International Journal of Applied Econometrics and Quantitative Studies Vol.

4-1 (2007)

SAVINGS BEHAVIOUR IN THE INDIAN ECONOMY

UPENDER, M. *
REDDY, N.L.
Abstract
An attempt has been made in the present study to examine the savings
behaviour in the Indian Economy in terms of shift in the growth rates of
domestic savings, and in magnitude of income elasticity of the
domestic savings at the aggregate and disaggregate levels during post
economic reform period. The results show that there is no shift in the
growth rate of the domestic savings both at aggregate and disaggregate
levels during post economic reform period. However there has been
acceleration in the growth rates of domestic savings of household and
private sectors and deceleration in public sector during 1950--2002.
The estimate of constant income elasticity of household savings is found
to be more than unity implying that the marginal propensity to save is
higher than the average propensity to save, all else equal. Further the
constant income elasticity of household savings is moderately higher
than that of the income elasticities of domestic savings estimated for
private and public sectors during pre economic reform period. The
results point out that there is no shift in the magnitude of income
elasticity of savings of household, private and public sectors during post
economic reform period showing the homogeneity in the size of the
income elasticity of domestic savings. Thus the economic reforms that
have been initiated in 1992 could not bump up the growth rate of
savings and magnitude of the income elasticity of domestic savings
both at aggregate and disaggregate levels in the Indian Economy during
post economic reform period.
JEL code: O53
Keywords: Savings, India
*

Dr M Upender, Professor of Economics, Department of Economics,

Osmania University,Hyderabad-500 007, Andhra Pradesh, India,
[email protected], and Dr N Laxma Reddy, Branch Manager,
Bank of Maharashtra[ Gopvernment of India] Hyderabad-500 072, Andhra
Pradesh, India.
Note: This article is based on the thesis entitled Savings Behaviour in the
Indian Economy submitted by Dr N Laxma Reddy to Osmania University,
for the award of PhD degree [Awarded in 2007] under the guidance of Prof
M Upender.

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

1. Introduction
It is well-known fact that the rate of savings has been an
important economic variable for economic development of,
particularly, the countries like India. The extent of domestic savings
is only ultimate source for capital formation, which is indispensable
for rapid economic development in India. The excess of income over
consumption expenditure is referred to as savings. The Policy of the
government of India has been to promote savings and capital
formation. Increased savings can be used for financing required
investment. It is also known fact that an increase in the rate of
investment is essential for rapid development. Increase in investment
is possible only by increase in savings rate. Therefore the extent of
domestic savings is an imperative factor for attaining high rate of
investment. The gross savings in the economy can be increased by
increasing the national income. Therefore propensity to save
depends,inter alia,on the national income. Thus the aggregate
savings,inter alia, is a function of national income. The generation of
the theoretical savings function depends on the aggregate
consumption function as the sum of the aggregate savings [GDS] and
consumption expenditure [GC] is the aggregate income [GDP].
GDP=C+GDS
GDS=GDP- GC
The general equation for the linear consumption function is
GC=c0+c1GDP
Where c0 is autonomous consumption expenditure and c1,
dGC/dGDP, is the marginal propensity to consume.
The general equation for the linear savings function is
GDS=s0+s1YGDP
Where s0 is the amount of the savings at the theoretical zero level of
Income and s1, dGDS/dGDP is the marginal propensity to save.
Substituting equation [GC=c0+c1GDP] in equation [GDS=GDPGC], we have the following
GDS=GDP- (c0+c1GDP)
GDS= - c0 - c1GDP +GDP)
GDS=- c0 + (1- c1) GDP where 1- c1= s1
36

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

Thus the domestic savings would,inter alia, depend on income. In

view of the importance of the domestic savings, the empirical on the
behaviour of the savings have been carried out in India using the
macro time series data. In India domestic savings originate from
three principal sectors namely: (i) household sector, (ii) private
corporate sector and (iii) Public sector. The present study on the
savings behaviour in the Indian Economy will be an extension to the
empirical literature in terms of [1] using the macro time series data
available till date [ 53 years] [2] searching the structural change in
the savings function during the post economic reforms period and [3]
estimating the speed of adjustment between the actual change in the
domestic savings and desired change in the domestic savings in
India. More specifically the present study has been carried out with
the following objectives
1. To find out the presence of acceleration/deceleration in the
growth rates of the gross domestic savings
2. To estimate the degree of income elasticity of gross
domestic savings and
3. To examine the extent speed of adjustment between the
actual change and desired change in gross domestic savings
The above objectives have been examined in the luminosity of the
economic reforms that have been initiated in 1992 at the aggregate
and disaggregate levels.
Hypotheses of the Empirical Study. Keeping the objectives of the
present Study in view, the empirical strength of the following
hypotheses have also been examined
1. There has been acceleration in the growth rate growth rate of
savings of the household sector
2. The marginal propensity to save of the household sector is
relatively higher than that of the private and public sectors
3. The income elasticity of domestic savings of the household
sector is relatively higher than that of the private and
government sectors
4. The speed of adjustment between actual change and desired
change in domestic savings of household sector is not quick
37

