted averages, and measures
ofdispersion), and (v) a large set of saving-related variables and
potential savingdeterminants.
This paper is meant to be descriptive and focuses on time trends
and pairwisecorrelations based. In a separate paper, the same
database is used to estimate savingmodels in a multivariate
framework (Loayza, Schmidt-Hebbel, and Servn 1998). Thispaper is
organized as follows. Section 2 of this paper briefly discusses the
issues of datacoverage and definitions -- we refer to our companion
paper for a detailed description ofthe data. In Section 3, we focus
on the time trends observed in different saving measuresduring the
last three decades in the world at large as well as in relevant
country groupsand regions. Section 4 examines the association
between various saving measures andsaving-related variables, both
across countries and over time, using pairwise correlations.Section
5 concludes briefly. A separate Annex (available on request)
presents figures ontime trends and bi-variate correlations that
complement the information in the tablesincluded in the main
text.
-
2
2. Data definitions and coverage
2.1 Income and saving measures
Our preferred measure of aggregate income, consistent with gross
national saving,is gross national disposable income inclusive of
all current and capital transfers fromabroad (GNDI). Gross national
saving (GNS) is defined accordingly, as the differencebetween GNDI
and aggregate (private and public) consumption expenditure.
Bothvariables are in current-price units of national currency.
These comprehensive measuresdominate both the related variables
that exclude transfers and the corresponding domesticvariables that
reflect geographical boundaries (Gross domestic saving and GDP).
Therationale for our choice is that the national variables
inclusive of all transfers are morerelevant to the agents or
sectors for which we identify specific saving hypotheses
anddeterminants. In fact the national sector simply represents the
aggregation of the privateand public sectors. Thus GNS/GNDI is the
relevant aggregate or national saving rate (incurrent prices) used
in this paper.1
At the government level, four saving measures are used,
corresponding to thecombinations of two alternative government
coverage definitions (consolidated centralgovernment and public
sector) and two alternative statistical measures (unadjusted
andadjusted for capital gains from inflation). Gross private saving
(GPS) is obtainedresidually by subtracting any of the four previous
government saving measures fromnational saving. Hence there are
four residual private-sector saving (and income)measures that can
be derived. We denote them respectively: CU (central
government,unadjusted), CA (central government, adjusted), PU
(public sector, unadjusted), and PA(public sector, adjusted). The
coverage of these series varies. The preferred measures
forgovernment and private saving are the PA measures, which better
capture the savingeffort of the relevant economic unit. GPS/GPDI
represents the private saving rate, whereGPDI is the gross private
disposable income measure consistent with the selecteddefinition of
GPS. GPDI itself is computed as GNDI minus the relevant definition
ofgovernment saving plus public consumption expenditure.
A disaggregation of net private sector saving between corporate
and householdsaving is used for a smaller available sample.
All saving variables (and the corresponding income measures used
as thedenominator to construct saving ratios) are in current
prices. Consumption andinvestment data (and the corresponding
ratios) are in constant prices.
1 For a discussion of measurement issues and problems that
affect aggregate saving data see Schmidt-Hebbel and Servn (1999),
chap. 2.
-
3
2.2 Country and time coverage
Country aggregation is defined in terms of country groups and
geographicalregions. The five country groups are: industrial OECD
countries (IND), China (alone),other takeoff developing countries,
other developing countries, and the world. Takeoffcountries are
defined a priori as those developing countries that have achieved
high andsustained saving and growth rates during the last two
decades.2
The country regions are defined as: industrial OECD (IND),
China, other EastAsia and Pacific (EAP), Middle East and North
Africa (MENA), Sub-Saharan Africa(SSA), South Asia (SA), and Latin
America and Caribbean (LAC).
Time coverage is dictated by data availability within the period
1960-94. Thestarting year of the data varies across countries, but
is dated mostly between 1960 and1975.
We use three different samples: the fixed and the full sample
for national andprivate saving data, and a smaller household sample
for data on household saving.3 Thefixed sample is comprised of a
total of 85 countries, distributed by country regions andcountry
groups. Each country in a specific region has a full sample of
annualobservations for every year from the starting year for each
region (for instance, 1960 foreach IND country or 1967 for each LAC
country) through 1994. In other words, eachregion in the fixed
sample is a balanced panel even though the fixed sample is not.
The85 countries distributed by the seven country regions are
summarized in Box 1.
2 This group includes China (considered separately) and 9 market
economies: Hong Kong, Indonesia,Korea, Malaysia, Singapore, Taiwan
(China), and Thailand in East Asia; Mauritius in Africa; and Chile
inLatin America. Botswana -- another high-saving and high-growth
economy -- has been excluded fromour sample because of lack of
acceptable data.
3 Due to data problems or lack of data, the following ten
countries are omitted from all three samples:Bhutan, Brazil,
Israel, Nicaragua, the unified post-1991 Germany, United Arab
Emirates, Zambia, Somalia,Gabon, and Yemen (former Arab Republic
of).
-
4
Box 1Distribution of Countries in the Fixed Sample by
Regions
The seven regional groups are the following:
IND: 1960-1994.Australia, Austria, Belgium, Canada, Denmark,
Finland, France, Greece, Iceland,
Ireland, Italy, Japan, Luxembourg, Netherlands, New Zealand,
Norway, Portugal, Spain,Sweden, Switzerland, United Kingdom, and
United States (22 countries).
China: 1966-1994.
EAP (without China): 1966-94.Fiji, Hong Kong, Indonesia, Korea,
Malaysia, Philippines, Singapore, Taiwan
(China), and Thailand (9 countries).
MENA: 1967-1994.Algeria, Cyprus, Egypt, Iran, Jordan, Morocco,
Saudi Arabia, Tunisia, and Turkey
(9 countries).
SA: 1968-1994.India, Pakistan, and Sri Lanka (3 countries).
SSA: 1970-1994.Cote dIvoire, Benin, Burkina Faso, Central
African Republic, Gambia, Ghana,
Kenya, Madagascar, Malawi, Mali, Mauritania, Mauritius, Niger,
Nigeria, Rwanda,Senegal, Sierra Leone, South Africa, Togo, and
Uganda (20 countries).
LAC: 1967-1994.Argentina, Bolivia, Chile, Colombia, Costa Rica,
Ecuador, El Salvador,
Guatemala, Haiti, Honduras, Jamaica, Mexico, Panama, Paraguay,
Peru, Trinidad andTobago, Uruguay, and Venezuela (18
countries).
The three developing-country groups are the following:
All developing countries: 1970-1994 (63 countries, including
China).
Take-off developing countries: 1970-1994 (9 countries, without
China).
Non take-off LDCs (53 countries, without China): 1970-94.
-
5
Box 2Sample Sizes
GNS/GNDI GPS/GPDI (PA) NHHS/GPDIFixed SampleCountries 82 60Years
25-35 25-35Country-years 2665 1427
Full SampleCountries 126 82Years 2-35 4-35Country-years 3160
1700
Household SampleCountries 31Years 5-32Country-years 644
The full sample comprises the fixed sample, all annual country
observations thatare available before the start of the fixed sample
periods, and all annual observations forany available years for
countries excluded from the fixed sample. Hence the full sampleis
unbalanced also at the regional level. Finally the household sample
comprises a muchsmaller unbalanced data panel for net household
saving. The number of countries, yearsand country-year observations
for each sample is summarized in Box 2.4
Tables 1.1 through 4.3 report summary statistics averages,
medians, andmeasures of dispersion of saving rates and related
variables for country regions andcountry groups. The first digit of
table numbers (1, 2, 3, and 4) denotes respectivelyinformation for
gross national saving, gross private saving, net household
(and/orcorporate) saving, and aggregate demand components. The
summary informationreported in the tables is based on the above
mentioned samples for the latter variables.
4 The unbalanced panel of 31 countries and 644 country-year
observations is for net household saving data(as used in tables 3.2
and 8, and in figures 3.2 and 8.1-8.10). For both net household and
net corporatesaving data (i.e., for disaggregated net private
saving) a smaller sample is available, comprising 31countries and
580 country-year observations (as used in Table 3.1 and Figure
3.1). An even smaller sampleof 430 country-year observations (not
used in this paper) is available in the World Bank Saving
Databasefor both gross household and corporate saving data (i.e.,
for a disaggregated gross private saving). Finallynote that total
net private saving reported in Table 3.1 as Sum (of net household
and corporate savingfigures) is different from any of the four
measures of gross private saving reported elsewhere.
-
6
The tables are generally reported for both the comprehensive
cross-country time-series (orpooled-data) sample and the
cross-country sample of country averages. In addition,separate
summary statistics are reported in most cases for three different
sub-periods thatroughly correspond to decades before and after the
first oil shock: 1965-73, 1974-84, and1985-94. 5
Tables 5-8 report full-sample correlations between each of four
saving variables(national saving, private saving CU, private saving
PA, and net household saving,respectively) and a number of
saving-related variables, among them potential savingdeterminants.
Correlations are reported for three country regions in both the
panel andlong-run cross-section samples. Table 9 reports
full-sample correlations among saving-related variables for both
the world panel and world long-run cross-section samples.
