SYMPOSIUM

Financial Sector Development and Economic

Growth: The South-Asian Experience
S. M. A H M E D A N D M. I. A N S A R I

This article investigates the relationship between financial sector development and eco-
nomic growth for three majors South-Asian economies, namely, India, Pakistan, and Sri
Lanka. The standard Granger causality tests are employed to determine the pattern of
causal linkage between various measures of financial sector development and economic
growth. Also, several regression equations are estimated, using the Cobb-Douglas produc-
tion function framework, in order to analyze the impact of financial sector development on
economic growth. Results from causality analysis indicate that financial sector develop-
ment causes economic growth in the Granger sense. This seems to validate the supply-lead-
ing hypothesis. The regression results, using pooled data based on time-series and cross-
sectional observations, reinforce the findings of the causality analysis, suggesting that
financial sector development has indeed made a significant contribution to economic
growth in these countries.

INTRODUCTION

The importance of financial sector development in promoting economic growth has

been studied extensively in the literature (e.g., Goldsmith, 1966, 1969; Gurley &
Shaw, 1955, 1960; Patrick 1966; Porter, 1966; Khatldaate, 1972; McKinnon, 1973;
and Bhatia & Khatkhate, 1975). Economic history contains many examples of link-
ages between financial sector development and economic growth. Financial sector
development includes both financial widening and financial deepening. Financial
widening refers to the expansion of financial services and growth of financial institu-
tions, while financial deepening refers to either an increase in per capita amount of

S. M. Ahmed • Red Deer College, Red Deer, Alberta, T4N 5H5, Canada. M. I. Ansari • Athabassca University, Box 10000,
Athabasca, Alberta TOG 2R0, Canada.
Journal of Asian Economics, Vol. 9, No. 3, 1998 pp. 503-517 Copyright © 1998 by JAI Press Inc.
ISSN: 1049-0078 All rights of reproduction in any form reserved.

503
504 JOURNAL OF ASIAN ECONOMICS 9(3), 1998

financial services and institutions or an increase in the ratio of financial assets to

income. Financial sector development, by widening and diversifying the financial
markets, leads to a superior allocation of resources. Increased savings and their chan-
nelization into more productive uses in the economy should result in a superior allo-
cation of resources. More savings are generated through availability of a wide variety
of financial instruments and the lengthening of maturities of various assets. Besides,
financial sector development tends to increase investors' borrowing options allowing
investors to choose an optimal debt structure for a given project.
Many developing economies have recently implemented a wide variety of finan-
cial liberalization measures that have led to significant changes in their financial
structures. At the same time, some of these economies have managed to achieve high
rates of economic growth, while others have lagged behind. Although differences in
their economic performance can be attributed to many factors, there is a growing
belief among economists that financial development has played an important role in
promoting this high rate of economic growth. The objective of this paper is to evalu-
ate the role of financial sector development in economic growth in India, Pakistan,
and Sri Lanka. Since the role of financial sector development in economic growth has
not been investigated sufficiently for these economies, this study is intended to make
a significant contribution in this respect. 1 Because of these countries have had com-
mon colonial pasts and share similar socio-economic characteristics, it would be use-
ful to compare the interaction between their financial sector development and
economic growth.
The article is organized as follows. Section II provides an overview of the theo-
retical and empirical literature on the relationship between financial sector develop-
ment and economic growth. In section III we analyze the process of monetization in
these countries. Section IV describes data and the methodology and presents the
results of the Granger causality analysis. We present the results of the pooled regres-
sion equations in section V. Section VI contains some concluding remarks.

II. FINANCIAL SECTOR DEVELOPMENT AND

ECONOMIC GROWTH: THEORY AND EVIDENCE

There is a large body of theoretical and empirical literature on the role of financial
sector development in economic growth. Some economists hold the view that finan-
cial development is a necessary condition for achieving high rate of economic growth
(Goldsmith, 1969; McKinnon, 1973; and Shaw, 1973). This is what Patrick (1966)
refers to as the "supply-leading" role of financial development. Financial develop-
ment contributes to economic growth in the following ways: 1) financial markets
enable small savers to pool funds, 2) savers have access to a wider range of instru-
ments which stimulate savings, 3) efficient allocation of capital is achieved as the pro-
portion of financial saving in total wealth rises, 4) more wealth is created as financial
intermediaries redirect savings from individuals and the slow-growing sectors to the
The South-Asian Experience 505

fast-growing sectors, 5) financial intermediaries partially overcome the problem of