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

The generation of empirical knowledge out of the present exercise

on the estimates of constant/variable growth rates of domestic
savings and responsiveness of domestic savings to the changes in
income at the aggregate and disaggregate levels will be handy in
understanding the impact of the economic reforms on the behaviour
of the domestic savings in India .
2. A Concise Review of Earlier studies on Savings Behaviour in
India
There have been a plethora of empirical studies on the savings
behaviour in India based on both cross-sectional and time series data.
Some of the time series empirical studies related to the behaviour of
savings in the Indian Economy have been examined to view the
present study in a wider perspective . Of them the most important
study that was carried out by krishnamurthy and Saibaba1 The
important contribution of this study was that the rate of savings in
India was rising during 1952-53 to 1978-79 with year to year
fluctuations. In an other study, Majumdar2 et al examined the
behaviour of savings in the Indian Economy. They observed that the
net income and the share of non-agricultural income in the total
income were the important factors in explaining the variations in
savings. Joshi3 analyzed the savings behaviour in India over a period
of thirteen years. He observed that savings out of incremental income
plays the key role in raising the rate of savings. Shetty4 Reviewd the
trends in domestic saving rates in India. He observed the changes in
consumption patterns seem an obvious explanation for the absence of
any buoyancy in household savings. In an other study Krishnamurthy
1

K Krishnamurthy and P Saibaba, Savings Behaviour in India, Institute of

Economic Growth, Occasional Paper:New Series, No.6, Hindustan
Publishing Corporation, Delhi,1982
2
N A Majumdar, T R Venkatachalam and M V Raghavachary, The high
Saving Phase of the Indian Economy:1976-79-An Explorative
Interpretation,Occational Papers, Reserve Bank of India,1980
3
Joshi, Saving Behaviour in India,Indian Journal of Economics, Vol.50,
April-June 1970,
4
S L Shetty ,Savings Behaviour in India Some Lessons,Economic and
Political Weekly, Vol.XXV,No.11,March 17, 1990
38

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

et al5 examined the trends in savings and its composition in India. In

an empirical exercise, Upender6 estimated the elasticity of gross
domestic savings with respect to gross domestic product during
1950-90 using the linear regression model. The numerical value of
the elasticity was turned out to be unity.This study did not attempt to
see the presence of acceleration and deceleration in the growth rate
of the domestic savings and speed of adjustment. Thus the time
series studies examined the savings behaviour till 1990 only in the
Indian Economy. They have not attempted to examine the speed of
adjustment between the actual change and desired change in the
domestic savings. No study has been undertaken to scan the
structural change in the savings function after the economic reforms
initiated in India using the most recent data. There fore there is a
need to generate empirical information on these aspects for the
Indian Economy. The present Study has been an endeavor in this
direction.
3. Empirical Methodology
3.1 Data Base. In order to examine the above objectives and test the
empirical validity of the hypotheses of the present study, the required
time series data on domestic savings by houshold,private and public
sectors, domestic income and Gross Domestic Product at market
prices have been collected from various issues of the Economic
Survey published by the ministry of finance [Economic Wing]
National Accounts Statistics published by Central Statistical
Organization , Ministry of planning and the Basic statistics relating
to the Indian Economy by Reserve Bank of India for the period from
1950-51 to 2002- 03.
3.2. Empirical Techniques.
3.2.1. Growth rates of domestic savings. The shift in the growth
rates of the domestic savings during post economic reform period has

5
K Krishnamurthy, KS Krishna Swamy and P D Sharma, Saving Behaviour in
India An Overview in Development Process of Indian Economy edited by P R
Brahmananda and V R Panchamukhi, Himalaya Publishing House, Delhi, 1982
6
M Upender, Estimation of Propensities to save in the Indian Econmy using the
time series data 1950-51- 1989-90, Finance India, Vol.VII,No.2, June 1994,
pp.365-370

39

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

been scanned by fitting the following form of the specification with

dummy and interaction variables to the time series data.
log GDS = log 0 + 1 dummy + 2 time+ 3 (dummy*time)

[I]