3. Saving trends and facts
Next we discuss the main regional and world saving trends and
cross-regionalsaving differences, complemented by some information
on consumption and investmenttrends.
On the basis of the information described earlier, we can
identify several stylizedfacts and trends concerning saving across
the world:
(1) The worlds average national saving rate has declined since
the early 1970s.(Tables 1.1 - 1.2; figures 1.1 - 1.4c)
The worlds average gross national saving rate (the ratio of
gross national saving to grossnational disposable income, GNS/GNDI)
has been declining for the last three decades.Looking across
decades, the median (fixed-sample) world saving rate fell from
20.7% in1965-73 to 20.3% in 1974-84, and further to 18.7% in
1985-94. Indeed, the world savingrate attained a peak of 23% in
1973-74, at the time of the first oil shock. From then on
itdeclined steadily to reach a trough of 18% in 1992-93. This trend
is consistently observedacross different statistics (medians,
unweighted averages, GNP-weighted averages),samples (fixed and
full), and aggregate saving measures (national and domestic).6
(2) Saving rates show a large international dispersion.
5 Note that the data reported for the first sub-period (1965-73)
comprise only observations that areconsistent with the
corresponding country or region group (for example, SSA
observations start only in1970). Also note that the summary
statistics for the pooled-data sample, for example the medians,
arecalculated from the full pool of data for the corresponding
period. For instance SSAs median saving ratefor 1965-73 is the
median of all corresponding observations for all countries during
this period -- it is notthe median of each countrys median or
mean.
6 Recent and incomplete data for the years after 1994 (not
reported in this paper) suggest that a possiblebreak is observed
after 1992-93, as many countries have adopted more orthodox fiscal
policies reflected inhigher public and hence higher national saving
rates. This is the case of European countries workingtoward meeting
Maastricht criteria for monetary union and Latin American countries
strengthening theirmacroeconomic policies. Time will tell if this
is starting a sustained trend break from the past trend declinein
saving rates.
-
7
(Tables 1.1 3.2)
In the fixed pooled-data sample, the standard deviation of
GNS/GNDI is 9.2%,with the minimum at 30.6% and the maximum at
77.9%. This reflects a huge variationacross space and time, with
the bulk of the variability stemming from the
cross-sectiondimension. If we control for time variation by
focusing on the cross-country sample, thestandard deviation of the
saving rate declines to 7.6%, with the minimum long-runcountry
saving rate at 3.7% and the maximum at 38%.7 This large dispersion
reflectsmostly a large variance of saving rates among developing
countries. By contrast, savingrates are much more homogeneous
across industrial countries (as shown by theconsiderably lower
standard deviation for this sample). Similar differences are
observedin the case of the gross private saving rate (the ratio of
gross private saving to grossprivate disposable income, GPS/GPDI)
and the net household saving rate (the ratio of nethousehold saving
to gross private disposable income, NHS/GPDI).
(3) Saving rates show divergent patterns across developing
regions.(Tables 1.1 - 1.2; figures 1.0 - 1.4c)
The median (fixed-sample) GNS rate in developing countries rose
from 16.7% in1965-73 to 19.0% in 1974-84, declining subsequently to
17.7% in 1985-94. However,this aggregate behavior disguises widely
different and even diverging saving patternswithin the developing
world. Saving rates have risen rapidly in China and the nine
othertakeoff countries since the 1970s. In fact, the median
national saving ratio in the ninetakeoff countries increased
gradually and systematically from 20% in 1970-1972 to 34%in
1992-94. Chinas already high saving rate (at 29% during 1970-77)
was boosted by theeconomic reforms started in 1978, rising to 41%
in 1993-94.
By contrast, saving rates in other developing countries and
regions stagnated oreven declined during the last three decades.
The divergent saving trends between takeoffand non-takeoff
countries was also reflected across regional aggregates. As
takeoffeconomies are highly represented in the East Asia and
Pacific region, the latters mediansaving rate (fixed sample) shows
a spectacular rise from 20% in 1966-68 to 33% in 1992-94. South
Asias median saving rate also shows a substantial rise, from 15% in
1968-70to 22% in 1992-94. In Latin America and the Caribbean the
median national saving ratedoes not show a clear trend (it hovers
around an average 16% during 1967-94) but showsswings, increasing
after the first oil shock (1973-80) and declining with the debt
crisis(1982-84). Nor is a clear trend observed in the Middle East
and North Africa althoughswings correlated with the price of oil
are observed in this region around its 1967-94median saving rate of
22%. In Sub-Saharan Africa the median saving rate
declinesprecipitously from 14% in 1970-76 to a trough of 10% in
1980-86 but recovers to the14% level in 1987-94. Notwithstanding
this recovery, SSAs saving rate is the lowestacross all
regions.
7 The dispersion measures in the full sample (Table 1.2) are
similar to those in the fixed sample (Table 1.1).
-
8
(4) Industrial-country saving rates have declined strongly since
the early 1970s.(Tables 1.1 - 1.2; figures 1.0 - 1.4c)
The median (fixed-sample) GNS rate in industrial countries
increased graduallyfrom 25% in the early 1960s to a historical peak
of 27.5% in 1972-73, just before the firstoil shock. Since then, it
has declined systematically, falling to a (median) level of 19%
in1993-94.
(5) Average saving rates are negatively correlated with standard
deviations ofsaving rates over time
(Tables 1.1, 2.3, and 3.2; figures 1.5, 2.10, 2.11, and 3.2)
Countries with lower saving rates also show a larger dispersion
over time in theirsaving rates. For instance, while the unweighted
average (fixed-sample) GNS rate inindustrial countries attained
23.4% in 1965-94, the standard deviation over the sameperiod of
industrial-country saving rates was on average 5.5%. However in
developingcountries the unweighted average (fixed-sample) GNS rate
was 18.1% in 1965-94, withan average standard deviation of 10.2%.
Therefore the coefficient of variation of saving(the ratio of its
standard deviation and its average) attained 0.56, more than twice
thelevel of 0.24 observed in industrial economies. A similar
negative correlation betweenfirst and second moments is observed in
the case of gross private saving rates, but not inthe case of net
household saving rates. This reflects that low-saving countries
show adismal saving behavior not only in their low saving effort
but also in their high savingvolatility.
(6) Developing-country public saving has been on the rise since
the early 1980s.(Table 2.1; figures 2.1-2.4)
Public saving is low in both industrial and developing countries
typicallybetween one-tenth and one-fourth of aggregate national
saving. The median (fixed-sample) public saving rate in developing
countries was halved between the late 1970s andearly 1980s. Since
then fiscal adjustment has been successful in raising public
savingrates back to (or surpassing) their previous peaks. This
trend is common to the fourdifferent public saving measures
although their levels differ considerably from each
other.Unadjusted central government saving stood at only 2% of GNDI
in the early 1990s whilethe adjusted public sector saving measure
attained 7% in the early 1990s. The data showsthat approximately
half of the difference between both measures is due to higher
savingby the non-central government levels of the public sector (as
reflected by the differencebetween CU and PU, or between CA and PA
saving levels) while the other half reflectscapital gains from
inflation, that lead to higher adjusted central government or
publicsaving levels (the difference between CU and CA, or between
PU and PA saving levels).
(7) Industrial-country public saving has declined significantly
since 1975.(Table 2.1; figures 2.1-2.4)
-
9
The median (fixed-sample) public saving rate in industrial OECD
countries fellsteadily from the mid-1970s to the mid-1980s,
recovered partly in the late 1980s, andthen fell precipitously to
reach negative levels in 1992-93. As in the case of
developingcountries, this trend is common to all four public saving
measures.
(8) Capital gains from inflation accruing to the public sector
are of similarmagnitude in developing and industrial economies.
(Table 2.2)
Computing the difference between adjusted and unadjusted public
(private) savingyields the capital gain (loss) accruing to the
public (private) sector due to inflation.8 Inthe world at large the
inflation tax thus computed has ranged from 1.1% to 1.5% of
GNDIduring the last three decades, slightly larger in industrial
than in developing countries.However while the inflation tax shows
a trend decline in industrial countries, it hasincreased in
developing countries during 1985-94.
(9) Private saving has declined in developing countries.(Tables
2.0, 2.1 and 2.3; figures 2.3 - 2.9)
Private saving is obtained residually from national saving and
any one measure ofpublic saving. Whatever measure is used, the
median private saving rate in developingcountries declined between
1974-84 and 1985-94. 9
As in the case of the national saving rate, the trend decline in
private savingconceals large differences within developing regions.
During the last three decades theregional median private saving
rates increased substantially in East Asia and Pacific, theMiddle
East and North Africa in South Asia, declined in Latin America and
theCaribbean, and fell precipitously in Sub-Saharan Africa.
(10) Private saving is roughly constant in industrial
countries.(Tables 2.1 and 2.3; figures 2.3 - 2.9)
The median private saving rate in industrial countries remained
virtually constantduring the last two decades hovering around 23%
when defined as consistent with 8 The calculation of the public
(private) sectors capital gains (loses) from inflation is based on
theoutstanding stocks of central government or public-sector debt
and the monetary base. Public-sector capitalgains from inflation
exceed those to the central government if state-owned enterprises
or sub-national levelsof government are net debtors to the private
sector. Notice that the net capital gains from inflation reportedin
Table 2.2 differ from the corresponding differences between
adjusted and unadjusted saving ratesreported in Table 2.1 because
of differences in samples.