adverse selection in credit markets, and 6) financial markets encourage specialization
in production, development of entrepreneurship and adoption of new technology.
Levine (1997) breaks down the functions of a financial system into five basic tasks:
"financial systems 1) facilitate the trading, hedging, diversifying, and pooling of risk,
2) allocate resources, 3) monitor managers and exert corporate control, 4) mobilize
savings, and 5) facilitate the exchange of goods and services." (Levine, 1997, p. 691).
There is, however, a considerable debate on how exactly financial development
affects economic growth (Gupta, 1984; Spears, 1992). The structuralists, led by Gold-
smith (1969), are of the view that financial development directly increases savings in
the form of financial assets, thereby, encouraging capital formation and economic
growth. Empirical support for this hypothesis can be found in the works of Tun Wai
(1972), Sinai and Stokes (1972), and Wallich (1969). The financial repressionists, led
by McKinnon (1973) and Shaw (1973), on the other hand, believe that the rate of
retum on real cash balances is the key determinant of the rate of capital formation and
therefore, for achieving a high rate of growth. According to this view, the effect of
financial development on the economic growth depends on the effects of financial
development on the interest rate. This leads the financial repressionists to advocate
measures of financial liberalization, interest rates deregulation being the most promi-
nent one. Using a traditional aggregate saving equation, Fry (1988), in his study of
Asian LDC's, finds that the interest rate has a significant positive effect on the aggre-
gate savings function, although the magnitude of this effect is small for most coun-
tries. Most other empirical studies have not found any strong statistical association
between the real interest rate and domestic savings. Gupta (1987), in his study of
selected Asian and Latin American countries, found some evidence in support of the
repressionist view, that liberalization of interest rates is beneficial in generating
higher levels of savings in developing economies.
A distinctly opposite view has emerged in the literature which Patrick (1966)
terms as the "demand-following" role of financial development. In this view, financial
development is the 'handmaiden' of economic development, reacting passively to the
demands for new financial services by a growing economy (Robinson, 1952; Stern,
1989). The development in the financial sector is facilitated by the growth in the real
sector of the economy. As the economy grows, it needs a wider variety of financial
services and a growing number of institutions to provide those services. According to
this view, a scarcity of financial institutions and services reflects a low demand for
these services. Since financial intermediation helps transferring resources from the
slow-growing sectors to the fast-growing sectors of the economy, the need for finan-
cial intermediation depends also on the variance in growth rates among different sec-
tors of the economy (see Patrick, 1966). However, Patrick has noted that the response
of the financial sector to the growing demand for financial services originating in the
real sector "may not be automatic, flexible or inexpensive in underdeveloped coun-
tries" (Patrick, 1966, p. 175).
506 JOURNAL OF ASIAN ECONOMICS9(3), 1998

A synthesized view of the above two approaches is that there is actually a two-
way relationship between financial sector development and economic growth. Patrick
(1966) suggests that the nature of the relationship between financial sector develop-
ment and economic growth depends on the stage of economic development. During
the early stages of development, financial sector expansion plays a growth-augment-
ing role through the creation of financial institutions and the supply of financial
assets. This is consistent with the "supply-leading" view, as explained above. How-
ever, during the more advanced stages of development, financial sector expansion
plays a "demand-following" role.
More recently, there have been some studies that have adopted endogenous
growth approach. Bencivenga and Smith (1990) have shown that the existence of
financial intermediaries reduces investment in liquid assets with low rates of return.
Under certain assumptions about risk aversion, an equilibrium with financial interme-
diation yields higher growth rate than an equilibrium without financial intermedia-
tion. Cooley and Smith (1991) suggest that financial markets promote specialization
and technological innovation. A support for this synthesized view is also found in
some of these endogenous growth models (e.g., Greenwood & Jovonovic, 1990; King
& Levine, 1993a). In an endogenous growth model that used a cross-country regres-
sion and case studies of micro and macroeconomic effects of financial development,
King and Levine find that financial development increases productivity by choosing
high quality entrepreneurs and superior projects. A developed financial system helps
capital accumulation and economic growth by increasing the savings rate and/or by
encouraging technological innovations (Grossman & Helpman, 1991; Aghion &
Howitt, 1992). In other studies, King and Levine (1993a, 1993b), find evidence of 1)
a strong, positive relationship between each of the four financial development indica-
tors and three growth indicators, 2) a significant correlation between financial depth
and each of the growth indicators, 3) initial level of financial development being a
good predictor of subsequent rates of economic growth. King and Levine conclude
that finance does not merely follow economic activity, but plays an active role in eco-
nomic growth and capital accumulation.