Where GDS = Gross Domestic Savings of the Household sector /

Private sector / Government sector [Rs.crore], Time = Time in [53]
years [1950-51 to 2002-2003], Dummy [proxy for policy variable] =
0 for the years from 1950-51 to 1991-92 [pre economic reform
period] and 1 for the years from 1992-93 to 2002-03 [post economic
reform period]
If 1 is positively significant then there will be an upward
shift in the average domestic savings during post economic reform
period. If 1 is negatively significant then there will be a downward
shift in the average domestic savings during post economic reform
period. If 1 is negatively/positively insignificant then there will be
no shift in the average domestic savings during post economic
reform period
Annual growth rate of domestic savings during pre economic
reform period= d log GDP / dTime = 2
The differential growth rate of domestic savings during post
economic reform period = 3 < 0 [downward shift] or > 0 [upward
shift] subject to test of significance.
Growth rate of the domestic savings during post economic reform
period = 12
If 1 and 3 in [I] are not significant,the following regression
model will be fitted to know the variable growth in terms of
acceleration /deceleration during th period under consideration
log GDS = log 0 + 1 time b2 time2

[II]

Annual growth rate of domestic savings= dlogY/d time=1 22*

time. If 2 is significantly positive/negative then there wil be
acceleration/deceleration in the growth rate of domestic savings.
If 2 is not significant then there will be constant growth rate in the
domestic savings

40

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

3.2.2: Income elasticity of Domestic Savings: Shift in income

elasticity of domestic savings. The impact of economic reforms on
domestic savings in terms of a shift in the income elasticity of
domestic savings has been examined by fitting the following form of
specification with an interaction variable7 [D*logGDPt]
log GDSt = log 0 + 1 log GDPt + 2 D + 3 ( D * log GDPt ) + error III

GDS = Gross Domestic Savings of the Household [GDS1]/

Private[GDS2] / Government sector[GDS3][Rs crore]
GDPt = Gross Domestic Product at market prices [Rs crore]
0 =Intercept during pre economic reform period [D = 0], 2 =
Differential intercept during post economic reform period [D = 1],
1 =Magnitude of income elasticity of domestic savings during pre
economic reform period (D = 0); > 0, 3 = Magnitude of
differential income elasticity of domestic savings during post
economic reform period (D = 1) ; (3 more than or less than zero
subject to the significance, (1 3 ) = Magnitude of income
elasticity of domestic savings during post economic reform period
(D = 1), 3 = differential coefficient of income elasticity of
domestic savings [3 more than or less than 0] that allows a shift [an
upward / a downward] in the income elasticity of domestic savings
during post reform period [1992 to 2002 ] when D =1. As the
interaction variable [D*logGDPt] enters the equation in dichotomous
form [i.e.,D = 0 in pre economic reform period and D = 1 in post
economic reform period] the derivative of logGDSt with respect to
[D*log GDPt] does not exist. Instead, the coefficient of
[D*logGDPt] subject to statistical significance, measures the
discontinuous effect of the presence the attribute [D = 1] represented
by an interaction variable on the domestic savings .The variable
[D*log GDPt], which is called an interaction variable, has been
introduced in model to capture the interaction effect of economic
reforms and income on the domestic savings .The interaction
7

The methodology followed in the present study is the similar to the methodology
followed by M Upender., Estimates of Coefficients of Economic Relationships:
Some Exercise for India, Manak Publications 2002, New Delhi

41

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

variable takes a value equal to log GDPt during post economic

reform period [when D = 1] and 0 during pre economic reform
period [when D = 0] ; If [ 1* 3* ] more than or less than 1*
then there will be an upward or a downward shift in the degree of
income elasticity of domestic savings during post reform period ; If
[1* + 3**] = (1*, then there will be a homogeneity in the
magnitude of income elasticity of domestic savings i.e., magnitude
of income elasticity of domestic savings remains the same in pre and
post economic reform periods implying the absence of differential
income elasticity of domestic savings . Where * and ** denote
statistically significant and insignificant respectively.
3.3.3:Distributed lag Domestic Savings Function: Speed of
Adjustment. In order to examine the speed of adjustment between
actual change in domestic savings and desired change in domestic
savings a distributed lag model with the partial adjustment
mechanism has also been estimated at aggregate and disaggregate
levels.
log GDSt* = log 0 + 1 log GDP t + error

(IV)

Where GDSt*=Desired level [long run/equilibrium level] of domestic

savings in current [t] year, GDPt= Gross domestic product at market
prices [Income] in current year
The following mechanism [Marc Nerlovians partial adjustment
model] has been adopted to estimate the equation [IV] to know the
speed of adjustment between actual change in domestic savings and
desired change in domestic savings
(GDSt/GDSt-1) = (GDSt* / GDSt-1) is transformed into log linear
form
(log GDSt - log GDSt-1) = ( log GDSt* - log GDSt-1)
(V)
Actual change in domestic savings = [Desired change in domestic
savings], = Actual change in domestic savings/ Desired change in
domestic savings, where = Speed of adjustment <1
42