9 Notice that the differences betwen regional median private
saving rates (consistent with the PA definition)reported in Tables
2.1 and 2.3 is due to the use of GNDI in Table 2.1 and GPDI in
Table 2.3 asdenominators.
-
10
unadjusted central-government saving or just over 18% when
defined as consistent withadjusted public-sector saving.
(11) For our much smaller sample of net household and corporate
saving rates, bothrates are larger in developing countries than in
industrial economies.
(Table 3.1; figure 3.1)
Both household and corporate saving rates are generally higher
in developing thanin industrial countries in our smaller sample for
disaggregated net private saving rates.For 1985-94 the median net
household saving rate (as a ratio of gross private
disposableincome) is 8.4% in developing and 6.7% in industrial
countries. For the same decade theregional median net corporate
saving rate (as a ratio of gross private disposable income)is 5.6%
in developing and 3.9% in industrial economies.
(12) World private consumption ratios to GDP are stable over
time but there arediffering trends across regions.
(Table 4.1; figures 4.1 - 4.1b)
Median private consumption to GDP ratios (at constant prices) do
not show muchvariation since 1970 -- for the world at large (at
64%), industrial countries (at 60%), andother developing economies
(71%). However private consumption ratios showsubstantial declines
in China and other takeoff countries, and in the regions of East
Asiaand Pacific, South Asia, and Latin America and the Caribbean.
In Sub-Saharan Africathe median private consumption ratio to GDP
rises during the last three decades. Thesepatterns are generally --
but not always -- consistent with the regional trends in
privatesaving rates. The differences in time trends between
current-price private saving toincome ratios and constant-price
private consumption to GDP ratios (for instance in LatinAmerica and
the Caribbean) reflect trend changes in the relative price of
consumption(The ratio of the private consumption deflator relative
to the GDP deflator) as well astrend changes in the difference
between GDP and gross private disposable income.
(13) Private consumption levels generally follow random walk
processes
In order to assess persistence of the time-series process of
constant-price privateconsumption (in levels, not as a ratio to
GDP), for each country in the sample wecompute the persistence
parameter corresponding to the spectral density at the
zerofrequency. The average of the resulting point estimates of the
persistence parameter andits standard error, as well as the group
average of the first-order autocorrelationcoefficients (of the
first difference of consumption), are reported below for the
worldsample as well as various regional groups.
Persistenceparameter
Standard error First autocorrelationcoefficient
World 1.05 0.098 0.04Industrial 1.55 0.09 0.34
-
11
China 0.025 0.003 -0.22EAP 1.21 0.14 0.13MENA 0.85 0.12 -0.1LAC
0.76 0.07 -0.07SA 1.28 0.17 0.06SSA 0.92 0.07 -0.04
The estimates reveal that the spectral density at the zero
frequency for real privateconsumption is generally consistent with
a random walk process (i.e., a persistenceparameter of 1), which in
itself is consistent with the simple version of the
permanent-income hypothesis for consumption. The estimate for the
world sample is 1.05 (with astandard error of 0.04) and not
significantly different from 1. In four regions it is also
notsignificantly different from 1, also implying that the
random-walk hypothesis cannot berejected. However the estimate is
1.55 (significantly larger than 1) for industrialcountries. In
contrast, the estimate is 0.76 (significantly smaller than 1) for
LatinAmerica and Caribbean, suggesting that consumption is
stationary in the latter region.Finally in China the persistence
parameter estimate is extremely low (0.025), implyingconsumption
stationarity even more strongly than in LAC.
(14) World government consumption ratios are higher in
industrial countries thanin developing countries.
(Table 4.2; figures 4.2 - 4.2b)
Median real government consumption ratios to GDP (at constant
prices) have notchanged much at world or regional levels of
aggregation. The median 1965-94 ratio is12.4% in developing
countries and significantly higher -- at 17.1% -- in
industrialcountries. This is consistent with Wagners Law, the
empirical observation thatgovernment size tends to increase with
per capita income levels.
(15) World investment rates have declined since the early
1970s.(Table 4.3; figures 4.3 - 4.3b)
The median world gross domestic investment ratio to GDP (at
constant prices)shows a steep decline after the seventies. This
trend reduction is observed in bothindustrial and other developing
countries although the latter exhibited a temporaryinvestment boom
during 1974-1982. By contrast median investment rates have
increasedin takeoff countries (China) from 24% (30%) during 1970-74
to 36% (37%) in 1990-94.
4. Correlation between saving and related variables
Now we turn to the correlation between saving and key
saving-related variables,using both the long-run cross-country and
cross-country time-series dimensions of thedata. Tables 5-8 report
the correlation between a set of saving related variables and
thegross national saving rate (Table 5), the gross private saving
rate (in two versions, Tables6 and 7), and the net household saving
rate (Table 8). The cross-correlations among
-
12
saving-related variables are reported in Table 9. Figures 5.1
through 8.10 in the Annexdepict cross-country, pair-wise
correlations that supplement the correlation coefficientsreported
in the tables.
Before turning to the specific results we want to point out two
general findings.First, in most cases the bi-variate correlations
between saving rates and saving-relatedvariables are similar for
the gross national saving rate and the gross private saving
rate.However they tend to be quite different for the much smaller
sample of net householdsaving rates. We attach more weight to the
correlations observed between gross nationaland private saving
rates (on the one hand) and saving-related variables (on the other
hand)than to the correlations observed between the net household
saving rate and relatedvariables, as the latter are based on a much
smaller country sample.
Second, the cross-country correlations are typically larger in
magnitude than thepooled-data correlations -- an observation common
to national, private, and householdsaving rates. This suggests that
the annual time-series data are noisier than the cross-country
observations.
The data reveal the following specific findings.
(i) Saving rates and income levels are positively
correlated.
Most saving measures show a significant and positive
cross-country correlationwith real per capita income levels. For
the world sample, the correlation coefficientsbetween gross
national saving rate (in Table 5) and the gross private saving rate
(in tables6 and 7), on the one hand, and real per capita income
levels, on the other, are ca. 0.50 forthe cross-country sample and
ca. 0.40 for the pooled-data sample and are significant at the5%
level. The correlations are higher in developing countries than in
industrialeconomies. For the household saving sample, however, the
correlation is close to zero(Table 8).
Fig. 5.1 confirms the correlation for the GNS rate at the
cross-country level. Thefigure also shows that takeoff developing
countries are located at low to middle-incomelevels with abnormally
high saving rates. In contrast, industrial countries seem
totypically save less than what should be expected given their high
income levels. Thissuggests that the saving-income relation tapers
off at high income levels. Non-takeoffdeveloping countries are
clustered at low income and low saving levels. Separateregression
lines for developing and industrial economies confirm the
cross-regionalheterogeneity in the saving-income relation, showing
a larger positive correlation fordeveloping countries.
(ii) Saving rates and growth rates are positively
correlated.
All saving measures show positive correlations with real per
capita incomegrowth rates. For the world sample, the correlation
coefficients between gross national
-
13
and gross private saving rates and real per capita growth rates
are ca. 0.40 for the cross-country sample and ca. 0.29 for the
pooled-data sample and are significant at the 5%level. The
correlations between household saving and growth are lower, at ca.
0.37 and0.17, respectively. As opposed to the preceding correlation
between saving rates andincome levels, the saving-growth
correlation is higher for industrial than for
developingeconomies.
Fig. 5.2 confirms the world correlation at the cross-country
level. The figure alsoshows that industrial economies are closely
clustered at moderately high saving andgrowth rates. Takeoff
countries are clustered in the upper right-hand corner of the
figures,at very high saving and growth rates. Non-takeoff LDCs show
a larger dispersion andthey are centered at low saving and growth
rates.
(iii) National saving and domestic investment rates are
positively correlated.
The correlation between the GNS rate and the gross domestic
investment (GDI)rate is large and significant -- ca. 0.70 for both
the world at large and developingcountries, and close to 0.60 for
industrial economies. For private saving the correlation islower.
For household saving in industrial countries the correlation turns
to be negative.
The high cross-country correlation between long-run national
saving and grossdomestic investment rates is confirmed in Fig. 5.3.
OECD countries are more frequentlylocated above the 45-degree line
(along which the current account is balanced), reflectingthat they
tend to be capital exporters, while most developing economies, that
tend to becapital importers, are located below the line.
(iv) National saving and public saving rates are positively
correlated.
The correlation between the national saving rate and the public
saving rate ispositive and high. For the world at large, the
correlation coefficient is ca. 0.35 for thecentral government
unadjusted (CU) saving rate and ca. 0.40 for the
public-sectoradjusted (PA) saving rate; both are significant at the
5% level . At the sub-sample levelsof developing and industrial
countries, the correlations are even larger, in the range of0.47 to
0.60. These figures suggest a partial offsetting of private-public
saving levels, anissue addressed next.