III. MONETIZATION AND ECONOMIC GROWTH IN

INDIA, PAKISTAN, AND SRI LANKA

For the purposes of analyzing the financial sector growth, we have used the following
three measures of financial development.2

1. The ratio of broad money (currency plus demand deposits and quasi-money)
to nominal GDP (M2/YN). This measure has been widely used in the litera-
ture as a monetization variable.
2. The ratio of quasi-money (time and savings deposits) to nominal GDP (QM/
YN). Quasi money is generally considered to be a better measure of financial
The South-Asian Experience 507

intermediation than M2 because it excludes the currency and the demand

deposit components of the aggregate money supply. These two components
of monetary aggregates are seen to represent the transactions demand for
money rather than the asset demand.
. The ratio of domestic credit to nominal GDP (DC/YN). In underdeveloped
countries, domestic credit is mainly used to finance investment activities of
the firms and it represents the single most important source of investible
funds. Thus, domestic credit is considered to be an important productivity-
enhancing financial service, which is crucial for achieving a high rate of eco-
nomic growth (King & Levine, 1993).

We have used real GDP (Y) and per capital real GDP (PY) as the two income
variables in our study. In order to shed light on the magnitude of financial develop-
ment in the three countries over the period under study, we have summarized the
major financial and other macroeconomic indicators for each country in Tables 1
through 3. Following the 1971 India-Pakistan war,
Bangladesh became an independent state in December of 1971. Allowing for
some post-war adjustments, we have chosen 1973 as the initial year for our study. As
shown in Table 1, money supply, as measured by various monetary aggregates, has
shown a phenomenal growth in India. For instance, each one of M1, quasi-money,
and M2 in nominal terms has increased ten, twenty-six, and seventeen fold, respec-
tively, during the period from 1973 to 1991. Domestic credit has expanded from
178.6 billion rupees in 1973 to 3347.4 billion rupees in 1991, representing a nineteen
fold increase. Similarly, real money supply (M1) has more than doubled, quasi-
money in real terms has increased six fold, and M2 in real terms has increased four

TABLE 1. Major Financial and Other Economic Indicators for India

Indicators 1973 1977 1981 1985 1989 1991
Money, MI 101.0 178.5 232.5 412.4 746.9 1039.1
Quasi-money 71.5 143.6 385.5 740.1 1402.5 1857.9
Broad money, M2 172.5 322.1 618.0 1152.5 2149.4 2897.0
Domestic credit 178,6 324.0 675.5 1351.4 2531.2 3347.4
GDP deflator (1987=100) 34.2 43.9 64.2 86.4 117.2 147.1
Real money I 295,3 406.6 362.2 477.3 637.3 706.4
Real quasi-money 209,1 327.1 600.5 856.6 1196.7 1263.0
Real broad money 504.4 733.7 962.6 1333.9 1834.0 1969.4
Nominal GDP 1620,0 960.7 1598.2 2622.4 4539.9 6086.7
Real GDP 1814.9 2188.1 2490.4 3035.2 3872.7 4138.1
Real GDP per capita 3097 3408 3542 3968 4649 4773
Real GDP growth rate 3.4 7.26.5 5.4 5.4 1.2
Gross dom. invest. (%) 18.3 19.8 25.0 24.0 23.3 24.1
Note: 1Realmoneysupplyfiguresare obtainedby dividingthe nominalmoneysupplyby GDP deflators.All moneysupplyand
GDP figuresare in billionsof Indianrupeesexceptreal GDP per capita.
Source: All data seriesexceptdomesticcredit are taken fromthe WorldBank,WorldTables.The serieson domesticcreditis
taken fromthe IMF,International Financial Statistics Yearbook.
508 JOURNAL OF ASIAN ECONOMICS 9(3), 1998

TABLE 2. Major Financial and Other Economic Indicators for Pakistan

Indicators 1973 1977 1981 1985 1989 1991
Money, M1 22.2 40.0 72.3 123.1 217.0 306.0
Quasi-money 9.4 19.0 36.3 68.9 83.0 92.5
Broad money, M2 31.6 58.9 108.5 192.0 300.1 398.5
Domestic credit 32.2 66.0 127.0 242.8 395.0 520.3
GDP deflator (1987=100) 26.5 50.4 70.2 92.6 119.0 143.6
Real money 1 83.7 79.3 103.0 132.9 182.4 213.1
Real quasi-money 35.4 37.6 51.6 74.4 69.8 64.4
Real broad money 119.1 116.9 154.6 207.3 252.2 277.5
Nominal GDP 66.9 149.8 278.2 472.2 769.8 1022.3
Real GDP 251.9 296.9 396.5 509.8 646.8 711.8
Real GDP per capita 3776 3933 4659 5301 5939 6144
Real GDP growth rate 7.1 4.0 8.0 7.6 4.8 5.5
Gross dom. invest. (%) 12.9 19.3 18.8 18.3 18.9 18.7
Note: IReal money supply figures are obtained by dividing the nominal money supply by GDP deflators. All money supply and
GDP figures are in billions of Pakistani rupees except real GDP per capita.
Source." All data series except domestic credit are taken from the World Bank, World Tables. The series on domestic credit is
taken from the IMF, International Financial Statistics Yearbook.