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

Substituting the equation [IV] in equation [V] we obtain the

following
(log GDSt - log GDSt-1) = (log 0 + 1 log GDPt - log GDSt-1)
log GDSt = (log 0 + 1 log GDPt - log GDSt-1) + log GDSt-1
log GDSt = ( log 0 + b1 log GDPt - log GDSt-1) + log GDSt-1
log GDSt = log b0 + 1 log GDPt + (1- ) log GDSt-1
(VI)
The equation [VI] is known as the short run domestic savings
function [to examine short run economic relationships]. The long run
domestic savings function [IV] will be estimated by deflating the
short run domestic savings function [VI] by the coefficient of
adjustment [], ratio of Actual change in domestic savings to
Desired change in domestic savings, and skip the log GDSt-1. as
shown below
(log GDSt / ) = ( log 0 / ) + ( 1 log GDPt) /
(VII)
log GDSt* = log 0 + 1 log GDP t
Where log GDSt*=log GDSt /
4. Savings Behaviour in the Indian Economy :Empirical Results
4.1: A Visual Plot of the Time series Data. In order to get some
insight in to the behavior of the tendency of the time series data on
domestic savings at the aggregate and disaggregate levels in India
the following graph is plotted for the period under consideration.
MOVEMENTS IN DOMESTIC SAVINGS DURING 1950-2002
700000
600000
500000
400000
300000
200000
100000
0
-100000
50

55

60

65

70

75

80

85

90

95

00

Y1 = GROSS DOMESTIC SAVINGS

Y2=GROSS HOUSEHOLD SECTOR SAVINGS
Y3=GROSS PRIVATE CORPORATE SECTOR SAVINGS
Y4=GROSS PUBLIC SECTOR SAVINGS

43

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

It can be perceived from the plot that there is an increasing tendency

in the aggregate savings and increasing tendency with fluctuations in
household, private and public sectors. In order to have a precise
statistics on the growth rates, an attempt has also made fitting
regression equations to the data [summary statistics are furnished in
appendix]
4.2: Growth Rate Of The Domestic Savings at The Aggregate
Level. With a view to scan the presence/absence of an upward or
downward shift in the growth rate of the domestic savings at the
aggregate and disaggregate levels during post economic reform
period the regression model with dummy and interaction variables
has been fitted to the macro time serried data.The results are
furnished in the following tables. The estimate of the regression
coefficient of the interaction variable is not significant showing
that the absence of shift in the growth rate of the gross domestic
savings during post economic reform period. The absence of shift in
the growth rate shows that the results presented in table - 1 are
valid for the entire period 1950-2002.Therfore the regression model
with time and square of time has been attempted to see the presence
of acceleration / deceleration, known as variable growth, in the
growth rate of the gross domestic savings. The results of the same
are set out in table -2. The estimates of the regression coefficients of
time and that of its square are positively significant implying the
presence of acceleration in the growth rate of the gross domestic
savings during the period under consideration.
Table 1.Search For Shift In The Growth Rate Of Gross Domestic Savings
Dependent Variable: Gross Domestic Savings= Log(Gds1)
Method: Least Squares Sample: 1950 2002. Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
6.383195
0.042896
148.8047
Time*
0.128535
0.001801
71.35991
Dummy
0.516732
0.636936
0.811278
Dummy*time
-0.004252
0.013610
-0.312416
R-squared
0.995649
Durbin-Watson statistic
0.522005
Adjusted R-sq.
0.995383
Notes:* Significant at one percent level

44

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

Table 2. Acceleration/Deceleration In The Growth Rate Of Gross

Domestic Savings
Dependent Variable: Gross Domestic Savings= Log(Gds1)
Method: Least Squares Sample : 1950 2002. Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
6.560347
0.043679
150.1962
Time*
0.103357
0.003884
26.60788
Time2*
0.000597
7.22E-05
8.263268
Durbin-Watson statistic
R-squared
0.997316
0.632431
Adjusted R-sq.
0.997208
Notes:* Significant at one percent level

4.3: Growth Rate Of The Domestic Savings At The Disaggregate

Level. The regression results furnished in table 3 show that the
coefficients of dummy and interaction variable are not significant
showing the absence of the shift in the growth rate of the gross
domestic savings by the household sector during post economic
reform period. The absence of shift during post economic reform
period shows that the empirical estimates presented in table- 3 are
applicable for the entire period 1950-2002.In view of this the
regression model with time and square of time has been estimated
to observe the presence of acceleration / deceleration in the growth
rate of the gross domestic savings by household sector [ table 4] .
Table 3. Search For Shift In The Growth Rate Of Gross Domestic
Household Sector Savings
Dependent Variable: Gross Domestic Household Sector Savings=
Log(Gds2)
Method: Least Squares Sample: 1950 2002. Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
5.974600
0.052429
113.9558
Time*
0.131677
0.002201
59.81268
Dummy
-0.291701
0.778478
-0.374707
Dummy*time
0.014604
0.016635
0.877904
R-squared
0.993953
Durbin-Watson statistic
0.449136
Adjusted R-sq.
0.993582
Notes:* Significant at one percent level