(v) The correlation between private and public saving rates are
not robust acrossdifferent saving measures and regional
samples.
Private and public saving rates, consistent with unadjusted
measures for thecentral government (CU), are mostly uncorrelated
(with exception of a 0.13 correlationfor both the cross-country
developing and industrial country sub-samples). The cross-country
correlations for private and public saving, consistent with
adjusted measures for
-
14
the public sector (PA), are positive and moderate (but not
significant at the 5% level),ranging from 0.12 for the world sample
to 0.22 for developing countries. At thehousehold level, the
cross-country correlation is 0.4 for the world (significant at
5%level), -0.58 for industrial economies (significant at 5% level),
and 0.126 for developingcountries (not significant at 5% level).
Therefore little can be concluded about private-public saving
offset coefficients from these simple correlations.
(vi) Saving and the terms of trade are uncorrelated.
The correlation between saving rates and the terms of trade (the
relative price of acountrys exports and imports) is generally not
significantly different from zero andfrequently changes sign across
different saving measures and data samples.
(vii) The correlation between saving and the real interest rate
is not robust acrossdifferent saving measures and samples.
The world correlation between national or private saving rates
and the real interestrate ranges from 0.31 to 0.34 (significant at
a 5% level) at the cross-country level.However it is much smaller
for the pooled-data samples (and mostly negative for theindustrial
countries). At the household level the cross-country correlation
turns negativeand is ca. 0.35.
(viii) The correlation between saving and the inflation rate is
not robust acrossdifferent saving measures and samples.
The correlation between saving rates and inflation rates
displays a pattern that issimilar but of opposite sign to the one
observed between saving and the real interest rate.This would not
be surprising when considering that the real interest rate is
constructed asthe nominal rate minus inflation. However the
correlation between the two lattervariables is not very high,
varying between -0.42 for the cross-country sample and -0.18for the
pooled-data sample (Table 9).
The correlation between national or private saving rates and the
inflation rate isgenerally negative and often not robust, varying
between 0 and 0.50. It tends to belarger in magnitude for
industrial than for developing countries. However at thehousehold
level the correlation turns positive and varies significantly, not
exceeding 0.35.Therefore across all samples and saving measures the
saving-inflation correlation is notrobust.
(ix) Saving and money stocks are positively correlated.
All saving measures show positive correlations with the M2 money
stock ratio toincome. For the world sample, the cross-country
correlation coefficients for grossnational or gross private saving
rates and M2 are ca. 0.5. The correlation is larger in
-
15
industrial and lower in developing economies, and larger in the
cross-section than in thepanel dimension.
(x) Saving and private financial wealth are also positively
correlated.
The private financial wealth rate is the ratio of the sum of the
monetary base,private net foreign debt, and public net domestic
debt to national disposable income. Thesaving-private wealth
correlations are positive and tend to be significant and larger
forindustrial countries.
(xi) The correlation between saving rates and private credit
flows is not robust acrossdifferent saving measures and
samples.
Correlations between saving rates and private credit flows vary
with savingmeasures and samples. An extreme case of lack of
robustness across samples is observedfor household saving, where
cross-country correlations with private credit flows rangefrom 0.42
in developing countries to +0.63 in industrial countries.
(xii) The correlation between saving rates and social security
spending flows is notrobust across different saving measures and
samples.
The cross-country correlation between national or private saving
rate and theshare of social security spending in total public
expenditure is close to 0.60 in industrialcountries (significant at
the 5% level). However it is not significantly different from
zerofor developing economies. At the household level the
correlation is negative across bothregional groups and is larger
for developing countries.
(xiii) The correlation between saving and the old-age dependency
ratio is not robustacross different saving measures and
samples.
The world correlation between the GNS ratio and the old-age
dependency ratio ispositive and ca. 0.34 (significant at the 5%
level) for GNS and GPS. However a verydifferent picture emerges at
sub-sample levels: correlations change in magnitude and signfor
developing and industrial countries and across national and private
saving measures.For household saving the correlations are
systematically negative.
(xiv) Saving rates and young-age dependency ratios are
negatively correlated.
A clearer picture is obtained for the correlation between saving
measures and theyoung-age dependency ratio. The world cross-country
correlation between saving and theyoung-age dependency ratio is
negative and ca. -0.50 (significant at the 5% level) for bothGNS
and GPS. This negative correlation is generally maintained for
developing andindustrial countries separately. For household
saving, however, the correlation is low andunstable in sign.
-
16
(xv) The correlation between saving rates and the urbanization
ratio is not robustacross different saving measures and
samples.
At the world level the cross-country correlation between saving
and theurbanization rate is ca. 0.50 (significant at the 5% level)
for GNS and GPS. However atthe sub-sample level the correlations
are of opposite sign: for developing countries it isca. 0.4
(significant at the 5% level) and for industrial economies it is
ca. 0.25(significant at the 5% level). The negative correlation is
confirmed for industrialcountries for the household saving
sample.
(xvi) The correlation between saving and income concentration is
negative in the worldat large but not robust across different
regional samples.
For the world at large, the cross-country correlations between
national and privatesaving (on one hand) and the Gini coefficient
for personal income concentration (on theother hand) are always
negative, ranging from 0.11 for household saving to -0.42
forprivate saving (the latter is significant at a 5% level).
However the correlations are highlyunstable at sub-sample levels,
ranging from moderate positive to large negative values.
(xvii) Household and corporate saving rates are negatively
correlated.
A cross-country world correlation of 0.49 (significant at a 10 %
level) isobserved between household and corporate saving rates for
our small country sample.The coefficient rises in magnitude to
about 0.63 at the sub-sample levels for developingand industrial
countries. This suggests high household-corporate saving
offsetting.
(xviii) Real gross domestic investment rates and real per capita
GDP growth rates arepositively correlated.
There is a high correlation between real GDI ratios (to real
GDP) and per capitagrowth rates of real GDP, attaining .23 in the
world at large and significant at the 5%level. In developing
countries, the correlation is .23, while in industrial countries it
is .33,both significant at 5%.
Finally we report cross-correlations among saving-related
variables (Table 9),with cross-country correlations shown in the
lower half and pooled-data correlationsreported in the upper half
of the table. Of the large number of results we want to singleout
only one: the high correlation (in magnitude) observed between the
level ofdevelopment (reflected by per capital real income levels)
and saving-related variables thatalso proxy structural indicators
of development, including financial depth (M2/income),social
security expenditure, demographic dependency ratios, urbanization,
and incomeconcentration. While this result is not unexpected, it is
important to keep it in mind when
-
17
estimating saving functions in a multivariate framework that
accounts for the influence ofboth per capita income and structural
indicators.
5. Conclusions
This paper has documented the main saving trends and
correlations betweensaving and other variables observed in
industrial and developing countries during the lastthree decades.
This has been possible by using the recent Saving Project World
Databaseput together by the World Bank. The paper has provided a
comprehensive description offacts, trends, and correlations based
on different saving aggregates, alternative publicsaving measures,
various world regions and country groups, different
regionalaggregation measures (unweighted averages, medians, and
GNP-weighted averages), anda large set of saving-related variables
and potential saving determinants.
Among the main findings reported here are the following. The
worlds averagesaving rate has declined for the last two decades but
country saving rates exhibit a largedispersion, especially in
developing regions. While in a small number of developingcountries
saving has risen substantially (together with growth), saving has
stagnated oreven declined in most other developing countries.
Countries that save more exhibit, onaverage, more stable saving
rates over time. Country saving rates tend to rise with percapita
real income levels and growth rates, domestic investment, public
saving, moneystocks, and private financial wealth. However
substantial heterogeneity in the correlationbetween saving and
related variables is observed between industrial and
developingcountries.
References
Feldstein, M. and C. Horioka (1980): Domestic Savings and
International CapitalFlows, Economic Journal 90: 314-329.
Loayza, N., H. Lpez, K. Schmidt-Hebbel and L. Servn (1998): The
World SavingData Base, World Bank manuscript, January.
Loayza, N., K. Schmidt-Hebbel, and L. Servn (1998): What drives
Saving across theWorld?, World Bank manuscript, September.
Masson, Bayoumi, and Samiei (1996): Saving Behavior in
Industrial and DevelopingCountries, International Monetary
Fund.
Schmidt-Hebbel, K. and L. Servn (1997): Saving across the World:
Puzzles andPolicies, World Bank Discussion Paper No. 354.
-
18
Schmidt-Hebbel. K. and L. Servn (eds.) (1999): The Economics of
Saving and Growth.Cambridge University Press, forthcoming.