TABLE 3. Major Financial and Other Economic Indicators for Sri Lanka
Indicators 1973 1977 1981 1985 1989 1991
Money, M1 2.9 5.3 10.0 18.7 35.1 46.6
Quasi-money 2.5 5.9 21.0 45.2 64.2 92.3
Broad money, M2 5.3 l 1.2 31.0 63.9 99.3 139.0
Domestic credit 4.5 10.6 39.4 65.7 117.2 150.2
GDP deflator (1987=100) 17.8 29.3 54.6 86.7 121.8 163.0
Real money 1 15.5 18.2 18.2 21.5 28.8 28.6
Real quasi-money 14.1 20.0 38.5 52.2 52.7 56.6
Real broad money 29.6 38.2 56.7 73.7 81.5 85.2
Nominal GDP 18.4 36.4 85.0 165.1 255.8 381.0
Real GDP 103.3 124.3 155.7 190.4 210.0 233.7
Real GDP per capita 7894 8920 10388 12017 12493 13596
Real GDP growth rate 9.5 5.1 5.6 9.8 2.1 4.7
Gross dom. invest. (%) 13.8 14.5 27.8 23.4 21.4 22.7
Note." IReal money supply figures are obtained by dividing the nominal money supply by GDP deflators. All money supply and
GDP figures are in billions of Sri Lankan rupees except real GDP per capita.
Source." All data series except domestic credit are taken from the World Bank, World Tables. The series on domestic credit is
taken from the IMF, International Financial Statistics Yearbook.

fold over this period. Over the same period, real G D P has doubled while per capita
G D P has increased by fifty percent. The ratio o f M 2 to nominal G D P (M2/YN) has
increased from 28 percent in 1973 to 47 percent in 1991. The quasi-money as a per-
centage o f nominal G D P (QM/YN), on the other hand, has increased from 4.4 percent
to 30.5 percent over this period. Similarly, the ratio o f domestic credit to n o m i n a l
G D P (DC/YN) has increased from 11 percent in 1973 to 55 percent in 1991.
Pakistan has experienced similar increases in different measures o f both nominal
and real m o n e y supply, as reported in Table 2. For instance, M1 in nominal terms has
The South-Asian Experience 509

increased about fourteen fold, quasi-money ten fold, and M2 has increased thirteen
fold during the period from 1973 to 1991. Similarly, domestic credit has shown a six-
teen-fold increase. Real money supply (M1), has grown two and a half times and both
quasi-money and M2 in real terms have doubled over this period. The ratio of quasi-
money to nominal GDP (QM/YN) has registered a decline from 14 percent in 1973 to
9 percent in 1991, while the ratio of M2 to nominal GDP (M2/YN) has declined from
47 percent to 39 percent over this period. The ratio of domestic credit to nominal GDP
(DC/YN), however, has increased from 48 percent in 1973 to 51 percent in 1991. Real
GDP has almost tripled, while per capita GDP in real terms has increased by 63 per-
cent over this period.
For Sri Lanka, M1 has increased seventeen fold, quasi-money thirty-seven fold,
and M2 has increased twenty-six fold. Measured in real terms, these three monetary
aggregates have increased by almost two, four, and three times, respectively. Domes-
tic credit, on the other hand, has shown a thirty-three fold increase, from 4.5 billion
rupees in 1973 to 154.20 billion rupees in 1991. During the same period, real GDP
has almost doubled. The ratio of M2 to nominal GDP (M2/YN) has increased from 29
percent to 36 percent with some fluctuations. The ratio of quasi-money to nominal
GDP (QM/YN), on the other hand, has increased from 14 percent to 24 percent. Sim-
ilarly, the ratio of domestic credit to nominal GDP (DC/YN) has increased from 25
percent to 39 percent.
It is evident from the foregoing discussion that there has indeed been a phenom-
enal growth in all the major indicators of financial development in these countries.
Further, the size of the financial sector has grown faster than the real sector, resulting
in significant increases in all financial indicator/GDP ratios.