45

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

The estimates of the regression coefficients of time and that of its

square are positively significant implying that there has been an
acceleration in the growth rate of the gross domestic savings by the
household sector during the period under consideration. The
regression results furnished in table 5 illustrate that the
coefficients of dummy and
interaction variable are also not
significant in case of private corporate sector
illustrating the
absence of the shift in the growth rate of the gross domestic
savings by the private corporate sector during post economic reform
period. The absence of shift in case of private corporate sector
during post economic reform period shows that the empirical
estimates presented in table- 5 are pertinent for the entire period
under consideration.
Table 4 Acceleration/Deceleration In The Growth Rate Of Gross Domestic
Household Sector Savings
Dependent Variable: Gross Domestic Household Sector Savings=
Log(Gds2)
Method: Least Squares. Sample: 1950 2002. Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
6.213857
0.046403
133.9111
Time*
0.097764
0.004127
23.69060
Time2*
0.000796
7.68E-05
10.37162
R-squared
0.995649
Durbin-Watson statistic
0.522005
Adjusted R-squared
0.995383
Notes:* Significant at one percent level
Table 5 Search For Shift In The Growth Rate Of Gross Domestic Private
Corporate Sector Savings
Dependent Variable: Gross Domestic Private Corporate Sector Savings=
Log(Gds3)
Method: Least Squares Sample: 1950 2002, Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
4.168093
0.074428
56.00199
Time*
0.124079
0.003125
39.70270
Dummy
0.537193
1.105117
0.486096
Dummy*time
0.008183
0.023615
0.346530
R-squared
0.988391
Durbin-Watson statistic
1.064254
Adjusted R-sq.
0.987680
Notes:* Significant at one percent level
46

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

For that reason the regression model with time and square of
time has been fitted to the time series data to make out the presence
of acceleration / deceleration in the growth rate of the gross domestic
savings by private corporate sector. The results of the same are
furnished in table -6
Table 6. Acceleration/Deceleration In The Growth Rate Of Gross Domestic
Private Corporate Sector Savings
Dependent Variable: Gross Domestic Private Sector Savings= Log(Gds3)
Method: Least Squares. Sample: 1950 2002. Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
4.496744
0.090690
49.58363
Time*
0.072424
0.008065
8.979684
Time2*
0.001325
0.000150
8.832906
R-squared
0.989743
Durbin-Watson statistic
0.808024
Adjusted R-sq.
0.989332
Notes:* Significant at one percent level

The estimates of the regression coefficients of time and that of its

square are positively significant implying that there has been an
acceleration in the growth rate of the gross domestic savings by the
private corporate sector during the period under consideration
Table 7. Search For Shift In The Growth Rate Of Domestic Public Sector
Savings
Dependent Variable: Gross Domestic Public Sector Savings= Log(Gds4)
Method: Least Squares. Sample: 1950 2002. Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
4.930187
0.092671
53.20096
Time*
0.115123
0.003891
29.58505
Dummy
-4.638735
3.255258
-1.424998
Dummy*time
0.094657
0.073172
1.293625
R-squared
0.964351
Durbin-Watson statistic
1.008623
Adjusted R-sq.
0.961921
Notes:* Significant at one percent level

The regression results relating to the growth rate of the savings by

the public sector furnished in table 7 demonstrate that the
coefficients of dummy and the interaction variable are also not
47

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

significant in case of public sector showing the absence of the

shift during post economic reform period. In view of the absence of
shift the regression model with time and square of time has also
been attempted to see the presence of acceleration / deceleration in
the growth rate of the gross domestic savings by public sector .The
regression results of the same are furnished in table -8
Table 8. Acceleration/Deceleration In The Growth Rate Of Gross Domestic
Public Sector Savings
Dependent Variable: Gross Domestic Public Sector Savings= Log(Gds4)
Method: Least Squares.Sample: 1950 2002.Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
4.742408
0.126216
37.57370
Time*
0.144338
0.012420
11.62096
Yime2*
-0.000742
0.000256
-2.903182
R-squared
0.964006
Durbin-Watson statistic
0.945342
Adjusted R-sq.
0.962406
Notes: * Significant at one percent level

The regression results furnished in the above table show the

regression coefficient of square of time is negatively significant
implying that there has been deceleration in the growth rate of
domestic savings by public sector during the period under
consideration.
4.4: responsiveness of the domestic savings to the changes in
domestic income. With a view to examine the degree of
responsiveness of the domestic savings to the changes in income the
domestic savings function with log linear specification has been
estimated at the aggregate level and disaggregate levels. The results
have been furnished in the following tables.
The numerical values of the regression results presented in table9 illustrate that the estimate of constant income elasticity of savings
at the aggregate level is more than unity and significant during pre
economic reform period revealing that, on the average, a one percent
increase in domestic product will lead to increase the domestic
savings by 1.22 percent, all else equal. The regression coefficient of
interaction variable, which is differential income elasticity of gross
48