-
List of Tables
1.1 Gross National Saving/Gross National Disposable Income
ratios by CountryRegions/Groups and Periods (Fixed Sample)
1.2 Gross National Saving/Gross National Disposable Income
ratios by CountryRegions/Groups and Periods (Full Sample)
2.1 Private and Public Saving ratios to GNDI (regional medians)
by Public SavingMeasures and Periods (Fixed Sample)
2.2 Net Capital Gains to the Government due to Inflation
(regional medians) by Periods(Fixed Sample)
2.3 Gross Private Saving/Gross Private Disposable Income by
Country Regions/Groupsand Periods (adjusted, public sector
definition) (Fixed Sample)
3.1 Household and Corporate Saving ratios to GPDI (regional
medians) by Periods(Household and Corporate Sample)
3.2 Household Saving ratio to GPDI by Country Regions/Groups and
Periods(Household Sample)
4.1 Real Private Consumption/Real GDP ratios by Country
Regions/Groups and Periods(Fixed Sample)
4.2 Real Government Consumption/Real GDP ratios by Country
Regions/Groups andPeriods (Fixed Sample)
4.3 Real Domestic Investment/Real GDP ratios by Country
Regions/Groups and Periods(Fixed Sample)
5. Cross-country and panel correlations between GNS/GNDI and
saving-related variablesby Country Regions (Full sample)
6. Cross-country and panel correlations between GPS/GPDI
(consistent with non-adjusted CG) and saving-related variables by
Country Regions (Full sample)
7. Cross-country and panel correlations between GPS/GPDI
(consistent with adjustedPS) and saving-related variables by
Country Regions (Full sample)
8. Cross-country and panel correlations between HHS/GPDI and
saving-related variablesby Country Regions (Full sample)
-
1
9. Cross-country and panel correlations among saving-related
variables in the World(Full sample)
-
SAVING IN THE WORLD: Annex of Figures
Norman LoayzaHumberto Lpez
Klaus Schmidt-HebbelLuis Servn
November 1998
This Annex to our paper Saving in the World: Facts, Trends, and
Correlationspresents figures for time trends and bi-variate
correlations that complement theinformation in the tables included
in the main text.
Figures 1.0 through 1.4 depict time charts for Gross National
Saving ratios toGross National Disposable Income (GNS/GPDI) for
country groups (for 1970 through1994, series 1.1 and 1.3) and
country regions (for 1960 or later through 1994, series 1.2and
1.4). Fixed country samples are considered in series 1.1 and 1.2,
and full countrysamples are considered in series 1.3 and 1.4.
Medians are depicted in figures 1.i,unweighted averages in figures
1.ib, and GNP-weighted averages are depicted in figures1.ic, for
i=1,..4. Figure 1.5 depicts the bi-variate cross-country relation
between nationalsaving rates and the standard deviation the various
saving measures and various saving-related variables for the world
sample.
Figures 2.0 through 2.8 depict time charts for Gross Public and
Gross PrivateSaving ratios according to various definitions of
public saving, for country groups andcountry regions. Figure 3.1
shows a time chart for net corporate and household saving
inindustrial and developing countries. Figures 4.1 through 4.3b
depict time charts forprivate consumption, public consumption, and
domestic investment ratios to GDP.
Figures 1.5, 2.10, 2.11, and 3.2 depict bi-variate cross-country
relations betweenvarious saving rates and their standard deviations
for the world sample.
Figures 5.1 through 8.11 depict bi-variate cross-country
relations between varioussaving measures and various saving-related
variables for the world sample. Each figuredepicts two correlations
(for the OECD and LDC samples) between average long-run
country observations for a given saving measure and a
saving-related variable. The linesrepresent the fitted values of an
OLS regression for the corresponding sample.
-
1
Annex of Figures, Table of Contents
Figure: Title:
1.0 GROSS NATIONAL SAVING RATES BY WORLD REGIONS: Gross national
saving rateincluding net current transfers as percent of gross
national disposable income (Regional medians atcurrent prices,
1965-1994)
1.1 GROSS NATIONAL SAVING RATES BY COUNTRY GROUPS: Gross
national saving aspercent of gross national disposable income
(Group medians at current prices, fixed countrysample,
1965-1994)
1.1b GROSS NATIONAL SAVING RATES BY COUNTRY GROUPS: Gross
national saving aspercent of gross national disposable income
(Unweighted group averages at current prices, fixedcountry sample,
1965-1994)
1.1c GROSS NATIONAL SAVING RATES BY COUNTRY GROUPS: Gross
national saving aspercent of gross national disposable income
(GNP-weighted group averages, fixed country sample,1965-1994)
1.2 GROSS NATIONAL SAVING RATES BY WORLD REGIONS: Gross national
saving as percentof gross national disposable income (Regional
medians at current prices, fixed country sample,1965-1994)
1.2b GROSS NATIONAL SAVING RATES BY WORLD REGIONS: Gross
national saving as percentof gross national disposable income
(Unweighted regional averages at current prices, fixed
countrysample, 1965-1994)
1.2c GROSS NATIONAL SAVING RATES BY WORLD REGIONS: Gross
national saving as percentof gross national disposable income
(GNP-weighted regional averages, fixed country
sample,1965-1994)
1.3 GROSS NATIONAL SAVING RATES BY COUNTRY GROUPS: Gross
national saving aspercent of gross national disposable income
(Group medians at current prices, full sample,1965-1994)
1.3b GROSS NATIONAL SAVING RATES BY COUNTRY GROUPS: Gross
national saving aspercent of gross national disposable income
(Unweighted group averages at current prices, fullsample,
1965-1994)
1.3c GROSS NATIONAL SAVING RATES BY COUNTRY GROUPS: Gross
national saving aspercent of gross national disposable income
(GNP-weighted group averages at current prices, fullsample,
1965-1994)
1.4 GROSS NATIONAL SAVING RATES BY WORLD REGIONS: Gross national
saving as percentof gross national disposable income (Regional
medians at current prices, full sample, 1965-1994)
1.4b GROSS NATIONAL SAVING RATES BY WORLD REGIONS: Gross
national saving as percentof gross national disposable income
(Unweighted regional averages at current prices, full
sample,1965-1994)
-
2
1.4c GROSS NATIONAL SAVING RATES BY WORLD REGIONS: Gross
national saving as percentof gross national disposable income
(GNP-weighted regional averages at current prices, fullsample,
1965-1994)
1.5 GROSS NATIONAL SAVING RATES LEVELS AND VARIATION: Gross
national savingrate as percent of gross national disposable income
and the standard deviation of the gross nationalsaving rate
(Country averages at current prices, fixed sample, 1965-1994)
2.0 GROSS PRIVATE SAVING RATES BY WORLD REGIONS: Gross private
saving rate aspercent of gross private disposable income (Adjusted,
public sector definition; Regional mediansat current prices,
1965-1994)
2.1 ALTERNATIVE MEASURES OF PUBLIC SAVING IN DEVELOPING
COUNTRIES (Percentof gross national disposable income, fixed
country sample, 1970-1994)
2.2 ALTERNATIVE MEASURES OF PUBLIC SAVING IN INDUSTRIAL
COUNTRIES (Percentof gross national disposable income, fixed
country sample, 1970-1994)
2.3 ALTERNATIVE MEASURES OF PRIVATE SAVING IN INDUSTRIAL AND
DEVELOPINGCOUNTRIES (Percent of gross private disposable income;
Medians of fixed country sample,1970-1994)
2.4 COMPOSITION OF GROSS NATIONAL SAVING: Public and private
saving as percent of grossnational disposable income (Unadjusted,
central government definition; Medians of fixed countrysample,
1970-1994)
2.5 COMPOSITION OF GROSS NATIONAL SAVING: Public and private
saving as percent of grossnational disposable income (Adjusted,
public sector definition; Medians of fixed country
sample,1970-1994)
2.6 GROSS PRIVATE SAVING RATES BY COUNTRY GROUPS: Gross private
saving as percentof gross private disposable income (Adjusted,
public sector definition; Group medians at currentprices, fixed
country sample, 1970-1994)
2.7 GROSS PRIVATE SAVING RATES BY WORLD REGIONS: Gross private
saving as percent ofgross private disposable income (Adjusted,
public sector definition; Group medians at currentprices, fixed
country sample, 1970-1994)
2.8 GROSS PRIVATE SAVING RATES BY COUNTRY GROUPS: Gross private
saving as percentof gross private disposable income (Adjusted,
public sector definition; Group medians at currentprices, full
country sample, 1970-1994)
2.9 GROSS PRIVATE SAVING RATES BY WORLD REGIONS: Gross private
saving as percent ofgross private disposable income (Adjusted,
public sector definition; Group medians at currentprices, full
country sample, 1970-1994)
2.10 PRIVATE SAVING RATES LEVELS AND VARIATION: Private saving