IV. FINANCIAL DEVELOPMENT AND ECONOMIC GROWTH:

A CAUSALITY ANALYSIS

Although there have been many studies on the relationship between financial devel-
opment and economic growth, they do not provide an unambiguous answer as to the
direction of this relationship. Perhaps the most thorough study employing causality
tests is Jung (1986). The study, based on data for 56 countries (of which 19 are indus-
trial economies), has found support for the synthesized view. The empirical evidence
has supported the supply-leading hypothesis for the LDCs, while it has supported the
demand-following hypothesis for the developed countries. In his study of 10 sub-
Saharan countries, Spears (1992) has applied the Granger causality tests and has
found that financial intermediaries, as measured by M2/YN, cause economic growth.
In order to gain some preliminary insights into the strength of the relationship
between financial development and economic growth, we have first carried out a pair-
wise correlation analysis. Table 4 shows the correlation coefficients between different
measures of financial development and per capita nominal and real GDP. For all three
countries, the correlation coefficients are positive as expected and are found to be sta-
510 JOURNAL OF ASIAN ECONOMICS 9(3), 1998

TABLE 4. Results of Correlation Analysis

Financial Per capita Per capita
Country indicators nominal GDP real GDP
India M2/YN .98 .99
QM/YN .96 .97
DC/YN .88 .91
Pakistan M2/YN .96 .99
QM/YN .78 .87
DC/YN .70 .77
Sri Lanka M2/YN .89 .98
QM/YN .57 .77
DC/YN .59 .76

TABLE 5. Results of Causality Tests for India

Lags Directionof causality F-value Accept or Reject Ho:
2 M2/YN~Y .91 Accept
Y~M2/YN 6.25 Reject at 5%
2 QM/YN~,Y 4.47 Reject at 5%
Y~QM/YN 1.21 Accept
1 DC/YN-~Y 3.58 Reject at 10%
Y~DC/YN .80 Accept
2 DC/YN~Y 3.04 Reject at 10%
Y-~DC/YN 2.60 Accept
2 M2/YN-~PY 1.24 Accept
PY-~M2/YN 7.49 Reject at 1%
3 M2/YN~PY .42 Accept
PY~M2/YN 2.96 Reject at 10%
2 QM/YN-~PY 6.38 Reject at 5%
PY-~QM/YN 1.45 Accept
1 DC/YN~PY 4.09 Reject at 10%
PY-~DC/YN .54 Accept

tistically significant at the one percent level. But a correlation analysis does not indi-
cate the d i r e c t i o n o f causality. Thus, we h a v e c o n d u c t e d c a u s a l i t y tests b e t w e e n
measures o f financial d e v e l o p m e n t and e c o n o m i c growth for each o f the three coun-
tries.
In the G r a n g e r model, the procedure is to determine whether the inclusion o f the
causal variable reduces significantly the forecast error. It uses only lagged values o f
both the causal and the dependent variable. A null hypothesis o f no causality is tested
in a joint test that coefficients o f the lagged causal variable are significantly different
f r o m zero. To test for causality, we have used the following two regression equations:

F t = ZC~liYt_i + El~2iFt_i + V t ... (1)

Yt = ~ l i F t - i + Y~i2iYt-i + Ut -" (2)

The South-Asian Experience 511

TABLE 6. Results of Causality Tests for Pakistan

Accept or
Lags Directionof causality F-value Reject Ho:
2 M2/YN-*Y 4.43 Reject at 5%
Y-~M2/YN .45 Accept
3 M2/YN-~Y 4.36 Reject at 5%
Y-*M2/YN .47 Accept
3 QM/YN-~YN 3.18 Reject at 10%
YN~QM/YN 1.34 Accept
1 DC/YN-*Y 4.42 Reject at 10%
Y--*DC/YN 3.51 Reject at 10%
2 DC/YN~Y 7.67 Reject at 1%
Y~,DC/YN 1.35 Accept
3 DC/YN-*Y 5.35 Reject at 5%
Y~DC/YN 1.12 Accept
4 DC/YN-*Y 3.24 Reject at 10%
Y-*DC/YN 1.93 Accept
2 M2/YN--*PY 4.02 Reject at 5%
PY--*M2/YN .57 Accept
3 M2/YN-*PY 3.80 Reject at 10%
PY~,M2/YN .53 Accept
4 QM/YN~PY .88 Accept
PY~QM/YN 3.35 Reject at 10%
1 DC/YN-,PY 3.68 Reject at 10%
PY-*DC/YN 4.08 Reject at 10%
2 DC/YN-~PY 7.00 Reject at 1%
PY-*DC/YN 1.81 Accept
3 DC/YN-+PY 4.06 Reject at 5%
PY-*DC/YN 1.35 Accept
4 DC/YN~PY 2.62 Accept
PY-,DC/YN 3.80 Reject at 10%

where Y is real gross domestic product (or per capita real income, PY) and F is s o m e
indicator o f financial development. Equations (1) and (2) provide the following four
possible causal relationships.

a. Unidirectional causality f r o m Y t to F t exists if Y~(~li ~: 0 and ~ l i = 0.

b. Unidirectional causality exists from F t to Yt if Y~li 4:0 and Y~(~li = 0.
c. Bidirectional causality between F t and Yt exists if both Y~Otli ~ 0 and Y~li ~
0.
d. No causality is established between F t andY t if both ~(~li = 0 and ~ l i = 0.