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

domestic savings, is significantly negative showing that the income

elasticity of gross savings is approximately unity during post
economic reform period. The estimate of the income elasticity of
gross domestic savings, which is just above the unity during pre
economic reform period, has come down by 0.18 points during post
economic reform period. The size and sign of the differential
coefficient of income elasticity of gross savings show the absence of
an upward shift in the degree of income elasticity of gross domestic
savings during post economic reform period.
Table 9. Gross Domestic Savings Function - Income Elasticity Of Savings
Dependent Variable: Gross Domestic Savings= Log(Gds1)
Method: Least Squares. Sample: 1950 2002. Included Observations: 53
Variable
Coefficient
Standard
t-Statistic
Error
Constant*
-4.284424
0.138841
-30.85850
log(GDP)*
1.215723
0.012602
96.46912
Dummy
2.391573
1.176332
2.033077
Dummy*log(GDP)**
-0.184689
0.083228
-2.219082
R-squared
0.997610
Durbin-Watson statistic
0.782627
Adjusted R-squared
0.997463
Notes: * Significant at one percent level &** Significant at 5 percent level
Table 10.Gross Domestic Household Sector Savings Function - Income
Elasticity Of Savings
Dependent Variable: Gross Domestic Household Sector Savings=
Log(Gds2)
Method: Least Squares.Sample: 1950 2002.Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
-4.979960
0.152045
-32.75313
log(GDP)*
1.247834
0.013801
90.41804
Dummy
0.556306
1.288205
0.431846
Dummy* log (GDP)
-0.051395
0.091143
-0.563896
R-squared
0.997333
Durbin-Watson statistic
1.074421
Adjusted R-sq.
0.997169
Notes:*Significant at one percent level

The numerical values of the regression results presented in

table- 10 illustrate that the estimate of constant income elasticity of
49

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

household sector savings is more than unity and significant during

pre economic reform period illuminating that a one percent increase
in income augments, on the average, the gross domestic savings by
household sector by 1.25 percent, all else equal. The regression
coefficient of interaction variable, which is differential income
elasticity of gross domestic savings, is negatively insignificant
showing that the income elasticity of gross domestic savings by
household sector is stable both in pre and post economic reforms
period. Thus the estimates of income elasticity of gross domestic
savings are akin both in pre and post economic reform periods
Table 11. Gross Domestic Private Corporate Sector Savings Function Income Elasticity Of Savings
Dependent Variable: Gross Domestic Savings= Log(Gds2)
Method: Least Squares. Sample: 1950 2002. Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
-6.157206
0.266644
-23.09146
log (GDP)*
1.176091
0.024203
48.59370
Dummy
0.960648
2.259146
0.425226
Dummy* log (GDP)
-0.040495
0.159839
-0.253351
R-squared
0.992185
Durbin-Watson statistic
1.118271
Adjusted R-sq.
0.991707
Notes:* Significant at one percent level

The estimates of the regression results presented in table- 11

illustrate that the regression coefficients on dummy and interaction
variable are not significant showing the absence of any shift in the
private corporate sector savings function during post economic
reform period. The regression coefficient of income, estimate of
constant income elasticity of private corporate sector savings, is
more than unity and significant during pre economic reform period
illuminating that a one percent increase in income augments, on the
average, the gross domestic savings by private sector
by 1.18
percent, all else equal. The regression coefficient of interaction
variable is negatively insignificant showing that the income
elasticity of gross domestic savings by private sector is stable both
in pre and post economic reforms period.

50

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

The estimates of the regression results pertaining to public

sector savings function presented in table- 12 illustrate that the
regression coefficients on dummy and interaction variable also not
significant showing the absence of any shift in the public sector
savings function during post economic reform period. The regression
coefficient of income, is unity and significant during pre economic
reform period illuminating that a one percent increase in income
augments, on the average, the gross domestic savings by household
sector by 1.07 percent, all else equal. The regression coefficient of
interaction variable is insignificant
Table 12. Gross Domestic Public Sector Savings Function -Income
Elasticity of Savings
Dependent Variable: Gross Domestic Public Sector Savings= Log(Gds4)
Method: Least Squares. Sample: 1950 2002. Included Observations: 53
Variable
Coefficient
Standard Error
t-Statistic
Constant*
-4.461101
0.486797
-9.164192
log (GDP)*
1.073952
0.044185
24.30571
Dummy
-6.669153
8.370938
-0.796703
Dummy* log (GDP)
0.419883
0.602945
0.696386
R-squared
0.948412
Durbin-Watson statistic
0.735745
Adjusted R-sq.
0.944894
Notes:* Significant at one percent level