rate as percent ofgross private disposable income and the standard
deviation of the private saving rate (Unadjusted,central government
definition; Country averages at current prices, fixed sample)
2.11 PRIVATE SAVING RATES LEVELS AND VARIATION: Private saving
rate as percent ofgross private disposable income and the standard
deviation of the private saving rate (Adjusted,public sector
definition; Country averages at current prices, fixed sample)
-
3
3.1 COMPOSITION OF PRIVATE SAVING: Net household saving and net
corporate saving, ratiosto private disposable income (Net household
and corporate saving data sample, 1965-1992)
3.2 HOUSEHOLD SAVING RATES LEVELS AND VARIATION: Household
saving as percent ofgross private disposable income and the
standard deviation of the household saving rate (Nethousehold data
sample)
4.1 PRIVATE CONSUMPTION/GDP RATIOS BY COUNTRY GROUPS (Group
medians atconstant 1987 prices, fixed country sample,
1970-1994)
4.1b PRIVATE CONSUMPTION/GDP RATIOS BY WORLD REGIONS (Regional
medians atconstant 1987 prices, fixed country sample,
1970-1994)
4.2 PUBLIC CONSUMPTION/GDP RATIOS BY COUNTRY GROUPS (Group
medians at constant1987 prices, fixed country sample,
1970-1994)
4.2b PUBLIC CONSUMPTION/GDP RATIOS BY WORLD REGIONS (Regional
medians at constant1987 prices, fixed country sample,
1970-1994)
4.3 DOMESTIC INVESTMENT/GDP RATIOS BY COUNTRY GROUPS (Group
medians atconstant 1987 prices, fixed country sample,
1970-1994)
4.3b DOMESTIC INVESTMENT/GDP RATIOS BY WORLD REGIONS (Regional
medians atconstant prices, fixed country sample, 1970-1994)
5.1 LONG-TERM NATIONAL SAVING AND REAL INCOME: Gross national
saving rate and realper capita national disposable income (Country
averages, 1965-1995)
5.2 LONG-TERM NATIONAL SAVING AND REAL INCOME GROWTH: Gross
national savingrate and growth of real per capita national
disposable income (Country averages, 1965-1995)
5.3 LONG-TERM NATIONAL SAVING AND DOMESTIC INVESTMENT: Gross
national savingand gross domestic investment ratios to national
disposable income (Country averages, 1965-1995)
5.4a LONG-TERM NATIONAL SAVING AND PUBLIC SAVING: Gross national
saving and grosspublic saving ratios to national disposable income
(Unadjusted, central government definition;Country averages,
1965-1995)
5.4b LONG-TERM NATIONAL SAVING AND PUBLIC SAVING: Gross national
saving and grosspublic saving ratios to national disposable income
(Adjusted, public sector definition; Countryaverages,
1965-1995)
5.5 LONG-TERM NATIONAL SAVING AND REAL INTEREST RATE: Gross
national saving rateand average of ex-ante and ex-post real
interest rates (Country averages, 1965-1995)
5.6 LONG-TERM NATIONAL SAVING AND INFLATION: Gross national
saving and ex-post (log)inflation rate (Country averages,
1965-1995)
5.7 LONG-TERM NATIONAL SAVING AND MONEY STOCK: National saving
and broad money(M2, year average), ratios to gross national
disposable income (Country averages, 1965-1995)
5.8 LONG-TERM NATIONAL SAVING AND PRIVATE WEALTH: Gross national
saving and netprivate financial wealth (year average), ratios to
gross national disposable income (Countryaverages, 1965-1995)
-
4
5.9 LONG-TERM NATIONAL SAVING AND DOMESTIC CREDIT FLOWS: Gross
nationalsaving and domestic credit flows, ratios to gross national
disposable income (Country averages,1965-1995)
5.10 LONG-TERM NATIONAL SAVING AND SOCIAL SECURITY: National
saving rate and socialsecurity expenditure relative to public
expenditure (Country averages, 1965-1995)
5.11 LONG-TERM NATIONAL SAVING AND TOTAL DEPENDENCY RATIO:
National savingrate and population aged 65 and over and 15 and
under relative to working-age population(Country averages,
1965-1995)
5.12 LONG-TERM NATIONAL SAVING AND OLD DEPENDENCY RATIO:
National saving rateand population aged 65 and over relative to
working-age population (Country averages, 1965-1995)
5.13 LONG-TERM NATIONAL SAVING AND YOUNG DEPENDENCY RATIO:
National savingrate and population aged 15 and under relative to
working-age population (Country averages,1965-1995)
5.14 LONG-TERM NATIONAL SAVING AND URBANIZATION: National saving
rate and urbanpopulation relative to total (Country averages,
1965-1995)
5.15 LONG-TERM NATIONAL SAVING AND INCOME DISTRIBUTION: National
saving ratioand Gini coefficient (Country averages, 1965-1995).
5.16 LONG-TERM REAL GROSS DOMESTIC INVESTMENT AND REAL PER
CAPITA GDPGROWTH: Real gross domestic investment ratio to real GDP
and growth of real per capita GDP(Country averages, 1965-1995).
6.1 LONG-TERM PRIVATE SAVING AND REAL PRIVATE INCOME: Gross
private saving rateand real per capita private disposable income
(Unadjusted, central government definition; Countryaverages,
1965-1995)
6.2 LONG-TERM PRIVATE SAVING AND REAL PRIVATE INCOME GROWTH:
Gross privatesaving rate and growth rate of real per capita private
disposable income (Unadjusted, centralgovernment definition;
Country averages, 1965-1995)
6.3 LONG-TERM GROSS PRIVATE AND PUBLIC SAVING: Gross private
saving rate and grosspublic saving rate (Unadjusted, central
government definition; Country averages, 1965-1995)
6.4 LONG-TERM GROSS PRIVATE SAVING AND REAL INTEREST RATE: Gross
private savingrate and real interest rate (Unadjusted, central
government definition; Country averages, 1965-1995)
6.5 LONG-TERM GROSS PRIVATE SAVING AND MONEY STOCK: Gross
private saving rateand broad money (M2, year average), ratios to
gross private disposable income (Unadjusted,central government
definition; Country averages, 1965-1995)
6.6 LONG-TERM GROSS PRIVATE SAVING AND DOMESTIC CREDIT FLOWS:
Gross privatesaving rate and domestic credit flows, ratios to gross
private disposable income (Unadjusted,central government
definition; Country averages, 1965-1995)
6.7 LONG-TERM GROSS PRIVATE SAVING AND OLD AGE DEPENDENCY RATIO:
Grossprivate saving rate and population aged 65 and over relative
to working-age population(Unadjusted, central government
definition; Country averages, 1965-1995)
-
5
6.8 LONG-TERM GROSS PRIVATE SAVING AND YOUNG DEPENDENCY RATIO:
Grossprivate saving rate and population aged 15 and under relative
to working-age population(Unadjusted, central government
definition; Country averages, 1965-1995)
6.9 LONG-TERM GROSS PRIVATE SAVING AND URBANIZATION RATE: Gross
private savingrate and urbanization rate (Unadjusted, central
government definition; Country averages, 1965-1995)
7.1 LONG-TERM PRIVATE SAVING AND REAL PRIVATE INCOME: Gross
private saving rateand real per capita private disposable income
(Adjusted, public sector definition; Countryaverages,
1965-1995)
7.2 LONG-TERM GROSS PRIVATE SAVING AND REAL PRIVATE INCOME
GROWTH: Grossprivate saving rate and growth rate of real per capita
private disposable income (Adjusted, publicsector definition;
Country averages, 1965-1995)
7.3 LONG-TERM GROSS PRIVATE AND PUBLIC SAVING: Gross private
saving rate and grosspublic saving rate (Adjusted, public sector
definition; Country averages, 1965-1995)
7.4 LONG-TERM GROSS PRIVATE SAVING AND REAL INTEREST RATE: Gross
private savingrate and real interest rate (Adjusted, public sector
definition; Country averages, 1965-1995)
7.5 LONG-TERM GROSS PRIVATE SAVING AND MONEY STOCK: Gross
private saving rateand broad money (M2, year average), ratios to
gross private disposable income (Adjusted, publicsector definition;
Country averages, 1965-1995)
7.6 LONG-TERM GROSS PRIVATE SAVING AND DOMESTIC CREDIT FLOWS:
Gross privatesaving rate and domestic credit flows, ratios to gross
private disposable income (Adjusted, publicsector definition;
Country averages, 1965-1995)
7.7 LONG-TERM GROSS PRIVATE SAVING AND OLD AGE DEPENDENCY RATIO:
Grossprivate saving rate and population aged 65 and over relative
to working-age population (Adjusted,public sector definition;
Country averages, 1965-1995)
7.8 LONG-TERM GROSS PRIVATE SAVING AND YOUNG AGE DEPENDENCY
RATIO: Grossprivate saving rate and population aged 15 and under
relative to working-age population(Adjusted, public sector
definition; Country averages, 1965-1995)
7.9 LONG-TERM GROSS PRIVATE SAVING AND URBANIZATION RATE: Gross
private savingrate and urbanization rate (Adjusted, public sector
definition; Country averages, 1965-1995)
8.1 LONG-TERM NET HOUSEHOLD SAVING AND REAL PRIVATE INCOME: Net
householdsaving rate and real per capita private disposable income
(Unadjusted, central governmentdefinition; All countries,
1965-1995)
8.2 LONG-TERM NET HOUSEHOLD SAVING AND REAL PRIVATE INCOME
GROWTH: Nethousehold saving rate and growth of real per capita
private disposable income (Unadjusted, centralgovernment
definition; All countries, 1965-1995)