T h e results o f the causality tests for India are reported in Table 5. In all estima-
tions we have used a four-period lag structure. In order to e c o n o m i z e space, we have
reported results for only those lags which have b e e n found to be statistically signifi-
cant in at least one direction. As the table shows, the evidence o f uni-directional cau-
sality running f r o m Q M / Y N to Y and PY at lag two and f r o m D C / Y N to Y and PY at
512 JOURNAL OF ASIAN ECONOMICS 9(3), 1998

TABLE 7. Results of Causality Tests for Sri Lanka

Direction of Accept or
Lags causality F-value Reject 11o:
4 M2/YN-~Y 6.98 Reject at 5%
Y~M2/YN .30 Accept
4 QM/YN~Y 3.67 Reject at 10%
Y-~QM/YN .10 Accept
4 M2/YN~PY 6.31 Reject at 5%
PY-~M2/YN .32 Accept
4 QM/YN~PY 3.30 Reject at 10%
PY~QM/YN .10 Accept

lag one is very strong. These results support the hypothesis that financial develop-
ment, as measured by these two variables, causes economic growth. On the other
hand, there is strong evidence of a reverse causation suggesting that both Y and PY
lead to growth in M2/YN.
Causality results for Pakistan, as reported in Table 6, indicate that M2/YN
Granger causes real income (Y) and per capita real income (PY) at lags two and three
and QM/YN Granger causes YN at lag three. Similarly, DC/YN seems to Granger
cause Y at lags one, two, three, and four, and PY at lags one, two, and three. There is
also some evidence of bi-directional causality between DC/YN and both Y and PY at
lag one. Results of causality tests for Sri Lanka are reported in Table 7. As the table
shows, both M2/YN and QM/YN Granger causeY and PY at lag four. There is no evi-
dence of causality running from domestic credit to either of the two measures of
income.
Thus, these results seem to be reasonably robust, leading us to conclude that in
each of the three cases financial sector development has led to a growth in income.
Since a causality analysis does not indicate the magnitude of the underlying relation-
ships, it is no substitute for an impact study. We, therefore, estimate regression equa-
tions incorporating important explanatory variables including the measures of
financial development in the following section.

V. F I N A N C I A L D E V E L O P M E N T AND E C O N O M I C GROWTH:
AN E C O N O M E T R I C ANALYSIS

In this section we rely on the basic determinants of growth model which has been
used extensively in empirical studies. We begin with the following Cobb-Douglas
type equation.

Y = e h K ~ L ft F r ... (3)

Taking a natural log of both sides, we get,

The South-Asian Experience 513

lnY = h + odnK + ~lnL + FlnF ... (4)

Differentiating equation (4) with respect to time, we get,

dlnY _ h+ dK1 ,dL1 FdF1

dt ct -d-[ K + t~~ r Z + -d t F "'" (5)

Adding an error-term that satisfies the standard assumptions, equation (5) can be
written as follows:

YT t = h + txKT t + ~LT t + 1-TTt + u t ... (6)

where

dL 1 dF 1
dlnY yT t , dK 1 = KT t • = L T t and . . . . FT t (7)
d--~ = d---['K, ' - ~ r -L ' dt F

The four terms in equation (6) represent, respectfully, growth rates of income,
capital, labour, and financial sector development. Since the equation is in a double-
log form, the coefficients 'ct', '13' and 'F' are elasticities of output with respect to cap-
ital, labor, and financial sector development. The constant 'h' is expected to capture
possible productivity effects of Hicks-neutral technological progress. Since data on
growth of capital stock is not available, we have used I/Y as the proxy variable for KT
(See Park, 1992). The following is a brief description of all the variables used in the
model and their sources.

Y = GDP in 1987 prices, collected from World Bank, World Tables.

I/Y = Investment as proportion of GDP. Both investment and GDP are in the
1987 prices, collected from World Bank, World Tables.
L = Population has been used as the proxy for labour force for each country.
Population figures are from World Bank, World Tables.
F = Three alternative measures of financial development defined in section
III.

The sample period, 1973-91 yields only 19 observations on each country.