4.5: Distributed lag domestic savings function : speed of

adjustment. With a view to examine the extent of speed of
adjustment between the actual change and desired change in the
domestic savings, a distributed lag model has also been estimated at
aggregate and disaggregate levels. The estimates of the same are
presented in the following tables.
The regression results of log linear distributed lag domestic
savings function for the Indian Economy show that the regression
coefficient of lagged domestic savings is statistically significant
evincing the presence of lag in the adjustment of actual domestic
savings to its desired level. The value of the coefficient of partial
adjustment or speed of adjustment implies that eighteen percent of
the discrepancy [disequilibrium] between actual change and desired
51

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

change in the gross domestic savings in the Indian Economy can be

eliminated in a year, all else equal.
Table 13. Distributed Lag Gross Domestic Savings Function - Regression
Results
Dependent Variable: Gross Domestic Savings= Log(Gds1)
Method: Least Squares. Sample: 1950 2002. Included Observations: 52
Variable
Coefficient
Standard Error
t-Statistic
Constant
-0.628859
0.457933
-1.373255
log(GDP)**
0.206618
0.136732
1.511117
log(GDS1(-1))*
0.829715
0.117456
7.064024
R-squared
0.998306
= Speed of adjustment=0.18
Durbin-Watson statistic 1.628006
Adjusted R-sq.
0.998237
Notes:* Significant at one percent level &** Significant at 5 percent level

Table-14. Distributed Lag Gross Domestic Household Sector Savings

Function-Regression Results
Dependent Variable: Gross Domestic Household Savings= Log(Gds2)
Method: Least Squares. Sample: 1950 2002. Included Observations: 52
Variable
Coefficient
Standard Error
t-Statistic
Constant*
-1.602218
0.604020
-2.652593
log(GDP)*
0.426114
0.155477
2.740694
log (GDS2(-1))*
0.655971
0.129073
5.082188
R-squared
0.997911
= Speed of adjustment=0.35
Durbin-Watson statistic 1.810173
Adjusted R-sq.
0.997826
Notes:*Significant at one percent level

The regression results of log linear distributed lag domestic

savings function estimated for household sector
show that the
regression coefficient of lagged domestic is statistically significant
revealing that there is a significant lag in the adjustment of actual
domestic household sector savings to its desired [long run] level.
The value of the coefficient of partial adjustment, known as speed of
adjustment, is 0.38 implying that, all else equal, thirty eight percent
of the discrepancy [disequilibrium] between actual change and
desired change in the gross domestic savings by household sector in
the Indian Economy can be eliminated in a year.

52

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

The regression results of log linear domestic savings function

estimated for private household sector [Table-15] show that the
regression coefficient of lagged domestic savings is statistically
significant evincing that there is a significant lag in the adjustment
of actual domestic private sector savings to its desired[long run]
level. The value of the speed of adjustment implying that forty two
percent of the discrepancy [disequilibrium] between actual change
and desired change in the gross domestic savings by private
corporate sector in the Indian Economy can be eliminated in a year,
all else equal.
Table 15. Distributed Lag Gross Domestic Private Sector
Savings
Function - Regression Results
Dependent Variable: Gross Domestic Private Sector Savings= Log(Gds3)
Method: Least Squares. Sample: 1950 2002. Included Observations: 52
Variable
Coefficient
Standard Error
t-Statistic
Constan*t
-2.837170
0.778766
-3.643163
log (GDP)*
0.520575
0.137628
3.782480
log (GDS3(-1))*
0.586922
0.111351
5.270940
R-squared
0.993327
= Speed of adjustment=0.42
Durbin-Watson statistic 1.854318
Adjusted R-sq.
0.993055
Notes:*Significant at one percent level
Table 16. Distributed Lag Gross Domestic Public Sector Savings
Function -Regression Results
Dependent Variable: Gross Domesticpublic Sector Savings= Log(Gds4)
Method: Least Squares. Sample: 1950 2002. Included Observations: 52
Variable
Coefficient
Standard Error
t-Statistic
Constant
-0.648634
0.455235
-1.424832
log (GDP)**
0.219191
0.099903
2.194038
log (GDS4(-1))*
0.768970
0.100409
7.658360
R-squared
0.968192
= Speed of adjustment=0.24
Durbin-Watson statistic 2.044703
Adjusted R-sq.
0.966747
Notes:* Significant at one percent level,** Significant at five percent level

The regression results based on log linear distributed lag

domestic savings function for public sector
show that the
regression coefficient of lagged domestic savings is statistically
significant implying the presence of significant
lag in the
53