8.3 LONG-TERM NET HOUSEHOLD SAVING AND PUBLIC SECTOR SAVING: Net
householdsaving and public sector saving, as percent of gross
private disposable income (Adjusted, publicsector definition;
Country averages, 1965-1995)
-
6
8.4 LONG-TERM NET HOUSEHOLD SAVING AND CORPORATE SAVING: Net
householdsaving and net corporate saving ratios to private
disposable income (Unadjusted, centralgovernment definition; All
countries, 1965-1995)
8.5 LONG-TERM NET HOUSEHOLD SAVING AND REAL INTEREST RATE: Net
householdsaving rate and real interest rate (Country averages,
1965-1995)
8.6 LONG-TERM NET HOUSEHOLD SAVING AND MONEY STOCK: Net
household saving rateand broad money (M2, year average), ratios to
gross private disposable income (Country averages,1965-1995)
8.7 LONG-TERM NET HOUSEHOLD SAVING AND DOMESTIC CREDIT FLOWS:
Nethousehold saving and domestic credit flows, ratios to gross
private disposable income (Countryaverages, 1965-1995)
8.8 LONG-TERM NET HOUSEHOLD SAVING AND OLD AGE DEPENDENCY RATIO:
Nethousehold saving rate and population aged 65 and over relative
to working-age population(Country averages, 1965-1995)
8.9 LONG-TERM NET HOUSEHOLD SAVING AND YOUNG DEPENDENCY RATIO:
Nethousehold saving rate and population aged 15 and under relative
to working-age population(Country averages, 1965-1995)
8.10 LONG-TERM NET HOUSEHOLD SAVING AND URBANIZATION RATE: Net
householdsaving rate and urbanization rate (Adjusted, public sector
definition; Country averages, 1965-1995)
-
Figure 1.0
China
EAP
MENA
INDSA
SSA
LAC
1965-73
1974-84
1985-940
5
10
15
20
25
30
35
40
GN
S/G
ND
I (%
)
GROSS NATIONAL SAVING RATES BY WORLD REGIONSGross national
saving rate including net current transfers as percent of gross
national disposable income
(Regional medians at current prices, 1965-1994)
-
Figure 1.1
GROSS NATIONAL SAVING RATES BY COUNTRY GROUPSGross national
saving as percent of gross national disposable income
(Group medians at current prices, fixed country sample,
1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Med
ian
GN
S/G
ND
I
China TAKEOFF otherLDC WORLD IND
-
Figure 1.1b
GROSS NATIONAL SAVING RATES BY COUNTRY GROUPSGross national
saving as percent of gross national disposable income
(Unweighted group averages at current prices, fixed country
sample, 1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Ave
rag
e G
NS
/GN
DI
China TAKEOFF otherLDC WORLD IND
-
Figure 1.1c
GROSS NATIONAL SAVING RATES BY COUNTRY GROUPSGross national
saving as percent of gross national disposable income
(GNDI-weighted group averages, fixed country sample,
1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Wei
gh
ted
Ave
rag
e G
NS
/GN
DI
China TAKEOFF otherLDC WORLD IND
-
Figure 1.2
GROSS NATIONAL SAVING RATES BY WORLD REGIONSGross national
saving as percent of gross national disposable income(Regional
medians at current prices, fixed country sample, 1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Med
ian
GN
S/G
ND
I
China MENA EAP SA SSA LAC IND
-
Figure 1.2b
GROSS NATIONAL SAVING RATES BY WORLD REGIONSGross national
saving as percent of gross national disposable income
(Unweighted regional averages at current prices, fixed country
sample, 1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Ave
rag
e G
NS
/GN
DI
China MENA EAP SA SSA LAC IND
-
Figure 1.2c
GROSS NATIONAL SAVING RATES BY WORLD REGIONSGross national
saving as percent of gross national disposable income(GNDI-weighted
regional averages, fixed country sample, 1965-1994)
0%
10%
20%
30%
40%
50%
60%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Ave
rag
e G
NS
/GN
DI
China MENA EAP SA SSA LAC IND
-
Figure 1.3
GROSS NATIONAL SAVING RATES BY COUNTRY GROUPSGross national
saving as a percentage of gross national disposable income
(Group medians at current prices, full sample, 1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Med
ian
GN
S/G
ND
I
China TAKEOFF otherLDC WORLD IND
-
Figure 1.3b
GROSS NATIONAL SAVING RATES BY COUNTRY GROUPSGross national
saving as percent of gross national disposable income(Unweighted
group averages at current prices, full sample, 1965-1994)
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Ave
rag
e G
NS
/GN
DI
China TAKEOFF otherLDC WORLD IND
-
Figure 1.3c
GROSS NATIONAL SAVING RATES BY COUNTRY GROUPSGross national
saving as percent of gross national disposable income
(GNDI-weighted group averages, full sample, 1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Wei
gh
ted
Ave
rag
e G
NS
/GN
DI
China TAKEOFF otherLDC WORLD IND
-
Figure 1.4
GROSS NATIONAL SAVING RATES BY WORLD REGIONSGross national
saving as percent of gross national disposable income
(Regional medians at current prices, full sample, 1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Med
ian
GN
S/G
ND
I
China MENA EAP SA SSA LAC IND
-
Figure 1.4b
GROSS NATIONAL SAVING RATES BY WORLD REGIONSGross national
saving as percent of gross national disposable income
(Unweighted regional averages at current prices, full sample,
1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Ave
rag
e G
NS
/GN
DI
China MENA EAP SA SSA LAC IND
-
Figure 1.4c
GROSS NATIONAL SAVING RATES BY WORLD REGIONSGross national
saving as percent of gross national disposable income
(GNDI-weighted regional averages, full sample, 1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991 1993
Wei
gh
ted
Ave
rag
e G
NS
/GN
DI
China MENA EAP SA SSA LAC IND
-
Figure 1.5
GROSS NATIONAL SAVING RATES - LEVELS AND VARIATIONGross national
saving rate as percent of gross national disposable income and the
standard deviation of the
gross national saving rate(Country averages at current prices,
fixed sample, 1965-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
0% 2% 4% 6% 8% 10% 12% 14% 16%
Standard Deviation of GNS/GNDI
Ave
rag
e G
NS
/GN
DI
LDC INDDeveloping Industrial
-
Figure 2.0
EAP
MENA
INDSA
SSA
LAC
1970-79
1980-89
1990-950
5
10
15
20
25
30
35
40
GP
riS
/GP
DI (
%)
GROSS PRIVATE SAVING RATES BY WORLD REGIONSGross private saving
rate as percent of gross private disposable income
(Adjusted, public sector definition; Regional medians at current
prices, 1965-1994)
-
Figure 2.1
ALTERNATIVE MEASURES OF PUBLIC SAVING IN DEVELOPING
COUNTRIES(Percent of gross national disposable income, fixed
country sample, 1970-1994)
0%
1%
2%
3%
4%
5%
6%
7%
8%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Pu
blic
Sav
ing
/GN
DI
CU PA PU CA
-
Figure 2.2
ALTERNATIVE MEASURES OF PUBLIC SAVING IN INDUSTRIAL
COUNTRIES(Percent of gross national disposable income, fixed
country sample, 1970-1994)
-4%
-2%
0%
2%
4%
6%
8%
10%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
PS
Sav
/GN
DI,
%
CU PA PU CA
-
Figure 2.3
ALTERNATIVE MEASURES OF PRIVATE SAVING IN INDUSTRIAL AND
DEVELOPING COUNTRIES (Private saving as a percent of gross private
disposable income; Medians of fixed country sample, 1970-1994)
10%
12%
14%
16%
18%
20%
22%
24%
26%
28%
30%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Pri
vate
Sav
ing
/GP
DI (
%)
CU_IND PA_IND CU_LDC PA_LDC
-
Figure 2.4
COMPOSITION OF GROSS NATIONAL SAVINGPublic and private saving as
percent of gross national disposable income
(Unadjusted, central Government definition; Medians of fixed
country sample, 1970-1994)
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Pu
bic
an
d P
riva
te S
avin
g/G
ND
I (%
)
Pub_IND Priv_IND Pub_LDC Priv_LDC
-
Figure 2.5
COMPOSITION OF GROSS NATIONAL SAVINGPublic and private saving as
percent of gross national disposable income
(Adjusted, public sector definition; Medians of fixed country
sample, 1970-1994)
-5%
0%
5%
10%
15%
20%
25%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Pri
vate
Sav
ing
/GN
DI
Pub_IND Priv_IND Pub_LDC Priv_LDC
-
Figure 2.6
GROSS PRIVATE SAVING RATES BY COUNTRY GROUPSGross private saving
as percent of gross private disposable income
(Adjusted, public sector definition; Group medians at current
prices, fixed country sample, 1970-1994)
0%
5%
10%
15%
20%
25%
30%
35%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Pri
vate
Sav
ing
/GP
DI
TAKEOFF otherLDC WORLD IND
-
Figure 2.7
GROSS PRIVATE SAVING RATES BY WORLD REGIONSGross private saving
as percent of gross private disposable income
(Adjusted, public sector definition; Group medians at current
prices, fixed country sample, 1970-1994)
0%
5%
10%
15%
20%
25%
30%
35%
40%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Pri
vate
Sav
ing
/GP
DI
MENA EAP SA SSA LAC IND
-
Figure 2.8
GROSS PRIVATE SAVING RATES BY COUNTRY GROUPSGross private saving
as percent of gross private disposable income
(Adjusted, public sector definition; Group medians at current