Recognizing that the inadequacy of sample size can seriously undermine the reli-
ability o f these results, we have a d o p t e d the p o o l i n g of data a p p r o a c h and
employed the combined cross-section and time-series model, obtaining a total of
57 observations.
Equation (6) can be re-written as follows:

YTit = hi + °ti KTit + [~i LTit + l"i F'Fit + uit ... (7)
514 JOURNAL OF ASIAN ECONOMICS 9(3), 1998

where i=1,2,3 and t=l, 2 ..... 19. This model can be estimated either as a random error
components model or as a cross-sectionally correlated and time-wise autoregressive
model (Kmenta, 1971). We have used the later approach under which the error term,
Uit has the following structure.

E(u2it) = (Yii (heteroscedasticity)

E(uit ujt) = (Jij (contemporaneously correlated)
Uit = Pi ui, t-1 + (Yit (autoregressive)

where

E it ~ N(0,(~ii), E(ui, t - l e j t ) = 0),

E ( e itEjt) = {~ij,
E ( E itEjs) = 0 (t 4: s), i,j = 1,2,3.

By assumption, the initial value of u is set to have the following properties:

ui0 ~ N [ 0 , {~ij/(1-pi2)],
E(uio Ujo) = ¢ij/(1-piPj).

We have used Shazam version 6.1 to estimate the combined equation. This model
has been estimated in three specifications, one for each of the three measures of finan-
cial development. Table 8 presents the estimated results of the cross-sectionally het-
eroscedastic, time-wise autoregressive model, as suggested by Kmenta. Since we
have no a priori knowledge of the nature of cross-sectional disturbances, we have esti-
mated the model assuming both cross-sectional independence and cross-sectional

TABLE 8. Results of the pooled time-series and cross-section model

Specification (1) Specification (2) Specification (3)
N=3, T=19, K=3 N=2, T=19, K=3 N=3, T=19, K=3
L .556 .421 .207
(9.47)* (12.37)* (5.04)*
I/Y .010 .014 .019
(.70) (.95) (.73)
M2 .562 -- --
(7.93)*
QM -- .831 --
(20.55)*
DC -- -- .967
(22.12)*
CONS. -.001 -.042 -.061
(-.03) (-.46) (-.83)
Buse R 2 .98 .99 .96
Notes: T-values are in the parentheses.
*significant at the one percent level using a two-tailed test.
The South-Asian Experience 515

dependence. Since the results do not seem to differ we have reported the results of the
three specifications assuming cross-sectional dependence only. Judging by the size of
the Buse R 2, all three specifications have produced very good fits of the model. The
coefficient of the labor force is positive and statistically significant in each case. How-
ever, the coefficient of I/Y, even though has the positive sign in each specification, has
very low statistical significance. This might imply that I/Y is not a good proxy for
investment/capital ratio (or the growth rate of capital stock). But more important for
our present study, the coefficients of all three measures of financial development have
the expected signs and are statistically significant. For example, one percent increase
in the broad money supply leads to about .56 percent increase in the rate of economic
growth. The robustness of these results indicates that regardless of the measures cho-
sen, financial development had a significant impact on the rate of economic growth in
these countries.

VI. CONCLUSIONS

In this article we have investigated the relationship between alternative measures of

financial development and economic growth for three South-Asian countries, namely,
India, Pakistan, and Sri Lanka. We have used three approaches. First, results of the
correlation analysis indicate a high degree of association between financial sector
development and economic growth. Second, the results from the Granger causality
tests provide strong support for the supply-leading hypothesis, i.e., financial develop-
ment causes economic growth. This is consistent with Patrick's view that financial
development contributes to economic growth during the early stages of development.
Since causality tests do not take into account other relevant variables explaining eco-
nomic growth, we have estimated a Cobb-Douglas production function type equation
pooling the data series for three countries. The results from the pooled regression
model further reinforce the findings of the causality analysis. The main policy impli-
cation that follows from this study is that governments in these countries can help
promote economic growth by encouraging financial sector development through fur-
ther liberalization measures.

Acknowledgment: We gratefullyacknowledgeseveralvaluablecommentsand suggestionsfrom Tor-

ben Andersenand refereesof this journal. Authors alone are responsible for any remainingerrors.