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

adjustment of actual domestic public sector savings to its desired

[long run] level. The empirical value of the speed of adjustment
implies that twenty four percent of the discrepancy [disequilibrium]
between actual change and desired change in the gross domestic
savings by public sector in the Indian Economy can be eliminated in
a year, all else equal
5. Conclusions
On the basis of the results the empirical validity of the hypotheses
that There has been acceleration in the growth rate growth rate of
savings of the household sector, The marginal propensity to save
of the household sector is relatively higher than that of the private
and government sectors and The elasticity of domestic savings of
the household sector with respect to the changes in income is
relatively higher than that of the private and government sectors
can be accepted. However the empirical validity of the hypothesis
that The speed of adjustment between actual change and desired
change in domestic savings of household sector is not quick can not
be accepted as the coefficient of speed of adjustment in case of
household sector is somewhat smaller as compared to private sector.
The savings behaviour in the Indian Economy has been
empirically
examined
in
terms
of
presence
of
acceleration/deceleration in the growth rates of domestic savings,
responsiveness of the domestic savings to the changes in gross
domestic product and extent of discrepancy between actual change
and desired change in domestic savings at the aggregate and
disaggregate levels during 1950 2002.The empirical results show
that there is no shift in the growth rate of the domestic savings both
at aggregate and disaggregate levels during post economic reform
period. However there has been acceleration in the growth rates of
domestic savings of household and private sectors and deceleration
in public sector during the period under consideration. The estimate
of constant income elasticity of household savings is found to be
more than unity and relatively higher than the private and public
sectors. The results point out that there is no shift in the magnitude of
income elasticity of savings of household, private and public sectors
54

Upender, M., Reddy, N.L.

Savings Behaviour in the Indian Economy

during post economic reform period showing the homogeneity in the

size of the income elasticity of domestic savings. The extent of
discrepancy that can be eliminated in a year between actual change
and desired changes in the domestic savings is ranged from
eighteen percent to forty two percent at the aggregate and
disaggregate levels. In the light of the results it can be understood
that the economic reforms that have been initiated in 1992 could not
augment the growth rate of savings and income elasticity of domestic
savings both at aggregate and disaggregate levels in the Indian
Economy.It should be noted that the results emerged out the present
empirical exercise though useful to understand the behaviour of the
domestic savings in terms of shift in growth and income elasticity of
domestic savings at the aggregate and disaggregate levels during post
economic reform period, they are subjected to the specification of the
relationship between savings and income and data used in the study.
Bibliography
Ashok Rudra, (1976). Consumption, Saving and Capital Formation
A trend Report in A Survey of Research in Economics, Vol II, Macro
Economics, ICSSR, Allied.
Brahmananca, P.R. (1995), Planning for a Wage Goods Economy,
Himalaya Publishing House, Mumbai.
Chakarvarty (1990): Overall Aspects of Saving in Real Terms, Datta
Roy Choudar & Bagchi, (eds.), VikasPublication, New Delhi, 95162.
Charan D Wadhava [ed.] (1978), Some problems of Indias
Economic Policy, Tata McGraw-Hill, New Delhi.
Joshi, V.H. (1970). Saving Behaviour in India, Indian Economic
Journal, Vol. 15, 1969-70.
Pandit, B.L. (1985). Saving Behaviour and Choice of Indian
Households The Indian Economic Review, Vol. XX.

55

International Journal of Applied Econometrics and Quantitative Studies Vol.4-1 (2007)

Appendix-1
Domestic Savings By Sector.Summary Statistics: Whole Period :1950-2002
Summary
Gross
Gross
Gross
Gross
statistic
Domestic
Domestic
Domestic
Domestic
Savings
Household
Private
Public Sector
Savings
Sector Savings
Corporate
Sector Savings
Mean
95758
69796
23532
843
Median
17408
9743
1413
1379
Maximum
597697
519040
559258
24065
Minimum
861
55
64
-62704
Std. Dev.
157072
123676
78830
15146
Domestic Savings By Sector. Summary Statistics: Pre-Economic Reform Period

Summary
statistic

Gross
Domestic
Savings

Mean
Median
Maximum
Minimum
Std. Dev.

24448
7008
143908
861
35928

Gross
Domestic
Household
Sector Savings
18669
4926
110736
583
28827

Gross
Domestic Private
Corporate
Sector Savings
2509
720
20304
64
4279

Gross
Domestic
Public Sector
Savings
3270
1361
12868
143
3454

Domestic Savings By Sector. Summary Statistics: Post-Economic Reform Period

Summary
statistic

Gross
Domestic
Savings

Gross Domestic
Household
Sector Savings

Gross Domestic
Private Corporate
Sector Savings

Mean
Median
Maximum
Minimum
Std. Dev.

368034
352178
597697
162906
141755

265007
233252
519040
55
152430

103801
63486
559258
19968
152626

Journal published by the EAAEDS: http://www.usc.es/economet/ijaeqs.htm

56

Gross
Domestic
Public
Sector
Savings
-8421
5445
24065
-62704
32012