prices, full country sample, 1970-1994)
0%
5%
10%
15%
20%
25%
30%
35%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Pri
vate
Sav
ing
/GP
DI
TAKEOFF otherLDC WORLD IND
-
Figure 2.9
GROSS PRIVATE SAVING RATES BY WORLD REGIONSGross private saving
as percent of gross private disposable income
(Adjusted, public sector definition; Group medians at current
prices, full country sample, 1970-1994)
0%
5%
10%
15%
20%
25%
30%
35%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Pri
vate
Sav
ing
/GP
DI
MENA EAP SA SSA LAC IND
-
Figure 2.10
PRIVATE SAVING RATES - LEVELS AND VARIATIONPrivate saving rate
as percent of gross private disposable income & the standard
deviation of the private saving
rate(Unadjusted, central government definition; Country averages
at current prices, fixed sample)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
0% 2% 4% 6% 8% 10% 12% 14% 16% 18%
Standard Deviation of GPriS/GPDI
Ave
rag
e G
Pri
S/G
PD
I
LDC INDDeveloping Industrial
-
Figure 2.11
PRIVATE SAVING RATES - LEVELS AND VARIATIONPrivate saving rate
as percent of gross private disposable income & the standard
deviation of the private saving
rate(Adjusted, public sector definition; Country averages at
current prices, fixed sample)
0%
5%
10%
15%
20%
25%
30%
35%
0% 5% 10% 15% 20% 25%
Standard Deviation of GPriS/GPDI
Ave
rag
e G
Pri
S/G
PD
I
LDC IND
Developing Industrial
-
Figure 3.1
COMPOSITION OF PRIVATE SAVINGNet household saving and net
corporate saving, ratios to private
disposable income(Net household and corporate saving data
sample, 1965-1992)
0%
2%
4%
6%
8%
10%
12%
14%
1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989
1991
Co
rpo
rate
an
d H
ou
seh
old
Sav
ing
/ GP
DI (
%)
HH_IND CO_IND HH_LDC CO_LDC
-
Figure 3.2
HOUSEHOLD SAVING RATES - LEVELS AND VARIATIONHousehold saving as
percent of gross private disposable income and the standard
deviation of the household
saving rate(Net household data sample)
-5%
0%
5%
10%
15%
20%
0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10%
Standard Deviation of Net HHS/GPDI
Ave
rag
e N
et H
HS
/GP
DI
LDC IND
Developing Industrial
-
Figure 4.1
PRIVATE CONSUMPTION/ GDP RATIOS BY COUNTRY GROUPS(Group medians
at constant 1987 prices, fixed country sample, 1970-1994)
40%
45%
50%
55%
60%
65%
70%
75%
80%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Rea
l Pri
vate
Co
nsu
mp
tio
n/R
eal G
DP
China TAKEOFF otherLDC WORLD IND
-
Figure 4.1b
PRIVATE CONSUMPTION/ GDP RATIOS BY WORLD REGIONS(Regional
medians at constant 1987 prices, fixed country sample,
1970-1994)
35%
40%
45%
50%
55%
60%
65%
70%
75%
80%
85%
90%
95%
100%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Rea
l Pri
vate
Co
nsu
mp
tio
n/ R
eal G
DP
China MENA EAP SA SSA LAC IND
-
Figure 4.2
PUBLIC CONSUMPTION/ GDP RATIOS BY COUNTRY GROUPS(Group medians
at constant 1987 prices, fixed country sample, 1970-1994)
6%
8%
10%
12%
14%
16%
18%
20%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Rea
l Go
vern
men
t C
on
sum
pti
on
/Rea
l GD
P
China TAKEOFF otherLDC WORLD IND
-
Figure 4.2b
PUBLIC CONSUMPTION/GDP RATIOS BY WORLD REGIONS(Regional medians
at constant 1987 prices, fixed country sample, 1970-1994)
6%
8%
10%
12%
14%
16%
18%
20%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Rea
l Go
vern
men
t C
on
sum
pti
on
/ Rea
l GD
P
China MENA EAP SA SSA LAC IND
-
Figure 4.3
DOMESTIC INVESTMENT/ GDP RATIOS BY COUNTRY GROUPS(Group medians
at constant 1987 prices, fixed country sample, 1970-1994)
15%
20%
25%
30%
35%
40%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Rea
l GD
I/Rea
l GD
P
China TAKEOFF otherLDC WORLD IND
-
Figure 4.3b
DOMESTIC INVESTMENT/ GDP RATIOS BY WORLD REGIONS(Regional
medians at constant prices, fixed country sample, 1970-1994)
10%
15%
20%
25%
30%
35%
40%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
1994
Med
ian
Rea
l GD
I/Rea
l GD
P
China MENA EAP SA SSA LAC IND
-
Figure 5.1
LONG-TERM NATIONAL SAVING AND REAL INCOMEGross national saving
rate and real per capita national disposable income
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000
Real GNDI Per Capita
GN
S/G
ND
I (%
)
LDC IND
Developing Industrial
-
Figure 5.2
LONG-TERM NATIONAL SAVING AND REAL INCOME GROWTHGross national
saving rate and growth of real per capita national disposable
income
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Growth Rate of Real Per Capita GNDI (%)
GN
S/ G
ND
I (%
)
LDC IND
Developing Industrial
-
Figure 5.3
LONG-TERM NATIONAL SAVING AND DOMESTIC INVESTMENTGross national
saving and gross domestic investment ratios to national disposable
income
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
10% 15% 20% 25% 30% 35% 40%
GDI/GNDI (%)
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.4a
LONG-TERM NATIONAL SAVING AND PUBLIC SAVINGGross national saving
and gross public saving ratios to national disposable income
(Unadjusted, central government definition; Country averages,
1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
-15% -10% -5% 0% 5% 10% 15%
Unadjusted CCG Saving/GNDI (%)
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.4b
LONG-TERM NATIONAL SAVING AND PUBLIC SAVINGGross national saving
and gross public saving ratios to national disposable income
(Adjusted, public sector definition; Country averages,
1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
-5% 0% 5% 10% 15% 20%
PS Saving/ GNDI (%)
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.5
LONG-TERM NATIONAL SAVING AND REAL INTEREST RATEGross national
saving rate and average of ex-ante and ex-post real interest
rates
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
-50% -40% -30% -20% -10% 0% 10% 20% 30% 40%
Real Interest Rate
GN
S/G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.6
LONG-TERM NATIONAL SAVING AND INFLATIONGross national saving and
ex-post (log) inflation rate
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
0% 20% 40% 60% 80% 100% 120%
Inflation rate
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.7
LONG-TERM NATIONAL SAVING AND MONEY STOCKNational saving and
broad money (M2, year average), ratios to gross national disposable
income
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
0% 20% 40% 60% 80% 100% 120%
M2/GNDI (%)
GN
S/ G
ND
I (%
)
LDC IND
Developing Industrial
-
Figure 5.8
LONG-TERM NATIONAL SAVING AND PRIVATE WEALTHGross national
saving and net private financial wealth (year average), ratios to
gross national disposable income
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
-100% -50% 0% 50% 100% 150% 200% 250% 300%
Private Financial Wealth/GNDI (%)
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.9
LONG-TERM NATIONAL SAVING AND DOMESTIC CREDIT FLOWSGross
national saving and domestic credit flows, ratios to gross national
disposable income
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
-5% 0% 5% 10% 15% 20% 25%
Domestic Credit Flow/GNDI (%)
GN
S/G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.10
LONG-TERM NATIONAL SAVING AND SOCIAL SECURITYNational saving
rate and social security expenditure relative to public
expenditure
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%
Public Social Security (% of Public Expenditure)
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.11
LONG-TERM NATIONAL SAVING AND TOTAL DEPENDENCY RATIONational
saving rate and population aged 65 and over and 15 and under
relative to working age population
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
40% 50% 60% 70% 80% 90% 100% 110% 120%
Total Age Dependency Ratio
GN
S/G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.12
LONG-TERM NATIONAL SAVING AND OLD DEPENDENCY RATIONational
saving ratio and population aged 65 and over relative to
working-age population
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
0% 5% 10% 15% 20% 25% 30%
Old Dependency Ratio
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.13
LONG-TERM NATIONAL SAVING AND YOUNG DEPENDENCY RATIONational
saving ratio and population aged 15 and under relative to
working-age population
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
Young Dependency Ratio
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.14
LONG-TERM NATIONAL SAVING AND URBANIZATIONNational saving ratio
and urban population relative to total
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
0% 20% 40% 60% 80% 100%
Urban Population Ratio
GN
S/ G
ND
I (%
)
LDC INDDeveloping Industrial
-
Figure 5.15
LONG-TERM NATIONAL SAVING AND INCOME DISTRIBUTIONNational saving
ratio and Gini coefficient
(Country averages, 1965-1995)
0%
5%
10%
15%
20%
25%
30%
35%
40%
25 30 35 40 45 50 55 60 65