NOTES

1. The only notable exception is Jung (1986) who has studied the causal relationship between
financial development and economic growth for 56 countries including India, Pakistan, and Sri
Lanka.
2. Goldsmith(1969) has discussed alternativemeasuresof financialdevelopment.
516 JOURNAL OF ASIAN ECONOMICS 9(3), 1998

REFERENCES

Aghion, E, and Howitt, E 1992. "A Model of Growth through Creative Destruction," Econometrica,
60:323-351.
Bencivenga, V., and Smith, B. D. 1991. "Financial Intermediation and Endogenous Growth," Review of
Economic Studies, 58:195-209•
Bhatia, R.J., and Khatkhate, D. R. 1975. "Financial Intermediation, Savings Mobilization and
Entrepreneurial Development: The African Experience," IMF Staff Papers, 22:132-158.
Christian, J., and Pagoulatos, E. 1973. "Domestic Financial Markets in Developing Economies: An
Econometric Analysis," Kyklos, 26:75-85.
Cooley, T. E, and Smith, B. D. 1991. "Financial Markets Specialization and Learning by Doing,"
Mimeo.
Fischer, B. 1981. "Interest Rates Ceilings, Inflation and Economic Growth in Developing Countries,"
Economics, 23:75-93.
Fry, M. J. 1988. Money, Interest and Banking in Economic Development. Baltimore: Johns Hopkins
University Press•
Galbis, V. 1977. "Financial Intermediation and Economic Growth in Less-Developed Countries: A
Theoretical Approach," Journal of Development Studies, 13:58-72.
Gertler, M., and Rose, A. 1991. "Finance, Growth and Public Policy" World Bank, Washington, DC,
working paper no. 814.
Goldsmith, R. W. 1966. The Determinants of Financial Structure. Paris: The OECD.
• (1969), Financial Structure and Development• New Haven: Yale University Press,
Greenwood, J., and Jovonovic, B. 1990. "Financial Development, Growth and the Distribution of
Income," Journal of Political Economy, 98:1076-1107.
Grossman, G., and Helpman, E. 1991. "Quality Ladders in the Theory of Economic Growth," Review of
Economic Studies, 58:43-61.
Gupta, K. L. 1984. Finance and Economic Growth in Developing Countries. London: Croom-Helm).
• (1987) "Aggregate Savings, Financial Intermediation and Interest Rate," Review of Economics
and Statistics, 69:303-311.
Gurley, J. G., and Shaw, E. S. 1955. "Financial Aspects of Economic Development," American
Economic Review, 45:515-538.
. 1960. Money in a Theory of Finance. Washington, D. C.: Brookings Institution.
Jung, W. S. 1986. "Financial Development and Economic Growth: International Evidence," Economic
Development and Cultural Change, 34:333-346.
Khatldaate, D. R. 1972. "Analytic Basis of the Working of Monetary Policy in Less Developed
Countries," IMF Staff Papers, 19.
King, R. G., and Levine, R. 1993a. "Finance and Growth: Schumpeter Might be Right," Quarterly
Journal of Economics, 108:717-738.
. 1993b. "Finance, Entrepreneurship and Growth: Theory and Evidence," Journal of Monetary
Economics, 32:515-542.
Kmenta, J. 1971. Elements of Econometrics. New York: MacMillan.
Levine, R. 1992. "Financial Structures and Economic Development," World Bank, Washington, DC.
working paper no. 849.
. 1997. "Financial Development and Economic Growth: Views and Agenda," Journal of
Economic Literature, 35:688-726.
McKinnon, R. I. 1973. Money and Capital in Economic Development• Washington, DC: The Brookings
Institution.
Park, J. H. 1992. "Exports and Economic Growth in Developing Countries: The Case of Latin
America," Rivista lnternazionale di Scienze Economiche e Commerciali, 39:239-258.
The South-Asian Experience 517

Patrick, H. T. 1966. "Financial Development and Economic Growth in Underdeveloped Countries,"

Economic Development and Cultural Change, 34:174-189.
Porter, R. C. 1966. "The Promotion of the 'Banking Habitat' and Economic Development," Journal of
Development Studies, July
Robinson, J. 1952. "The Generalization of the General Theory," in J. Robinson (ed.), The Rate of
Interest and Other Essays. London: MacMillan.
Shaw, E. S. 1973. Financial Deepening in Economic Development. New York: Oxford University Press.
Sinai, A., and Stokes, H. H. 1972. "Real Money Balances: An Omitted Variable in the Production
Function?" Review of Economics and Statistics.
Spears, A. 1992. "The Role of Financial Intermediation in Economic Growth in Sub-Saharan Africa,"
Canadian Journal of Development Studies, 13:361-380.
Stem, N. 1989 "The Economics of Development: A Survey" The Economic Journal, 100:597-685.
Tun Wai, U. (1972), Financial lntermediation and National Savings in Developing Countries, (New
York: Praeger).
Wallich, H. C. 1969. "Money and Growth: A Country Cross-Section Analysis," Journal of Money,
Credit, and Banking.

Received March 1997; Revised July 1998